Age Tenure General Self-Efficacy and Sales Performance of Sale

Section 1: Foundation of the Study

Spending millions of dollars on sales training, companies worldwide often fail to

address the potential effect of personal attributes of salespeople on high sales

performance (Atefi, Ahearne, Maxham, Donavan, & Carlson, 2018; Guenzi, Sajtos, &

Troilo, 2016). Despite significant advances in sales performance research since

Churchill's seminal work in 1977, many sales managers lack understanding of the

correlation between personal attributes of salespeople and sales performance within their

own industry (Carter, Dixon, & Moncrief, 2008; Walker, Churchill, & Ford, 1977). As

concluded by Hamstra et al. (2015), personal attributes of salespeople can predict up to

38% variance in sales performance, and, therefore, understanding these correlates is

important for sales managers. When searching for predictors for high sales performance,

personal attributes of salespeople are important research topics because these attributes

are typically longitudinally stable, and, therefore, significant correlates have high

predictive validity (Cobb-Clark & Schurer, 2012; Grether, Sowislo, & Wiese, 2018;

Wihler, Meurs, Momm, John, & Blickle, 2017).

The information technology (IT) industry is global highly competitive but

growing industry generating worldwide revenue of $4.8 trillion and employing 5.3

million people in the United States alone (Comptia, 2018). Using sales performance

correlates is of significant importance for sales managers within the IT industry for

several reasons. First, understanding the predictors of high sales performance might help

IT industry sales managers to improve sales performance. Second, high sales

performance directly affects companies’ share price, and, thus, the success of the

company (Dyer, Godfrey, Jensen, & Bryce, 2015; Kirkpatrick & Dahlquist, 2007). Third,

sales managers within the IT industry might improve human and social conditions by

applying findings of this study to a salespeople recruitment process, thereby reducing the

risk of discrimination based on age or experience.

Background of the Problem

A well-functioning sales organization is crucial for a contemporary company

wishing to achieve competitive advantage over its rivals (Guenzi et al., 2016). Scientific

research for sales performance and drivers for high sales performance have become

widely studied areas, and companies often rapidly adopt scientifically proven methods to

improve their sales performance (Hamstra et al., 2015; Yang, Kim, & McFarland, 2011).

Some IT companies have failed in improving their sales performance while others have

flourished (Akhter, Rahman, & Rahman, 2014; Kelly, 2018; Marley & Mooney, 2014).

Lack of understanding of basic correlates of sales performance may be one contributing

factor to decreased sales performance and industry-specific age-discrimination issues

(Fisher, Truxillo, Finkelstein, & Wallace, 2017; Hamstra et al., 2015; Quan, Dattero, &

Galup, 2010).

One of the most studied generalizable factors predicting sales performance is

individuals’ level of general self-efficacy (Goad & Jaramillo, 2014; Joseph, Jin,

Newman, & O’Boyle, 2015; Verbeke, Dietz, & Verwaal, 2011). General self-efficacy

construct has theoretical underpinnings in cognitive psychology, and researchers use it to

measure individuals’ overall belief in achieving goals in life despite unexpected

challenges or hurdles (Luszczynska, Scholz, & Schwarzer, 2005). High general self-

efficacy helps individuals to cope with high or unexpected personal and job stress (Lu,

Du, & Xu, 2016). Within the highly competitive IT industry, sales jobs are often

strenuous, and setbacks because of lost sales are frequent (Micevski, Dewsnap, Cadogan,

Kadic-Maglajlic, & Boso, 2019). Understanding the correlation between specific personal

attributes of salespeople with high sales performance may help managers to improve

sales performance.

Problem Statement

Companies across the globe spend millions of dollars on sales training but often

fail to address the potential effect of personal attributes of salespeople on high sales

performance (Atefi et al., 2018; Guenzi et al., 2016). However, these personal attributes

can account for up to 38% of the variance in monthly sales performance (Hamstra et al.,

2015). The general business problem was that some managers do not understand the

relationship between personal attributes of salespeople and sales performance. The

specific business problem was that some sales managers of technology hardware, storage,

and peripherals (THSP) IT companies do not understand the relationship between age,

tenure, general self-efficacy, and sales performance of salespeople.

Purpose Statement

The purpose of this quantitative correlational study was to examine the

relationship between age, tenure, general self-efficacy, and sales performance of

salespeople. The independent variables were age, tenure, and general self-efficacy. The

dependent variable was sales performance. The targeted population consisted of face-to-

face salespeople of THSP IT companies. The implications for positive social change

include the potential to better understand the correlates of sales performance, thus

contributing to the reduction of discrimination in the recruiting salespeople of various

ages and experience.

Nature of the Study

The three methods for solving research problems are quantitative, qualitative, and

mixed (Creswell, 2014). I used the quantitative method to examine the relationship

between age, tenure, general self-efficacy, and sales performance of salespeople.

Researchers use a quantitative method to test hypotheses using numerical data and

examine the relationship between variables and a qualitative method to gain an

understanding of underlying reasons, opinions, and motivations (Creswell, 2014;

Creswell & Poth, 2018). Mixed methods research involves integrating qualitative and

quantitative analysis in mixed methods research to solve research problems (Creswell,

2014). This study did not require qualitative inputs; therefore, neither qualitative nor

mixed methods were appropriate for this study.

Quantitative studies can be descriptive, correlational, quasi-experimental, or

experimental (Creswell, 2014). Researchers use descriptive designs to report basic

statistical characteristics of a sample, such as mean, standard deviation, or range of

analyzed variables of the sample (Creswell, 2014). The descriptive design did not meet

the standard for this study, as this study required an examination of the relationship

between variables. Researchers use the quasi-experimental and experimental designs to

study causal relationships between variables (Campbell & Stanley, 2010), making both

inappropriate for this study as the goal was to examine correlational relationships. In this

study, I used correlational design to allow an examination of noncausal relationships

between the variables.

Research Question and Hypotheses

What is the relationship between the age, tenure, general self-efficacy, and sales

performance of salespeople?

H0: There is no statistically significant relationship between age, tenure, general

self-efficacy, and sales performance of salespeople.

H1: There is a statistically significant relationship between age, tenure, general

self-efficacy, and sales performance of salespeople.

Theoretical Framework

The theoretical framework for this study was the theory of self-efficacy, originally

developed by Bandura in 1977 and later advanced as general self-efficacy by Sherer et al.

(1982). Bandura (1977) formulated the theory of self-efficacy to define drivers for human

behavior from the perspective of cognitive processing. Bandura asserted that the level of

a person’s own self-efficacy governs how that person copes with challenges and the

person’s willingness to expend effort to overcome the given challenge. Bandura

originally stressed that a person’s self-efficacy is situationally specific, but later scholars

expounded the theory of self-efficacy to include two dimensions: situational self-efficacy

and general self-efficacy (Luszczynska et al., 2005; Sherer et al., 1982).

Multiple researchers have associated a high level of salespeople’s self-efficacy

with increased sales performance in studies spanning different cultures and contexts

(Barbaranelli, Fida, Paciello, & Tramontano, 2018; Lu et al., 2016; Vieira, Perin, &

Sampaio, 2018). Because the relationship between general self-efficacy and sales

performance was significant in multiple earlier studies, I wanted to examine if a high

level of self-efficacy predicts high sales performance in a large-scale, globally operating

IT company regardless of age or length of tenure.

Operational Definitions

This section includes a definition of terms used in this study.

Face-to-face salespeople: Individuals working with THSP companies sales

organizations with a nominated set of customers. The responsibility of face-to-face

salespeople is to drive sales and ensure business performance and customer satisfaction

of their nominated set of customers (Ingram, 2015).

General self-efficacy: An individual’s overall belief in achieving goals in life

despite unexpected challenges or hurdles (Luszczynska et al., 2005; Sherer et al., 1982).

Homoscedasticity: A feature of statistical analyses indicating that the

homogeneity of variance of independent variables is similar (Mertler & Vannatta, 2017).

Sales performance (SP): A percent value of annual quota attainment, based on the

last four quarters’ average percent value. It is one typical method to measure sales

performance within the IT industry (Dearborn, 2015; Tuggle, 2014).

Self-efficacy (SE): Self-efficacy construct that refers to either task-specific self-

efficacy (TSSE) or general self-efficacy (Grether et al., 2018).

Task-specific self-efficacy (TSSE): An individual’s belief in his or her ability to

succeed in specific situations or accomplish a certain task (Bandura, 2012).

Technology hardware, storage, and peripherals (THSP): A subsector of the

global IT industry consisting of companies that manufacture and sell information

technology equipment (industry classification code 452020; S&P Global Market

Intelligence, 2018).

Tenure (TE): The length of time an employee worked with the company in the

same or similar job role (Hyatt & Spletzer, 2016). In this research, earlier consecutive job

roles account for tenure if a person was part of their current sales organization with a

face-to-face sales role.

Assumptions, Limitations, and Delimitations

Rigor research involves management of assumptions, limitations, and

delimitations to ensure the quality of research (Dane, 2018). By describing assumptions,

limitations, and delimitations, a researcher helps other scholars to identify areas of future

study as well as to critically analyze the conclusions of the research (Dane, 2018). This

section contains the description of the assumptions, limitations, and delimitations of this

research.

Assumptions

Assumptions are beliefs and pieces of information related to the study that the

researcher accepts as true but lack validation (Nkwake & Morrow, 2016). The

participants responded honestly to the survey and understood the general self-efficacy

questions similarly, despite lingual differences. The quotas of salespeople at the target

companies were equally fair, and quota setting did not induce bias to statistical results.

The instrument used to measure general self-efficacy measured the same variable

regardless of the participants’ cultural background. Participants’ mood, fatigue, and

attention span effect on questionnaire results distributed evenly. There was no attrition

bias caused by missing responses of salespeople who would have been part of the

population but who left the companies before answering the survey.

Limitations

Limitations are aspects of the study that may lessen the validity of the results of

the study but are not controllable by the researcher (Aguinis, Ramani, & Alabduljader,

2018). The short time for data collection may have limited the number of participants in

this study. Generalizability of this study may be limited to similar companies and to sales

roles that use a similar performance construct. Participants may have felt obliged to

answer positively to the general self-efficacy survey, and I did not provide the means to

measure such bias. Participants responded to general self-efficacy survey using the GSES

scale developed by Sherer et al. (1982), whereas more recent instruments could have

yielded different results (see Barahona, González García, Sánchez-García, Barba, &

Galindo-Villardón, 2018). Compared to other established instruments to measure general

self-efficacy, GSES has the widest adoption rate in multicultural studies, and multiple

authors concluded that the instrument provides a unidimensional measure of general self-

efficacy (Nel & Boshoff, 2016; Scholz, Gutiérrez Doña, Sud, & Schwarzer, 2002). In this

study, I used convenience sampling because sales performance research is prone to low

survey response rates (see Carter et al., 2008). Thus, the participants may not have been

representative of the whole population. Nonresponse bias may have limited the credibility

of statistical conclusions, as the methods to estimate bias were limited (Mullinix, Leeper,

Druckman, & Freese, 2015). The method to measure sales performance limited

comparison of quota fairness or difficulty between the participants.

Delimitations

Delimitations are intentional parameters that limit the boundaries of the study

(Newman, Hitchcock, & Newman, 2015). This study was limited to face-to-face

salespeople who, at the time of the study, (a) worked for a THSP IT company, (b) had a

tenure of at least 1 year, (c) were at least 20, (d) could read and understand English, and

(e) had an annual quota attainment result. Limiting participants to a single industry

subsector increased comparability between the dependent variable and allowed industry-

relevant analysis of the results. As this study was cross-sectional, potential longitudinal

changes in the individuals’ level of general self-efficacy were not within the scope of this

research. This research covered only certain characteristics of salespeople (age, length of

tenure, and general self-efficacy), and other personal characteristics that may correlate

with sales performance, such as big-five personality traits or emotional intelligence, were

not included in this research. The relationship between the variables used in this study

may also be nonlinear. However, no such relationship exists in the reviewed literature,

and the statistical methods used in this study could only provide inferential results with

the assumption of a linear relationship between the variables.

Significance of the Study

Contribution to Business Practice

Understanding the significance of the correlation between specific personal

attributes of salespeople with high sales performance may help managers improve sales

performance through informed recruitment. The population of this study consisted of

face-to-face salespeople of THSP IT companies with a global market capitalization value

of $981 billion (Fidelity Investments, 2018). As earlier studies indicated, the personal

characteristics of salespeople predict up to 38% variance in sales performance.

Consequently, an informed recruiting of sales personnel based on scientifically analyzed

sales performance predictors may yield significant revenue and share price increases.

Implications for Social Change

The people, planet, profit model developed by Elkington (1998) aligns business

objectives with an effect on people and their environment. This study contributed to

positive social change by examining attributes of salespeople that may correlate with high

sales performance. High age is often seen as linked to a decrease in sales performance

within the IT industry, which contributes to age discrimination as an industry-wide

problem (Fisher et al., 2017; Quan et al., 2010; Shaun, 2017). Sales managers of the IT

companies studied might improve human and social conditions by applying findings of

this study to a salespeople recruitment process, thereby reducing the risk of

discrimination based on age or experience.

A Review of the Professional and Academic Literature

The strategy used for the literature review was to systematically find English-

language, peer-reviewed scientific articles, doctoral-level studies, and business books

with verifiable source data. I searched ABI/INFORM, Academic Search Complete,

Business Source Complete, Emerald Management Journals, ProQuest Central,

PsycARTICLES, PsycINFO, SAGE Premier, Taylor & Francis Online, and Thoreau

databases. The search terms included (but were not limited to) sales performance, self-

efficacy, general self-efficacy, job performance in sales, sales performance antecedents,

self-efficacy construct, and self-efficacy instrument. Studies that addressed the effect of

self-efficacy (either general or task-specific) only in the academic world were not

included the literature review because success in academia is inherently different than

success in business-to-business sales (although both venues could share the same

antecedent factors). In total, the reviewed literature consists of 124 articles, of which 107,

or 86.29%, are from within 5 years (see Table 1).

Table 1

The Type and Date of Literature Reviewed

Literature t

Publication Yea

Book

sectio

Dissertatio

Peer reviewed

ournal article

Grand

Total

Recency

cumulative

percent

1977 1 1 100.00 %

1982 1 1 99.19 %

1988 1 1 98.39 %

1995 1 1 97.58 %

2001 1 1 96.77 %

2002 1 1 95.97 %

2005 1 1 95.16 %

2008 1 1 94.35 %

2009 1 1 93.55 %

2013 1 1 6 8 92.74 %

2014 2 11 13 86.29 %

2015 3 19 22 75.81 %

2016 1 5 21 27 58.06 %

2017 5 24 29 36.29 %

2018 15 15 12.90 %

2019 1 1 0.81 %

Grand Total 2 16 106 124

I categorized articles under eight main themes, used in this section: theory of self-

efficacy, general self-efficacy, alternative theories relating to predictors of sales

performance, sales performance research, performance in sales contexts, instruments to

measure general self-efficacy, and finally relationship between self-efficacy and sales

performance in which I compared and contrasted findings in chronological order of

existing studies, examining the relationship between at least two of the variables used in

this study.

The objective of this quantitative correlational study was to examine the

relationship between age, tenure, general self-efficacy, and sales performance of

salespeople. The independent variables were age, tenure, and general self-efficacy. The

dependent variable was sales performance. The null hypothesis in this study was there is

no statistically significant relationship between age, tenure, general self-efficacy, and

sales performance of salespeople. An alternative hypothesis was there is a statistically

significant relationship between age, tenure, general self-efficacy, and sales performance

of salespeople. Later advancement of Bandura’s (1977) theory of self-efficacy, general

self-efficacy by Sherer et al. (1982), was the theoretical framework for this study.

Theory of Self-Efficacy

Bandura, who is one of the most influential psychologists of all time (Diener,

Oishi, & Park, 2014), began to criticize behaviorism in the early 1970s, which seemed

incomplete in explaining psychological processes occurring before human actions. Based

on earlier work with social learning theory, Bandura used deductive reasoning and

empirical analysis to formulate a theory of self-efficacy in 1977 (Bandura, 1977). The

theory aimed to define drivers for human behavior from the perspective of cognitive

processing. Differing from previous motivational theories, such as Maslow’s (1943)

hierarchy of needs and Vroom’s (1995) expectancy theory, Bandura asserted that the

level of person's self-efficacy governs how that person copes with challenges and is

willing to expend effort to overcome the given challenge.

In an original empirical study, Bandura (1977) conducted two separate tests with

a group of volunteers who had a severe snake phobia. In both tests, Bandura randomly

assigned participants to three groups (participant modeling [G1], modeling alone [G2],

and control group [G3]). In the tests, participants completed increasingly frightening

tasks with boa constrictor snakes. The G1 group received therapeutic help anytime they

needed help (Bandura, 1977). The G2 group only observed the therapist performing the

tasks, and the G3 group did not receive help (Bandura, 1977). After the first testing, the

participants performed the same tests again, without external help (Bandura, 1977).

Before both tests, participants estimated their ability (i.e., perceived self-efficacy) to

perform the test with a 100-point probability scale (Bandura, 1977).

The tests confirmed Bandura’s (1977) expectation that participant modeling and

modeling alone increased self-efficacy levels, and self-efficacy was significantly

associated with the ability to perform the test. Hence, these tests confirmed Bandura’s

hypothesis that an individual’s perception of his or her ability to accomplish a certain task

at least partially explains the actual capability of performing the task. Bandura named this

perception of ones’ ability to accomplish a task as self-efficacy. Bandura also listed four

major sources for self-efficacy: performance accomplishments, vicarious experience,

verbal persuasion, and emotional arousal. Bandura (1997) repeatedly stressed that a

person’s self-efficacy is situationally specific. However, later scholars expounded the

theory of self-efficacy to include two dimensions: situational self-efficacy and general

self-efficacy (Luszczynska et al., 2005; Sherer et al., 1982).

General Self-Efficacy

Bandura (1997) originally stressed that a person’s self-efficacy always depends on

a specific situation, but Sherer et al. (1982) advanced the theory to two dimensions:

situational self-efficacy and general self-efficacy. Noting significant differences between

then-current behavioral theories, Sherer et al., who developed the construct of general

self-efficacy, still considered the construct as part of the theory of self-efficacy. Sherer et

al. specifically mentioned that Bandura predicted the presence of some general factor of

self-efficacy, but the theory did not include a construct to measure it. Sherer et al.

deduced that because self-efficacy is a product of past experience (both own and

observed), the persons who have endured multiple different experiences should logically

have a higher general level of self-efficacy than the persons with no such experiences.

Sherer et al. deduced that individuals’ anxiety (a facet of emotional arousal) in new

situations is also affected by prior experience and one’s personality. With these

observations, Sherer et al. hypothesized that at least three sources that Bandura listed for

self-efficacy together with one’s personality would contribute to a general form of self-

efficacy.

The results of some recent research about general self-efficacy indicated that the

level of general self-efficacy is mainly genetic and a relatively stable personality trait

through the life of an individual (Gottschling, Hahn, Maas, & Spinath, 2016; Waaktaar &

Torgersen, 2013). Waaktaar and Torgersen (2013) conducted a cross-sectional cohort

study among Norwegian twins born between 1988 and 1994 (7 cohorts, N = 1,394), and

measured variance in general self-efficacy caused by genetics and environmental factors.

Waaktaar and Torgersen measured general self-efficacy with the Children’s Perceived

Self-Efficacy Scale (developed by Pastorelli et al. [2001] in co-operation with Bandura).

The original Children’s Perceived Self-Efficacy Scale instrument consists of 37

questions, of which Waaktaar and Torgersen chose to use 12 questions in the study.

Waaktaar and Torgersen confirmed participant zygosity by DNA samples (15% of the

sample) and with discriminant questionnaire analysis with an estimated margin of error

<2%. Because of the ability to determine participants’ zygosity, Waaktaar and Torgersen

provided more accurate information on genetic hereditary of traits and genetic differences

between the participants (as cited in Cutler et al., 2015).

Waaktaar and Torgersen (2013) fitted covariances of raw data to the structural

equation model using maximum likelihood estimation and subsequently presented the full

psychometric model. Waaktaar and Torgersen concluded that 75% of the variation in

general self-efficacy is a result of genetic factors, and 25% of the variation is a result of

nonshared environmental causes (i.e., different education or hobbies). As expected,

shared environmental causes did not cause variance in self-efficacy between twins

(Waaktaar & Torgersen, 2013). However, as Waaktaar and Torgersen noted, their

research design was limited in separating the genetic factor from participants’ early

childhood experiences. Experiences during the first 5 years of childhood may

significantly affect personality (Kail & Cavanaugh, 2016; Soto & Tackett, 2015).

Therefore, Waaktaar and Torgersen’s research conclusions mean that genetic factors,

together with early childhood experiences, account for 75% of the variance in general

self-efficacy.

The question whether genetic effect and early childhood experiences on general

self-efficacy persist to adult life was partially answered by Gottschling et al. (2016) who

conducted cross-sectional correlational analysis among 579 people who participated in

German Twin Study research in 2006 measuring the effect of optimism (O), self-

regulation (SR), and self-efficacy (SE) to coping mechanisms at work. Gottschling et al.

used primary data and constructs from the German Twin Study on Personality and

Wellbeing from 2006. Because of twin study data containing participant DNA zygosity,

Gottschling et al. measured variance caused by genetic differences to each relationship

between measured variables.

To measure the genetic effect on each variable, Gottschling et al. (2016)

performed genetic mediation analysis, which revealed that genetic factors affecting SE

and SR substantially overlap with those affecting neuroticism (54%) and ReS (26%).

This result indicated that SE and SR reduce the negative effect of neuroticism on ReS

(Gottschling et al., 2016). Gottschling et al. also concluded that resistance to stress (ReS)

and occupational attitudes toward life (OcA) had significant relationship to SR (ReS β =

.57, p < .001, OcA β = .25, p < .001) and SE (ReS β = .54, p < .001, OcA β = .37, p <

.001).

Because Waaktaar and Torgersen (2013) concluded that inherited genetic factors

and early childhood experiences affect to the development of general self-efficacy, and

Gottschling et al. (2016) noted that the effect persists to adulthood and job contexts, one

open question remains: Does the level of general self-efficacy fluctuate during

adulthood? This question is important from the managerial perspective for at least two

reasons: If the level of general self-efficacy remains stable during adulthood,

measurement of general self-efficacy during the recruitment process could predict future

success of the recruits in sales roles. Second, if the level of general self-efficacy remains

stable during adulthood, the activities (such as coaching or training) have little or no

effect on one of the potential correlates of sales performance.

According to Bandura’s (1997) theory of self-efficacy, the four sources of self-

efficacy continue contributing to adult’s task-specific self-efficacies. This conclusion is

logical because adults can learn new skills, and acquired proficiency increases

individuals’ self-efficacy with the task that requires using newly learned skills

(Barbaranelli et al., 2018). However, the task-specific self-efficacy differs from the

general form of self-efficacy (i.e., general self-efficacy), which appears to be

longitudinally relatively stable (Grether et al., 2018; Schwarzer, 2006). Schwarzer’s

(2006) general self-efficacy measurements from over 19000 individuals from 26

countries indicate that median scores of general self-efficacy are within 1% in all age

groups between 20 and 70 years old adults, and that age is significant, but very weak

predictor for general self-efficacy (n = 6220, F = 7.81, r2 = .001, p < .01). This result

indicated that as individuals mature, their level of general self-efficacy continues to

increase, but the increase remains negligible. This construct stability over individuals’

lifetime indicates that the general self-efficacy is a trait-like characteristic of an

individual. Thus, several scholars consider general self-efficacy as one of the trait -

theoretical attributes with a potential of predicting future job or sales performance

(Grether et al., 2018; Smith, Kass, Rotunda, & Schneider, 2006).

Traumatic life events, diseases, chronic stress, and sleep deprivation may cause

neuronal and behavioral changes, and also affect to the level of general self-efficacy

(Cyniak-Cieciura, Popiel, & Zawadzki, 2015; Fuchs & Flügge, 2014). Although a high

level of general self-efficacy buffers against the negative effect of traumatic events,

serious illness, and stressful events, these may still cause a decrease in the level of

general self-efficacy of an individual (Cyniak-Cieciura et al., 2015; Welsh, Olson, &

Perkins, 2018). From a managerial perspective, these types of negative events may not be

visible at the workplace, but because of the relationship with general self-efficacy, they

may influence job performance. Thus, an assumption of longitudinal stability of once

measured general self-efficacy may not hold true if an individual experienced significant

negative events. However, informed sales managers could frequently measure the general

self-efficacy of salespeople to reveal factors that relate to sales performance but would

otherwise remain hidden.

Socioeconomic and demographic factors may correlate with adults’ level of

general self-efficacy. For instance, Bonsaksen et al. (2018) measured the relationship

between general self-efficacy (GSE) and sociodemographic characteristics among a large

group of Norwegian people (n = 1787) during Norwegian Population Study (NorPop). To

measure general self-efficacy, Bonsaksen et al. used Schwarzer’s scale (full ten-item

version), and to measure sociodemographic background, they collected participants’ age,

sex, education, employment status, relationship status, and the size of the city of

residence. Bonsaksen et al. then used analysis of variance (ANOVA) to measure the

difference of GSE among different sociodemographic groups, and linear regression

analysis to analyze total explained variance to GSE by sociodemographic factors.

Bonsaksen et al. (2018) concluded that male gender and being employed were

significantly related to higher GSE (Gender-GSE β = -.12, p < .001, being employed-

GSE β = .14, p < .001), and that age moderated this relationship so that the relationship

was stronger with young people. They also noted that all sociodemographic factors

explain 6.6% variance in GSE (7.6% with 1st tier interaction effects included; Bonsaksen

et al., 2018). Bonsaksen et al. noted that their research results should not be considered

causal. In particular, being employed could increase one’s general self-efficacy, but being

employed could also be a result of an individual’s higher general self-efficacy

(Bonsaksen et al., 2018).

Alternative Theories Relating to Predictors of Sales Performance

Sales function, whether a separate organizational entity or not, is an essential part

of any for-profit company (Guenzi et al., 2016). Sales function, or process, results in

revenue, which equals to the monetary value of goods and services that the company

produced (Brealey, Myers, & Marcus, 2018). Therefore, theoretical frameworks that

address organizational performance, job performance, or sales job performance, are valid

for researching correlates of sales performance. Because the purpose of this study was to

examine the relationship between age, tenure, general self-efficacy, and sales

performance of salespeople, alternative theoretical frameworks analysis remained limited

to the theories that address job performance or sales job performance. Also, because all

predictor variables in this research are characteristics of salespeople, theoretical

frameworks that address interpersonal factors (such as leadership theories, group

dynamics theories, or theories addressing the relationship between the salespeople and

the customer) were not covered in this research.

Motivational theories. Motives of human behavior (i.e., why people choose to

act as they do) have interested researchers since the days of early philosophy (Pinder,

2014). In an organizational context, researchers retain primary interest in how to measure

someone’s level of motivation, how to influence motivation level, and how the

motivation level affects work performance (Pinder, 2014). Early organizational theorists,

like Frederick Winslow Taylor (1856-1917), considered humans as mechanical actors

whose actions are guided by reward and punishment (Skinner, 2017). Elton Mayo (1880-

1949) extended the understanding of human motivation in an organizational context by

adding the need for social wellbeing as one antecedent for high motivation (Dagher,

Chapa, & Junaid, 2015).

Abraham Maslow (1908-1970) developed hierarchy of needs -theory, in which

description included antecedents of human motivation in a five-level hierarchical model

with the principle that all humans need to fulfill lower levels (such as physiological, and

safety) needs before higher-order needs (like self-esteem and self-actualization; Kanfer &

Chen, 2016). Frederick Herzberg (1923-2000) added the understanding of two types of

motivation: hygiene factors, which can only lower motivation if not fulfilled, and

motivators, which can increase motivation, if and when also fulfilling hygiene factors

(Kanfer & Chen, 2016). Victor Vroom (1964-) built upon the existing motivational

theories and added the intended outcome as one factor of motivation (Kanfer & Chen,

2016; Vroom & Jago, 2007). In Vroom’s expectancy theory, motivation is a function of

three factors: expectancy (understanding that one’s effort results expected job outcome),

instrumentality (understanding that a job outcome will result in personal reward), and

valence (understanding that the personal reward has personal value; Kanfer & Chen,

2016).

Motivation is one of the primary predictors of job performance, and motivation is

one of the top predictors of sales job performance (Cerasoli, Nicklin, & Ford, 2014;

Verbeke et al., 2011). When searching for how much motivation predicts job

performance in sales contexts, Walker et al. (1977) hypothesized that sales performance

is a product function of motivation, aptitude, and role perceptions of salespeople. Walker

et al. did not include any primary data analysis, which would have supported the

hypothesis. Extending the work of Walker et al., Churchill et al. (1985) conducted an

extensive meta-analysis of sales performance determinants by reviewing 116 published

and unpublished studies dated between 1918 and 1982. Churchill et al. concluded that

three additional factors affect sales performance: skill level, personal factors, and

organizational factors. Despite vast statistical analysis presented in their research,

Churchill et al. found no high correlations between sales performance and any of the

factors analyzed – and all single factors contributed to less than 10% variance in sales

performance. Therefore, Churchill et al.’s (1985) research seemed to indicate that no

generalizable common factors exist which could be used to predict sales performance

across industries. The reason for that, almost nullifying, conclusion may lie in the

methodology of meta-research itself: as authors noted: “determinants of sales

performance are job-specific” [and] “hidden company studies may be more positive,” and

as such, most fruitful research would be within a specific industry, or environment

(Churchill et al., 1985, p. 117). Other weak points (noted, but affected the research

nonetheless) missing were a coherent construct of the performance itself, and the

apparent volatility of motivation (Churchill et al., 1985).

Later, scholars developed instruments aiming to measure more stable levels of job

motivation. For example, Gagné et al. (2015) developed a cross-culturally valid 19-item

Multidimensional Work Motivation Scale (MWMS) to measure individuals’ perceived

level of work motivation. However, as Gagné et al. (2015) noted, multiple factors affect

an individuals’ level of motivation, and the level of motivation varies depending on both

intra-organizational and personal situations. Because of these variations, the level of

motivation measure lacks suitability for predicting future sales performance; especially

considering that individuals’ level of motivation during recruitment process may

significantly differ from the subsequent level of motivation after period of working time

(Chemolli & Gagné, 2014; Van Iddekinge, Aguinis, Mackey, & DeOrtentiis, 2018).

An example of this lack of predictive validity from recruitment process to job

contexts can be observed with the study of Bodla and Naaem (2014), who concluded that

among employees (n = 688) of fast-moving consumer goods companies, creative

performance (CP) fully mediated the relationship between intrinsic motivation (IS) and

sales performance (SP) (IS -> CP, β = .54, p < .01; CP -> SP, β = .52, p < .01). Bodla and

Naaem’s instrument for measuring intrinsic motivation was a six-item Likert scale

survey, but the questions, such as “I wish I didn't have to retire someday so I could

always continue selling for the pleasure of it” cannot be valid if a person does not have

prior expertise from sales jobs. Also, responses to question “Becoming successful in sales

is something that I want do for me.” could be significantly different if a person is just

starting the career in sales function compared to a person with 10 years of expertise in

sales, even though the instrument should measure the same latent construct.

Trait theories. Trait theories pose an interesting possibility to study correlates of

sales performance. Since Allport’s (1961) groundwork scientific study of personality,

multiple scholars have explored identifiable personality traits and their relationship to

success in academic, military, and job contexts (Wihler et al., 2017). Scholars identified

several stable personality traits using lexical analysis, and later, with factor analysis

techniques (Plouffe, 2018; Wihler et al., 2017). The five-factor model (or Big Five) is one

of the most studied personality trait structure in recent history (Widiger, 2015). Five-

factors (openness to experience, conscientiousness, extraversion, agreeableness, and

neuroticism) seem to be relatively stable over individuals’ adult life (Cobb-Clark &

Schurer, 2012). Some researchers concluded that two of the five factors,

conscientiousness and extraversion, link to high sales performance (Hamstra et al., 2015;

Wihler et al., 2017). For instance, Hamstra et al. (2015) examined the relationship

between Big Five factors, regulatory focus, and sales performance among salespeople (n

= 80) of Dutch event organizer operating in Greece. Using hierarchical multiple

regression analysis, Hamstra et al. concluded that Big Five personality traits in total

explain 38% of variance in sales performance, and that three of the five factors had

significant relationship with sales performance (extraversion-sales performance, β = .44,

p < .01; conscientiousness-sales performance, β = .34, p < .01; agreeableness-sales

performance β = -.28, p < .01).

Mahlamäki et al. (2018) obtained similar results when they examined the

relationship between Big-Five factors and sales performance among business-to-business

key account managers (n = 180) of Finnish companies. Differing from Hamstra et al.’s

results, Mahlamäki et al. noted that two other constructs fully mediated the relationship

between extraversion, conscientiousness, and sales performance (learning orientation,

and performance orientation) and there was no significant direct relationship between any

of the five factors and sales performance. Mahlamäki et al. also noted that the

performance orientation mediated an additional relationship between agreeableness and

job performance (agreeableness -> performance orientation β = .21, p < .01; performance

orientation -> sales performance β = .18, p < .01), and that the relationship was positive,

instead of negative as in Hamstra et al.’s results.

Frieder, Wang, and Oh (2018) noted the relationship between some personality

factors (as indicated by the Five-Factor IPIP scale) and sales performance. They

examined the relationship between personality traits, leadership style, perceived

meaningfulness, and sales performance among sales representatives (n = 496) of an

educational services company in South Korea (Frieder et al., 2018). Frieder et al.

concluded that two of the Big-Five factors had significant relationship with sales

performance (conscientiousness-sales performance, β = .22, p < .05; openness to

experience-sales performance β = .08, p < .05) but extraversion relationship with sales

performance was non-significant. Unfortunately, Frieder et al. did not measure

agreeableness nor neuroticism, and this omittance of factors limits the comparison to

other studies.

In addition to personality traits, researchers analyzed if individuals possess other

traits that correlate with success in life. One of the most studied, yet controversial traits,

is the general mental ability (GMA), or intelligence quotient (IQ), which also seem to be

relatively stable over individuals’ adult life (Kirkpatrick & Dahlquist, 2007; Lyons et al.,

2017; Rönnlund, Sundström, & Nilsson, 2015). Many scholars concluded that high GMA

predicts success in academia, military, and job contexts and is even linked to lower

mortality (Čukić, Brett, Calvin, Batty, & Deary, 2017; Gottfredson, 2002; Joseph et al.,

2015). General mental ability is also one of the few traits analyzed longitudinally over

extended time. For example, Lewis Terman (1877-1956) from Stanford University began

a series of studies among high-IQ children in 1921, and subsequent scholars have

continued the series for over 75 years (as cited in Beauvais, 2016).

Similarly, governmental researchers in Scotland began nationwide IQ tests for 11-

year old students in 1947, and since then, over 100 researchers examined the study

(SMS1947) data with subsequent measures, and additional correlates (Čukić et al., 2017).

However, Richardson and Norgate (2015) concluded that most meta-analytic studies

covering the relationship between GMA and job performance (let alone any other type of

success in life) employ so different instruments that the reliability of conclusions is very

low. Also, the effect of GMA is not always positive with sales jobs, especially when the

emotional intelligence skills of salespeople are deficient (Truninger, Fernández-i-Marín,

Batista-Foguet, Boyatzis, & Serlavós, 2018; Verbeke, Belschak, Bakker, & Dietz, 2008).

Verbeke et al. (2008) conducted two studies among salespeople (n = 171, 107) of

Dutch companies to examine the assumed relationship with high GMA (general mental

ability) and sales performance. Their research generalized modern sales role as a

knowledge-brokering role, and although research repeatedly included reference to GMA

as hardware and thinking styles as software, they did not cover changing role of sales

when the Internet serves the knowledge brokering role (Verbeke et al., 2008). This

generalized assumption of sales as a knowledge-brokering role might have basis in the

same authors’ later meta-analytic study of sales performance predictors, which partially

included the same primary data as for their 2008 article (Verbeke et al., 2011).

Despite a narrow approach to a sales role, researchers did find a significant

relationship between high GMA and high sales performance (Verbeke et al., 2008). This

effect was curvilinear and was moderated with salespeople’s social competence so that

high GMA combined with high social skills resulted in best sales performance, and high

GMA combined with low social skills resulted in lowest sales performance - “competent

jerks,” as Verbeke et al. (2008, p. 50) described. The empiric studies included in Verbeke

et al.’s research had some limitations, namely Study 1 did not assess profitability, but

revenue only – and Study 2 used managerial assessment of salespeople’s performance,

which is subject to managerial bias, as noted in some previous research (Lilly, Porter, &

Meo, 2003; Schoorman, 1988).

Sales Performance Research

Walker et al. (1977) and their colleagues begun initial systematic research of

salesmen personality factors (and, more generally, a wide number of antecedents)

influencing the sales performance in 1977. Walker et al. concluded that then-current sales

performance research was practically non-existent, and that each sales executive relied on

their own expertise on the success factors. To alleviate the situation, Walker et al. created

a framework of determinants of performance of salespeople. Their initial framework

consisted of three salespeople related factors: personal, organizational, and

environmental, which affect sales performance via motivation, aptitude, and role

perceptions of the salespeople (Walker et al., 1977). The research also indicated that

while achieving high sales performance was critical for the corporations, academia still

relegated the study of this field to second-class status (Walker et al., 1977). Thus,

recommendations included the need for future scholars to test their hypothesized

framework and to search for generalizable predictors for sales performance (Walker et

al., 1977).

Continuing the work of Walker et al. (1977) on the search for generalizable

predictors for sales performance, Churchill et al. (1985) conducted an extensive meta-

analysis of sales performance determinants by reviewing 116 published and unpublished

studies dated between 1918 and 1982. Based on their meta-analytical approach, Churchill

et al. (1985) re-categorized the factors influencing sales performance to six broad

categories: personal factors (total T = .043), skills (total T = .037), role (total T = .02),

aptitude (total T = .018), motivation (total T = .017), and organizational factors (total T =

.01). However, Churchill et al. noted that the lack of a consistent method for measuring

sales performance or the predictor variables was a significant limitation of their study.

Therefore Churchill et al. called for future scholars to develop and use standardized

measures in analyzing sales performance predictors.

Krishnan et al. (2002) responded to this call for research and examined the

relationship between self-efficacy, competitiveness, salesperson effort, and self-reported

sales performance (Krishnan et al., 2002). Their research consisted of two quantitative

inquiries amongst two US companies and 273 salespeople, concluding that salesperson

effort partially mediated the relationship between salesperson self-efficacy, and sales

performance and that salesperson effort fully mediated the relationship between

salesperson competitiveness, and sales performance (Krishnan et al., 2002). Krishnan et

al. (2002) hypothesized the causality of events using Vroom’s expectancy theory and

Maslow’s hierarchy of needs theory (as cited in Maslow, 1943; Vroom, 1995).

The latest, and simultaneously, the most advanced current meta-analytic research

of sales-performance antecedents is the work by Verbeke, Dietz, and Verwaal (2011).

They used the original antecedent factor classification by Walker et al. (1977) and

Churchill et al. (1985) and utilized the most current body of knowledge from

motivational and psychological theories to construct a multidimensional model of

generalizable antecedents for high-performance sales (Verbeke et al., 2011). Based on

this research, high self-efficacy of salespeople has been predominantly associated with

high sales performance (Verbeke et al., 2011). Since similar findings have emerged from

significantly differing culture and business context by Yang et al. (2011), the findings

may apply in multiple business contexts and should be further researched with companies

willing to advance their sales performance (GLOBE Foundation, 2007; Panagopoulos et

al., 2011).

Job Performance in Sales Contexts

Although multiple scholars researched the determinants of performance of

salespeople, the whole concept of performance within job contexts is worth further

examination. Combining motivational theories with organizational research, Campbell et

al. (1993) created a job performance theory in 1993. Campbell et al. (1993) theory of job

performance defined performance as a sum of all behaviors an individual engages within

a job. Campbell et al. also defined eight dimensions by which the scholars and

practitioners should measure performance in job contexts: job-specific task proficiency,

maintaining personal discipline, demonstrating effort, facilitating peer and team

performance, non-job-specific task proficiency, communication task proficiency,

supervision, and management. Also, Campbell et al. defined three common antecedents

for individuals’ job performance: declarative knowledge, procedural knowledge and

skills, and motivation. Campbell et al.’s theory did not include instruments to measure

any of the dimensions of job performance nor the antecedents of the performance.

Within sales contexts, this lack of universal instrument to measure job

performance remains prevalent in current research (Johnson & Jaramillo, 2017).

According to three recent meta-analytic studies, job performance in sales contexts is

typically measured either by supervisory rating, self-rating, revenue generated, or by

quota-attainment (Johnson & Jaramillo, 2017; Joseph et al., 2015; Plouffe, 2018). Quotas

of salespeople typically consist of a composite target level of sales based on revenue, and

profit, or a combination of these (Fu, 2015). Salespeople within the IT industry typically

receive quotas for three, six, or 12 months at a time (Fu, 2015; Tuggle, 2014). Because

salespeople are responsible for various sizes of customers, the sizes of quotas are

different between the salespeople (Benson, 2015; Kräkel & Schöttner, 2016).

Each of these sales performance measures has some advantages and

disadvantages: For example, supervisory ratings are prone to biases relating to the

relationship between the supervisor and the salespeople (Lilly et al., 2003). Self-report

ratings are prone to social desirability bias and imbue the risk of subjective comparability

between participants (Bellizzi & Bristol, 2005). Measures using absolute revenue (or

margin) have an inherent assumption of equal opportunity between the participants (i.e.,

there is measurement difference caused by working hours, set of customers, or other

conditions that the salespeople cannot directly influence; Carter et al., 2008). This lack of

participant equivalence in measuring absolute sales volumes may also cause type II errors

in research.

For instance, Osborne (2015) examined the relationship between emotional

intelligence, cognitive intelligence, personality traits, and sales performance among

salespeople (n = 35) of U.S. based outdoor advertising company. He concluded that none

of the examined independent variables had a significant relationship with neither sales

revenue (P1) or a number of sales contracts (P2) (Osborne, 2015). However, closer

examination of Osborne’s (2015) data reveal that both performance constructs had very

high standard deviation compared to mean (P1 Mean = $1,112,318, SD = $875,262; P2

Mean = 142.40, SD = 96.62). High deviation indicates that the spread of absolute

performance between salespeople is significantly higher than any trait theory would

predict, and thus variance in performance is more likely a result of different customer set

(or other significant differentiating factors between the salespeople) (Osborne, 2015).

Studies employing quota-setting as a measure of sales performance are prone to

quota-setting bias (which is often a result of managerial bias; Lilly et al., 2003). Although

no universally accepted method exists to measure job performance in sales contexts,

researchers of sales performance still need to choose an appropriate measure. The

participants for this research are face-to-face salespeople working with THSP IT

companies. The quota-attainment directly affects each participants’ personal salary on a

quarterly and annual basis through annual merit increases (Martin, 2013). The THSP

companies use annual quota-attainment measure for employee appraisals (Quan et al.,

2010). Therefore, this study used an annual quota-attainment percentage to measure sales

performance.

Instruments to Measure General Self-Efficacy

Since Sherer et al. (1982) coined the general self-efficacy construct, multiple

scholars contributed to the development of instruments to measure general self-efficacy

(Barahona et al., 2018). These instruments are similar in terms of administration

(participants answer to multiple questions based on their own perception of the item) and

response format (participants answer to questions using Likert-type scale). The

differences between the instruments lie in question-wording, number of questions, and in

scale granularity. As Schwarzer and Jerusalem (1995) noted, the instrument question

wording significantly affects the generalizability of the measure over cultural and lingual

barriers. The number of questions in the survey instrument is important because the aim

of the instrument is to measure one construct (general self-efficacy), and additional

questions may capture elements of other latent psychological constructs (Barahona et al.,

2018; Schwarzer & Jerusalem, 1995).

Similarly, too few questions in the instrument may result in an insufficient or

inadequate measure of the whole construct (Barahona et al., 2018). Contrarily, the high

number of questions may decrease survey response rates (Allen, 2016). Scale granularity

refers to the number of possible answer items for each survey question (Cai, Lin, &

Zhang, 2016). Bipolar scales have a neutral center point, whereas unipolar scales start

from zero value (Cabooter, Weijters, Geuens, & Vermeir, 2016).

Sherer General Self-Efficacy Scale (SGSE), 1982. Sherer et al. (1982)

conducted two studies to construct an instrument to measure individuals’ level of general

self-efficacy, one among university students (n = 376) and one among patients (n = 150)

from the Tuscaloosa Veterans Administration Medical Center who were in the

alcoholism treatment unit. In the first study, research included an initial scale of 36 items

with a factor analysis using a scree test with the varimax method (Sherer et al., 1982).

The resulting analysis confirmed two dimensions of self-efficacy, which Sherer et al.

named as general self-efficacy (α = .86) and social self-efficacy (α = .71). Sherer et al.

also discarded items that had a factor loading less than .40 or more than .40 but for

multiple factors. The resulting scale consists of 23 survey items (17 to measure general

self-efficacy and 6 to measure social self-efficacy) with a 14-point bipolar Likert-type

scale. In the second study, Sherer et al. used a new scale to measure the relationship

between the level of general self-efficacy and (a) status of employment, (b) a number of

jobs quit, (c) number of times fired, (d) educational level (measured by highest achieved

education), (e) military rank. Conclusion included that general self-efficacy has

significant relationship with each item: A (β = .278, p < .05), B (β = -.240, p < .05), C (β

= -.226, p < .01), D (β = .268, p < .05), E (β = .218, p < .05). Although the Sherer et al.’s

effort in conducting statistical analysis of the SGSE scale was substantial, the initial

reasoning for each item was very limited. In specific, Sherer et al. did not disclose items

in the original scale (36 items), and they did not establish nomological validity for the

new scale (SGSE) either. For example, SGSE item 9 “When I decide to do something, I

go right to work on it” can be conceived as a tendency for lack of preparation. Second, as

Sherer et al. used a unique construct of performance (status of employment, number of

jobs quit, number of times fired, educational level and military rank among patients

treated for alcoholism), the generalizability of findings to business context may be

limited (Carter et al., 2008).

General Self-Efficacy Scale (GSES), 1995. Acknowledging the importance of a

new psychometric construct of general self-efficacy, Schwarzer and Jerusalem (1995)

became interested in cross-cultural validity of the construct. They integrated cross-

cultural understanding with Bandura’s theoretic model and developed a new scale to

measure the general self-efficacy of a person (GSES; Schwarzer & Jerusalem, 1995).

Schwarzer and Jerusalem originally presented a 20 item scale in 1979 but reduced it to 10

items (with a 4-point unipolar Likert-type scale for each) in 1993 (Schwarzer &

Jerusalem, 1995). Because of the ease of administration, multiple available translations,

and cross-cultural validation studies, Schwarzer’s and Jerusalem’s (1995) scale is the

most used instrument to measure general self-efficacy and it has been used in over 1000

studies since introduction (Barahona et al., 2018; Luszczynska et al., 2005). Schwarzer’s

instrument to measure general self-efficacy is publicly available, published in the largest

available database consisting of GSES measures from over 18000 individuals from 25

countries (Schwarzer, 2006). According to that data, median scores of general self-

efficacy (measured with GSES) are within 1% in all age groups between 20 and 70 years

old adults, and that age is significant, but very weak predictor for general self-efficacy (n

= 6220, F = 7.81, r2 = .001, p < .01) (Schwarzer, 2006).

New General Self-Efficacy Scale (NGSE), 2001. Chen et al. (2001) noted that

then-current literature prominently used generalized self-efficacy scale (SGSE) by Sherer

et al. (1982). SGSE is a 17-item scale where participants respond to each item using a 14-

point Likert scale (from strongly disagree to strongly agree). According to Chen et al.

(2001), SGSE has high internal consistency (α = .76 to .89) and high predictive validity,

but it captures other constructs and sometimes negatively correlates with situational self-

efficacy (SSE). Because of the weaknesses of SGSE, Chen et al. (2001) developed a new

general self-efficacy scale (NGSE) to measure generalized self-efficacy.

First, Chen et al. (2001) combined their earlier scale (GSE) with SGSE and

eliminated identical items. The resulting scale consisted of 14 items with five-point

bipolar Likert-type scales. As a second step, the Chen et al. took the new scale as a

starting point and removed six scale items, which linearly correlated with other items of

scale. To ensure content validity, they asked independent panels to analyze the scale

items of three different scales (SGSE [17 items], Rosenberg Self-Esteem Scale [10 items]

and new scale [8 items]) and arranged the items to three constructs (generalized self-

efficacy, self-esteem or other; Chen et al., 2001). The panel results confirmed the

discriminant and content validity of GSE and indicated that NGSE is substantially more

consistent with GSE compared to SGSE (Chen et al., 2001). Subsequent testing (3

studies, n = 316, 323 and 54) indicated that new 8-item model has high internal validity

(α = .87, .88, and .85) (Chen et al., 2001).

Chen et al. (2001) also noted that in their three studies, the SGSE scale yielded

three dimensions with eigenvalues greater than 1, while the NGSE scale resulted in only

one dimension indicating one latent construct. Also, as Chen et al. noted, the NGSE scale

remains more appealing for organizational studies, since several scholars have concluded

that a reduced number of scale items increase response rates (Sheehan, 2006). Although

Chen et al. conducted three surveys confirming the results of the hypothesized new scale,

the context of surveys was limited to the academic world. In specific, they used mid-term

exam results as a construct of performance in studies 1 and 2 (Chen et al., 2001). Chen et

al.’s third study compared the content-related validity, reliability, dimensionality, and

predictive validity of the Hebrew versions of the NGSE scale and the SGSE scale among

Israeli managers.

Although past literature indicates earlier SGSE, and GSES measures are culture

and context-independent instruments to measure generalized self-efficacy (Luszczynska

et al., 2005; Sherer et al., 1982), research did not include compelling argumentation for

similar applicability for new NGSE instrument (Chen et al., 2001). Second, as Chen et al.

used mid-term exam results as a construct of performance, the generalizability of findings

to business context may be limited.

Interested in potential differences of instruments measuring general self-efficacy,

Scherbaum et al. (2006) analyzed all three general self-efficacy instruments (SGSE,

GSES, and NGSE) using Samejima’s graded response model (GRM). They conducted a

cross-sectional survey among students (n = 606) at a large northeastern university and

collected all three measures from each participant (Scherbaum et al., 2006). Scherbaum et

al. concluded that each measure demonstrated satisfactory internal consistency, and all

measures were positively correlated. After comparing instruments with test information

function (TIF), they concluded that NGSE is a preferred choice for testing general self-

efficacy because it is shorter and results in nearly the same information as the longer

measures (Scherbaum et al., 2006). Because Scherbaum et al. analyzed only one sample,

and the population consisted of students from a large northeastern university, the

generalizability of their findings remain limited, especially to other cultures or business

contexts.

General Self-Efficacy Criticism

Self-efficacy, general self-efficacy and their relationship to other variables such as

job performance, have become one of the most widely studied variables in the

educational, psychological, and organizational sciences (Scherbaum et al., 2006).

Bandura’s original theory of self-efficacy described the construct as domain-specific, and

the shortest instruments to measure self-efficacy contain only one survey item (Pajares &

Urdan, 2006). More generalizable instruments, such as SGSE, attracted scholars to

research, validate, and criticize the instruments (Luszczynska et al., 2005).

Bandura (2012) critically analyzed the status of self-efficacy literature in 2012

and concluded that bipolar scales with a neutral center and most trait-like constructs to

measure self-efficacy are fundamentally non-scientific. In specific, Bandura argued that

one’s perceived ability to complete some task could never be neutral. Also, in Bandura’s

view, self-efficacy, without the surrounding social cognitive theory, is not sufficient

measure in analyzing human behavior. Also noted were several other flaws in earlier

experimental studies, such as deficient assessment of self-efficacy, misguided goal

setting, and defective control measures (Bandura, 2012). However, Bandura did not

completely disregard the notion of general self-efficacy, but they warned that scholars

should analyze such a construct within the social cognitive theory framework, and not

separately.

The latest criticism for general self-efficacy (as described by Schwarzer and

Jerusalem (1995)) was provided by Barahona et al. (2018), who performed secondary

analysis of Schwarzer’s (2006) general efficacy data which consists GSES measures of

19719 individuals from 26 countries. Barahona et al. used seven statistical tests to study

latent factors of GSES: factor analysis (FA), principal component analysis (PCA), Sparse

PCA, Dual Statis, Item Response Theory (IRT), differential item functioning (DIF), and

finally, Multiple Group Confirmatory Factor Analysis (MGCFA). According to Barahona

et al. test results, GSES represents perceived general self-efficacy with an explained

variability of 43.7%. However, the GSES item distribution pattern differences between

the countries indicate that GSES does not measure a universal, nor one-dimensional

construct (Barahona et al., 2018). However, Barahona et al. did not discuss the limitations

of their research arising from the differences of data sources in Schwarzer’s (2006)

original data. Schwarzer’s original data consists of measures from multiple studies, and

each country data contains a different set of demographic groups. Therefore, Barahona et

al. statistical analysis aiming to find factorial differences between the countries, will

logically exhibit the differences between sample groups. For example, Barahona et al.’s

notion of arbitrarily high GSES among the German group of respondents is not a

limitation of the construct, but the effect of sample (n = 106) respondents from the

German army having statistically higher GSES measures than mean demographic groups

from other countries.

Although the previously discussed original data sampling method may have

caused Barahona et al. (2018) findings revealing multiple latent constructs within GSES

measures, some similar findings emerged from a study by Zhou (2016). Interested in the

validity and factorial construct of Schwarzer’s general self-efficacy scale (GSES), Zhou

(2016) wanted to explore if the scale measures a unidimensional construct among

university students in China. Recruitment included university students (n = 185) from

three randomly selected Chinese universities to participate in the study (Zhou, 2016). To

analyze possibly overlapping constructs, Zhou also measured responses to the Nario-

Redmond scale (8 items) of individualism and Lai’s life-orientation test. Zhou then tested

all scale results for normality and then performed exploratory factor analysis measuring

the underlying factors of the GSES scale. Zhou then compared six different models with

CFA and concluded that the second-order factor model with one correlated error was the

most appropriate explaining the latent factors of general self-efficacy.

This model explained 54.5% of variance of general self-efficacy and that first

factor also predicted optimism (β = 0.18, p < .05) (Zhou, 2016). Although Zhou’s (2016)

factor analysis indicated the presence of a hierarchical factorial model (first-order coping

self-efficacy and action self-efficacy constituting to second-order general self-efficacy),

their results were logical and expected. As Schwarzer and Jerusalem (1995) noted, the

general self-efficacy scale includes design to assess an individual’s optimistic self-beliefs

in coping with difficulties in life, and the definition does not rule out additional sub-

factors.

Similarly, as Zhou (2016), Nell and Boshoff (2016) acknowledged that scholars

frequently use Sherer’s (1982) General Self-Efficacy Scale in clinical, personality, and

organizational research, but that the researchers have not sufficiently tested the

instrument for latent factors. To address this lack of existing research, Nell and Boshoff

(2016) examined the factor structure of Sherer’s General Self-Efficacy Scale (SGSE)

among chartered accountants (n = 295) as part of their compulsory professional exam.

Nell and Boshoff collected SGSE responses from 295 participants and then performed

exploratory factor analysis minimum average partial test and parallel analysis and

subsequently used confirmatory factor analysis to compare two measurement models

(unidimensional, CFI = .967 and three-dimensional, CFI = .969). Because both models

exhibited a similar level of fit, Nell and Boshoff used the Schmid-Leiman solution, which

indicated that SGSE is unidimensional, and the general factor explains 76% of the

variance. However, Nel’s and Boshoff’s did not fully justify decision to use the Schmid-

Leiman solution for factor rotation. More advanced (and reliable) methods, such as bi-

quartimin, bi-geomin (by Jennrich-Bentler) or direct oblimin, could have indicated

several different latent factorial structures of SGSE (Mansolf & Reise, 2016).

Relationship Between Self-Efficacy and Sales Performance

Fu et al. (2009) noted that self-set goals and level of effort mediate the

relationship between self-efficacy and sales performance in their longitudinal study of

U.S. and Canada -based salespeople (n = 802) working for global construction and

building maintenance company. The study included seemingly unrelated regression

analysis to examine the relationship between assigned goals, task-specific self-efficacy,

self-set goals, effort, and new product sales (Fu et al., 2009). Fu et al. concluded that

relationship between task-specific self-efficacy of salespeople, and new product sales

performance is non-significant (β = .043, n.s.) but that task-specific self-efficacy of

salespeople has significant relationship with self-set goals (β = .19, p < .05) and with

effort (β = .25, p < .01). Furthermore, Fu et al. noted that self-set goals were strongest

predictor for new product sales performance (stage 1, β = .34, p < .01; stage 2, β = .44, p

< .01) and that the relationship is not linear, but an inverted U-shaped curve. These

results indicated that self-set goals mediate the relationship between task-specific self-

efficacy of salespeople and subsequent sales performance, but setting the target too high

results in decreased new product sales performance (Fu et al., 2009). Although Fu et al.

used quota attainment instrument to measure new product sales (with objective measures

from target companies), their instrument to measure self-efficacy was unique to the

research, which may limit external validity.

Gupta et al. (2013) hypothesized the relationship between sales self-efficacy

(SSE) and job performance and compared the results to personality traits (Big Five)

relationship with job performance among job applicants (n = 14.666), and sales associates

(n = 479) working for nine retail stores across the United States. Of particular note was

that recruitment decisions often involve personality analysis (using Big-Five), and

therefore Gupta et al. wanted to test whether SSE has a different relationship with job

performance than Big-Five personality traits. The first study involved measuring SSE and

a portion of big-five results from the job applicants to analyze the predictive power of

measures to subsequent sales performance (Gupta et al., 2013). The second study

involved measuring SSE and past four months sales performance among sales associates

who had been working in the company for at least four months (Gupta et al., 2013).

To test the relationship between SSE, Big-five, and sales performance Gupta et al.

(2013) developed a modified instrument to test SSE and used Goldberg’s (1999) standard

five-factor scale to measure personality traits. They measured job performance with two

customized measures: sales performance (SP) (sales per hour for each employee,

averaged to monthly values) and informal appraisal (IA) (supervisors’ feedback regarding

sales associate’s performance using 17 item scale) (Gupta et al., 2013). Gupta et al. then

performed a correlation analysis between SSE results and external general self-efficacy

results and found that SSE had a significant correlation with GSE (β = .40, p < 0.01).

Gupta et al.’s (2013) examination included an analysis of the relationship between

big-five items and SSE. They found that conscientiousness, extraversion, and openness

were significantly associated with SSE (β = .28, .42, .19, p < 0.05; Gupta et al., 2013).

The analysis included a correlation analysis of the predictive study and regression

analysis of concurrent study (Gupta et al., 2013). Predictive study results indicated that

there were no significant correlations between SSE, big-five factors and subsequent sales

performance among applicants who subsequently were recruited (except the very limited

relationship between SSE (skill) with month four sales results (β = .08, p < .05; Gupta et

al., 2013)). Concurrent study results indicated that SSE had significant effect to SP (β =

.28, p < .01) and to IA (β = .16, p < .01) and out of five personality factors, only

conscientiousness had significant positive effect to SP (β = .11, p < .05) and IA (β = .12,

p < .05; Gupta et al., 2013). Gupta et al. also noted that supervisors’ rating (IA) correlated

with actual sales performance (SP), but the correlation was relatively low (β = .42, p <

.01).

Gupta et al.’s (2013) results indicated that task-specific self-efficacy and past

sales performance have a significant relationship, but that the same task-specific self-

efficacy (measured among job applicants) did not significantly predict future sales

performance. The weakness in Gupta et al.’s research lies in two separate participant

groups (employees and job applicants). According to Bandura’s (1977) theory of self-

efficacy, task-specific self-efficacy is a product of four factors: performance

accomplishments, vicarious experience, verbal persuasion, and emotional arousal.

Because job applicants did not have prior experience from the job they applied for,

participants’ SSE measure was either a result of their past unknown experience or, more

likely, the result of social desirability bias during the job application process (Gupta et

al., 2013). Gupta et al.’s findings highlight the importance of selecting appropriate

constructs when researching correlates of sales performance. Task-specific self-efficacy

(as in Gupta et al. research), by definition, is a volatile construct that researchers cannot

use for predictive analysis unless they ensure construct stability (Talsma, Schüz,

Schwarzer, & Norris, 2018). Gupta et al.’s results regarding the limited correlation

between IA and SP are similar to Benson’s (2015) and highlight discrepancy between the

objective, and managers’ subjective assessment of sales performance.

Pettijohn et al. (2014) analyzed the relationship between role-ambiguity,

autonomy, task-specific self-efficacy, and self-reported sales performance among

salespeople (n = 245) of two large US-based real estate agency companies. Pettijohn et al.

presented descriptive data as three tables, including respondent demographics, result

means, alpha-coefficients and correlations, and finally, beta-coefficient and F-values for

hypothesis confirmation. Similarly, as Gupta et al. (2013), Pettijohn et al.’s findings

indicate that task-specific self-efficacy and autonomy are predictors for high sales

performance together, explaining 17% variance in sales performance, whereas role-

ambiguity is both predictor and product of low sales performance. Pettijohn et al.’s

statistical analysis of data and argumentation for findings are strong, and there is a very

small room for alternative interpretations. However, they used self-reporting data for

assessing sales performance (Pettijohn et al., 2014). Other scholars concluded that this

method might induce an additional bias to the research as self-reporting may be affected

by evaluative bias (Ferrando & Lorenzo-Seva, 2010; Leavitt, 1977; Ray, 1990). Also,

Pettijohn et al. did not discuss the limitations of their research (and thus did not list

performance construct as a limitation).

Lu et al. (2016) analyzed the relationship between general self-efficacy, challenge

stressors, hindrance stressors, and job performance among salespeople (n = 164) of a

Chinese insurance company. They sent a questionnaire to 215 participants and received

164 valid responses (76.3% response rate; Lu et al., 2016). Compared to a similar study

by Pousa and Matthieu (2016), the response rate in Lu et al.’s study was exceptionally

high. High response rate, together with performance self-assessment, may indicate the

risk of acquiescence bias (Ray, 1990).

Lu et al. (2016) used three well-established instruments to measure challenge-

hindrance stressors, job performance, and general self-efficacy (GSES). They concluded

that general self-efficacy significantly moderated the relationship between challenge

stressors and job performance (β = .14, p < .05), and moderating effect between

hindrance stressors and job performance was nonsignificant (β = .-08, p = n.s.; Lu et al.,

2016). Although the study provides rigorous statistical analysis which supports the

hypothesized effect, the study has some weaknesses (Lu et al., 2016). First, Lu et al.

measured sales performance using two adapted instruments (self-assessment and

managerial assessment). Although this adaptation may be suitable in a single study, it

limits the generalizability of findings, especially because Lu et al. did not disclose the

final instrument. Also, multiple researchers concluded that self-reported sales

performance often includes bias – even when the data is supplemented with managerial

assessment (Jaramillo, Carrillat, & Locander, 2003; Kaplan, Petersen, & Samuels, 2018;

Lilly et al., 2003). Also, the study was conducted in China and one company (Lu et al.,

2016). Therefore cultural differences such as variance in uncertainty avoidance and

power-distance may limit the generalizability of findings across different cultures and

contexts.

Interested in the effect of self-efficacy on sales performance within insurance

sales, Cheng and Chiou (2016) wanted to research if self-efficacy is associated with

increased sales performance within the Taiwanese insurance sales sector while taking

into account the psychological aspect of positive illusion. To test the hypothesis, Cheng

and Chiou conducted three repeated online surveys among insurance salespeople (n =

160) of two Taiwanese insurance companies. Cheng and Chiou received a 94% response

rate (151 usable responses), which is exceptionally high and could indicate forced

answering (Hammer, 2017).

Noteworthy was Cheng’s and Chiou’s (2016) use of the GSES to measure general

self-efficacy and custom instruments to measure positive illusion (the difference between

salespeople’s expectancy of future sales performance, and actual sales performance), and

sales performance. Cheng and Chiou performed a correlation analysis between general

self-efficacy, positive illusion, and sales performance. They also conducted ANOVA

between high- and low-self efficacy groups. Similarly as Gupta et al. (2013), Cheng and

Chiou concluded that general self-efficacy of the salespeople has significant relationship

with sales performance at each survey interval (1: β = .66, p < .01; 2: 1: β = .74, p < .01;

3: β = .66, p < .01) and that group of salespeople whose general self-efficacy was above

the average had significantly higher sales performance than a group that had lower than

average general self-efficacy (t(149) = 5.90, p < .01).

Cheng and Chiou (2016) developed two unique instruments to measure constructs

in their study: they measured positive illusion using delta value between sales-persons

own expectation of future performance and subsequent achievement. They also measured

sales performance using monthly commission, which is a product of multiple factors,

including individual target setting (Cheng & Chiou, 2016). Because of instrument

uniqueness, the generalizability of Cheng’s and Chiou’s study may be limited (Benson,

2015). Finally, Cheng and Chiou conducted the study in one industry and one

geographical location. Therefore, bias may arise from cultural norms (such as

performance orientation and uncertainty avoidance, as defined by Globe research (2007)),

which affect or limit the generalizability of findings to other cultures or contexts.

Singh et al. (2017) reported similar, although stronger then Cheng and Chiou

(2016), relationship between general self-efficacy, and sales performance in their study

among salespeople (n = 297) of pharmaceutical companies in Asian countries measuring

the relationship between self-efficacy (NGSE), thought self-leadership (TSL), selling

skills (SS), adaptive selling (AS), and sales performance (SP). Interested in the effect of

TSL to JP, Singh et al. hypothesized a model where TSL has a relationship with JP,

mediated by SE, AS, and SS. To test the hypothesis, Singh et al. conducted an online

survey among salespeople of pharmaceutical companies in Asian countries and received

297 usable responses. They then performed correlation analysis and structural equation

modeling (SEM) statistical analysis to measure the relationship between the variables.

SEM indicated three viable, but contradicting effect pathways with similar comparative

fit index, and root mean square error of approximation (Singh et al., 2017). Singh et al.

concluded that in full structural model, TSL significantly predicted SE (β = .61, p < .01),

and that relationship between SE and SP was significant (β = .52, p < .01), and partially

mediated by AS and SS (Singh et al., 2017). Singh et al. used only one data source, which

might have increased the risk of common method bias and acquiescence bias (Ylitalo,

2009). Also, Singh et al. did not discuss the potential implications of either risk. Singh et

al.’s research contain two additional limitations arising from using a shortened instrument

(3-item variation of NGSE) to measure general self-efficacy, and from using a self-report

measure of sales performance. The shortened instrument may not capture the whole

construct of general self-efficacy, and a self-report measure of sales performance is prone

to bias (Jaramillo et al., 2003; Zhou, 2016).

Interested in the antecedents of performance of the salespeople, Carter et al.

(2016) examined extant literature covering individual factors predicting high job

performance. They also conducted longitudinal correlational analysis among employees

(n = 64) of Australian financial services organization measuring the relationship between

task-specific self-efficacy (TSSE), employee engagement (EE), and job performance (JP)

(Carter et al., 2016). Guided by prior studies examining correlates of sales performance,

they selected two factors (employee engagement, EE, and task-specific self-efficacy,

TSSE) for predictor variables (Carter et al., 2016). Their data analysis indicated that both,

TSSE and EE have significant relationship JP (TSSE-JP β = .54, p < .01, EE-JP β = .53, p

< .01) (Carter et al., 2016). As Carter et al. did not test the reliability of a custom

instrument for measuring TSSE, the generalizability of the study remains limited (Gupta

et al., 2013; Mullinix et al., 2015).

Pousa and Mathieu (2016) analyzed the relationship between supervisory

coaching, task-specific self-efficacy, and performance of the salespeople among financial

services institution salespeople (n = 133) in Canada. They received 121 usable responses

(32.7% response rate), which is typical with surveys among salespeople (Pedersen &

Nielsen, 2016). In their survey, Pousa and Mathieu measured supervisory coaching with

Ellinger’s scale, task-specific self-efficacy of the salespeople using Sujan’s scale, and

performance of the salespeople using Fang’s scale. Pousa and Mathieu then verified the

effect path by using structural equation modeling and concluded that supervisory

coaching increased task-specific self-efficacy of the salespeople (β = .45, p < .01), and

task-specific self-efficacy fully mediated the effect of supervisory coaching to sales

performance. Similarly, as with Gupta et al. (2013), Pousa and Mathieu concluded that

task-specific self-efficacy of the salespeople had a significant relationship with behavior

performance (β = .51, p < .01), and with sales performance (β = .68, p < .01).

Relationship Between Self-Efficacy and Job Performance

Multiple scholars observed a significant relationship between employees’ level of

self-efficacy and job performance. For example, Tims et al. (2014) concluded that self-

efficacy increases job performance both directly, and indirectly via an increased level of

crafting variety, crafting opportunities for development, and work enjoyment. In their

study, Tims et al. hypothesized mechanism of self-efficacy to job performance using Job

Demands-Resources (JD-R) theory by Bakker and Demerouti (2014). Based on the

theory and review of prior literature, Tims et al. presented five hypotheses: H1. Day-level

self-efficacy has a positive relationship with day-level performance, H2. Day-level self-

efficacy is positively associated with day-level job crafting, H3. Day-level job crafting

mediates the relationship between day-level self-efficacy and day-level work enjoyment,

H4. Day-level work enjoyment mediates the relationship between day-level job crafting

and day-level performance, and H5. Day-level self-efficacy is positively related to day-

level performance via day-level job crafting and work enjoyment. Tims et al. study

hypotheses highlight their view of the volatility of the self-efficacy and it significantly

differs from the stable nature of general self-efficacy (Waaktaar & Torgersen, 2013).

Because of this difference, some scholars have concluded that the two facets of self-

efficacy (general and specific) are not comparable measures (Grether et al., 2018).

Tims et al. (2014) conducted a longitudinal study among volunteered employees

(n = 47) from small companies operating within the IT sector. Each participant responded

to the same set of questions for five consecutive days yielding a total sample size of 215

(Tims et al., 2014). Tims et al. used adapted instruments to measure self-efficacy

(Schwarzer & Jerusalem, 1995), day-level job crafting (Tims, Bakker, & Derks, 2012),

day-level work enjoyment (Bakker, 2008) and day-level job performance (Williams &

Anderson, 1991). Tims et al. used multilevel structural equation modeling statistical

analysis to measure the effect of each variable to another and presented path analysis

indicating the effect sizes. They concluded that the relationship between self-efficacy and

job performance was significant (β = .21, p < 0.01) (Tims et al., 2014).

Lisbona et al. (2018) obtained similar results as Tims et al. (2014) in

acknowledging the importance of work-engagement (WE), and self-efficacy (SE) on the

personal initiative (PI), and self-reported job performance (PE). Lisbona et al. wanted to

examine the relationship between WE, SE, PI, and PE in various organizations in Spain

and Mexico, and they conducted two independent studies: one cross-sectional and one

repeated measures longitudinal study. They first identified possible performance

antecedents constructs based on extant literature and chose the instruments to measure

each construct. Lisbona et al. then conducted both studies and measured SE, WE, PI, and

PE using surveys. They then performed correlation analysis and structural equation

modeling (SEM) statistical analysis to measure the relationship between the variables.

Cross-sectional study SEM indicated three significant effects between the variables (WE-

PI, β = .42, p < .01; SE-PI, β = .60, p < .01; PI-PE, β = .13, p < .01) and longitudinal

study SEM indicated six significant effects between the variables (T1 SE- T1 PI, β = .91,

p < .05; T1 PI - T1 PE, β = .67, p < .05; T1 SE – T2 SE, β = .60, p < .05; T2 SE – T2 PI,

β = .64, p < .05; T2 PI – T2 PE, β = .45, p < .05; T1 PE – T2 PE, β = .49, p < .001)

(Lisbona et al., 2018).

Lisbona et al. (2018) concluded that WE and SE lead to higher PI, which, in turn,

leads to higher PE. However, they used only one data source, which might have increased

the risk of common method bias and acquiescence bias (Ylitalo, 2009). Also, Lisbona et

al. used custom instruments to measure all constructs and did not disclose the final

instruments. Although their verbal conclusions of the study are logical, they did not

discuss the implications of SEM indicating five different models with a similar fit

(Lisbona et al., 2018). This limitation is significant since Lisbona et al.’s proposed model

(model 5) has the lowest CFI score of all models and omits the analysis of the inverse

relationship between performance and self-efficacy. Prior research has indicated that

mastery experiences (achieving high performance at work) increase contextual self-

efficacy (Talsma et al., 2018). Future studies should use generalizable instruments to

measure all constructs and analyze alternative factorial models in explaining the role of

SE, WE, and PI on performance.

Similarly, as Tims et al. (2014), Miraglia et al. (2017) concluded that job-crafting

is a mediator for the relationship between self-efficacy and job performance. Miraglia et

al. (2017) conducted repeated measures longitudinal analysis among white-collar workers

(n = 465) of one large Italian service organization measuring the relationship between

self-efficacy (SE), job crafting (JC), and job performance (JP). Miraglia et al. conducted

two online surveys to measure SE and JC and obtained JP data from a company HR

(Miraglia et al., 2017). They then performed correlation analysis and structural equation

modeling (SEM) statistical analysis to measure the relationship between the variables

(Miraglia et al., 2017). SEM indicated nine significant effects between the variables in

measure points one and two (Miraglia et al., 2017). Miraglia et al. concluded that JC fully

mediated the relationship between SE (JC-SE β = .74, p < .01) and JP (JC-JP β = .11, p <

.01) and that the effect persisted over time. However, SE's direct relationship with JP was

non-significant at both measure points, and JP at measure point one did not predict SE at

measure point two (Miraglia et al., 2017). Miraglia et al. also noted that employee age

and tenure had a significant negative relationship with SE, JC, and JP at both measure

points.

Noting the potential mediating effect, such as Miraglia et al. (2017) found, and

interested in the direct and indirect effect of self-efficacy on job-performance, De Clercq

et al. (2018) proposed a hypothetical path effect in which job-related anxiety mediates the

effect between self-efficacy and job-performance and perceived workplace incivility

moderates the relationship between self-efficacy and job-related anxiety. To test the

hypotheses, De Clercq et al. conducted a survey among 1000 employees of Pakistani

organizations and received 454 usable completed responses. Using correlation analysis

and multiple regression analysis, they concluded that employees’ self-efficacy had a

significant direct relationship with job performance (β = .346, p < .001) and with job-

related anxiety (β = -.095, p < .05) (De Clercq et al., 2018).

Although high performance in job contexts is different phenomena than high

performance among entrepreneurs, some comparison might be useful, especially

considering sole-proprietors (whose success is directly related to their own work)

(Campbell et al., 1993; Dyer et al., 2015). Hallak et al. (2018) recognized this similarity

and conducted cross-sectional correlational analysis among tourism business owners (n =

298) in Australia measuring the relationship between entrepreneurs’ place identity (PI),

entrepreneurial self-efficacy (ESE), and business performance (BP). Hallak et al.

identified 957 tourism entrepreneurs in Australia and sent them a mail invitation to

participate in the study, with the survey to measure PI, ESE, and BP. They received 298

completed surveys yielding a 31% response rate (Hallak et al., 2018).

Hallak et al. (2018) then performed correlation analysis and structural equation

modeling (SEM) statistical analysis to measure the relationship between the variables

(Hallak et al., 2018). Hallak et al. concluded that SEM indicated three significant effects

between the variables and that PI had significant, and positive relationship with ESE (β =

.36. p < .01), and that ESE is a direct predictor of BP (β = .57, p < .01) (and not vice

versa [β = -.14, p < .01]), for both male and female entrepreneurs. Hallak et al.’s

conclusions regarding the relationship between self-efficacy and performance are similar

to Gupta et al.’s (2013) and Cheng’s and Chiou’s (2016) even though the constructs and

sample demographics vary significantly.

Self-Efficacy Moderating Effects

High level of general self-efficacy buffers against the negative effect of job stress,

job ambiguity, and role conflicts (Joseph et al., 2015; Schwarzer & Warner, 2013;

Sitzmann & Yeo, 2013; Theorell et al., 2015). Earlier scholars, such as Thompson and

Gomez (2014), examined this buffering effect by analyzing the relationship between

negative stressors, self-efficacy, and job performance.

Interested in buffering against negative stressors -effect of self-efficacy,

Thompson and Gomez (2014) measured the relationship between role ambiguity (RA),

role conflict (RC), self-esteem (SE), general self-efficacy (SEF), and depression, anxiety,

and stress (DASS) among a diverse group of employees (n = 78) of Australian

organizations. To conduct the study, Thompson and Gomez (2014) recruited a diverse

group of 78 participants from various Australian organizations, to participate in an online

survey. Thompson and Gomez measured RA (Breaugh & Colihan Scale), RC (Rizzo,

House & Lurtzman Scale), SE (Rosenberg Scale), SEF (Schwarzer & Born Scale), and

DASS (Lovibond Scale) in the survey. Thompson and Gomez then performed correlation

analysis and hierarchical regression analysis to measure the relationship between the

variables. Thompson and Gomez concluded that that SEF moderated the relationship

between role ambiguity and depression and between performance role ambiguity and

stress, while self-esteem moderated the relationship between role ambiguity and anxiety.

Thompson and Gomez (2014) of the study used only one data source, which

might have increased the risk of common method bias and acquiescence bias (Ferrando &

Lorenzo-Seva, 2010; Ylitalo, 2009). Discussion did not include the potential implications

of either risk (Thompson & Gomez, 2014). Thompson and Gomez used custom

instruments to measure DASS and did not fully disclose the contents of the instrument.

Thompson’s and Gomez’s use of custom instruments may limit the external validity of

the results (Pearl & Bareinboim, 2014). Thompson and Gomez conducted the study in

one geographical location, and therefore bias may arise from cultural norms (such as

performance orientation and uncertainty avoidance, as defined by Globe research (2007)),

which affect or limit the generalizability of findings to other cultures or contexts.

Although Thompson and Gomez (2014) concluded that self-efficacy buffers

against the effect of negative stressors, they did not discuss how and why such an effect

occurs. Interested in how self-efficacy actuates in such situations, Delahaij and Van Dam

(2017) examined the effect of coping style, coping self-efficacy, and appraisal emotions

to acute stress among military personnel (n = 648) in the Netherlands. Acknowledging

the effect of coping behavior on stressful situations among military and police personnel,

Delahaij and Van Dam designed a study to test the effect of coping style (CS), coping

self-efficacy (CSE), and appraisal emotions (AE) on coping behavior (CB). To test the

effect, Delahaij and Van Dam conducted a cohort study (three cohorts with the length of

18 to 33 weeks) and measured CS, CSE, AE, and CB for each participant in stressful

training situations. Delahaij and Van Dam then performed confirmatory factor analysis

(to verify that AE and CB are distinct constructs) and used structural equation modeling

to verify the effect model. They concluded that CS had significant relationship with CB

(emotion-oriented β = .18, p < .01, task-oriented β = .38, p < .01), and that CSE had

significant positive relationship with challenge emotions (β = .23, p < .001) and

subsequent task-focused CB (β = .23, p < .001) (Delahaij & Van Dam, 2017). They also

noted that CSE had significant negative relationship with threat emotions (β = -.11, p <

.001), again which has significant relationship to emotion-focused CB (β = .43, p < .001)

(Delahaij & Van Dam, 2017).

In summary, Delahaij and Van Dam (2017) concluded that CSE plays an

important role in shaping individuals’ responses to acute stress situations. However, as

81% of study participants were men, and all participants were Dutch, the external validity

of Delahaij’s and Van Dam’s research remains limited to a demographic group that the

sample of this study represents (Pearl & Bareinboim, 2014). Second, it is not clear how

much of the reported coping self-efficacy was a result of participants’ earlier experiences

of similar stressful situations (Delahaij & Van Dam, 2017). Because of this limitation, the

effect of participants’ past experiences may partially explain Delahaij’s and Van Dam's

conclusions regarding the positive effect of self-efficacy in stressful situations.

Relationship Between Age, Tenure, and Sales Performance of Salespeople

Only a few researchers examined the relationship between age, length of tenure,

and sales performance of salespeople and the results of existing studies remain

contradictory (Kwak, Anderson, Leigh, & Bonifield, 2019). Wihler et al. (2017)

conducted such research examining the relationship between salespeople’s

conscientiousness, extraversion, age, length of tenure, and objectively reported sales

performance among insurance agents (n = 114) of one large German insurance company.

Wihler et al. concluded that conscientiousness and extraversion together had significant

relationship with sales performance (β = .42, p < .05, r2 = .08) and that age had

significant negative relationship with sales performance (β = -.33, p < .01) whereas

tenure had significant positive relationship with sales performance (β = .38, p < .05).

Stajkovic et al. (2018) obtained similar results when analyzing the relationship between

salespeople’s age, tenure, general self-efficacy, and sales performance among sales

associates (n = 142) employed by car retail group which operates in 16 cities in US and

Canada. Stajkovic et al. concluded that general self-efficacy had nonsignificant

relationship with past (β = -.03, n.s.), and future sales performance (β = .14, n.s.) , but

salespeople’s age and length of tenure had significant relationship with past performance

(age->past performance β = -.20, p < .05; tenure->past performance β = .25, p < .01) and

future performance (age->future performance β = -.30, p < .01; tenure->future

performance β = .31, p < .01). Interestingly, length of tenure and age seemed to affect the

contents of sales so that those salespeople who had long tenure sold more profitable cars

regardless of age (Stajkovic et al., 2018). Stajkovic et al. also measured salespeople’s

hope, optimism, and resilience and noted that tenure had a significant negative

relationship with hope (β = .26, p < .01) but that general self-efficacy had a significant

positive relationship with hope (β = .72, p < .01). However, regression analysis revealed

that in total age, gender, length of tenure, years of industry, and past sales performance

explained 42% variance in future sales performance and that the strongest predictor for

future sales performance was past sales performance (β = .52, p < .01) (Stajkovic et al.,

2018). Unfortunately, Stajkovic et al.’s (2018) instrument for measuring future sales

performance consisted only of sales volume and commission payment of 1 month, and

therefore additional latent factors, such as working times and seasonal changes, could

affect the results.

Feng and Fay (2016) obtained significantly different results while examining the

relationship between salespeople’s capabilities and future sales performance among

salespeople (n = 1049) of one Chinese insurance company. Feng and Fay used

salespeople’s age and length of tenure as factors of unique construct salespersons

capability and found that although salesperson’s capability had a significant relationship

with future sales performance, both salespeople’s age and length of tenure relationship

with sales performance were insignificant. In fact, the only individual elements of

salespersons capability which had a significant relationship with sales performance were

salespeople’s intention to quit a job; and an average age of customers (Feng & Fay,

2016).

Summary

Bandura (1977), who developed the theory of self-efficacy, considered then-

current behavioral theories insufficient in explaining human behavior. The construct self-

efficacy is an individuals’ perception of his or her ableness to accomplish a certain task

(Bandura, 1977). After Bandura, multiple researchers concluded that an individual’s self-

efficacy is significantly related to the actual capability of accomplishing a particular task

(Dagher et al., 2015; Sitzmann & Yeo, 2013). Sherer et al. (1982) expanded the theory of

self-efficacy and introduced the construct of general self-efficacy, which describes an

individuals’ overall optimistic self-belief in coping with difficulties in life (Sherer et al.,

1982). Waaktaar and Torgersen (2013) concluded that the level of the individual’s

general self-efficacy depends on genetics and experiences during childhood, and the level

of general self-efficacy remains relatively stable during adulthood. Traumatic events,

serious illness, and chronic stress may cause degradation in the levels of individuals’

general self-efficacy, although the general self-efficacy also seem to buffer against, and

speed recovery of these types of events (Cyniak-Cieciura et al., 2015; Smith et al., 2006).

Many scholars also noted that both task-specific self-efficacy, and general self-efficacy

significantly relate to success in academia, military and in job contexts – either directly,

or through moderating the effect of other predictor constructs (Churchill et al., 1985;

Stajkovic et al., 2018; Verbeke et al., 2011).

When searching for predictors for sales performance, researchers and practitioners

are interested in the factors with generalizable properties (Carter et al., 2008; Dyer et al.,

2015; Lu et al., 2016). These properties are important because sales jobs are often

strenuous, and situations vary significantly (Lu et al., 2016; Plouffe, 2018). If some

characteristic of salespeople has significant, and known relationship to sales performance,

sales managers can use measures of that characteristic in the recruitment process. Most

researchers have used task-specific self-efficacy in measuring the relationship between

self-efficacy and sales performance, but that construct lacks the predictive power of

general self-efficacy as the task-specific self-efficacy is volatile and constantly

developing construct (Grether et al., 2018). Also, because of significant variations of

sales situations, once measured task-specific self-efficacy may not relate to the new sales

situations, and therefore the validity of such research might be lower compared to those

employing a more generalizable measure of self-efficacy.

Transition

Section 1 covered the foundations of this study. Section 1 began with a

description of the background of the study, followed by the problem statement, the

purpose statement, and the nature of the study. I then presented the research questions,

hypotheses, and theoretical framework that guided the study. Section 1 also included

definitions of operational terms, the assumptions, limitations, and delimitations

underlining the study, as well as the significance of the study. Finally, Section 1

contained a literature review, which is a critical analysis and synthesis of the literature

related to the study and the constructs of self-efficacy and sales performance. The

literature review consisted of the introduction for the sales performance research topic,

discussion about the construct of performance in sales contexts, explanation of the theory

of self-efficacy as well as the alternative theories predicting success in sales. I also

elaborated on the criticism toward general self-efficacy and analyzed the articles

measuring the relationship between either task-specific self-efficacy, or general self-

efficacy and sales (or job) performance. Also included were introductions of the other

independent variables (age, and length of tenure of salespeople) to justify their

importance in this research.

Section 2 covers the nature and structure of the research study and its design,

including the steps involved in collecting, validating, and analyzing the data and

protecting the participants. I describe the role of the researcher and the population and

justified sampling method, instrument selection, data collection techniques, and data

analysis methods. The end of Section 2 includes the discussion of the internal and

external validity of this study.

Section 3 contains the presentation and analysis of the results and findings.

Discussion also includes the application to professional practice, implications for social

change, and recommendations for professional practice and future research. Finally, I will

provide a study summary and conclusions.

Section 2: The Project

This section includes a description of my role as a researcher, the process for

finding participants to volunteer, as well as expounding on the research method and

design and presenting the methods used to ensure ethical research. This section also

includes a discussion of the data collection, analysis, and validation of results processes.

Section 2 is the foundation for Section 3, includes a presentation of the results,

implications for social change, and recommendations for action and further research.

Purpose Statement

The purpose of this quantitative correlational study was to examine the

relationship between age, tenure, general self-efficacy, and sales performance of

salespeople. The independent variables were age, tenure, and general self-efficacy. The

dependent variable was sales performance. The targeted population consisted of face-to-

face salespeople of THSP IT companies. The implications for positive social change

include the potential to understand the correlates of sales performance better, thus

contributing to the reduction of discrimination in the recruiting salespeople of various

ages and experience.

Role of the Researcher

As the researcher, my primary role in this quantitative correlational study was to

design the study, identify and recruit participants, collect the data, and analyze the data.

The researcher’s role in the quantitative research process is to collect the data and

objectively analyze gathered data accurately with appropriate tools and statistical

methods (Childs, McLeod, Lomas, & Cook, 2014; Moon, 2015). As noted by Creswell

(2014), the level of researchers’ interaction with the participants is usually lower in

quantitative studies compared to qualitative or mixed-method studies. In this study, my

interaction with the participants was limited to sending the invitation to the survey via a

LinkedIn post. Although my relationship with the topic involved managing salespeople in

the IT industry, I did not have a direct relationship with the targeted participants in this

research.

Ethical principles protect the rights and well-being of human participants

(Bromley, Mikesell, Jones, & Khodyakov, 2015). Thus, this study followed the ethical

practices and protocols articulated in the Belmont Report (1979) to ensure that the

participants provided informed consent before participating; all responses remained

confidential (Miracle, 2016). To help ensure participants’ protection in conducting this

study, the Walden University Institutional Review Board (IRB) granted approval

(approval number 07-30-19-0561026) as part of the doctoral study process before the

data collection began. I also met the required training from the National Institutes of

Health on human subject protection. This study included the use of validated instruments

to collect the data to avoid any personal bias. Earlier researchers concluded that the

chosen instruments in this study are reliable and valid (Nel & Boshoff, 2016; Zhou,

2016). I did not (a) conduct research in my direct professional setting, (b) employ any

custom instruments in the research, or (c) have a conflict of interest in this research.

Participants

The targeted population consisted of face-to-face salespeople of THSP IT

companies. Because the exact size of the population was unknown, and participants

worked for multiple different companies, this study included the use of the IT Sales

Global Community LinkedIn group to contact participants. The IT Sales Global

Community LinkedIn group is the largest professional social network group for IT sales

professionals, with 41.050 members at the time of data collection of this study. A

professional social media group was the most appropriate method to contact the

participants as scholars estimated that over 80% of salespeople frequently use LinkedIn

and that the usage was highest among IT industries (Agnihotri, Dingus, Hu, & Krush,

2016).

Creswell (2014) noted that research participants in quantitative research should be

part of the research population, and if a researcher uses sampling, the chosen sample

should be representative of the total population (Walker, 2014). Walker (2014)

emphasized the importance of sampling diversity and randomization, especially with

binary variables, such as gender. In this study, participants had to be face-to-face

salespeople who, at the time of the study, (a) worked for a THSP IT company, (b) had a

tenure of at least 1 year, (c) were at least 20, (d) could read and understand English, and

(e) had an annual quota attainment result.

The IT Sales Global Community LinkedIn group administrators allowed posting

an invitation to this study on the group’s front page. The study included a research

preannouncement and the group administrators’ championing to ensure successful

recruiting for research. Both of these techniques increase the salespeople’s willingness to

participate in studies (Panagopoulos et al., 2011; Pedersen & Nielsen, 2016).

Newington and Metcalfe (2014) concluded that researchers should inform

participants of the nature of the research in an e-mail. Because reaching participants via

email was not feasible with this study, a LinkedIn group announcement contained a

description of the nature of this study. To reduce the time burden for participants, the

LinkedIn group announcement contained a link to the actual survey. Some scholars noted

that researchers could increase participants’ willingness to participate in surveys by

providing comprehensive instructions, personalized emails, or the possibility to win

prizes after completion of the survey (Pedersen & Nielsen, 2016; Trespalacios & Perkins,

2016). However, Trespalacios and Perkins (2016) concluded that neither the length of the

description of research purpose nor informing of the possibility of winning prizes after

responding increased the survey response rates.

Consequently, the LinkedIn group announcement was short and participants could

not win prizes. Pedersen and Nielsen (2016) concluded that an egoistic text appeal of an

email invitation to the survey had a significant positive relationship with the response

rates (β = .45, p < .01) among working-age adults (N = 6,162) who participated in a

Danish survey panel. This study included the same principles in constructing the

LinkedIn group announcement.

To establish working relationships with study participants, researchers should

clearly explain their research and rationale for the research design to get participants

involved (Pedersen & Nielsen, 2016). Subsequently, the LinkedIn group announcement

included the purpose of this research as well as the support from the groups’

administrators. Ward and Meade (2018) noted the risk of participants responding

carelessly to online surveys. Therefore, the researcher should build and maintain a

relationship with participants and consistency of communication as well as emphasize the

importance of careful responding with the survey (Trespalacios & Perkins, 2016; Ward &

Meade, 2018). In this study, participants saw a group announcement, as well as additional

reminders, and received a thank you notification after completing the survey. Following

the guidance from Ward and Meade, the survey contained guidance for the participants to

respond to each survey question carefully.

Research Method and Design

Research Method

This study included the use of the quantitative method to examine the relationship

between age, tenure, general self-efficacy, and sales performance of salespeople.

Edmondson and McManus (2007) introduced a contingency framework to help new

researchers in selecting an appropriate research approach and methodology based on

existing scholarly knowledge about the research topic. Noteworthy is that quantitative

and qualitative research approaches are not distinct and mutually exclusive, but rather

represent the different ends of a continuum with varying level of method mixing

(Edmondson & McManus, 2007). Two dimensions in Edmondson and McManus’s

framework govern the most appropriate research approach: the maturity of existing

theory (nascent, intermediate, and mature) and the type of data (qualitative, hybrid, and

quantitative).

At one end of the continuum (nascent theory with qualitative data), the research

should be qualitative only, whereas, at the other end (mature theory with quantitative

data), the research should be quantitative only (Edmondson & McManus, 2007). The

theoretical framework for this research is mature, and the researchers have been

developing instruments to measure general self-efficacy for nearly 40 years (Bandura,

2012; Johnson & Jaramillo, 2017). Similarly, the research domain of sales performance

relies on quantitative outputs (whether the researchers measure output as self-rating,

supervisory rating, or by objective measure). Hence, the quantitative method was

appropriate in this study.

The quantitative method allows researchers to examine the relationship between

variables (Creswell, 2014; Lund Research, 2016). Quantitative studies are a common

approach to study predictors of sales performance (Joseph et al., 2015; Wihler et al.,

2017). Previous researchers have used the quantitative method to conduct similar

research among sales professionals (see Beltrán-Martín, Bou-Llusar, Roca-Puig, &

Escrig-Tena, 2017; Bonney, Plouffe, & Wolter, 2014; Hallak et al., 2018; Miraglia et al.,

2017; Ozyilmaz, Erdogan, & Karaeminogullari, 2018). The use of psychometric

constructs with a quantitative method follows the principles of a postpositivist approach

to sales performance research (Carter et al., 2008; Heale & Twycross, 2015; Phillips &

Burbules, 2000). As positivists, postpositivist researchers test hypotheses based on

theories using experimental, archival, or survey data (Phillips & Burbules, 2000).

However, differing from a positivistic stance in which a researcher considers variables as

absolute measures, the postpositivistic stance is conjectural and recognizes that constructs

(such as general self-efficacy) can be measured only indirectly and thus only partially

reflect the particular construct (Phillips & Burbules, 2000).

This study involved the testing of the hypothesis if a statistically significant

relationship existed between age, tenure, general self-efficacy, and sales performance of

salespeople. Therefore, the quantitative method was appropriate for this study.

Researchers use qualitative studies to answer questions of how and why (Yin, 2014), but

the qualitative method did not meet the needs for the examination of relationships

between variables (see Creswell, 2014; Creswell & Poth, 2018). Therefore, a qualitative

method was not appropriate for this study. Mixed-methods studies are useful when a

quantitative or a qualitative method alone is not sufficient to address the research

problem (Johnson & Onwuegbuzie, 2004; Morse & Niehaus, 2009). This study involved

testing hypotheses based on established theory. Because no need existed to explore the

problem qualitatively, a mixed-methods study was not appropriate.

Research Design

In this study, I used a quantitative method with correlational design. Quantitative

studies can be descriptive, correlational, quasi-experimental, or experimental (Creswell,

2014). Researchers use descriptive designs to report basic statistical characteristics of a

sample, such as mean, standard deviation (SD), or range of analyzed variables of the

sample (Creswell, 2014). The descriptive design was not sufficient for this study as the

study requires an examination of relationships between variables. Researchers use the

quasi-experimental and experimental designs to study causal relationships between

variables (Campbell & Stanley, 2010). Quasi-experimental designs are appropriate when

the researcher examines the relationship between variables among nonrandom

participants, and experimental designs are appropriate when the researcher examines the

relationship between variables among random participants (Creswell, 2014). The study

included an examination of the correlational relationship between the variables, and,

therefore, neither the quasi-experimental nor experimental design were appropriate for

this study. I used a correlational design because of allowing the examination of noncausal

relationships between the variables and requiring no manipulation of variables.

Population and Sampling

The targeted population consisted of face-to-face salespeople of THSP IT

companies. Because the exact size of the population is unknown, the study used a

representative LinkedIn group (The IT Sales Global Community) with 41.050 members

to gain access to the population. Because accessing all members of the population is often

impractical, researchers use a sampling of the total population and employ various tools

to calculate the required sample size (Creswell, 2014; Rahi, 2017). Required sample size

in quantitative correlational research depends on the statistical method, error tolerance,

number of measured variables, and assumed dependent variable effect size (Bosco,

Aguinis, Singh, Field, & Pierce, 2015; Martínez-Mesa, González-Chica, Bastos,

Bonamigo, & Duquia, 2014).

One of the most used tools is the G*Power application, which is a statistical

software package used to conduct a priori sample size analysis (Faul, Erdfelder, Buchner,

& Lang, 2009). I conducted a power analysis using G*Power version 3.1.9.2 software to

determine the appropriate sample size for the study. An a priori power analysis, assuming

small to medium effect size (f2 = .15), α = .05, and three predictor variables, indicated the

requirement of a minimum sample size of 77 participants to achieve a power of .80.

Increasing the sample size to 118 would have increased power to .95. Therefore, the goal

was to seek between 77 and 118 participants for the study (see Figure 1). Because this

study used multiple hierarchical regression analysis, I also calculated the required sample

size for R2 increase using three total predictor variables, and 1 or 2 tested predictors. This

analysis resulted in the same sample size requirements (77 for the power of .80 and 118

for the power of .95).

Figure 1. Power as a function of sample size.

The use of medium effect size (f2 = .15) was appropriate for this proposed study. Twenty-

seven articles (see Appendix A), where sales performance was the outcome

measurement, supported the use of medium effect size for sample size calculation.

Researchers use two primary types of sampling methods: probabilistic (e.g.,

random) and non-probabilistic (e.g., non-random). With probabilistic sampling, each

participant has an equal non-zero opportunity to be selected in the sample, and with non-

Total sample size

probabilistic sampling, the researcher controls the sample-inclusion probability of each

participant (Davies & Hughes, 2014). The different classifications of probabilistic

sampling used in quantitative studies are (a) simple random sampling (b) stratified

random sampling, and (c) cluster sampling; the non-probabilistic methods are (a)

convenience sampling, (b) quota sampling, and (c) purposive sampling (Davies &

Hughes, 2014).

This study included non-probabilistic purposeful sampling because (a) typical

survey response rates among salespeople are below 30%, (b) general self-efficacy data is

not available by other means than the survey, and (c) applying probabilistic sampling

would not reduce the potential risk of bias caused by characteristical differences between

participants and non-participants (Allen, 2016). Multiple earlier scholars examining the

predictors of sales performance used the same sampling method (Bonney et al., 2014; Fu

et al., 2009; Lu et al., 2016; Pettijohn et al., 2014).

Purposeful sampling results risk the sample not being representative of the whole

population (Wagner, 2014). Several statistical techniques can counter this risk, including

normal distribution analysis, and comparison between the participant and non-participant

data (Martínez-Mesa, González-Chica, Duquia, Bonamigo, & Bastos, 2016). This study

included an examination of normal distribution analysis of all measured variables to

mitigate the risk of the sample being not representative of the whole population.

Ethical Research

The principles of ethical research require a researcher to maintain the credibility

of the research process and protect the participants as well as the participating

organization from harm (U. S. Department of Health & Human Services, 1979). It is the

responsibility of the researcher to follow the principles of the Belmont Report with this

study. To comply with these requirements, I completed the Protecting Human Research

Participants training by The National Institutes of Health (NIH) Office of Extramural

Research and received certificate number 2399671 on May 21, 2017 (see Appendix B).

As part of the doctoral study process, Walden University IRB reviewed compliance of

this study by the university’s ethical standards and U.S. Federal regulations and granted

approval for the study (approval number 07-30-19-0561026). The data collection for this

study begun after IRB approval.

The principles of the Belmont Report include informing participants of their

rights and preserving their confidentiality (U.S. Department of Health & Human

Services, 1979). The online survey contained an informed consent form. The consent

form included the background and purpose of the study, guidance for completing and

submitting the survey, and instructions for voluntary participation and withdrawal

process. The survey introduction web page contained a statement of confidentiality, as

well as risks and benefits for the participants. As the study included no compensation nor

prizes for participation, participants did not receive any extrinsic rewards. The online

survey contained contact information for the researcher and Walden University, should

participants have had any additional research-related questions.

Some studies, especially those relating to medical care or legal procedures,

require written consent form from the participants (U. S. Department of Health & Human

Services, 1979). However, researchers conducting survey research can often use implied

consent, in which the participant a) receives information about the study and b)

voluntarily completes the survey (Hammer, 2017). The study included implied consent,

which means participants indicated their consent by clicking the link to the survey on the

LinkedIn group page, completing the survey, and finally, submitting the survey.

The online survey contained an option for participants to save a copy of the

consent form. The use of implied consent and not including any clear-text identifiable

information in the online survey helped to maintain the confidentiality of participants

(Evans & Mathur, 2018). The survey was anonymous, therefore, the study data does not

contain participants’ contact information nor any company affiliation information. I

downloaded and then deleted cloud-based (SurveyMonkey) data after completion of the

survey as recommended by Vitak et al. (2016). After 5 years, I will discard the raw data

by deleting all copies.

Data Collection Instruments

The study included the use of a standardized instrument to measure age, length of

tenure, general self-efficacy, and sales performance to measure identified variables. All

variables of this study (age, length of tenure, general self-efficacy, and sales

performance) are interval type metric variables, and the data for these variables came

from the survey.

Earlier researchers predominantly used three different instruments to measure

general self-efficacy (Barahona et al., 2018; Zhou, 2016). Each of these instruments has

some advantages over others, but both SGSE and NGSE are bipolar scales with lack of

large multi-cultural validation studies (Barahona et al., 2018; Luszczynska et al., 2005).

Bandura (2012) specifically alerted against using bipolar scales to measure any forms of

self-efficacy, arguing that a neutral level of self-efficacy, especially at the center of the

scale, is a logical fallacy. Also, both SGSE and NGSE consist of a larger number of

survey items compared to GSES and, therefore, might reduce the response rate and

accuracy of the answers (Allen, 2016; Peytchev & Peytcheva, 2017).

Finally, the GSES instrument is publicly available and is free to use without

explicit permission from the authors and existing studies indicated that the instrument

results in a unidimensional measure of general self-efficacy with high validity and

reliability (α between 76 and 90 depending on the study; Schwarzer, 2014; Zhou, 2016).

Because of these advantages of GSES over SGSE, and NGSE, the study used GSES (see

Appendix C) to measure general self-efficacy. Appendix D contains the permission to use

the GSES instrument.

Reliability and Validity of Measures

Several scholars concluded that general self-efficacy measured with GSES

instrument has high construct validity, test-retest reliability, and high internal consistency

(Barahona et al., 2018; Schwarzer, 2014; Zhou, 2016). Construct validity is an indicator

of how accurately a test measures what it claims to measure (Cronbach & Meehl, 1955).

Construct validity differs from internal validity as it concerns each construct in the

research (Straub, Boudreau, & Gefen, 2004). To ensure construct validity, Straub et al.

(2004) recommend testing discriminant and convergent validity, including the factorial

validity of each construct. The data in this study included one psychometric construct,

general self-efficacy for which the study included a test of the validity as suggested by

Straub et al. (2004). Because of measuring only one psychometric construct, general self-

efficacy, the examination of discriminant, and convergent validity were not in the scope

of this study. However, other scholars concluded that general self-efficacy, measured

with GSES, results in unidimensional measure with high discriminant and convergent

validity (Barahona et al., 2018; Scholz et al., 2002; Straub et al., 2004).

Reliability (internal consistency) measures the item correlation results between

different tests of measuring the same construct (Straub et al., 2004). Straub et al. (2004)

suggested using Cronbach’s α above .60 for exploratory research and above .70 for other

types of research. Because this study was not exploratory, .70 Cronbach’s α requirement

applied for constructs in this study. Earlier researchers, such as Schwarzer (2006) and

Zhou (2016), concluded that the chosen instrument to measure general self-efficacy has

high internal consistency (α > 76). This study included reporting the consistency

(Cronbach’s α) of the GSES as part of the Presentation of the Findings -section.

Reliability (internal consistency) did not apply to other measures in this study because of

the nature of the data.

Predictive validity concerns the level of confidence that change in the input

constructs in the study produces a measurable change in output construct (Straub et al.,

2004). As an example, Straub et al. (2004) described how schools use GMAT scores in

an academic setting to predict performance. An instrument may have high predictive

validity if a large number of similar, confirming results exist in the scholarly literature

(Stajkovic et al., 2018); however, it is the researcher's responsibility to analyze the

feasibility of such earlier findings.

As discussed in the review of the professional and academic literature -section,

general self-efficacy remains relatively stable construct over an individuals’ adult life.

Multiple scholars concluded that general self-efficacy relates to success in work,

academic, and military contexts (Sitzmann & Yeo, 2013). Similarly, multiple scholars

concluded that the other independent variables used in this study (age, and length of

tenure of salespeople) have a significant relationship with sales performance (Wihler et

al., 2017); and by inherent nature of these variables, they do not exhibit random

fluctuation. Because of the logical stability of the measures for salespeople’s age, length

of tenure, and general self-efficacy, this study has high predictive validity.

Unidimensional Reliability concerns the level of how many latent constructs

selected test item measures and it is related to discriminant and convergent validity

(Zhou, 2016). In a perfect test, all instruments measure only one construct, therefore

achieving unidimensionality (Zhou, 2016). However, in most tests, an instrument may

measure one or more latent constructs to varying degrees (Thompson, 2004). To improve

unidimensional reliability, Straub et al. (2004) recommended using factor analysis for

measured variables (such as GSES in this study); and in case multiple factors emerge,

structural equation modeling (SEM) with latent factor analysis available in LISREL and

SPSS.

Earlier scholars extensively studied the instrument used in this research to

measure general self-efficacy and found that the instrument results in a unidimensional

measure of general self-efficacy construct across disparate demographic groups (Zhou,

2016). The general self-efficacy instrument should have, therefore, resulted in a

unidimensional measure of the construct also in this study. Additionally, this study

included retesting the dimensionality of the measure by using factor analysis, which is a

method suggested by Straub et al. (2004). Unidimensional reliability did not apply to

other measures in this study as the nature of the data was inherently unidimensional.

Reliability (split halves) is a technique in which the researcher divides the sample

to two equally sized sub-samples, and the reliability of the results is improved by

comparing average correlations of each item (Straub et al., 2004). The complexity with

split-halves testing relates in splitting as the outcome varies depending on how the

sample is split into two (Straub et al., 2004). Split halves technique was not appropriate

for this research because of requiring equal probability for participant selection to groups,

and the sampling method used in this research did not fulfill the requirement.

Reliability (alternative forms) is a technique in which construct reliability is

increased by using several instruments to measure the same construct (Straub et al.,

2004). If different instruments (alternative forms) produce similar results for a given

construct, alternative forms reliability is high (Straub et al., 2004). Researchers can

measure individuals’ general self-efficacy with multiple instruments, such as with

Sherer’s (1982) GSE scale, Schwarzer’s and Jerusalem’s (1995) GSES scale, or with

Chen’s (2001) NGSES scale. Therefore, alternative forms could have been used in this

research to improve general self-efficacy construct reliability. However, each of these

instruments consists of 10 or more survey items. Scholars experienced with online

surveys, such as Allen (2016), concluded that survey item count increase correlates with

a decrease in survey response rates. To balance between sufficient survey response rate

and reliability, this study included the use of one previously validated instrument to

measure general self-efficacy. Reliability (alternative forms) did not apply to other

measures in this study, as other independent variables had no alternative forms, and no

comparable objective alternatives were available for the dependent variable.

Content validity is about selecting the best instrument to measure constructs in the

study (Straub et al., 2004). For example, Waaktaar and Torgersen (2013) measured

children’s self-efficacy levels in their study about the antecedents of self-efficacy.

Although they used a well-established instrument (Children’s Perceived Self-efficacy

Scale) to assess self-efficacy, they omitted 25 (of total 37) questions from testing

(Waaktaar & Torgersen, 2013). Because of omittance of questions, the content validity of

their research may be lower compared to the situation of using the full instrument, even

though Waaktaar and Torgersen examined the concurrent validity of the new scale.

Because this study used a non-modified version of GSES to measure individuals’ general

self-efficacy, and large scale studies indicated the advantages of the chosen instrument

over other instruments, the content validity is high for general self-efficacy (Zhou, 2016).

Content validity did not apply to other independent variables (age, and length of tenure of

salespeople) because of the nature of the data. Content validity is important for the sales

performance variable, but there was no universally accepted method to measure sales

performance, and many companies had their own measures for the construct. The quota

attainment was the best available tool to measure sales performance because the THSP

companies predominantly used quota attainment for employee appraisals and commission

payments (Inyang & Jaramillo, 2019).

Following Creswell’s (2014) guidance on storing research data, I will store the

raw data in a secure place for 5 years; the data is available for other scholars by request.

As noted by Tsai et al. (2016), storing the raw data increases research credibility as other

scholars can independently verify research conclusions. Tourangeau (2018a)

recommended pseudonymization or anonymization of the data to protect participant

confidentiality. As the raw data in this study did not contain any identifiers for the

participants or the companies that the participants worked for during the survey, other

scholars’ access to the raw data does not pose a risk for confidentiality.

Data Collection Technique

The data collection for this study began with an analysis of appropriate tools to

reach the population. Use of online software platforms surpassed traditional mail, and

face-to-face survey protocols platforms allow participants to respond at their convenience

and require minimum intervention to participants’ daily job (Evans & Mathur, 2018).

Using an online software tool also increases the chances of reaching a large participant

pool compared to mail administered surveys or face-to-face survey interviews (Evans &

Mathur, 2018). Because the targeted population consisted of salespeople residing in a

geographically large area, and the survey was anonymous, an online software platform

was the most appropriate method for data acquisition.

Online software platforms include many suitable tools for anonymous surveys

(Bentley, Daskalova, & White, 2017). SurveyMonkey is an online third-party software

platform that researchers frequently use when collecting data for surveys (Bentley et al.,

2017). Evans and Mathur (2018) concluded that online surveys are superior compared to

traditional mail, or interview surveys, especially when the targeted population resides

within a large geographical area. Schoenherr, Ellram, and Tate (2015) noted that online

surveys could be complemented with pre-screening questions, therefore, ensuring

participants’ eligibility for the survey.

Researchers using online surveys typically invite participants either by email or

by posting the survey invitation on a platform accessible to the population members

(Evans & Mathur, 2018). As the survey in this study was anonymous and no contact

information was available for an email invitation, a professional social media platform

was the most appropriate method to reach the targeted population. The IT Sales Global

Community LinkedIn group was the largest professional social networks group for IT

sales professionals, and scholars estimated that over 80% of salespeople frequently use

LinkedIn and that the usage was highest among IT industries (Agnihotri et al., 2016). The

data for this research came from an anonymous survey executed via SurveyMonkey. The

survey of this study contained pre-screening questions to verify that the participants are

part of the population and are eligible to participate based on purposeful sampling

criteria.

Carter et al. (2016) and Pransky et al. (2006) recommended using objective sales

performance data with studies involving examination of sales performance predictors.

Using objective sales performance requires participant employer company managers to

provide some of the required data, such as quota attainment results. However, using an

objective sales performance approach had some severe disadvantages that would render

the study infeasible. First: matching the participant survey data with objective sales data

required maintenance of participants’ identifiers invalidating the anonymity. Second:

Obtaining objective sales performance data (quota attainment results) from multiple

companies and for multiple persons was not possible because companies consider such

information confidential (Deeter-Schmelz, 2016). Third: even if the companies and the

participants allowed non-anonymous combinatory approach, legal data protection

requirements would make obtaining the data infeasible (the people operating with the

data at each of the companies would need to have local legal right to process each

persons’ data; Greene, Shmueli, Ray, & Fell, 2019; Hintze, 2018). Because of the

aforementioned issues with using objective data, the data for the sales performance

variable came from the survey.

A pilot study was not part of this research because a pilot study may increase the

risk of social desirability bias in the participants’ responses (Babatunde, 2016; Cope,

2015). Pilot studies are scaled-down studies that researchers frequently use before actual

full-scale studies (Cope, 2015). Pilot studies allow researchers to test and improve study

protocol before conducting a full-scale study, thereby improving the quality of the whole

study (Cope, 2015; In, 2017). However, scholars rarely use pilot studies with

correlational sales performance research (Talsma et al., 2018).

Data Analysis

The research question for this study was as follows: What is the relationship

between age, tenure, general self-efficacy, and sales performance of salespeople? Based

on the theoretical framework of this study, this study required statistical analysis of the

effect of all independent variables on the dependent variable. The null hypothesis in this

study was that there is no statistically significant relationship between age, tenure,

general self-efficacy, and sales performance of salespeople. An alternative hypothesis

was that there is a statistically significant relationship between age, tenure, general self-

efficacy, and sales performance of salespeople. To analyze the data, I used version 25.0

of the Statistical Packages for Social Sciences software (SPSS). Researchers use SPSS to

analyze quantitative data and test the hypotheses (Rovai, Baker, & Ponton, 2013). R-

based ridge regression and robust regression plugins supplemented standard regression

analyses of SPSS because SPSS did not offer these methods directly (Astivia & Zumbo,

2019).

According to Garson (2013), researchers use multiple hierarchical linear

regression analysis to determine the correlations between two or more variables. In this

study, I used multiple hierarchical linear regression analysis on the interval data to test

the above hypotheses. Earlier scholars noted that each of the independent variables used

in this study may have a significant relationship with sales performance, but also that

each of the independent variables may moderate the relationship between other

independent variables and sales performance (Alessandri, Borgogni, Schaufeli, Caprara,

& Consiglio, 2015; Joseph et al., 2015). Thus, the need existed to conduct multiple

hierarchical linear regression analysis and test each combination of independent variables

(Grömping, 2015).

Researchers examining quantitative data select appropriate statistical tests based

on the nature of the research and scale of measurement of the variables (Creswell, 2014;

Davies & Hughes, 2014; Hox, Moerbeek, & van de Schoot, 2017). One of the most used

statistical methods in quantitative correlational studies is regression analysis, which can

be used to measure the relationship between one or more independent variables and a

dependent variable (Grömping, 2015; Mertler & Vannatta, 2017). Researchers use simple

linear regression analysis to measure the relationship between one independent and one

dependent variable, and multiple linear regression analysis to measure the relationship

between multiple independent variables and dependent variable (Hox et al., 2017).

Researchers use multiple hierarchical linear regression analysis for a step-wise

examination of the moderating effect of one or more variables for the relationship

between other independent variables and the dependent variable (Hox et al., 2017).

A simple linear regression model did not meet the needs of this study because this

study required examination of the relationship between three independent variables and

one dependent variable considering possible moderating effects (Mertler & Vannatta,

2017). Similarly, simple multiple regression linear regression did not meet the needs for

this study as this statistical method did not allow examination of the moderating effect

between the variables (Mertler & Vannatta, 2017). This study included the use of

multiple hierarchical linear regression analysis because of allowing examination of the

effect of each independent variable on the dependent variable.

Researchers also use several other types of statistical tests to examine quantitative

data (Mertler & Vannatta, 2017). Researchers use the Pearson correlation to measure

magnitude and direction between two variables, the t-test to compare means of variables

of two groups, ANOVA to compare means of variables of multiple groups, and Analysis

of covariance (ANCOVA) to compare the means of variables of multiple groups while

controlling for covariates (Mertler & Vannatta, 2017). ANOVA and ANCOVA are

limited to analysis of one dependent variable, and researchers extended both tests to

cover multiple dependent variables, hence multiple analysis of variance (MANOVA) and

multiple analysis of covariance (MANCOVA) (Mertler & Vannatta, 2017).

Because the Pearson correlation, t-tests, ANOVA, ANCOVA, MANOVA, and

MANCOVA necessitate normal distribution of the data, researchers use other types of

statistical tests with non-normal data, such as the Chi-square test for categorical data

analysis, and the Kendall or Spearman correlation test to examine the relationship

between rank-ordered data (Mertler & Vannatta, 2017). Following the recommendations

by Phillips et al. (2016) for single data source survey research, this study included a t-test

to compare the means of early and late respondents of the survey. Because the research

variables in this study were interval type metric variables, there was no further need for

statistical analyses for categorical or non-normal data.

Mertler and Vannatta (2017) stated that multiple hierarchical linear regression

analysis has the assumption of linearity, normality, orthogonality, and homoscedasticity

of the data. Outliers and missing or erroneous data can also negatively affect the

reliability of multiple hierarchical linear regression analysis (Wu, Jia, & Enders, 2015;

Zahari, Ramli, Moktar, & Zainol, 2014). The technique for data collection in this study

prevented the issue of missing data, as the survey required input to all questions before

allowing the participant to submit the survey. Following Curran’s guidelines (2016) for

survey data error analysis, statistical analysis excluded illogical data (length of tenure

exceeding the age) for the whole record.

The assumption of linearity means that the relationship between the independent

variable and the dependent variable is linear (Mertler & Vannatta, 2017). The assumption

of normality means that the data for each variable should follow a centered bell-shaped

curve when plotted on a graph (Mertler & Vannatta, 2017). Researchers examine

normality using quantile-quantile (Q-Q) plots or predicted probability (P-P) plot (Mertler

& Vannatta, 2017). An assumption of orthogonality means that the independent variables

are not significantly correlated (i.e., multicollinearity; Mertler & Vannatta, 2017).

Because independent variables in this study include age and length of tenure of

salespeople, it was logical to assume the possibility of multicollinearity between these

two variables.

This study included testing for multicollinearity using variance inflation factor

and tolerance analyses, and because multicollinearity was not present, there was no need

to address it with a robust ridge regression (Zahari et al., 2014). Similarly, as

recommended by Zahari et al., as the data included outliers, I supplemented multiple

hierarchical linear regression analysis with robust ridge regression method available via

the R-ridge regression plugin for SPSS. Because the chosen statistical method does not

involve addressing missing data, and the robust ridge regression tolerates outliers, there

were no additional needs for data cleaning (Mertler & Vannatta, 2017).

The assumption of homoscedasticity means that the homogeneity of variance of

independent variables is similar (Mertler & Vannatta, 2017). Residual scatterplot allows

visual examination of the assumption homoscedasticity between the predicted dependent

variable scores and the errors of prediction (Mertler & Vannatta, 2017). Therefore, this

study included residual scatterplot. Because residual scatterplot indicated that neither

homoscedasticity, normality nor linearity assumption was violated, I did not need to use

bootstrapping to address these violations. The bootstrapping technique is one of the data

resampling techniques, allowing researchers to mitigate data violations with multiple

hierarchical linear regression analyses (Sillabutra et al., 2016).

Study Validity

Researchers categorize quantitative studies validity into two broad categories:

internal validity and external validity (Neuman, 2014). Internal validity is an indicator of

how well research closes out alternative explanations of the results and minimizes the

risk of confounding (Yin, 2014). With correlational designs, researchers are not interested

in causal explanations nor perform any experiments (Neuman, 2014). Therefore, most

threats to internal validity do not apply to correlational designs. External validity is an

indicator of how well the findings of the research are generalizable to similar

environments outside the original scope of the research (Pearl & Bareinboim, 2014).

Because of these different aspects of internal and external validity, some scholars

assert that improving internal validity will limit external validity because of additional

controls (Daoud, 2019; Moser & Kalton, 2017). Similarly, increasing external validity

(ensuring study conclusions are generalizable over other populations) would require the

removal of context-specific controlling factors, which result in a decrease in internal

validity (Daoud, 2019). Some aspects of internal validity still apply for correlational

designs, such as nomological validity and statistical conclusion validity (Mitchell, 1985;

Neuman, 2014). Mitchell and Neuman also noted that selection bias, testing bias, and

common method bias might threaten the internal validity of correlation design studies.

The following discussion contains details of how this study included addressing threats to

internal and external validity.

Nomological validity ensures that the research uses the nomological network

(theoretical framework, methodological description, constructs, and relationships), which

have sound logical reasoning or well-established prior research (Straub et al., 2004). The

theoretical framework in this study was the self-efficacy theory, which is one of the most

studied psychological theories (Diener et al., 2014; Johnson & Jaramillo, 2017).

Similarly, numerous scholars interested in sales performance examined relationships

between quantifiable predictors and quantitative outputs (Verbeke et al., 2011). As

discussed in the research method -section, a mature theory with numerical data warranted

the use of the quantitative methodology. Also, earlier research findings indicated a

statistically significant relationship between the variables used in this study in other

settings (Bonney et al., 2014; Verbeke et al., 2011). Therefore, the nomological validity

of this study is high.

Statistical conclusion validity concerns the statistical reliability in the research

(Heale & Twycross, 2015). Straub et al. (2004) described three quality levels for

statistical conclusion validity. Scholars pursuing the highest level of statistical conclusion

validity should make sure that statistical conclusions in the research are confident,

alternative explanations are ruled out, and construct validity is high (Straub et al., 2004).

To ensure statistical conclusion validity, this study included the use of the same statistical

method as earlier scholars used for examining the relationship between sales performance

predictors and sales performance. Multiple hierarchical linear regression analysis with

three predictor variables reduced the risk of type I error and thus reduced the risk of

alternative explanations (Grömping, 2015; Zahari et al., 2014). I used the well-

established instrument to measure construct general self-efficacy and recommended

methods to test the factorial validity of the measure in this study. The justification of the

other independent variables (age and length of tenure of salespeople) included the basis

on prior research and their relationship to the general self-efficacy and sales performance.

I described different constructs that prior scholars used to measure construct sales

performance, and justified the use of annual quota attainment as a measure. Thus, the

statistical conclusion validity and construct validity are high.

Selection bias refers to the selection of participants to groups with experimental

studies (Moser & Kalton, 2017). However, it is very similar to nonrespondent-bias, which

refers to potential characteristical differences between participants and non-participants

of the research population (Schoenherr et al., 2015). Such bias may occur because

participation in the study is voluntary, and thus, the probability of inclusion in the study

is not the same for each member of the population (Schoenherr et al., 2015). Although the

risk of selection bias is a limitation with this study, several scholars concluded that

survey sampling results in a representative sample of the whole population if the sample

size is sufficient (af Wåhlberg & Poom, 2015; Coppock, 2019; Mullinix et al., 2015). I

calculated the required sample size for this study and obtained a sufficient number of

responses. Therefore, the risk of selection bias was low.

Testing bias or testing effect refers to the multiple measures affecting the outcome

of a test (Neuman, 2014). Although scholars often associate testing bias only with

experimental designs, a similar effect may occur with correlational designs if pilot study

participants are eligible for the subsequent main study (Vinson, Dale, & Jones, 2019). As

this study did not include a pilot study, testing bias was not an issue. Testing bias may

also occur if participants share their experiences of the completed survey with other

participants who have yet to complete the survey (Neuman, 2014). The survey invitation

did not reveal any participants’ contact information to other participants. Thus the risk of

testing bias was minimal with this study.

Common methods bias may occur if a researcher uses only one method for data

collection or if the data collection happens only at one point in time (Straub et al., 2004).

Therefore, the instrument usage may cause common method bias, rather than an actual

predisposition the instrument is measuring (Straub et al., 2004; Ylitalo, 2009). With

qualitative studies, researchers can reduce common method bias by complementing

interview data with naturalistic observations, group interviews, or document analysis

(Creswell, 2014).

Regarding quantitative studies, Ylitalo (2009) concluded that common method

bias is a frequently occurring problem, especially with survey studies, where common

method bias can inflate or attenuate the relationship between the variables. As noted by

Ylitalo, researchers cannot fully address common method bias with statistical remedies,

and thus, researchers should use multiple sources in data collection. If a researcher uses

only a single data source, Straub et al. (2004) recommended collecting data at two or

more points in time and using confirmatory factor analysis to reduce common methods

bias. However, as this study was cross-sectional, collecting the data at two or more points

in time was not optional. The guarantee for participant anonymity also prevented

longitudinal measures. To reduce the risk of common method bias, the survey question

order followed principles recommended by Tourangeau (2018b) to ensure that questions

for sales performance, age, and length of tenure each were on separate screen before

questions of general self-efficacy.

Some scholars recommend testing for common method bias with a one-factor test

(Fuller, Simmering, Atinc, Atinc, & Babin, 2016; Mertler & Vannatta, 2017). Because

this study contained only one variable with a latent factor (construct general self-

efficacy), one-factor testing was not a feasible method to analyze common method bias.

Furthermore, as the data for other variables (age, length of tenure, and sales performance

of salespeople) represented objective values rather than latent constructs, the risk of

common method bias was low. Phillips et al. (2016) recommended a comparison of early

and late respondents as a method to identify common method bias. I used a comparison

of responses between early and late respondents as an additional method to identify and

quantify common method bias.

The external validity of the research is a fundamental aspect of science (Pearl &

Bareinboim, 2014). A study with high external validity allows other scholars to apply and

generalize findings from that study (Pearl & Bareinboim, 2014). Consequently, from an

epistemological perspective, studies with high external validity help scholars in the

building of cumulative knowledge of the research topic (Kuhn, 1996). Although rigorous

statistical methods allow researchers to generalize findings of a sample over a particular

population, achieving high external validity requires thorough consideration of the whole

study (Pearl & Bareinboim, 2014).

With correlational research designs with high internal validity, the external

validity predominantly depends on the study population, and the representativeness of the

sample of the total population (Mitchell, 1985; Mullinix et al., 2015). Mullinix et al.

(2015) concluded that a sufficient sample size, combined with regression analysis with

multiple predictor variables, will yield similar results as using random samples. Because

this study included sufficient sample and regression analysis with multiple predictor

variables, non-probabilistic sampling was not an issue.

Because of using standard instruments to measure all variables, reported

descriptive statistics of the sample, and the population of the study consisted of a diverse

group of people working for THSP IT companies, I expected this study to have high

external validity among salespeople in the IT industry. However, the lack of universally

accepted construct of sales performance may limit the external validity of this study to

the settings where the construct is similar as in this study.

Transition and Summary

Section 2 covered the nature and structure of the research study and its design,

including the steps involved in collecting, validating, and analyzing the data and

protecting the participants. I described the role of the researcher and described the

population. I also justified the sampling method, instrument selection, data collection

techniques, and data analysis methods. Discussion included the typical limitations of

similar studies, namely the risk of common methods bias, and the reasons and risks of

using non-probabilistic sampling methods. I then provided a detailed explanation of

addressing the risks in this study and how prior scholars used similar techniques to

mitigate similar risks in their studies. At the end of Section 2, discussions included the

internal and external validity of this study.

Section 3 contained the presentation and analysis of the results and findings. The

section began by re-stating the purpose of the study, followed by a summary of the

findings. Additionally included were details of the data used in the study, including

descriptive statistics, statistical methods assumption test results as well as inferential

statistical results. I reported the results for null and alternative hypotheses and provided

an answer to the research question of this study. Followed by reporting the findings of

this study, I compared and contrasted the results of this study to the studies analyzed in

Section 2. Discussion also included the implication of the findings within the context of

the theoretical framework.

After analyzing this study results among other similar studies and within the

contextual framework, I provided a detailed discussion of how and why the findings of

this study may apply to both business practice improvement, as well as to positive social

change development. Supplementing this discussion were the recommendations for

action in business and further research. Section 3 concluded by reflecting my own

doctoral study process and the development of my scholarly thinking during the process.

Section 3 ended with the concluding statement of the study.

Section 3: Application to Professional Practice and Implications for Change

Introduction

The objective of this study was to examine the relationship between age, length of

tenure, general self-efficacy, and sales performance of salespeople. Many earlier scholars

concluded that high self-efficacy predicts high job performance within sales contexts;

however, studies addressing specifically the IT industry and measuring the general form

of self-efficacy are scarce (Joseph et al., 2015; Pousa & Mathieu, 2016). The research

question for this study addressed if there was a statistically significant relationship

between age, length of tenure, general self-efficacy, and sales performance of

salespeople.

Multiple hierarchical linear regression analysis indicated that a statistically

significant relationship exists between age, length of tenure, general self-efficacy, and

sales performance of salespeople. The robust ridge regression equation with all three

predictors had significant relationship to the sales performance R2 = .22, adjusted R2 =

.19, F(3,92) = 8.64, p < .001. Further hierarchical linear regression analysis indicated that

age and length of tenure had no significant relationship with sales performance and that

general self-efficacy was a sole statistically significant predictor for sales performance,

predicting a 19% variance of sales performance. Additionally, curve estimation indicated

that the relationship between general self-efficacy and sales performance is positive and

linear.

Presentation of the Findings

In this section, discussion includes the data collection and illustrate statistical

testing, the variables, the purpose of the tests, and their relationships to the hypotheses.

This section also includes descriptive statistics, evaluation of statistical assumptions, and

inferential statistical analyses. In analyzing the findings, I share how the findings answer

the research question.

Data Collection

The study included the use of SurveyMonkey to collect the survey data and

invited population members via LinkedIn group announcement posts about the survey.

Appendix E contains the LinkedIn post invitation to the survey. The survey was open

from July 31, 2019, through August 2, 2019, during which 103 people completed the

survey. Out of 103 responses, 96 survey responses were fully completed with no illogical

data. The survey design prevented the generation of missing data as each question was

mandatory, and if a participant canceled their participation, SurveyMonkey discarded the

data that the user had already filled in. The sample size of 96 fulfills the requirement of

this study (minimum sample size of 77) and results in statistical power of .88. Table 2

includes the descriptive statistics of the survey results.

Table 2

Descriptive Statistics of the Variables From the Survey

Mean SD

Sales performance 104.28% 16.12% 96

Age 38.92 8.740 96

Experience 7.35 4.693 96

GSES 30.08 3.227 96

Test of Assumptions

Linearity. A test for an assumption of data linearity included analysis of the P-P

plot (see Figure 2) and residual scatterplot (see Figure 3), as recommended by Mertler

and Vannatta (2017). Residual scatterplot indicating data homoscedasticity and P-P plot

indicating data normality confirmed the linearity of the data, and no need existed for

bootstrapping.

Normality. A test for normality included P-P plot analysis (see Figure 2).

Distribution of the residuals along the diagonal normality line confirmed the assumption

of data normality. There was no need to use bootstrapping because the data met the

assumption of normality.

Figure 2. P-P plot indicating the normality of the data.

Homoscedasticity. A residual scatterplot indicated that the homogeneity of the

variance of independent variables was similar (see Figure 3). There was no need to use

bootstrapping because the data met the assumption of homoscedasticity.

Figure 3. Residual scatterplot indicating homogeneity of variance.

Orthogonality. A variance inflation factor analysis confirmed the orthogonality

(i.e., absence of multicollinearity) assumption as all VIF values were significantly below

the threshold level of 10 (see Table 3). Mertler and Vannatta (2017) suggested the use of

a VIF threshold level of 10 to detect multicollinearity with regression analyses. Daoud

(2017) concluded that independent variables tolerance value below .10 also indicate

multicollinearity. Table 3 indicates that the tolerance values for each independent

variable exceed the threshold level of multicollinearity. In the absence of

multicollinearity, a standard multiple hierarchical linear regression method was

sufficient, and there was no need to address multicollinearity.

Table 3

Orthogonality Analysis Confirming Absence of Multicollinearity

Model

Unstandardized

coefficients

Standardized

coefficients

t Sig.

Collinearity

statistics

Std.

Error Beta Tolerance VIF

1 (Constant) 30.386 17.407 1.746 .084

Age .081 .248 .044 .328 .744 .469 2.134

Experience -.562 .484 -.164 -1.161 .249 .427 2.342

GSES 2.489 .495 .498 5.029 .000 .864 1.157

a. Dependent variable: Sales performance

Outliers. A case wise diagnostic test of outliers indicated the presence of two

outlier cases (data residing outside of three standard deviations from the mean), see Table

4. Because standard regression models do not tolerate outlier data, a robust ridge

regression analysis supplements the hierarchical linear regression analysis in this study.

Table 4

Case Wise Diagnostic Test Indicating Two Outliers in the Data

Case number Std. Residual

Sales_perform

ance

Predicted

value Residual

17 3.395 149.00% 99.86% 49.14%

94 -3.025 65.00% 108.77% -43.77%

a. Dependent variable: Sales performance

Reliability of General Self-Efficacy Measure

A Cronbach alpha test for general self-efficacy resulted in .70, which indicates

acceptable reliability for the scale (see Straub et al., 2004). However, subsequent factor

analysis indicated three latent factors, explaining a total of 54.6% of the variance of

general self-efficacy (FAC1: 29.3%, FAC2: 14.5%, FAC3: 10.8%; see Table 5). As

suggested by Straub et al. (2004), further analysis of the latent factors follows.

Table 5

actor Analysis of GSES Measure

Factor

Initial eigenvalues

Extraction sums of squared

loadings

Rotation sums of squared

loadings

Total

% of

variance

Cumulative

% Total

% of

variance

Cumulative

% Total

% of

variance

Cumulative

1 2.930 29.303 29.303 2.318 23.185 23.185 1.681 16.810 16.810

2 1.454 14.538 43.841 .790 7.900 31.085 1.054 10.539 27.349

3 1.080 10.804 54.645 .487 4.872 35.957 .861 8.608 35.957

4 .874 8.742 63.387

5 .838 8.383 71.770

6 .723 7.234 79.004

7 .613 6.131 85.135

8 .552 5.522 90.657

9 .497 4.969 95.626

10 .437 4.374 100.000

ote. Extraction method: Maximum likelihood.

Successful factor analysis necessitates an examination of sampling adequacy and

sphericity of the data (Mansolf & Reise, 2016; Thompson, 2004). Initial examination of

the factorability of the 10 GSES survey questions included use of the Kaiser-Meyer-

Olkin (KMO) test for the sampling adequacy and Bartlett’s test of sphericity. The KMO

measure of sampling adequacy was 0.748, above the commonly recommended value of

0.600 (Watkins, 2018). Bartlett’s test of sphericity was also significant (χ2 (45) = 155.99,

p < .01) which indicates presence of latent factors (rather than identities; Mertler &

100

Vannatta, 2017; Watkins, 2018). Watkins (2018) concluded that in the factor analysis,

each item must have communality (proportion of item’s variance explained by the

extracted factors) of at least .300, otherwise, the item should be discarded. After the

removal of three items with communality of below .300, the factor analysis still indicated

three latent constructs (Eigenvalues > 1). Because removal of items would deviate the

GSES measure from the standardized instrument, the factor analysis with all the 10 items

was appropriate in this study.

101

Table 6

GSES Item Communalities

Initial Extraction

I can always manage to solve difficult problems if I try hard

enough.

.356 .480

If someone opposes me. I can find the means and ways to get what

I want.

.251 .432

It is easy for me to stick to my aims and accomplish my goals. .231 .321

I am confident that I could deal efficiently with unexpected events. .304 .382

Thanks to my resourcefulness. I know how to handle unforeseen

situations.

.223 .261

I can solve most problems if I invest the necessary effort. .220 .459

I can remain calm when facing difficulties because I can rely on my

coping abilities.

.146 .197

When I am confronted with a problem. I can usually find several

solutions.

.181 .370

If I am in trouble. I can usually think of a solution. .194 .206

I can usually handle whatever comes my way. .370 .486

ote. Extraction method: Maximum likelihood.

A factor analysis of GSES measure with maximum likelihood extraction method

and direct oblimin factor rotation resulted in three factors with eigenvalue exceeding 1

(see Figure 4), which is a typical threshold value for identifying latent factors with factor

analysis (Watkins, 2018). Mansolf and Reise (2016) recommended oblique rotation

methods for factor analyses with internal item correlation. Thus, the factor analysis of

GSES measure used oblique rotation method (direct oblimin) and maximum likelihood

102

method for factor extraction. The maximum likelihood method was suitable for factor

extraction as it is more accurate than simpler methods, such as principal component

analysis (Mertler & Vannatta, 2017).

Figure 4. Scree plot of GSES latent factors.

Table 7 contains individual item factor loadings in an unrotated solution, and

Table 8 contains individual item factor loadings (after Kaiser normalization) in a rotated

solution. Kaiser normalization results re-normalized items after the rotation, which allows

item-level examination between rotated and unrotated solution (Mertler & Vannatta,

2017).

103

Table 7

Unrotated Factor Structure of GSES Measure

Factor

1 2 3

I can always manage to solve difficult problems if I try hard

enough.

.622 -.279 -.127

If someone opposes me. I can find the means and ways to get

what I want.

.520 -.048 .399

It is easy for me to stick to my aims and accomplish my

goals.

.386 .390 -.141

I am confident that I could deal efficiently with unexpected

events.

.602 .124 .070

I can solve most problems if I invest the necessary effort. .475 .097 -.473

When I am confronted with a problem. I can usually find

several solutions.

.001 .596 .118

If I am in trouble. I can usually think of a solution. .439 -.041 -.110

I can usually handle whatever comes my way. .683 -.013 .140

Thanks to my resourcefulness. I know how to handle

unforeseen situations.

.460 -.192 .110

I can remain calm when facing difficulties because I can rely

on my coping abilities.

.232 .372 .070

ote. Extraction Method: Maximum Likelihood.

a. 3 factors extracted. 5 iterations required.

104

Table 8

Structure Matrix of Rotated GSES Factors

Factor

1 2 3

I can always manage to solve difficult problems if I try hard

enough.

.578 -.145 -.524

If someone opposes me. I can find the means and ways to get

what I want.

.625 .137 -.107

It is easy for me to stick to my aims and accomplish my goals. .246 .441 -.403

I am confident that I could deal efficiently with unexpected

events.

.561 .269 -.401

I can solve most problems if I invest the necessary effort. .263 .128 -.675

When I am confronted with a problem. I can usually find

several solutions.

-.057 .592 .042

If I am in trouble. I can usually think of a solution. .375 .044 -.394

I can usually handle whatever comes my way. .682 .166 -.405

Thanks to my resourcefulness. I know how to handle

unforeseen situations.

.495 -.061 -.251

I can remain calm when facing difficulties because I can rely

on my coping abilities.

.177 .422 -.146

ote. Extraction Method: Maximum Likelihood.

Rotation Method: Oblimin with Kaiser Normalization.

Because factor analysis of GSES measure revealed three latent factors, a

structural equation modeling (SEM) would have been an appropriate method to further

analyze the relationship between independent variables of this study, GSES latent factors,

105

and sales performance. However, a SEM method requires at least 155 responses with the

number of identified latent factors and factor loadings (see table 7) (Wolf, Harrington,

Clark, & Miller, 2013). Therefore, this study did not include a SEM analysis, and the

discussion in the section Recommendations for Further Research includes implications of

excluded SEM analysis.

Common method bias analysis. After splitting the participants’ responses to two

groups: early and late respondents, I examined if the groups statistically differed from

each other, as suggested by Agnihotri et al. (2016). Table 9 includes descriptive statistics

of the early and late respondents.

Table 9

escriptive Statistics of Early and Late Respondents Data

Response Time

Mean

Std.

Deviation Std. Error Mean

Experience Early 48 7.17 5.269 .760

Late 48 7.54 4.084 .589

Age Early 48 38.00 9.351 1.350

Late 48 39.83 8.078 1.166

GSES Early 48 30.85 3.832 .553

Late 48 29.31 2.271 .328

Sales

erformance

Early 48 104.82% 17.71% 2.56%

Late 48 103.75% 14.53% 2.10%

106

Levene’s test for equal variances indicated that population variances between the

groups were similar for age, length of tenure, and sales performance of salespeople (see

Table 10). A statistically significant difference of GSES between early and late

respondents (t(94) = 2.4, p = .02)) indicated that late respondents’ mean GSES was 1.54

points lower than early respondents’ mean GSES. A subsequent Harman’s single factor

test indicated that the total explained variance with all predictor variables is .223, which

is significantly below the threshold level of .50 and indicates that there is no common

method bias (Fuller et al., 2016).

107

Table 10

ndependent Samples t Test

etween Earl

and Late Respondents

Levene's Test

for Equality of

Variances t test for equality of means

Sig. t df

Sig. (2-

tailed)

Mean

differ-

ence

Std.

error

differ-

ence

95% CI of the

difference

Lower Upper

Experience Equal

variances

assumed

.080 .778 -.390 94 .698 -.375 .962 -2.285 1.535

Equal

variances

not assumed

-.390 88.50 .698 -.375 .962 -2.287 1.537

Age Equal

variances

assumed

2.353 .128 -1.03 94 .307 -1.833 1.784 -5.375 1.708

Equal

variances

not assumed

-1.03 92.05 .307 -1.833 1.784 -5.376 1.709

GSES Equal

variances

assumed

5.348 .023 2.398 94 .018 1.542 .643 .265 2.818

Equal

variances

not assumed

2.398 76.39 .019 1.542 .643 .261 2.822

Sales

erformance

Equal

variances

assumed

1.239 .268 .322 94 .748 1.06% 3.31% -5.50% 7.63%

Equal

variances

not assumed

.322 90.54 .748 1.06% 3.31% -5.50% 7.63%

108

Inferential Results

Multiple hierarchical linear regression analysis indicated that a statistically

significant relationship exists between age, length of tenure, general self-efficacy, and

sales performance of salespeople. The robust ridge regression equation with all three

predictors was significantly related to the sales performance R2 = .22, adjusted R2 = .19,

F(3,92) = 8.64, p < .001 (see Table 11). Further hierarchical linear regression analysis

indicated that age and length of tenure had no significant relationship with sales

performance and that general self-efficacy was a sole statistically significant predictor for

sales performance, predicting a 19% variance of sales performance (see Table 12).

Because the age and length of tenure of salespeople had no significant relationship with

sales performance, they did not moderate the relationship between GSES and sales

performance. Additionally, curve estimation indicated that the relationship between

general self-efficacy and sales performance is positive and linear (see Figure 5).

109

Table 11

egression Analysis With All Predictor Variables

Model R R Square

Adjusted

R Square

Std. Error

of the

Estimate

Change Statistics

R Square

Change

Change df1 df2

Sig. F

Change

1 .469a .220 .194 14.47% .220 8.642 3 92 .000

a. Predictors: (Constant), GSES, Age, Experience

b. Dependent Variable: Sales performance

Table 12

ierarchical Regression Analysis Indicating Significance of GSES

Model R R Square

Adjusted

R Square

Std. Error

of the

Estimate

Change Statistics

R Square

Change

Change df1 df2

Sig. F

Change

1 .074a .005 -.016 16.25% .005 .253 2 93 .777

2 .469

.220 .194 14.47% .214 25.29 1 92 .000

a. Predictors: (Constant), Experience, Age

b. Predictors: (Constant), Experience, Age, GSES

c. Dependent Variable: Sales performance

Because of the presence of three latent factors with GSES measure, this study also

includes further stepwise regression analysis for each latent GSES factor (see Table 13).

110

Table 13

egression Results With GSES Latent Factors

Model R R Square

Adjusted R

Square

Std. Error

of the

Estimate

Change Statistics

R Square

Change F Change df1 df2

Sig. F

Change

1 .074a .005 -.016 16.25% .005 .253 2 93 .777

2 .311

.097 .067 15.57% .091 9.301 1 92 .003

3 .507c .258 .225 14.19% .161 19.713 1 91 .000

4 .508d .259 .217 14.26% .001 .122 1 90 .727

a. Predictors: (Constant), Experience, Age

b. Predictors: (Constant), Experience, Age, GSES latent factor FAC1 for analysis 1

c. Predictors: (Constant), Experience, Age, GSES latent factor FAC1 analysis 1, GSES

latent factor FAC2 for analysis 1

d. Predictors: (Constant), Experience, Age, GSES latent factor FAC1 for analysis 1, GSES

latent factor FAC2 for analysis 1, GSES latent factor FAC3 for analysis 1

e. Dependent Variable: Sales performance

A latent factor analysis indicated that general self-efficacy, measured with GSES,

resulted in three distinct latent factors: FAC1, FAC2, and FAC3. A stepwise regression

equation with age and length of tenure of salespeople added in step one, and each latent

factor (FAC1, FAC2, and FAC3) added in subsequent steps resulted the total effect of R2

= .258, adjusted R2 = .225, F(4,91) = 19.713, p < .001. Age or length of tenure of

salespeople had no significant relationship with sales performance. Similarly, a third

latent factor (FAC3) had no significant relationship with sales performance.

111

Figure 5. Scatterplot with sales performance and GSES.

Summary

The research question for this study was: what is the relationship between age,

tenure, general self-efficacy, and sales performance of salespeople? The null hypothesis

in this study was that there is no statistically significant relationship between the age,

tenure, general self-efficacy, and sales performance of salespeople. An alternative

hypothesis was that there is a statistically significant relationship between age, tenure,

general self-efficacy, and sales performance of salespeople. Multiple hierarchical linear

regression analysis indicated that a statistically significant relationship exists between

age, length of tenure, general self-efficacy, and sales performance of salespeople.

Therefore, the null hypothesis of this study was rejected, and an alternative hypothesis

was accepted.

112

The robust ridge regression equation with all three predictors was significantly

related to the sales performance R2 = .22, adjusted R2 = .19, F(3,92) = 8.64, p < .001.

Further hierarchical linear regression analysis indicated that age and length of tenure had

no significant relationship with sales performance and that general self-efficacy was a

sole statistically significant predictor for sales performance, predicting a 19% variance of

sales performance. Additionally, curve estimation indicated that the relationship between

general self-efficacy and sales performance is positive and linear.

Factor analysis of the general self-efficacy scale indicated that the GSES

instrument captured three latent constructs instead of one. Thus, the results of this study

indicate that the GSES measure is not unidimensional, at least with the sample of this

study. First, two of the latent factors of GSES fully explained the total GSES predicted

variance in sales performance. The third latent factor had strongest factor loadings for

GSES survey items #2 (If someone opposes me, I can find the means and ways to get

what I want, β = .399) and #5 (I can solve most problems if I invest the necessary effort,

β = -.473) but the delta effect (two versus three factors for GSES) was nonsignificant (see

table 13, model 4). The survey item #5 (I can solve most problems if I invest the

necessary effort) loading for third latent factor was almost the opposite than for the first

latent factor which indicates that at least within this sample, a variance of responses for

the survey item #6 significantly differed from the total GSES variance. These findings are

similar to those of Barahona et al. (2018), who noted that GSES survey item #5 has a

wide variance between different geographical samples. Further analysis of the GSES

latent factors is outside of the scope of this study.

113

Applications to Professional Practice

Companies strive to develop their sales function as the sales process directly

affects organizational performance, and ultimately, companies' success or failure (Guenzi

et al., 2016). Although scientific analysis of sales success factors has become

mainstream, very few IT companies apply the knowledge into their sales processes

(Akhter et al., 2014). Hamstra et al. (2015) concluded that personal characteristics of

salespeople predict up to 38% variance of sales performance, and the reviewed literature

in this study indicated that predictors used in this study explain up to 29% variance in

sales performance. Thus, understanding the correlates of sales performance may help

companies to improve their organizational performance.

The use of stable predictors, as in this study, provides significant benefits for the

business practice as the sales managers within the IT industry can measure the values

before making recruitment decisions (Brewster & Hegewisch, 2017). The predictors used

in this study included age, length of tenure, and general self-efficacy of salespeople, and

all these measures are stable over time. Only general self-efficacy was a statistically

significant predictor for sales performance, explaining a 19.4% variance in sales

performance. Because measuring general self-efficacy is a simple 10-item survey, it is

easy to administer for potential sales recruits within the IT industry (Tourangeau, 2018b).

Because the population in this study consisted of salespeople from the whole THSP IT

sector, it is logical to assume generalizability of the predictive power of general self-

efficacy within the THSP IT sector.

114

Implications for Social Change

This study has several significant implications for positive social change; the

following discussion includes a description of these implications by Elkington’s (1998)

categories of people, planet, and profit. First, the results of this study indicate no

relationship exists between age or length of tenure and sales performance of salespeople

within THSP companies of the IT industry. Therefore, an industry-wide phenomenon of

age discrimination has no basis from the perspective of sales performance. From an

individual salespeople’s perspective, the findings of this study are positive – neither the

inevitable accumulation of age nor the length of tenure, predict decreasing sales

performance results. Thus, the findings of this study provide a scientific basis for non-

discriminatory practices.

Second, mere knowledge of the non-relationship between the age and sales

performance may help salespeople to achieve higher sales results. This effect is a logical

result of antecedents for task-specific self-efficacy. As Bandura (1977) concluded, a

positive belief in one’s capability to perform certain task increases the likelihood that an

individual can perform the task. Therefore the results of this study are beneficial for

improving salespeople’s task-specific self-efficacy (of being able to accomplish their

sales quota). Based on recent analysis by Barbaranelli et al. (2018) of the predictors of

job performance, the predictive value of task-specific self-efficacy is nearly as high as

general self-efficacy (task-specific β = .324, p < .001; general β = .394, p < 001). Thus,

sales managers of THSP IT companies should not only apply the findings of this study in

the sales recruitment process but also for leading salespeople in their jobs.

115

Finally, supporting and keeping aging personnel within the company (instead of

letting go) may have a positive effect on the environment. For example, Meyer (2014)

examined the effect of unemployment to pro-environmental behavior among 29.539

European people and concluded that unemployment negatively predict purchase of

environmentally labeled products (β = -.051, p < .001), purchase of local products (β = -

.058, p < .001), and reduce of energy usage (β = -.078, p < .001). People aged over 50

who face involuntary job separation face significant issues in trying to get employed

again. Johnson and Gosselin (2018) concluded that 54% of the people aged over 50 who

lost their job were still unemployed after 6 months. If these results are generalizable on

salespeople who work with THSP companies, it means that discrimination by age is

illegal (in many countries; Doron & Georgantzi, 2018; Papke, 2018), unwarranted from a

sales performance perspective, and also environmentally harmful.

Recommendations for Action

The results of this study provide a scientific understanding of the relationship

between age, length of tenure, general self-efficacy, and sales performance among

salespeople working with THSP IT companies. This study supports three

recommendations for the sales managers of THSP IT companies: (a) an examination of

the use of general self-efficacy upon recruitment decisions, (b) support of non-

discriminatory practices based on age, or experience, and (c) increasing task-specific self-

efficacy of the salespeople. Following these recommendations increase salespeople’s

sales performance, and as a result, an organizational success. Similarly, following these

116

recommendations will help to avoid illegal, discriminatory practices still present within

the IT industry (Kelly, 2019; Snapp, 2019).

I will present the executive summary of this study at the LinkedIn group The IT

Sales Global Community, which is globally the largest professional community for the IT

salespeople and sales managers. The raw data used in this study is available for further

analysis and examination upon request.

Recommendations for Further Research

Future researchers may further explore the predictors of high sales performance in

several ways. For example, this study did not consist examination of the quota fairness

between the participants. Earlier scholars concluded that an assumption of quota fairness

may be unwarranted and that some managers tend to favor people they have recruited

themselves (Johnson & Jaramillo, 2017; Kwak et al., 2019). A further quota analysis

could also reveal significant changes in the composition of the quota setting. For

example, some companies set quotas by revenue only, whereas other companies construct

quotas combining elements such as revenue, profit, and strategic initiatives (Deeter-

Schmelz, 2016).

Similarly, the measure of general self-efficacy could be improved. This study

included the use of GSES instrument, whereas more recent instruments could yield

improved results, and a combination of several instruments could provide more insights

into the factor structure of general self-efficacy (Barahona et al., 2018). This study

indicated that general self-efficacy, measured with GSES, does not result in a

unidimensional measure. Thus, an in-depth factor analysis of the data obtained with this

117

study (or subsequent studies) could advance understanding within the theoretical

framework of general self-efficacy. Also, because GSES measure of general self-efficacy

may result multiple latent factors, future studies could include SEM analysis for

increased understanding of the relationship between latent factors of general self-

efficacy, and other research variables.

A question of prevalence is a frequent discussion topic with social sciences as

populations may have specific, prevalent characteristics. Although the statistical analysis

in this study did not suffer from multicollinearity, the descriptive statistics indicated that

variables age and length of tenure of salespeople were correlated. A visual inspection (see

Figure 6), shows that all participants of this study, who were at least 40 years old, had at

least four years of experience from the face-to-face sales job. This finding may indicate

that there were very few new salespeople over 40 years old in the population of this

study; or that they did not respond to the survey. To address any potential limitations of

prevalence or sample representativeness, future studies could include an examination of

populations with known demographics.

118

Figure 6. Scatterplot of experience and age of participants

Also, longitudinal studies with two or more measures of all variables could be

beneficial. As this study was cross-sectional, the assumption of the longitudinal stability

of self-efficacy had a basis on existing literature rather than the findings of this study.

Similarly, the assumption of the predictive power of self-efficacy had a basis on the

theory of self-efficacy and findings from earlier studies. Scholars interested specifically

of the predictive power of general self-efficacy should, therefore, measure participants’

sales performance, and general self-efficacy over longer periods.

The population of this study consisted of face-to-face sales professionals working

for THSP IT companies worldwide. Replication studies with different populations or

within different geographic boundaries could reveal different relationships between the

study variables. If multiple replication studies would result in similar findings as this

119

study did, the myth of the detrimental effect of old age to sales performance could be

dissolved. Finally, Piervincenzi et al. (2017) recently found that certain types of

physiological training might increase neurophysiological capabilities and also general

self-efficacy. If such increase of general self-efficacy holds true with larger populations,

it will open completely new research venues: what causes the effect in general self-

efficacy?, how permanent are the changes?, and does the relationship between general

self-efficacy and sales performance (or other outcome measures) change accordingly?

Reflections

When first starting doctoral studies in 2015, my main intention was to understand

predictors for high-performance sales scientifically. During years as a sales manager and

sales director, I observed that certain people seemed to perform better in sales than the

others consistently. As my interest in understanding that phenomena grew larger, I

wanted to research the topic as much as I could. Thus, the topic chosen for this study

emerged from my professional background and business need.

The journey of completing this research study was one of arduous work, frequent

eureka moments, and eventually joy and fulfillment. From the beginning of this research

project, my professional colleagues were interested in the outcomes of the research

project. Similarly, professional salespeople with whom I discussed the topic, shared their

personal experiences and views. However, very few people in sales were familiar with

the theoretical background of this study, and even fewer were familiar with the research

method and design details – or how to reach conclusions after analyzing some topic.

120

Although IT sales management professionals recognized the importance of self-

belief of salespeople, the thought of having a quantifiable and general measure of self-

efficacy was unknown and unaccounted variable to most. Hence, the results of this study

may be useful for IT sales managers for salespeople recruitment and for understanding

the complexity of success in sales.

Most of my professional colleagues agreed that companies do not frequently

recruit older people to sales who have no previous experience from sales. However, this

study indicated that there is no statistically significant relationship between sales

performance and age of salespeople. Thus, a belief that a salespeople’s age would predict

a decrease in sales performance is a false belief.

Conclusion

The search for predictors for high sales performance continues to interest both

scholars and professional practitioners (Inyang & Jaramillo, 2019). Companies whose

sales performance is higher than their competitors outperform their rivals (Johnson &

Jaramillo, 2017). Within highly competitive industries, such as the IT industry,

understanding the predictors for high sales performance is crucial for the sales managers.

This study indicated that the general self-efficacy of IT salespeople predicts a 19%

variance in sales performance. This study also indicated that neither age nor length of

tenure of salespeople had a significant relationship with sales performance. The findings

of this study may help sales managers of IT companies to improve their sales

performance by informed recruitment. Similarly, findings of this study highlighted that

discrimination based on age or experience of salespeople within the IT industry remains

121

both unwarranted and even illegal (in many countries; Doron & Georgantzi, 2018; Papke,

2018).

122

References

af Wåhlberg, A. E., & Poom, L. (2015). An empirical test of nonresponse bias in internet

surveys. Basic and Applied Social Psychology, 37, 336–347.

https://dx.doi.org/10.1080/01973533.2015.1111212

Agnihotri, R., Dingus, R., Hu, M. Y., & Krush, M. T. (2016). Social media: Influencing

customer satisfaction in B2B sales. Industrial Marketing Management, 53, 172–

180. https://dx.doi.org/10.1016/j.indmarman.2015.09.003

Aguinis, H., Ramani, R. S., & Alabduljader, N. (2018). What you see is what you get?

Enhancing methodological transparency in management research. Academy of

Management Annals, 12(1), 83–110. https://dx.doi.org/10.5465/annals.2016.0011

Akhter, S., Rahman, N., & Rahman, M. N. (2014). Competitive strategies in the

computer industry. International Journal of Technology Diffusion, 5(1), 73–88.

https://dx.doi.org/10.4018/ijtd.2014010106

Alessandri, G., Borgogni, L., Schaufeli, W. B., Caprara, G. V., & Consiglio, C. (2015).

From positive orientation to job performance: The role of work engagement and

self-efficacy beliefs. Journal of Happiness Studies, 16, 767–788.

https://dx.doi.org/10.1007/s10902-014-9533-4

Allen, D. S. (2016). The impact of shortening a long survey on response rate and

response quality (Doctoral dissertation). Retrieved from ProQuest Dissertations &

Theses Database. (UMI No. 10179189)

Allport, G. W. (1961). Pattern and growth in personality. New York, NY: Holt, Rinehart,

& Winston.

123

Astivia, O. L. O., & Zumbo, B. D. (2019). Heteroskedasticity in multiple regression

analysis: What it is, how to detect it and how to solve it with applications in R and

SPSS. Practical Assessment, Research & Evaluation, 24(1), 1–16. Retrieved from

http://pareonline.net/

Atefi, Y., Ahearne, M., Maxham, J. G., Donavan, D. T., & Carlson, B. D. (2018). Does

selective sales force training work? Journal of Marketing Research, 55, 722–737.

https://dx.doi.org/10.1177/0022243718803096

Babatunde, A. Y. (2016). Perceived self-efficacy and dispositional optimism in leaders’

behavioral escalation of commitment (Doctoral dissertation). Retrieved from

ProQuest Dissertations & Theses Database. (UMI No. 1784308954)

Bakker, A. B. (2008). The work-related flow inventory: Construction and initial

validation of the WOLF. Journal of Vocational Behavior, 72, 400–414.

https://dx.doi.org/10.1016/j.jvb.2007.11.007

Bakker, A. B., & Demerouti, E. (2014). Job demands-resources theory. In C. L. Cooper

(Ed.), Wellbeing (pp. 1–28).

https://dx.doi.org/10.1002/9781118539415.wbwell019

Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change.

Psychological Review, 84(2), 191. https://dx.doi.org/10.1037/0033-295x.84.2.191

Bandura, A. (1997). Self-efficacy: The exercise of control. New York, NY: W.H.

Freeman.

124

Bandura, A. (2012). On the functional properties of perceived self-efficacy revisited.

Journal of Management, 38(1), 9–44.

https://dx.doi.org/10.1177/0149206311410606

Barahona, G. V., González García, N., Sánchez-García, A. B., Barba, M. S., & Galindo-

Villardón, M. P. (2018). Seven methods to determine the dimensionality of tests:

Application to the General Self-Efficacy Scale in twenty-six countries.

Psicothema, 442–448. https://dx.doi.org/10.7334/psicothema2018.113

Barbaranelli, C., Fida, R., Paciello, M., & Tramontano, C. (2018). ‘Possunt, quia posse

videntur’: They can because they think they can. Development and validation of

the Work Self-Efficacy scale: Evidence from two studies. Journal of Vocational

Behavior, 106, 249–269. https://dx.doi.org/10.1016/j.jvb.2018.01.006

Beauvais, C. (2016). Californian genius: Lewis Terman’s gifted child in regional

perspective. Paedagogica Historica, 52, 748–765.

https://dx.doi.org/10.1080/00309230.2016.1243138

Bellizzi, J. A., & Bristol, T. (2005). Supervising the unethical selling behavior of top

sales performers: Assessing the impact of social desirability bias. Journal of

Business Ethics, 57, 377–388. https://dx.doi.org/10.1007/s10551-004-6589-9

Beltrán-Martín, I., Bou-Llusar, J. C., Roca-Puig, V., & Escrig-Tena, A. B. (2017). The

relationship between high performance work systems and employee proactive

behaviour: Role breadth self-efficacy and flexible role orientation as mediating

mechanisms: HPWS and proactive behaviour. Human Resource Management

Journal, 27, 403–422. https://dx.doi.org/10.1111/1748-8583.12145

125

Benson, A. (2015). Do agents game their agents’ behavior? Evidence from sales

managers. Journal of Labor Economics, 33, 863–890.

https://dx.doi.org/10.1086/681107

Bentley, F. R., Daskalova, N., & White, B. (2017). Comparing the reliability of Amazon

Mechanical Turk and Survey Monkey to traditional market research surveys.

Proceedings of the 2017 CHI Conference Extended Abstracts on Human Factors

in Computing Systems - CHI EA ’17, 1092–1099.

https://dx.doi.org/10.1145/3027063.3053335

Bodla, M. A., & Naeem, B. (2014). Creativity as mediator for intrinsic motivation and

sales performance. Creativity Research Journal, 26, 468–473.

https://dx.doi.org/10.1080/10400419.2014.961783

Bonney, L., Plouffe, C. R., & Wolter, J. (2014). “I think I can…I think I can”: The

impact of perceived selling efficacy and deal disclosure on salesperson escalation

of commitment. Industrial Marketing Management, 43, 826–839.

https://dx.doi.org/10.1016/j.indmarman.2014.04.010

Bonsaksen, T., Lerdal, A., Heir, T., Ekeberg, Ø., Skogstad, L., Grimholt, T. K., & Schou-

Bredal, I. (2018). General self-efficacy in the Norwegian population: Differences

and similarities between sociodemographic groups. Scandinavian Journal of

Public Health, 140349481875670. https://dx.doi.org/10.1177/1403494818756701

Bosco, F. A., Aguinis, H., Singh, K., Field, J. G., & Pierce, C. A. (2015). Correlational

effect size benchmarks. Journal of Applied Psychology, 100, 431–449.

https://dx.doi.org/10.1037/a0038047

126

Brealey, R. A., Myers, S. C., & Marcus, A. J. (2018). Fundamentals of corporate finance

(9th ed.). New York, NY: McGraw-Hill.

Brewster, C., & Hegewisch, A. (2017). Policy and practice in European human resource

management: The Price Waterhouse Cranfield survey (1st ed.). New York, NY:

Routledge.

Bromley, E., Mikesell, L., Jones, F., & Khodyakov, D. (2015). From subject to

participant: Ethics and the evolving role of community in health research.

American Journal of Public Health, 105, 900–908.

https://dx.doi.org/10.2105/AJPH.2014.302403

Cabooter, E., Weijters, B., Geuens, M., & Vermeir, I. (2016). Scale format effects on

response option interpretation and use. Journal of Business Research, 69, 2574–

2584. https://dx.doi.org/10.1016/j.jbusres.2015.10.138

Cai, M. Y., Lin, Y., & Zhang, W. J. (2016). Study of the optimal number of rating bars in

the Likert scale. Proceedings of the 18th International Conference on Information

Integration and Web-Based Applications and Services - IiWAS ’16, 193–198.

https://dx.doi.org/10.1145/3011141.3011213

Campbell, D. T., & Stanley, J. C. (2010). Experimental and quasi-experimental designs

for research. Mason, OH: Cengage Learning.

Campbell, J. P., McCloy, R. A., Oppler, S. H., & Sager, C. E. (1993). A theory of

performance. In N. Schmitt & W. C. Borman, Personnel Selection in

Organizations (pp. 35–70). San Francisco, CA: Jossey-Bass.

127

Carter, R. E., Dixon, A. L., & Moncrief, W. C. (2008). The complexities of sales and

sales management research: A historical analysis from 1990 to 2005. Journal of

Personal Selling and Sales Management, 28, 403–420.

https://dx.doi.org/10.2753/PSS0885-3134280405

Carter, W. R., Nesbit, P. L., Badham, R. J., Parker, S. K., & Sung, L.-K. (2016). The

effects of employee engagement and self-efficacy on job performance: A

longitudinal field study. The International Journal of Human Resource

Management, 1–20. https://dx.doi.org/10.1080/09585192.2016.1244096

Cerasoli, C. P., Nicklin, J. M., & Ford, M. T. (2014). Intrinsic motivation and extrinsic

incentives jointly predict performance: A 40-year meta-analysis. Psychological

Bulletin, 140, 980–1008. https://dx.doi.org/10.1037/a0035661

Chemolli, E., & Gagné, M. (2014). Evidence against the continuum structure underlying

motivation measures derived from self-determination theory. Psychological

Assessment, 26, 575–585. https://dx.doi.org/10.1037/a0036212

Chen, G., Gully, S. M., & Eden, D. (2001). Validation of a new general self-efficacy

scale. Organizational Research Methods, 4(1), 62–83.

https://dx.doi.org/10.1177/109442810141004

Cheng, P.-Y., & Chiou, T.-T. (2016). The relationship between positive illusion and self-

efficacy for life insurance salesman sales performance. Commerce &

Management Quarterly, 17(3), 269–286. Retrieved from

http://www.cm.yuntech.edu.tw/

128

Childs, S., McLeod, J., Lomas, E., & Cook, G. (2014). Opening research data: Issues and

opportunities. Records Management Journal, 24(2), 142–162.

https://dx.doi.org/10.1108/RMJ-01-2014-0005

Churchill, G. A., Ford, N. M., Hartley, S. W., & Walker, O. C. (1985). The determinants

of salesperson performance: A meta-analysis. Journal of Marketing Research,

103–118. https://dx.doi.org/10.2307/3151357

Cobb-Clark, D. A., & Schurer, S. (2012). The stability of big-five personality traits.

Economics Letters, 115(1), 11–15.

https://dx.doi.org/10.1016/j.econlet.2011.11.015

Comptia. (2018). IT industry outlook 2018. Retrieved from Comptia website

https://www.comptia.org/resources/it-industry-trends-analysis

Cope, D. (2015). Conducting pilot and feasibility studies. Oncology Nursing Forum,

42(2), 196–197. https://dx.doi.org/10.1188/15.ONF.196-197

Coppock, A. (2019). Generalizing from survey experiments conducted on mechanical

turk: A replication approach. Political Science Research and Methods, 7, 613–

628. https://dx.doi.org/10.1017/psrm.2018.10

Creswell, J. W. (2014). Research design: Qualitative, quantitative, and mixed methods

approaches (4th ed.). Thousand Oaks, CA: Sage.

Creswell, J. W., & Poth, C. N. (2018). Qualitative inquiry & research design: Choosing

among five approaches (4th ed.). Thousand Oaks, CA: Sage.

Cronbach, L. J., & Meehl, P. E. (1955). Construct validity in psychological tests.

Psychological Bulletin, 52(4), 281–302. https://dx.doi.org/10.1037/h0040957

129

Čukić, I., Brett, C. E., Calvin, C. M., Batty, G. D., & Deary, I. J. (2017). Childhood IQ

and survival to 79: Follow-up of 94% of the Scottish Mental Survey 1947.

Intelligence, 63, 45–50. https://dx.doi.org/10.1016/j.intell.2017.05.002

Curran, P. G. (2016). Methods for the detection of carelessly invalid responses in survey

data. Journal of Experimental Social Psychology, 66, 4–19.

https://dx.doi.org/10.1016/j.jesp.2015.07.006

Cutler, T. L., Murphy, K., Hopper, J. L., Keogh, L. A., Dai, Y., & Craig, J. M. (2015).

Why accurate knowledge of zygosity is important to twins. Twin Research and

Human Genetics, 18(3), 298–305. https://dx.doi.org/10.1017/thg.2015.15

Cyniak-Cieciura, M., Popiel, A., & Zawadzki, B. (2015). General self-efficacy level and

changes in negative post-traumatic stress disorder (PTSD) symptoms among

motor vehicle accident survivors after PTSD therapy. Studia Psychologiczne, 53,

18–29. https://dx.doi.org/10.2478/V10167-010-0117-8

Dagher, G. K., Chapa, O., & Junaid, N. (2015). The historical evolution of employee

engagement and self-efficacy constructs: An empirical examination in a non-

western country. Journal of Management History, 21(2), 232–256.

https://dx.doi.org/10.1108/JMH-05-2014-0116

Dane, F. C. (2018). Evaluating research: Methodology for people who need to read

research (2nd ed.). Los Angeles, CA: Sage.

Daoud, J. I. (2017). Multicollinearity and regression analysis. Journal of Physics:

Conference Series, 949, 012009. https://dx.doi.org/10.1088/1742-

6596/949/1/012009

130

Daoud, M. (2019). Improving external validity may jeopardize internal validity.

Anesthesiology, 130, 508–509.

https://dx.doi.org/10.1097/ALN.0000000000002575

Davies, M., & Hughes, N. (2014). Doing a successful research project: Using qualitative

or quantitative methods (2nd ed.). Houndmills, NY: Palgrave Macmillan.

De Clercq, D., Haq, I. U., & Azeem, M. U. (2018). Self-efficacy to spur job performance:

Roles of job-related anxiety and perceived workplace incivility. Management

Decision, 56, 891–907. https://dx.doi.org/10.1108/MD-03-2017-0187

Dearborn, J. (2015). Data driven: How performance analytics delivers extraordinary

sales results. Hoboken, NJ: Wiley.

Deeter-Schmelz, D. R. (2016). Personal selling and sales management abstracts. Journal

of Personal Selling & Sales Management, 36, 363–373.

https://dx.doi.org/10.1080/08853134.2016.1207542

Delahaij, R., & Van Dam, K. (2017). Coping with acute stress in the military: The

influence of coping style, coping self-efficacy, and appraisal emotions.

Personality and Individual Differences, 119, 13–18.

https://dx.doi.org/10.1016/j.paid.2017.06.021

Diener, E., Oishi, S., & Park, J. (2014). An incomplete list of eminent psychologists of

the modern era. Archives of Scientific Psychology, 2(1), 20–31.

https://dx.doi.org/10.1037/arc0000006

131

Doron, I., & Georgantzi, N. (Eds.). (2018). Ageing, ageism and the law: European

perspectives on the rights of older persons. Cheltenham, UK: Edward Elgar

Publishing.

Dyer, J. H., Godfrey, P., Jensen, R., & Bryce, D. (2015). Strategic management:

Concepts and tools for creating real world strategy. Hoboken, NJ: Wiley & Sons.

Edmondson, A. C., & McManus, S. E. (2007). Methodological fit in management field

research. Academy of Management Review, 32, 1246–1264.

https://dx.doi.org/10.5465/amr.2007.26586086

Elkington, J. (1998). Cannibals with forks: The triple bottom line of 21st century

business. Oxford, U.K.: Capstone.

Evans, J. R., & Mathur, A. (2018). The value of online surveys: A look back and a look

ahead. Internet Research, 28, 854–887. https://dx.doi.org/10.1108/IntR-03-2018-

0089

Faul, F., Erdfelder, E., Buchner, A., & Lang, A.-G. (2009). Statistical power analyses

using G*Power 3.1: Tests for correlation and regression analyses. Behavior

Research Methods, 41, 1149–1160. https://dx.doi.org/10.3758/BRM.41.4.1149

Feng, C., & Fay, S. A. (2016). Inferring salesperson capability using stochastic frontier

analysis. Journal of Personal Selling & Sales Management, 36(3), 294–306.

https://dx.doi.org/10.1080/08853134.2016.1166966

Ferrando, P. J., & Lorenzo-Seva, U. (2010). Acquiescence as a source of bias and model

and person misfit: A theoretical and empirical analysis. British Journal of

132

Mathematical and Statistical Psychology, 63, 427–448.

https://dx.doi.org/10.1348/000711009X470740

Fidelity Investments. (2018, December 26). Technology hardware, storage, &

peripherals. Retrieved from

https://eresearch.fidelity.com/eresearch/markets_sectors/sectors/industries.jhtml?t

ab=learn&industry=452020

Fisher, G. G., Truxillo, D. M., Finkelstein, L. M., & Wallace, L. E. (2017). Age

discrimination: Potential for adverse impact and differential prediction related to

age. Human Resource Management Review, 27(2), 316–327.

https://dx.doi.org/10.1016/j.hrmr.2016.06.001

Frieder, R. E., Wang, G., & Oh, I.-S. (2018). Linking job-relevant personality traits,

transformational leadership, and job performance via perceived meaningfulness at

work: A moderated mediation model. Journal of Applied Psychology, 103(3),

324–333. https://dx.doi.org/10.1037/apl0000274

Fu, F. Q. (2015). Motivate to improve salesforce performance: The sales training

perspective. Performance Improvement, 54(4), 31–35.

https://dx.doi.org/10.1002/pfi.21474

Fu, F. Q., Richards, K. A., & Jones, E. (2009). The motivation hub: Effects of goal

setting and self-efficacy on effort and new product sales. Journal of Personal

Selling and Sales Management, 29(3), 277–292.

https://dx.doi.org/10.2753/PSS0885-3134290305

133

Fuchs, E., & Flügge, G. (2014). Adult neuroplasticity: More than 40 years of research.

Neural Plasticity, 2014, 1–10. https://dx.doi.org/10.1155/2014/541870

Fuller, C. M., Simmering, M. J., Atinc, G., Atinc, Y., & Babin, B. J. (2016). Common

methods variance detection in business research. Journal of Business Research,

69, 3192–3198. https://dx.doi.org/10.1016/j.jbusres.2015.12.008

Gagné, M., Forest, J., Vansteenkiste, M., Crevier-Braud, L., van den Broeck, A., Aspeli,

A. K., . . . & Westbye, C. (2015). The Multidimensional Work Motivation Scale:

Validation evidence in seven languages and nine countries. European Journal of

Work and Organizational Psychology, 24(2), 178–196.

https://dx.doi.org/10.1080/1359432X.2013.877892

Garson, G. D. (Ed.). (2013). Hierarchical linear modeling: Guide and applications.

Thousand Oaks, CA: Sage.

GLOBE Foundation. (2007). Understanding the relationship between national culture,

societal effectiveness and desirable leadership attributes: A brief overview of the

globe project. Retrieved from GLOBE Foundation website

http://globeproject.com/study_2004_2007

Goad, E. A., & Jaramillo, F. (2014). The good, the bad and the effective: A meta-analytic

examination of selling orientation and customer orientation on sales performance.

Journal of Personal Selling & Sales Management, 34(4), 285–301.

https://dx.doi.org/10.1080/08853134.2014.899471

134

Gottfredson, L. S. (2002). Where and why G matters: Not a mystery. Human

Performance, 15(1–2), 25–46.

https://dx.doi.org/10.1080/08959285.2002.9668082

Gottschling, J., Hahn, E., Maas, H., & Spinath, F. M. (2016). Explaining the relationship

between personality and coping with professional demands: Where and why do

optimism, self-regulation, and self-efficacy matter? Personality and Individual

Differences, 100, 49–55. https://dx.doi.org/10.1016/j.paid.2016.03.085

Greene, T., Shmueli, G., Ray, S., & Fell, J. (2019). Adjusting to the GDPR: The impact

on data scientists and behavioral researchers. Big Data, 7(3), 140–162.

https://dx.doi.org/10.1089/big.2018.0176

Grether, T., Sowislo, J. F., & Wiese, B. S. (2018). Top-down or bottom-up? Prospective

relations between general and domain-specific self-efficacy beliefs during a work-

family transition. Personality and Individual Differences, 121, 131–139.

https://dx.doi.org/10.1016/j.paid.2017.09.021

Grömping, U. (2015). Variable importance in regression models. Wiley Interdisciplinary

Reviews: Computational Statistics, 7(2), 137–152.

https://dx.doi.org/10.1002/wics.1346

Guenzi, P., Sajtos, L., & Troilo, G. (2016). The dual mechanism of sales capabilities in

influencing organizational performance. Journal of Business Research, 69, 3707–

3713. https://dx.doi.org/10.1016/j.jbusres.2016.03.033

135

Gupta, N., Ganster, D. C., & Kepes, S. (2013). Assessing the validity of sales self-

efficacy: A cautionary tale. Journal of Applied Psychology, 98, 690–700.

https://dx.doi.org/10.1037/a0032232

Hallak, R., Assaker, G., O’Connor, P., & Lee, C. (2018). Firm performance in the upscale

restaurant sector: The effects of resilience, creative self-efficacy, innovation and

industry experience. Journal of Retailing and Consumer Services, 40, 229–240.

https://dx.doi.org/10.1016/j.jretconser.2017.10.014

Hammer, M. J. (2017). Ethical considerations for data collection using surveys. Oncology

Nursing Forum, 44(2), 157–159. https://dx.doi.org/10.1188/17.ONF.157-159

Hamstra, M. R. W., Rietzschel, E. F., & Groeneveld, D. M. (2015). To go or not to go for

the sell: Regulatory focus and personal sales performance. Journal of Personnel

Psychology, 14(2), 109–112. https://dx.doi.org/10.1027/1866-5888/a000134

Heale, R., & Twycross, A. (2015). Validity and reliability in quantitative studies.

Evidence Based Nursing, 18(3), 66–67. https://dx.doi.org/10.1136/eb-2015-

102129

Hintze, M. (2018). Comparing the benefits of pseudonymization and anonymization

under the GDPR. Retrieved from Privacy Analytics Inc. website:

https://iapp.org/media/pdf/resource_center/PA_WP2-Anonymous-pseudonymous-

comparison.pdf

Hox, J. J., Moerbeek, M., & van de Schoot, R. (2017). Multilevel analysis: Techniques

and applications (3rd ed.). https://dx.doi.org/10.4324/9781315650982

136

Hyatt, H. R., & Spletzer, J. R. (2016). The shifting job tenure distribution. Labour

Economics, 41, 363–377. https://dx.doi.org/10.1016/j.labeco.2016.05.008

In, J. (2017). Introduction of a pilot study. Korean Journal of Anesthesiology, 70, 601.

https://dx.doi.org/10.4097/kjae.2017.70.6.601

Ingram, T. N. (2015). Sales management: Analysis and decision making (8th ed.).

https://dx.doi.org/10.4324/9781315700915

Inyang, A. E., & Jaramillo, F. (2019). Salesperson implementation of sales strategy and

its impact on sales performance. Journal of Strategic Marketing, 1–19.

https://dx.doi.org/10.1080/0965254X.2019.1593223

Jaramillo, F., Carrillat, F. A., & Locander, W. B. (2003). Starting to solve the method

puzzle in salesperson self-report evaluations. Journal of Personal Selling & Sales

Management, 23, 369–377. https://dx.doi.org/10.1080/08853134.2003.10749010

Johnson, J. S., & Jaramillo, F. (2017). Meta-analyses in sales research. Journal of

Personal Selling & Sales Management, 37(2), 134–152.

https://dx.doi.org/10.1080/08853134.2017.1286596

Johnson, R. B., & Onwuegbuzie, A. J. (2004). Mixed methods research: A research

paradigm whose time has come. Educational Researcher, 33(7), 14–26.

Johnson, R., & Gosselin, P. (2018). How secure is employment at older ages? 1–30.

Retrieved from

https://www.urban.org/sites/default/files/publication/99570/how_secure_is_emplo

yment_at_older_ages_2.pdf

137

Joseph, D. L., Jin, J., Newman, D. A., & O’Boyle, E. H. (2015). Why does self-reported

emotional intelligence predict job performance?: A meta-analytic investigation of

mixed EI. Journal of Applied Psychology, 100(2), 298–342.

https://dx.doi.org/10.1037/a0037681

Kail, R. V., & Cavanaugh, J. C. (2016). Human development: A life-span view. Boston,

MA: Cengage Learning.

Kanfer, R., & Chen, G. (2016). Motivation in organizational behavior: History, advances

and prospects. Organizational Behavior and Human Decision Processes, 136, 6–

19. https://dx.doi.org/10.1016/j.obhdp.2016.06.002

Kaplan, S. E., Petersen, M. J., & Samuels, J. A. (2018). Further evidence on the

negativity bias in performance evaluation: When does the evaluator’s perspective

matter? Journal of Management Accounting Research, 30(1), 169–184.

https://dx.doi.org/10.2308/jmar-51698

Kelly, J. (2019, September 19). Google is once again accused of age discrimination.

Retrieved from Forbes website

https://www.forbes.com/sites/jackkelly/2019/09/19/google-is-once-again-

accused-of-age-discrimination/

Kelly, T. (2018). The British computer industry: Crisis and development. Retrieved from

http://public.eblib.com/choice/publicfullrecord.aspx?p=5330115

Kirkpatrick, C. D., & Dahlquist, J. R. (2007). Technical analysis: The complete resource

for financial market technicians. Upper Saddle River, NJ: FT Press Financial

Times.

138

Kräkel, M., & Schöttner, A. (2016). Optimal sales force compensation. Journal of

Economic Behavior & Organization, 126, 179–195.

https://dx.doi.org/10.1016/j.jebo.2016.03.015

Krishnan, B. C., Netemeyer, R. G., & Boles, J. S. (2002). Self-efficacy, competitiveness,

and effort as antecedents of salesperson performance. Journal of Personal Selling

& Sales Management, 22(4), 285–295.

https://dx.doi.org/10.1080/08853134.2002.10754315

Kuhn, T. S. (1996). The structure of scientific revolutions (3rd ed). Chicago, IL:

University of Chicago Press.

Kwak, H., Anderson, R. E., Leigh, T. W., & Bonifield, S. D. (2019). Impact of

salesperson macro-adaptive selling strategy on job performance and satisfaction.

Journal of Business Research, 94, 42–55.

https://dx.doi.org/10.1016/j.jbusres.2018.09.015

Leavitt, C. (1977). Response bias: A special opportunity. Advances in Consumer

Research, 4, 401–404. Retrieved from https://acrwebsite.org/default.aspx

Lilly, B., Porter, T. W., & Meo, A. W. (2003). How good are managers at evaluating

sales problems? Journal of Personal Selling & Sales Management, 23(1), 51–60.

https://dx.doi.org/10.1080/08853134.2003.10748987

Lisbona, A., Palaci, F., & Salanova, M. (2018). The effects of work engagement and self-

efficacy on personal initiative and performance. Psicothema, (30.1), 89–96.

https://dx.doi.org/10.7334/psicothema2016.245

139

Lu, C.-Q., Du, D.-Y., & Xu, X.-M. (2016). What differentiates employees’ job

performance under stressful situations: The role of general self-efficacy. The

Journal of Psychology, 150, 837–848.

https://dx.doi.org/10.1080/00223980.2016.1203277

Lund Research. (2016, March 23). Constructs in quantitative research. Retrieved from

Laerd dissertation website: http://dissertation.laerd.com/constructs-in-

quantitative-research.php

Luszczynska, A., Scholz, U., & Schwarzer, R. (2005). The general self-efficacy scale:

Multicultural validation studies. The Journal of Psychology, 139, 439–457.

https://dx.doi.org/10.3200/JRLP.139.5.439-457

Lyons, M. J., Panizzon, M. S., Liu, W., McKenzie, R., Bluestone, N. J., Grant, M. D., . . .

& Xian, H. (2017). A longitudinal twin study of general cognitive ability over

four decades. Developmental Psychology, 53, 1170–1177.

https://dx.doi.org/10.1037/dev0000303

Mahlamäki, T., Rintamäki, T., & Rajah, E. (2018). The role of personality and motivation

on key account manager job performance. Industrial Marketing Management.

https://dx.doi.org/10.1016/j.indmarman.2018.11.013

Mansolf, M., & Reise, S. P. (2016). Exploratory bifactor analysis: The schmid-leiman

orthogonalization and Jennrich-Bentler analytic rotations. Multivariate

Behavioral Research, 51, 698–717.

https://dx.doi.org/10.1080/00273171.2016.1215898

140

Marley, R., & Mooney, J. L. (2014). M&A: The incredible shrinking U.S. PC industry.

Journal of Corporate Accounting & Finance, 25(6), 39–48.

https://dx.doi.org/10.1002/jcaf.21987

Martin, S. W. (2013, December 9). The twelve sales metrics that matter most. Harvard

Business Review. Retrieved from https://hbr.org/2013/12/new-insight-into-key-

sales-metrics

Martínez-Mesa, J., González-Chica, D. A., Bastos, J. L., Bonamigo, R. R., & Duquia, R.

P. (2014). Sample size: How many participants do I need in my research? Anais

Brasileiros de Dermatologia, 89, 609–615. https://dx.doi.org/10.1590/abd1806-

4841.20143705

Martínez-Mesa, J., González-Chica, D. A., Duquia, R. P., Bonamigo, R. R., & Bastos, J.

L. (2016). Sampling: How to select participants in my research study? Anais

Brasileiros de Dermatologia, 91, 326–330. https://dx.doi.org/10.1590/abd1806-

4841.20165254

Maslow, A. H. (1943). A theory of human motivation. Psychological Review, 50, 370–

396. https://dx.doi.org/10.1037/h0054346

Mertler, C. A., & Vannatta, R. A. (2017). Advanced and multivariate statistical methods:

Practical application and interpretation (6th edition). New York, NY: Routledge,

Taylor & Francis Group.

Meyer, A. G. (2014). Is unemployment good for the environment? SSRN Electronic

Journal. https://dx.doi.org/10.2139/ssrn.2586248

141

Micevski, M., Dewsnap, B., Cadogan, J. W., Kadic-Maglajlic, S., & Boso, N. (2019).

Sales intra-functional flexibility: Its relationship to performance and moderating

effects on role stressors. Journal of Business Research, S0148296318306313.

https://dx.doi.org/10.1016/j.jbusres.2018.12.021

Miracle, V. A. (2016). The Belmont report: The triple crown of research ethics.

Dimensions of Critical Care Nursing, 35(4), 223–228.

https://dx.doi.org/10.1097/DCC.0000000000000186

Miraglia, M., Cenciotti, R., Alessandri, G., & Borgogni, L. (2017). Translating self-

efficacy in job performance over time: The role of job crafting. Human

Performance, 30(5), 254–271.

https://dx.doi.org/10.1080/08959285.2017.1373115

Mitchell, T. R. (1985). An evaluation of the validity of correlational research conducted

in organizations. Academy of Management Review, 10(2), 192–205.

https://dx.doi.org/10.5465/amr.1985.4277939

Moon, C. (2015). The (un)changing role of the researcher. International Journal of

Market Research, 57(1), 15–16. https://dx.doi.org/10.2501/IJMR-2015-002

Morse, J. M., & Niehaus, L. (2009). Mixed method design: Principles and procedures.

Walnut Creek, CA: Left Coast Press.

Moser, C., & Kalton, G. (2017). Survey methods in social investigation (2nd ed.). New

York, NY: Routledge Taylor & Francis Group.

142

Mullinix, K. J., Leeper, T. J., Druckman, J. N., & Freese, J. (2015). The generalizability

of survey experiments. Journal of Experimental Political Science, 2(02), 109–

138. https://dx.doi.org/10.1017/XPS.2015.19

Nel, P., & Boshoff, A. (2016). Evaluating the factor structure of the general self-efficacy

scale. South African Journal of Psychology, 46(1), 37–49.

https://dx.doi.org/10.1177/0081246315593070

Neuman, W. L. (2014). Social research methods: Qualitative and quantitative

approaches (7th ed.). Harlow, NY: Pearson.

Newington, L., & Metcalfe, A. (2014). Factors influencing recruitment to research:

Qualitative study of the experiences and perceptions of research teams. BMC

Medical Research Methodology, 14(1). https://dx.doi.org/10.1186/1471-2288-14-

Newman, I., Hitchcock, J. H., & Newman, D. (2015). The use of research syntheses and

nomological networks to develop HRD theory. Advances in Developing Human

Resources, 17(1), 117–134. https://dx.doi.org/10.1177/1523422314559810

Nkwake, A. M., & Morrow, N. (2016). Clarifying concepts and categories of

assumptions for use in evaluation. Evaluation and Program Planning, 59, 97–

101. https://dx.doi.org/10.1016/j.evalprogplan.2016.05.014

Osborne, P. (2015). Emotional intelligence, cognitive intelligence and personality as

predictors of sales performance in outdoor advertising sales (Doctoral

dissertation). Retrieved from ProQuest Dissertations & Theses Database. (UMI

No. 3706863)

143

Ozyilmaz, A., Erdogan, B., & Karaeminogullari, A. (2018). Trust in organization as a

moderator of the relationship between self-efficacy and workplace outcomes: A

social cognitive theory-based examination. Journal of Occupational and

Organizational Psychology, 91(1), 181–204.

https://dx.doi.org/10.1111/joop.12189

Pajares, F., & Urdan, T. C. (Eds.). (2006). Self-efficacy beliefs of adolescents. Greenwich,

CT: IAP - Information Age Publishing.

Panagopoulos, N. G., Lee, N., Pullins, E. B., Avlonitis, G. J., Brassier, P., Guenzi, P., …

Weilbaker, D. C. (2011). Internationalizing sales research: Current status,

opportunities, and challenges. Journal of Personal Selling and Sales

Management, 31(3), 219–242. https://dx.doi.org/10.2753/PSS0885-3134310302

Papke, D. R. (2018). Ageism: A powerful nemesis for effective workplace discrimination

law. SSRN Electronic Journal. https://dx.doi.org/10.2139/ssrn.3277566

Pastorelli, C., Caprara, G. V., Barbaranelli, C., Rola, J., Rozsa, S., & Bandura, A. (2001).

The structure of children’s perceived self-efficacy: A cross-national study.

European Journal of Psychological Assessment, 17(2), 87–97.

https://dx.doi.org/10.1027//1015-5759.17.2.87

Pearl, J., & Bareinboim, E. (2014). External validity: From do-calculus to transportability

across populations. Statistical Science, 29, 579–595.

https://dx.doi.org/10.1214/14-STS486

Pedersen, M. J., & Nielsen, C. V. (2016). Improving survey response rates in online

panels: Effects of low-cost incentives and cost-free text appeal interventions.

144

Social Science Computer Review, 34(2), 229–243.

https://dx.doi.org/10.1177/0894439314563916

Pettijohn, C. E., Schaefer, A. D., & Burnett, M. S. (2014). Salesperson performance:

Exploring the roles of role ambiguity, autonomy and self-efficacy. Academy of

Marketing Studies Journal, 18(1), 99. Retrieved from

https://www.abacademies.org/journals/academy-of-marketing-studies-journal-

home.html

Peytchev, A., & Peytcheva, E. (2017). Reduction of measurement error due to survey

length: Evaluation of the split questionnaire design approach. Survey Research

Methods, Vol 11, No 4 (2017)-. https://dx.doi.org/10.18148/srm/2017.v11i4.7145

Phillips, A. W., Reddy, S., & Durning, S. J. (2016). Improving response rates and

evaluating nonresponse bias in surveys: AMEE Guide No. 102. Medical Teacher,

38(3), 217–228. https://dx.doi.org/10.3109/0142159X.2015.1105945

Phillips, D. C., & Burbules, N. C. (2000). Postpositivism and educational research.

Lanham, MD: Rowman & Littlefield Publishers.

Piervincenzi, C., Ben-Soussan, T. D., Mauro, F., Mallio, C. A., Errante, Y., Quattrocchi,

C. C., & Carducci, F. (2017). White matter microstructural changes following

quadrato motor training: A longitudinal study. Frontiers in Human Neuroscience,

11, 590. https://dx.doi.org/10.3389/fnhum.2017.00590

Pinder, C. C. (2014). Work motivation in organizational behavior. Retrieved from

http://public.eblib.com/choice/publicfullrecord.aspx?p=1744138

145

Plouffe, C. R. (2018). Is it navigation, networking, coordination . . . or what?: A

multidisciplinary review of influences on the intraorganizational dimension of the

sales role and performance. Journal of Personal Selling & Sales Management,

38(2), 241–264. https://dx.doi.org/10.1080/08853134.2018.1450147

Pousa, C., & Mathieu, A. (2016). Increasing salesperson’s self-efficacy and performance

through coaching: A quantitative study in Canada. In M. D. Groza & C. B.

Ragland (Eds.), Marketing Challenges in a Turbulent Business Environment (pp.

103–107). https://dx.doi.org/10.1007/978-3-319-19428-8_27

Pransky, G., Finkelstein, S., Berndt, E., Kyle, M., Mackell, J., & Tortorice, D. (2006).

Objective and self‐report work performance measures: A comparative analysis.

International Journal of Productivity and Performance Management, 55, 390–

399. https://dx.doi.org/10.1108/17410400610671426

Quan, J., Dattero, R., & Galup, S. D. (2010). Impact of age on information technology

salaries. In M. Khosrow-Pour, Global, social, and organizational implications of

emerging information resources management: Concepts and applications (pp.

403–420). Hershey, PA: IGI Global.

R. Goldberg, L. (1999). A broad-bandwidth, public-domain, personality inventory

measuring the lower-level facets of several five-factor models. In Personality

Psychology in Europe (Vol. 7, pp. 7–28). Tilburg, Netherlands: Tilburg

University Press.

Rahi, S. (2017). Research design and methods: A systematic review of research

paradigms, sampling issues and instruments development. International Journal

146

of Economics & Management Sciences, 06(02). https://dx.doi.org/10.4172/2162-

6359.1000403

Ray, J. J. (1990). Acquiescence and problems with forced-choice scales. The Journal of

Social Psychology, 130, 397–399.

https://dx.doi.org/10.1080/00224545.1990.9924595

Richardson, K., & Norgate, S. H. (2015). Does IQ really predict job performance?

Applied Developmental Science, 19(3), 153–169.

https://dx.doi.org/10.1080/10888691.2014.983635

Rönnlund, M., Sundström, A., & Nilsson, L.-G. (2015). Interindividual differences in

general cognitive ability from age 18 to age 65 years are extremely stable and

strongly associated with working memory capacity. Intelligence, 53, 59–64.

https://dx.doi.org/10.1016/j.intell.2015.08.011

Rovai, A. P., Baker, J. D., & Ponton, M. K. (2013). Social science research design and

statistics: A practitioner’s guide to research methods and SPSS analysis (1st pbk.

ed.). Chesapeake, VA: Watertree Press.

Scherbaum, C. A., Cohen-Charash, Y., & Kern, M. J. (2006). Measuring General Self-

Efficacy: A comparison of three measures using item response theory.

Educational and Psychological Measurement, 66, 1047–1063.

https://dx.doi.org/10.1177/0013164406288171

Schoenherr, T., Ellram, L. M., & Tate, W. L. (2015). A note on the use of survey research

firms to enable empirical data collection. Journal of Business Logistics, 36(3),

288–300. https://dx.doi.org/10.1111/jbl.12092

147

Scholz, U., Gutiérrez Doña, B., Sud, S., & Schwarzer, R. (2002). Is General Self-Efficacy

a universal construct?: Psychometric findings from 25 countries. European

Journal of Psychological Assessment, 18(3), 242–251.

https://dx.doi.org/10.1027//1015-5759.18.3.242

Schoorman, F. D. (1988). Escalation bias in performance appraisals: An unintended

consequence of supervisor participation in hiring decisions. Journal of Applied

Psychology, 73(1), 58–62. https://dx.doi.org/10.1037/0021-9010.73.1.58

Schwarzer, R. (2006, November). Multicultural general self-efficacy validation data.

Retrieved from http://userpage.fu-berlin.de/health/selfscal.htm

Schwarzer, R. (2014, May 30). Documentation of the general self-efficacy scale.

Retrieved from http://userpage.fu-berlin.de/~health/faq_gse.pdf

Schwarzer, R., & Jerusalem, M. (1995). Generalized self-efficacy scale. In Measures in

health psychology: A user’s portfolio. Causal and control belief (pp. 35–37).

Retrieved from http://userpage.fu-berlin.de/~health/selfscal.htm

Schwarzer, R., & Warner, L. M. (2013). Perceived self-efficacy and its relationship to

resilience. In S. Prince-Embury & D. H. Saklofske (Eds.), Resilience in children,

adolescents, and adults (pp. 139–150). https://dx.doi.org/10.1007/978-1-4614-

4939-3_10

Shaun, N. (2017, July 25). HP Inc, HPE both slapped with racism, ageism lawsuit.

Retrieved from The Register website

https://www.theregister.co.uk/2017/07/25/hp_hpe_sued_discrimination/

Students also viewed