PS2

Batman007

Resource3.pdf

Home >Business & Finance homework help >Operations Management homework help >PS2

Hiring for Performance and Retention:

Examining the Relationship between Cognitive Fit

and Employee Turnover in the U.S. Navy

Dissertation Manuscript

Submitted to Northcentral University

Graduate Faculty of the School of Business

in Partial Fulfillment of the

Requirements for the Degree of

DOCTOR OF PHILOSOPHY

RENEE J. SQUIER

Prescott Valley, Arizona

November, 2016

ProQuest Number:

INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed,

a note will indicate the deletion.

ProQuest

Published by ProQuest LLC ( ). Copyright of the Dissertation is held by the Author.

Microform Edition © ProQuest LLC.

ProQuest LLC. 789 East Eisenhower Parkway

P.O. Box 1346 Ann Arbor, MI 48106 - 1346

10252272

2017

Approval Page

Hiring for Performance and Retention: Examining the Relationship between Cognitive Fit and

Employee Turnover in the U.S. Navy

Renee Squire

Approved by:

Dec. 26,2016

Chair: Dr. Linda Cummins Date

Certified by:

Dean of School: Dr. Peter Bemski Date

Abstract

Retaining top-performing talent is one of the most fundamental human resource

challenges facing organizations today. Strong retention is critical to workforce quality

and controlling human resource costs—especially in an entry-level hiring system like the

U.S. Navy. The purpose of this quantitative study was to determine if cognitive fit

predicts employee turnover by comparing U.S. Navy enlisted sailor Armed Services

Vocational Aptitude Battery test scores to the cognitive demands for their career fields in

the Navy. It includes an analysis of this measurement of cognitive fit with retention data

to ascertain if it predicts employee turnover. The mean value of cognitive fit for the Navy

was negative, and although cognitive fit was statistically significant for voluntary and

involuntary turnover in the full dataset, the effect sizes were very small. Further testing of

10%, 1%, and stratified random subsets of the data refuted the value of the significance

of the findings in the full dataset, indicating that cognitive fit and interactions with gender

and length of service are not important predictors of future employee turnover. This

research implies that the Navy is not placing sailors in best fit jobs, and that objective

measurements of cognitive fit are not enough to predict future employee turnover.

Recommendations include changing the Navy’s recruitment process to allow time to

match applicants to best fit jobs and utilizing other subjective measurements, such as an

interest inventory, with cognitive fit in job placement. The results of this study benefit the

U.S. Navy, and other military services and organizations by offering new ideas on how to

measure cognitive fit and exploring ways to improve the hiring process, optimizing

placement, utilization, and retention of personnel.

iii

Acknowledgements

I would like to thank my husband Rob, and children, Tara, Samantha, and

Danica, for their patience and support while I worked week by week, year by

year, towards this goal. Rob held my hand through the ups and downs, helping me

stay the course, and I hope I have inspired a love of learning in my children. I also

want to thank my parents—they are responsible for my belief in myself that I can

do anything I set out to do.

There are several Navy friends, mentors, and colleagues who have helped

me reach this goal. Most importantly I would like to thank Dr. Sofiya Velgach,

whose guidance and inspiration lit the way, and Mr. Rick Ayala and Mr. Earl

Salter, who love the work of managing Navy sailors as much as I do. I also could

not have done it without the support of Ms. Jennifer Bennett, who kept everything

running at work when I was not there. Additionally, I would like to thank my

mentor and chair Dr. Linda Cummins for her guidance, encouragement, and

support throughout this process.

Table of Contents

Chapter 1: Introduction ....................................................................................................... 1

Background ................................................................................................................... 2

Statement of the Problem .............................................................................................. 3

Purpose of the Study ..................................................................................................... 4

Theoretical Framework ................................................................................................. 5

Research Question ........................................................................................................ 9

Hypotheses .................................................................................................................... 9

Nature of the Study ....................................................................................................... 9

Significance of the Study ............................................................................................ 11

Definition of Key Terms ............................................................................................. 12

Summary ..................................................................................................................... 12

Chapter 2: Literature Review ............................................................................................ 14

Documentation ............................................................................................................ 15

Employee Turnover .................................................................................................... 15

Employee Turnover and Situational Antecedents....................................................... 16

Employee Turnover and Individual Attributes ........................................................... 18

Military Employee Turnover ...................................................................................... 22

Military Turnover and Situational Antecedents .......................................................... 23

Military Turnover and Individual Attributes .............................................................. 24

Employee Fit ............................................................................................................... 31

Cognitive Ability ........................................................................................................ 40

Cognitive Testing in the U.S. Military ....................................................................... 43

Navy’s Algorithm for Cognitive Fit ........................................................................... 46

Summary ..................................................................................................................... 48

Chapter 3: Research Method ............................................................................................. 51

Research Methods and Design .................................................................................... 53

Population ................................................................................................................... 56

Sample......................................................................................................................... 56

Materials/Instruments ................................................................................................. 60

Operational Definition of Variables ............................................................................ 62

Data Collection, Processing, and Analysis ................................................................. 63

Assumptions ................................................................................................................ 64

Limitations .................................................................................................................. 65

Delimitations ............................................................................................................... 67

Ethical Assurances ...................................................................................................... 67

Summary ..................................................................................................................... 68

Chapter 4: Findings ........................................................................................................... 70

Results ......................................................................................................................... 70

Evaluation of Findings ................................................................................................ 78

Summary ..................................................................................................................... 85

Chapter 5: Implications, Recommendations, and Conclusions ........................................ 87

Implications................................................................................................................. 91

Recommendations ....................................................................................................... 93

Conclusions ................................................................................................................. 95

References ......................................................................................................................... 97

Appendixes ..................................................................................................................... 108

Appendix A: Research Request and Approval ............................................................... 109

Appendix B: Research Variables .................................................................................... 111

Appendix C: Human Subjects Research Determination ................................................. 112

List of Tables

Table 1. Navy Active Component Continuation Rates from 2000-2011.......................... 22

Table 2. Armed Services Vocational Aptitude Battery (ASVAB) Sub-Tests .................. 44

Table 3. Paygrade Composition ........................................................................................ 57

Table 4. Rating Composition ............................................................................................ 58

Table 5. Reliability for Armed Forces Qualification Test Composite and Armed

Services Vocational Aptitude Battery Sub-Tests ............................................... 61

Table 6. Comparison of Predictor Variables by Turnover Outcome ................................ 72

Table 7. Multinomial Logistic Regression Results: Full Dataset ..................................... 77

Table 8. Multinomial Logistic Regression Results: 10% Dataset .................................... 78

Table 9. Multinomial Logistic Regression Results: 1% Dataset ...................................... 79

Table 10. Multinomial Logistic Regression Results: Stratified Dataset ........................... 80

vii

List of Figures

Figure 1. Likely relationship between cognitive fit and employee turnover ...................... 7

Figure 2. Employee fit—Types and relationships ............................................................ 33

Figure 3. Cognitive fit by gender. ..................................................................................... 74

Chapter 1: Introduction

Retaining high-performing employees is valuable to organizations and managers

because it reduces replacement costs for recruiting and training, and increases human

capital value by preserving institutional knowledge and retaining future leaders (George,

2015; Maltarich, Nyberg, & Reilly, 2010). Employee turnover represents a significant

loss of organizational effort and financial resources (Godlewski & Kline, 2012).

Retention is fundamental to the U.S. Navy’s personnel system to maintain the workforce

and to develop senior leaders, since most new employees join at entry level (Pinelis &

Huff, 2014; Rumsey & Arabian, 2014b). However, sailor retention rates are low; in 2014

only 59.1% of enlisted sailors completed their first enlistment contract and stayed in the

Navy (Center for Naval Analysis, 2014). This reality is costly—to counteract low

employee retention, the Navy recruits thousands of new sailors as replacements and/or

spends millions of dollars in reenlistment bonuses every year (Pinelis & Huff, 2014).

The aim of this study was to investigate cognitive ability, an attribute the U.S.

Navy already uses in employee selection as a predictor of job performance (Held,

Hezlett, et al., 2014), to determine if it is also useful in predicting future retention in the

U.S. Navy. Cognitive ability is a measurement of general intelligence and the ability to

learn (Ones & Viswesvaran, 2011), which the Navy can compare with the cognitive

demands of a job to determine compatibility—called cognitive fit (Maltarich et al., 2010).

The premise of this research is that strong cognitive fit will result in greater retention,

while poor cognitive fit will lead to higher employee turnover.

This chapter presents an introduction to the issue of U.S. Navy retention and the

concepts of employee turnover and cognitive ability. It begins with a brief background

that highlights the problem and includes an explanation of the purpose of the study. Next,

the chapter includes a discussion of employee-career fit as a theoretical framework for

the study. The chapter also sets out the research questions and hypotheses, as well as key

terminology and information on the nature and significance of the study.

Background

The broad definition of the concept of fit is the compatibility between an

individual and his or her work environment (Billsberry, Talbot, & Ambrosini, 2012).

However, there is no objective measurement of fit with utility for predicting future

employee turnover for use in hiring decisions. The basis of the majority of the literature

on fit, researchers obtain subjective measurements of the match between the individual

and the work environment by capturing an employee’s perceptions of how well he or she

feels like a good fit for the organization after hiring has taken place (Erdogan, Bauer,

Peiro, & Truxillo, 2011a; Freund & Kasten, 2012; Gabriel, Diefendorff, Chandler,

Moran, & Greguras, 2014). Developing an objective measure of fit, useful for employee

selection and retention, would add to the employee management literature and provide

actionable results.

Cognitive ability, or general intelligence, is knowledge, recall of knowledge, and

ability to work with knowledge (Mumford, Watts, & Partlow, 2015) and as the capacity

to problem-solve, plan ahead, and learn from experience (Oh et al., 2014). Cognitive

ability is a valuable predictor of job performance, and it is useful for selecting new

employees (Ones & Viswesvaran, 2011), with research results typically reporting a .20 or

greater correlation, and validity near .40 (Schmidt, 2014). Maltarich et al. (2010) reported

a curvilinear relationship between cognitive ability and voluntary turnover in jobs with

high cognitive demands. Their results indicated that voluntary turnover is most likely for

individuals with the lowest and highest cognitive ability, refuting the traditional belief

that hiring the individual with the highest cognitive ability for every job is the best course

of action (Maltarich et al., 2010). Maltarich et al. defined the concept of cognitive fit as

the match between cognitive ability and cognitive job requirements, and the results for

jobs with high cognitive demands indicated that the likelihood of voluntary turnover

increases in relation to greater distance above or below the cognitive mean (Maltarich et

al., 2010). Like the results from Maltarich et al.’s (2010) study, the Navy found that

sailors with a good match between their cognitive abilities and the cognitive requirements

of their jobs were more likely to complete their initial training successfully, more likely

to gain promotion, and less likely to leave (Department of the Navy, 2012).

Statement of the Problem

Employee turnover is a prime concern the U.S. Navy (Pinelis & Huff, 2014).

Failure to retain high-performing sailors in the U.S. Navy increases recruitment and

reenlistment costs, and results in the promotion of lower quality and less experienced

Navy personnel. The Navy uses monetary bonuses (with an average cost of $47,948.00

per enlisted sailor offered a bonus) as an incentive to encourage sailors to stay based on

their skill set and manning level, training costs, or criticality to the mission (Coughlan,

Gates, & Myung, 2014; Pinelis & Huff, 2014). When not enough sailors remain, it is

necessary to recruit and train additional sailors; however, they join the Navy at entry

level—leaving an experience gap. Additionally, the Navy promotes sailors according to

vacancies at the next higher paygrade (Arkes & Cunha, 2015; Kumazawa, 2010). The

Navy orders sailors by rank in a competitive group based on several factors including

advancement exam scores, performance evaluations, education, and awards to determine

their relative quality (Kumazawa, 2010). However, this only results in the promotion of

the best sailors if there are fewer vacancies than sailors eligible for promotion, because if

the number of vacancies is higher than the number eligible for promotion, the entire

competitive group will receive promotion to fill the Navy’s requirements, regardless of

their quality or experience. These undesirable outcomes highlight retention as

fundamental to workforce quality in an entry-level hiring system. As a potential strategy

for the U.S. Navy to reduce personnel costs and maintain a high-quality workforce, the

researcher examined the relationship between cognitive fit and employee turnover.

Purpose of the Study

The purpose of this non-experimental, quantitative study was to examine the

relationship between cognitive fit and employee turnover in the U.S. Navy. The U.S.

Navy measures cognitive ability through the Armed Services Vocational Aptitude Battery

(ASVAB) and it uses the results in the hiring process for those desiring to enlist. The

researcher used secondary, case-file data from the U.S. Navy’s Career Waypoints

personnel database for all enlisted sailor retention decisions in 2014, which included

ASVAB test scores and employee turnover outcomes, as well as the demographic factors

gender and length of service as potential covariates. The researcher used logistic

regression to examine the relationship between cognitive fit and U.S. Navy enlisted sailor

turnover decisions. The goal of this research was to determine if employee turnover

decreases when cognitive fit increases. The dataset for 2014 contained 56,847 case files

for sailors who made retention decisions in this one-year period.

Theoretical Framework

Although existing theory and empirical research do not directly explain the

relationship between cognitive ability and employee turnover (Maltarich et al., 2010), the

theory of employee fit and its key construct, demands-abilities fit, provide a basis for

considering why cognitive ability may relate to employee turnover. The theory of

employee fit started with Super’s (1953) theory of vocational development, which

theorized that people differ in their abilities and interests and they qualify for careers

based on these attributes, and with Holland’s (1959) theory of vocational choice to help

people to select jobs. Employee fit serves as the basic theoretical framework for this

research because cognitive ability is an important aspect of fit (Holland, 1959; Super,

1953). More recently the concept of employee fit has evolved to mean the alignment

between an individual and his or her work environment (Billsberry et al., 2012; Kristof-

Brown & Billsberry, 2012; Kristof-Brown & Guay, 2011; Maynard & Parfyonova, 2013;

Thompson, Sikora, Perrewé, & Ferris, 2015), from which several dimensions have

emerged including person-job fit, person-vocation fit, person-supervisor fit, person-

group/team fit, and person-organization fit (Kristof-Brown & Guay, 2011). Person-job fit

has two dimensions: demands-abilities fit and supplies-values fit (Kristof-Brown &

Guay, 2011).

The typical conceptualization of demand-abilities fit is the match between a

person’s knowledge, skills, and abilities and job tasks (Kristof-Brown & Guay, 2011),

which is similar to the match between an individual’s cognitive ability and the cognitive

demands of a job. Demand-abilities fit is relevant to several key employment outcomes

including job commitment, job satisfaction (Bogler & Nir, 2015; Kristof-Brown,

Zimmerman, & Johnson, 2005; McKee-Ryan & Harvey, 2011), job meaningfulness

(Tims, Derks, & Bakker, 2016), organizational commitment, professional commitment,

intrinsic satisfaction, and extrinsic satisfaction (Bogler & Nir, 2015). Additionally,

demand-abilities fit has a negative correlation with turnover intentions (r = -.16, p < .01;

J. Peng, Lee, & Tseng, 2014). These findings support the utility of demand-abilities fit as

a construct that may relate to employee turnover such that a poor match between an

individual’s cognitive ability and the cognitive demands of a job may lead to higher

employee turnover.

The Navy uses the ASVAB to measure cognitive ability and to place individuals

in career fields. To improve training success, in 2009, the Navy changed its placement

process from assigning individuals to jobs based on minimum requirements to matching

individuals to jobs based on cognitive fit (Watson, 2010). The Navy measures cognitive

fit by comparing sailor ASVAB test scores to the cognitive demands of specific jobs

within the Navy. The Navy developed this process based on the theoretical framework of

the Yerkes-Dodson law (Watson, 2010), which states that moderate levels of cognitive

stimulus are the most effective in rapid habit formation (Yerkes & Dodson, 1908). This

relationship also has applicability in the relationship between human performance and

cognitive arousal (Watson, 2010).

Figure 1 is a depiction of the likely curvilinear relationship between cognitive

ability and employee turnover. In addition to past findings on demand-abilities fit, two

other theoretical perspectives were useful in formulating the likely curvilinear

relationship between cognitive fit and employee turnover and the relevant control

variables: the push-pull model (Jackofsky, 1984), and the kaleidoscope career model

(Mainiero & Sullivan, 2005).

Figure 1. Likely relationship between cognitive fit and employee turnover. When an

employee’s cognitive ability is a good fit for the cognitive demands of a job, the likely

relationship is that employee turnover will be low. If an employee’s cognitive ability is

either over- or undermatched to the demands of a job, the likely relationship is that

employee turnover will be high. Author’s depiction based on the Yerkes-Dodson law

(Yerkes & Dodson, 1908) and inspired by Maltarich et al. (2010, p. 1061).

The push-pull model offers the rationale for a curvilinear relationship between

cognitive fit and employee turnover. The push-pull model includes three determinants of

voluntary turnover: intention to quit, ease of changing jobs, and desirability of changing

jobs, and it states that performance directly relates to perceptions about the ease of

changing jobs and indirectly relates to the desirability of changing jobs (Jackofsky,

1984). The push-pull model denotes a curvilinear relationship across the performance

spectrum, where an organization pushes out low-performing employees through negative

feedback and fewer rewards, and market-based forces pull high-performing employees

into other organizations (Becker & Cropanzano, 2011; Jackofsky, 1984). In a longitudinal

study that included 2,385 person-year observations between 2004 and 2006 from one

division of an engineering technology company, Cox regression results showed that

current performance was significant in predicting employee turnover (β. = -76, p < .01;

Becker & Cropanzano, 2011). The push-pull model implies that a similar relationship

between cognitive fit and turnover may exist since cognitive ability is a stronger predictor

of performance than other individual differences at work such as personality traits (Ones

& Viswesvaran, 2011). Measuring the cognitive fit gap between an individual’s ability

and job requirements may offer an explanation for top performer employee turnover and

an actionable plan to improve retention in the future.

Another theoretical perspective from recent career theory research is also valuable

in considering the linkage between cognitive ability and employee turnover. The

kaleidoscope career model (KCM) captures the evolution of career enactment over a

career lifespan by examining the importance of three key career issues: authenticity,

balance, and challenge (Mainiero & Sullivan, 2005). KCM findings identified different

career patterns based on gender (Sullivan & Mainiero, 2007), providing a theoretical

basis to include gender as a variable. Additionally, KCM treats challenge as the desire to

do stimulating work and to experience career advancement. Research has shown it is the

highest priority career influencer for both men and women at the beginning of their

careers (Carette, Anseel, & Lievens, 2013; Mainiero & Sullivan, 2005). Recent research

on job challenge has included cognitive elaboration as an aspect of challenge, and has

shown a more positive relationship between challenging assignments and performance in

early-career than mid-career employees, with an overall variance of 21% (Carette et al.,

2013). The concept of job challenge in KCM is applicable, since the Navy recruits sailors

early in their careers, and the research indicates that organizations may experience less

employee turnover when matching early-career hires with work they find challenging

(Cabrera, 2009), providing a theoretical basis for length of service as a second control

variable (Bernerth & Aguinis, 2016).

Research Question

The aim of this study was to investigate the relationship between cognitive fit and

retention trends of U.S. Navy sailors to determine the extent to which cognitive fit

predicts sailor retention. The researcher examined the relationship between cognitive fit

and employee turnover in the Navy quantitatively, guided by the following research

question:

Q1. To what extent does cognitive fit, gender, and length of service predict

employee turnover amongst U.S. Navy enlisted sailors?

Hypotheses

H10. Cognitive fit, gender, and length of service do not predict employee

turnover among U.S. Navy enlisted sailors.

H1a. Cognitive fit, gender, and length of service significantly predict employee

turnover among U.S. Navy enlisted sailors.

Nature of the Study

The researcher chose a quantitative design to explore the relationship between

cognitive fit and employee turnover because of its applicability to the research question.

The purpose of this study was to determine if cognitive fit (as the predictor variable)

predicts employee turnover (as the criterion variable), and covaries with gender and

length of service, other variables that others have related to employee turnover (Hoglin &

Barton, 2013). The researcher used logistic regression for this research question, since it

calls for analysis about a predictive relationship with a categorical outcome (Field, 2009;

C. Peng, Lee, & Ingersoll, 2002). The researcher used binary logistic regression to

consider the dichotomous employee turnover outcomes (separation or reenlistment). The

researcher used a separate multinomial model to distinguish the type of separation using

polytomous employee turnover outcomes: involuntary separation, voluntary separation,

or reenlistment. In the regression models, the researcher added interaction terms between

cognitive fit, gender, and length of service to examine the combined effect of these

variables.

The dataset included all active U.S. Navy enlisted sailors, paygrades E1 thru E6,

with up to 14 years of service who made a retention decision in 2014. The U.S. Navy

provided archival data for this research, and the data included a measurement of cognitive

fit, calculated using the sailor’s cognitive ability measured by his or her ASVAB test

results compared to the cognitive abilities of other sailors assigned to his or her career

field, and who have successfully completed their required initial training. This technique

is similar to the method prior researchers have used to compute cognitive fit by

comparing ASVAB test scores to the average level of ability required by occupation

computed using data from the Occupational Information Network website (Maltarich et

al., 2010).

The data the Navy provided also included employee turnover decisions with three

categorical outcomes: reenlistment, voluntary separation from Naval service, or

involuntarily separation. The push-pull model establishes a basis for operationalizing

employee turnover using voluntary and involuntary separation in a way that has links to

cognitive fit. Functional turnover is the removal of the lowest performers, and it is

beneficial to an organization (Becker & Cropanzano, 2011). The U.S. Navy initiates

functional turnover actions by involuntarily separating sailors who are lower performers

than their peers, or who are not eligible for reenlistment. On the other hand, individuals

across the performance spectrum may self-initiate voluntary turnover. It may be in the

organization’s best interest for them to leave if they are poor performers, but when top

performers voluntarily choose to leave, it can negatively affect organizational

performance (Becker & Cropanzano, 2011). Identifying a predictive relationship between

cognitive fit and retention, and further by the three types of retention outcomes

(reenlistment, voluntary separation, or involuntary separation) may signify an opportunity

to improve retention by improving cognitive fit.

Significance of the Study

Competition for talent is increasing, and retaining high performing employees is

valuable in preserving institutional knowledge and avoiding costs for recruiting and

training (George, 2015; Maltarich et al., 2010). This study contributes to the body of

knowledge on employee management by identifying a predictive relationship between

cognitive fit and employee turnover. Insights into Navy turnover trends support the

importance of cognitive ability as an objective measure of job qualification and explain

its relationship to employee turnover. Identifying cognitive fit as a measurable attribute

with utility for selecting and optimally placing new hires to maximize the probability of

future retention could fundamentally improve human capital utilization. The results of

this study could benefit the U.S. Navy and other military services and organizations by

improving hiring processes to match individuals better with jobs, optimizing placement,

utilization, and retention of personnel. This research may also lead to a recommended

cognitive fit measure to improve optimal placement, utilization and retention of Navy

personnel.

Definition of Key Terms

Cognitive ability. The term cognitive ability describes and quantifies an

individual’s ability to learn (Ones & Viswesvaran, 2011).

Cognitive demand. The term cognitive demand describes and quantifies the

cognitive requirements for a particular job or career field (Maltarich et al., 2010).

Cognitive fit. The term cognitive fit describes and quantifies demand-abilities fit

between an individual and a job based on a comparison of cognitive ability to cognitive

demands (Maltarich et al., 2010).

Demand-abilities fit. Demand-abilities fit is one aspect of person-job fit, namely

the match between a person’s knowledge, skills, and abilities, and job tasks (Kristof-

Brown, Zimmerman, & Johnson, 2005).

Person-job fit. Person-job fit is the relationship between the requirements of a

job and the characteristics of an employee (Boon, den Hartog, Boselie, & Paauwe, 2011;

C. Chen, Yen, & Tsai, 2014; Gabriel et al., 2014).

Reenlistment. Reenlistment is the renewal of a sailor’s employment contract.

Retention. Retention means keeping employees on the job—the opposite of

employee turnover (Hong, Hao, Kumar, Ramendran, & Kadiresan, 2012).

Sailor. The term sailor identifies individuals currently serving in the U.S. Navy.

Summary

The competition for talent in the workforce is increasing (Maltarich et al., 2010).

Failure to retain high-performing employees is a problem because it increases recruitment

and reenlistment costs, and it can result in the promotion of lower quality and less

experienced personnel. The focus of this study was to examine employee turnover of U.S.

Navy enlisted sailors to determine if there is a significant and measurable predictive

relationship between cognitive fit and employee turnover. The quantitative research

design uses multinomial logistic regression to determine if there is a systematic

relationship between U.S. Navy sailor cognitive fit (using ASVAB scores), and turnover

decisions. Cognitive fit predicted employee turnover, which has implications for future

hiring and placement processes that may need to incorporate this construct to maximize

human capital value in the U.S. Navy and other organizations, optimizing placement,

utilization, and retention of personnel.

Chapter 2: Literature Review

High-performing employees are key to organizational success (Crook, Todd,

Combs, Woehr, & Ketchen, 2011). In a recent meta-analysis, human capital related to

performance with an effect size of .21, demonstrating that acquiring top talent, nurturing

it, and retaining it relates strongly to achieving high performance in organizations (Crook

et al., 2011). The resource-based theory of organizational performance has also

highlighted the importance of human capital to competitive advantage as the most

valuable and least imitable resource (Crook et al., 2011; Shaw, Park, & Kim, 2013).

Hence, the success of an organization largely depends on its people; hiring the best and

keeping them on the job. Retaining talented people is especially important in

organizations like the U.S. Navy, where hiring takes place exclusively at entry level, and

promotion is the only mechanism for replacing experienced employees (Rumsey &

Arabian, 2014b). Discovering an overlap or commonality between employee selection

and employee turnover may have utility during the hiring process that can help

organizations to maximize human capital investment by ultimately reducing employee

turnover.

The purpose of this literature review is to examine the knowledge base on

employee selection and employee turnover through the lens of cognitive fit. Cognitive fit

may be a link between these two essential tenets of workforce management. This section

begins with a discussion of employee turnover and the relevant research on potential

antecedents to turnover for both civilian and military employees. Next there is an

overview of employee fit, with a more in-depth discussion of the cascading concepts of

person-environment fit, person-job fit, and demands-abilities fit. It also includes a

discussion of the current literature on the relationship between employee fit and

employee turnover. The final topic is cognitive ability, and it includes an examination of

how the U.S. Navy both measures it and uses it in the selection and placement of new

recruits. The literature review concludes with a summary of the key concepts highlighting

the potential utility of fit during employee selection to predict employee turnover.

Documentation

The search strategy the researcher used in developing this literature review started

with the three main topics the researcher addressed in the study: employee turnover,

cognitive ability, and employee fit. The researcher used several key terms to identify

relevant literature, including employee turnover, employee retention, cognitive ability,

cognitive aptitude, overqualification, underqualification, employee fit, person-

environment fit, person-vocation fit, person-job fit, demands-abilities fit, work

engagement, military retention, sailor reenlistment, sailor retention, and sailor promotion.

The search engines and databases the researcher used were Google Scholar, Northcentral

University Roadrunner search, and the Defense Technical Information Center database.

The second step in the search strategy was to identify articles cited in the relevant

literature from the first exploration. The researcher reviewed these articles individually

for additional information.

Employee Turnover

Employee turnover is a complex topic because people leave organizations for a

broad range of reasons, and the impact can range from harmful to beneficial (Al-Emadi,

Schwabenland, & Qi, 2015; Allen, Bryant, & Vardaman, 2010). Employee turnover takes

place when an individual moves out of an organization’s employee membership, and

other authors have described this using a variety of terms including attrition, exits, quits,

and employee mobility or migration (Rainayee, 2013). To understand the differences and

organizational implications better, there are several ways to examine employee turnover

(Al-Emadi et al., 2015). First, employee turnover can be voluntary or involuntary—

voluntary when initiated by the employee, and involuntary when initiated by the

organization (Al-Emadi et al., 2015; Allen et al., 2010). Since involuntary turnover

usually occurs due to low performance or downsizing, it can be beneficial, while

individuals who voluntarily leave may be the employees an organization would like to

retain, thus creating a negative impact (Allen et al., 2010). Another distinction between

turnover actions is functional versus dysfunctional (Al-Emadi et al., 2015; Allen et al.,

2010). Turnover is functional if the employee is easy to replace, and dysfunctional when

the employee is hard to replace, which again can cause a negative organizational impact

(Allen et al., 2010). Employee turnover can also be avoidable or unavoidable depending

on whether or not the organization could have influenced the outcome (Al-Emadi et al.,

2015; Allen et al., 2010). Retention efforts in an organization typically focus on

voluntary, dysfunctional, and avoidable employee turnover (Allen et al., 2010).

Employee Turnover and Situational Antecedents

Research on employee turnover has primarily focused on an individual’s current

situation (i.e., alternate job availability and job attitudes) rather than more enduring traits,

such as cognitive ability, that employers can determine and use in the hiring decision

process (Boudreau, Boswell, Judge, & Bretz, 2001; Hom, Mitchell, Lee, & Griffeth,

2012). Many turnover models of this type are process-oriented and examine

psychological antecedents of employee turnover, such as negative job satisfaction or

organizational commitment, which may spur thoughts about leaving or intent to leave

actions, such as searching for another job (Lytell & Drasgow, 2009). Additional

situational factors studied include the employee’s social environment and the human

resources value an organization places on its employees (Tzafrir, Gur, & Blumen, 2015).

Other studies focused on the intentions or actions (i.e., thoughts of quitting or job

searches) that often immediately precede a turnover event (Lytell & Drasgow, 2009).

These antecedents have time links to the actual turnover event, limiting their utility as

prehire predictors.

One of the situational factors that may influence or predict employee turnover is

an employee’s assessment of alternative employment opportunities (Lytell & Drasgow,

2009; Mafini & Dubihlela, 2013; Rainayee, 2013; Smith, Holtom, & Mitchell, 2011).

However, past research on employees’ comparisons of alternatives as a predictor of

employee turnover has had mixed results (Lytell & Drasgow, 2009), and reportedly

related more to the environment than the individual (Pinelis & Huff, 2014). However, a

recent study of U.S. Air Force service members supports a correlation between

alternative employment options and separation (r = .19, p < .01) or retirement (r = .10, p

< .01; Smith et al., 2011).

These types of situational predictors are similar because they can change over

time, and the time between when employers measure them and when a turnover event

occurs complicates the measurement of their impact (Lytell & Drasgow, 2009). Lytell

and Drasgow (2009) used data from a 1999 Department of Defense survey and employee

turnover events from 1999 to 2002. Withdrawal intentions were the strongest predictor

with a hazard ratio of 2.42, meaning that individuals one standard deviation from the

mean are 2.42 times more likely to leave the military than those at the mean, ranging

from a 65% to 142% increased risk of turnover depending on the model (Lytell &

Drasgow, 2009). Other factors associated with employee turnover in this study included

job withdrawal (hazard ratio 1.29 and 15% to 29% increased risk of turnover), and

organizational commitment (hazard ratio 0.58, 12% to 42% increased risk of turnover;

Lytell & Drasgow, 2009). Although satisfaction with the military and perceived job

opportunities have hazard ratios of 0.72 and 0.69 respectively, they did not consistently

predict employee turnover in each model (Lytell & Drasgow, 2009). Although these

predictors may have utility once an individual is already an employee, they are not

measurable as a part of the hiring process.

Employee Turnover and Individual Attributes

When focusing on individual attributes, there are conflicting views on whether

high-performing employees are more or less likely to leave voluntarily (Nyberg, 2010).

Nyberg (2010) examined two different explanations for the effect of performance on

voluntary turnover. In the first case, the theory predicted that higher performers would be

less likely to leave voluntarily when there was a clear link between performance and

rewards as explained by expectancy theory, and when the ratio between work input and

outcomes was good compared to others as proposed by equity theory (Nyberg, 2010). On

the other hand, economic labor market theory postulates that high performing employees

will have more outside employment opportunities, thus making them more likely to leave

voluntarily (Nyberg, 2010).

A greater understanding of the relationship between cognitive ability and

employee turnover has significant implications for how organizations select and retain

their human resources (Erdogan et al., 2011a). In fact, potential employees who

demonstrate high cognitive ability, indicating they are likely to be top performers, may be

the same people who will leave the organization voluntarily (Maltarich et al., 2010),

although unemployment rates can affect this relationship (Kulkarni, Lengnick-Hall, &

Martinez, 2015), and the organization may be willing to accept a higher turnover rate for

less demanding positions in order to benefit from the personal attributes of overqualified

employees (Feldman & Maynard, 2011). Past research has indicated a negative

correlation between cognitive overqualification and job satisfaction (r = -.44; Fine &

Nevo, 2008), and job satisfaction with voluntary turnover (Maltarich et al., 2010).

Overqualification describes an employment situation in which an employee

possesses greater knowledge, skills, and abilities than the job requires (Hu et al., 2015). It

is possible to measure overqualification objectively by assessing specific job

requirements versus employee qualifications, or subjectively based on the employee’s

assessment of his or her qualifications compared to job requirements (Hu et al., 2015).

The majority of the prior research on this topic used employees’ on-the-job perceptions

of overqualification rather than an objective measurement and focused on current

employees rather than job applicants (Fine & Nevo, 2011). Prior research on

overqualified workers has shown they are less satisfied with their jobs, more likely to

engage in counterproductive work behaviors, and more likely to leave (Liu, Luksyte,

Zhou, Shi, & Wang, 2015; Lobene & Meade 2013; Maynard & Parfyonova, 2013).

From these findings, the association between cognitive overqualification and

increased employee turnover seems straightforward, yet the results of one study

examining this link found that the relationship was more complex (Maltarich et al.,

2010). The research hypothesis predicted a U-shaped relationship between cognitive

ability and voluntary turnover when comparing to others in similar jobs. For jobs with

high cognitive demands, including cognitive ability in the model improved fit over the

baseline model (Δχ2 = 9.07, p < .01), and produced a statistically significant negative

coefficient (HR = 0.71, two-tailed p < .01), but did not yield a statistically significant

result for jobs with low or medium cognitive demands (Maltarich et al., 2010). This

finding suggests that some high-cognitive-ability employees may intentionally choose

jobs with low cognitive demands (Maltarich et al., 2010).

Another, more recent study found that perceived peer overqualification moderated

the relationship between employee overqualification and negative outcomes such as

increased turnover behavior (Hu et al., 2015). In Hu et al.’s (2015) study, if employees

perceived that their individual situation was commensurate with their peers who were

similarly overqualified, it had a positive moderating effect on the relationship between

overqualification and task significance (β = .15, p < .01) and task significance related

positively to performance (β = .11, p < .05; Hu et al., 2015). As these studies show,

although overqualification is a complex issue, the empirical evidence indicates that it has

potential for prehire testing and possible utility for reducing employee turnover.

Research on factors affecting employee turnover has also included tenure and

career stage. In a study on U.S. Army soldier retention, G. Chen and Ployhart (2006)

collected longitudinal data over two years, including two related variables (military

tenure and rank) to examine the impact of career stage on employee turnover decisions.

For these career variables, military tenure and social support predicted job involvement

(β = -583, p < .05), which can function as social support, and is more important in the

early career stages (Chen & Ployhart, 2006). Although this was the only significant

finding related to differences in career stage, it demonstrates a need to consider career

stage as a factor when seeking a predictor of employee turnover. Chen and Ployhart’s

results also indicated that turnover intentions and the predictors thereof changed over

time and varied by individual. This finding is important because it highlights the need for

a more static turnover predictor.

The literature on employee selection and turnover also includes the personal

attributes of vocational interests and personality. A 2011 meta-analysis by van Iddekinge,

Roth, Putka, and Lanivich used 74 studies (41 journal articles, 17 dissertations and

theses, 14 technical reports, and two book chapters) resulting in 141 distinct samples to

explore the relationship between vocational interests and both employee performance and

turnover. The results of the meta-analysis indicated that vocational interests have

predictive value for employee turnover (corrected validity = –.22, k = 15), meaning that

people who are interested in the type of work that they do are more likely to continue

doing that work (van Iddekinge et al., 2011). In a similar fashion, a meta-analysis on the

importance of personality in retaining productive employees used the five-factor model

of personality, which comprises conscientiousness, emotional stability, agreeableness,

extraversion, and openness, to predict two effectiveness outcomes: high performance at

one end of the spectrum and withdrawal behaviors including employee turnover at the

other end of the spectrum (Li, Barrick, Zimmerman, & Chiaburu, 2014). The results

showed that the validity of conscientiousness, emotional stability, and agreeableness,

combined on aggregated withdrawal behavior, increased by 37%-55% over their

individual impact, leading to the conclusion that personality as an individual attribute

considered in an aggregated fashion may have valuable utility in predicting employee

turnover (Li et al., 2014).

Research on employee turnover has shown that applicant biodata can be a useful

predictive tool (Breaugh, 2014). In a recent study, applicants who applied previously,

included optional personal history information on their application, already had jobs, and

who came via a referral from another employee were less likely to leave voluntarily

(Breaugh, 2014). Employers can determine these biodata factors as part of the hiring

process, and, based on this research, they may reduce employee turnover.

Military Employee Turnover

Although there has been some variation, employee turnover trends for enlisted

sailors in the U.S. Navy show the highest rates of turnover at the 4-year and 20-year

points, with turnover averaging 73% of sailors leaving the Navy after four years of

service, and 42% leaving after 20 years of service as shown in Table 1 (Department of

Defense, 2011a).

Table 1

Navy Active Component Continuation Rates from 2000-2011

Years of

Service 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Average

4 66.4 70.4 75.6 75.7 72.1 70.9 73.2 71.1 72.9 73.5 79.7 78.6 73.34

20 39.0 46.2 55.4 45.5 38.2 37.1 38.9 39.6 42.9 42.2 41.1 41.6 42.31

Low military continuation rates have compelled a significant body of research on

employee turnover in the military, often focused on losses that occur within or at the

completion of the first term of enlistment. Similar to research on employee turnover in

the civilian sector, military research also fits into two categories based on situational or

environmental antecedents and individual attributes.

Military Turnover and Situational Antecedents

Like the civilian sector, there are several situational antecedents with empirical

evidence of a relationship with employee turnover. Work climate is a situational

antecedent that affects military retention in the South African Air Force (Mafini &

Dubihlela, 2013). The economy is also a key situational factor that correlates with the

retention of military personnel (Pinelis & Huff, 2014). A recent study examined the

relationship between the economy and the retention of U.S. Navy enlisted personnel

between 1992 and 2012 by combining the 11 variables in the Blue Chip Economic

Indicators into three subsets: unemployment and Treasury rate, production growth, and

price index (Pinelis & Huff, 2014). The results of Pinelis and Huff’s (2014) study

indicated a close link between the unemployment and Treasury rate and the employment

decisions of U.S. Navy sailors, where an increase of one standard deviation in the

unemployment and Treasury rate correlated to an 8.4% increase in retention of male

sailors in their first term of enlistment. Another antecedent of turnover is job

embeddedness, including both organizational and community embeddedness with three

components: compatibility/fit, formal and informal networks, and sacrifice/costs of

leaving (Smith et al., 2011). Organizational embeddedness was the stronger predictor for

reenlistment (r = -.25, p < .01) versus retirement (r = -.19, p < .01; Smith et al., 2011).

Community embeddedness was only significant once an Air Force service member was

eligible to retire (r = -.09, p < .01; Smith et al., 2011). Although these factors

undoubtedly contribute to employee turnover, they are not measurable as part of the

hiring process, and, therefore, they do not have utility for reducing employee turnover as

a pre-hire construct.

Military Turnover and Individual Attributes

Military employment is unusual in the 21st century because of the training

investment afforded to new recruits. The military does not expect applicants already to

possess the knowledge, skills, and abilities required for a particular job before it hires

them; instead, it uses cognitive ability testing to place individuals into career fields in the

military based on the likelihood they can complete the required training to acquire the

necessary skills. Due to this fixed investment in training new recruits, there has been

widespread research on military turnover to determine ways to reduce attrition and its

associated costs.

A significant amount of military research on employee turnover has been on

demographic and psychosocial factors, which include ASVAB test results, likely because

of the availability of this kind of data (Knapik, Jones, Hauret, Darakjy, & Piskator, 2004).

Other factors under study relative to first-term attrition include mental health, general

health, and physical fitness (Knapik et al., 2004), gender, age, country of birth/ethnicity,

military service, military occupation, highest education level, aptitude and cognitive

ability, marital status, children (Hoglin & Barton, 2013), paygrade (Pinelis & Huff,

2014), preentry expectations, attitudes and intentions (Ford, Gibson, DeCesare, Marsh, &

Griepentrog, 2013), vocational aspirations (Marcus & Wagner, 2015), preentry

commitment, desire for a military career, and mental toughness (Godlewski & Kline,

2012), job satisfaction, organizational commitment, job embeddedness, and person-

organization fit (Holtom, Smith, Lindsay, & Burton, 2014). Many of these individual

factors relate to military employee turnover.

Based on a 2004 literature review, the main reasons for military employee

turnover during the first six months of service include performance issues,

medical/physical problems, and fraudulent enlistments where individuals were not

qualified for service (Knapik et al., 2004). Employee losses during the remainder of an

initial service obligation (typically four years) have included misconduct, physical

problems, drug use, performance issues, and character or behavioral disorders (Arkes &

Mehay, 2014; Knapik et al., 2004). From this past research, performance issues, which

may relate to cognitive ability, are a factor in the entire first term (typically four to six

years). They accounted for 34% of losses in the first six months and 8% during the

remainder of the first term of enlistment (Knapik et al., 2004). Also, according to this

literature review, higher Armed Forces Qualification Test scores had weak associations

with lower employee turnover in 26 prior research studies (Knapik et al., 2004). Another

important factor in military retention is educational attainment. The results of 40 studies

on military attrition indicated a correlation between low educational attainment and

increased employee turnover (Knapik et al., 2004). Military members without a high-

school diploma are twice as likely to separate during their first term of enlistment as

those who have earned a high-school diploma (Knapik et al., 2004).

In a recent study of first-term attrition of military personnel in the Australian

Defence Force, 69% of all military recruits did not complete their initial three-to-six-year

contract obligation (Hoglin & Barton, 2013). The aim of these studies was to analyze

preenlistment predictors of first-term attrition, including gender, age, country of

birth/ethnicity, military service, military occupation, highest education level, aptitude and

cognitive ability, and marital status and children, and Hoglin and Barton (2013) found

that aptitude score, psychologist interview, and preenlistment level of education were the

most significant measures for predicting employee turnover. Military members who

completed 12 years of education were 54% more likely to complete their first term of

enlistment than those with only 10 years of education (p < .01; Hoglin & Barton, 2013).

Recruits with aptitude scores of seven or less were 29% less likely than those with

aptitude scores of 10 to complete their enlistment (p < .01). In addition, recruits received

a pre-enlistment assessment to determine their suitability for military service (on a scale

of 1 to 7, where 1 means totally unacceptable and 7 is outstanding); those who received a

psychologist interview rating of 2 were 22% less likely to complete than those with a

rating of 4 (p < .01; Hoglin & Barton, 2013). In addition, recruits with above-average

aptitude scores did not complete their first term of enlistment at a higher rate than those

with average scores (Hoglin & Barton, 2013). These results are consistent with the

theoretical framework for this research, indicating a need to control for education level,

which relates to cognitive ability, and providing further evidence to support the theory

that the level of cognitive fit may be useful in predicting turnover outcomes.

Other military research on retention has included cognitive ability as a potential

factor. In a U.S. Army study, G. Chen and Ployhart (2006) developed a retention model

to integrate situational and personal factors to determine if job attitudes and motivation

mediate the impact of personal factors and situational variables on turnover intentions.

The personal factor they chose as a way to highlight individual differences was general

cognitive ability—based on its historical use as the main tool the U.S. Army utilized to

select and place new recruits (Chen & Ployhart, 2006).

One hypothesis Chen and Ployhart (2006) examined was that cognitive ability

would negatively predict job attitudes. The premise of this hypothesis was the relative

incidence of highly complex jobs in the Army, compared to less challenging jobs as the

basis for expecting highly intelligent individuals, would be less likely to find their jobs

motivating and challenging (Chen & Ployhart, 2006). A second hypothesis was that work

characteristics, such as job challenge, task significance, and social support, would

moderate the negative influence of cognitive ability on job attitudes (Chen & Ployhart,

2006). The results of Chen and Ployhart’s study did not support either hypothesis;

cognitive ability did not predict job attitudes or turnover intention. Although this result is

contrary to the theoretical framework and the likely relationship for this research, it was

not unexpected, since Chen and Ployhart only utilized general cognitive ability and did

not explore cognitive fit between the individual and the job requirements.

General cognitive ability and level of education have separately correlated with

employee turnover (Hoglin & Barton, 2013; Knapik et al., 2004). Based on their

predictive relevance, the U.S. military often combines these two factors to measure the

quality of accessions (Pinelis & Huff, 2014; White, Rumsey, Mullins, Nye, & LaPort,

2014). According to a recent study by Pinelis and Huff (2014) on the economy and U.S.

Navy enlisted retention, high-quality sailors, defined as those with a high-school diploma

and an Armed Forces Qualification Test (AFQT) score of 50 or higher, were less likely to

reenlist; men 5.1%, and women 3.3%. They noted this relationship as a concern because

the Navy’s percentage of sailors meeting the definition of high quality increased from

64.9% in 2007 to 87.4% in 2011 (Pinelis & Huff, 2014). Pinelis and Huff’s finding on

quality and retention is important because it highlights the need to explore the impact of

education level on the relationship between cognitive fit and employee turnover. The

increase in Navy accession quality is also relevant because, based on the results of Pinelis

and Huff’s study and some of the previous research, higher general cognitive ability

correlates to lower retention. However, since the Navy implemented its new process for

job placement based on cognitive fit in 2009, this result may no longer be valid. Since the

key question in this research was whether cognitive fit is more meaningful than general

cognitive ability for future retention, it is useful in assessing the impact of this Navy

policy change, and providing greater understanding regarding the correlation between

cognitive ability and retention.

Paygrade is another individual attribute with a relationship to reenlistment. In the

military enlisted ranks, E-1 is the most junior paygrade, and E-9 is the most senior

paygrade. In the same study, Pinelis and Huff (2014) found a relationship between

paygrade and reenlistment, where more junior E-3 sailors were 28.4% (male) and 23.9%

(female) less likely to reenlist at the end of their first term than those who moved two

paygrades higher during that same timeframe to the rank of E-5. Furthermore, E-4 sailors

were 9.2% (male) and 10.0% (female) less likely to reenlist than those who were one

paygrade higher at the rank of E-5 (Pinelis & Huff, 2014). Interestingly, E-6 sailors,

which is the highest paygrade possible for a sailor reach during a first enlistment, were

3.1% (male) and 5.5% (female) less likely to reenlist than E-5 sailors, which may indicate

a similarity between this result and the lower likelihood of high-quality sailors reenlisting

(Pinelis & Huff, 2014). Clearly an individual’s paygrade may also relate to cognitive fit

and employee turnover.

Preentry expectations, attitudes, and intentions have had predictive value for

determining military tenure (Ford et al., 2013). Using sample data from individuals

during late youth and early adulthood and Cox regression analysis, preenlistment

expectations regarding quality of life significantly predicted tenure (β = -.36, p < .05;

Ford et al., 2013). In Ford et al.’s (2013) study, preentry attitudes were also significant

predictors of military tenure (β = -.26, p < .05). Ford et al. asked participants about their

intent to join the military, with possible responses on a four-item Likert-type scale

including definitely not, probably not, probably, and definitely. Participants who chose

definitely, probably, and probably not were all less likely to leave the military than those

who indicated they were definitely not joining the military (β = -.34, -.24, and -.27

respectively, p < .05; Ford et al., 2013). These results imply that individuals who have

positive attitudes and expectations about the military, and intentions to join the military in

late youth or early childhood, are less likely to leave (Ford et al., 2013). Marcus and

Wagner (2015) obtained a related result when assessing the validity of vocational

aspirations in employment outcomes. Simply put, they found that individuals who

attained their aspired vocation—living out their personal answer to the question “what do

you want to be when you grow up,”—had greater job satisfaction and higher performance

than those who worked in a career field that matched their vocational interests (person-

vocation fit; Marcus & Wagner, 2015).

A longitudinal study of the Canadian armed forces (including individuals from the

Army, Navy, and Air Force) also focused on preentry employment factors including

preentry commitment, desire for a military career, and mental toughness (Godlewski &

Kline, 2012). The concepts of preentry commitment and desire for a military career were

similar to preentry expectations, attitudes and intentions, and vocational aspirations

(Godlewski & Kline, 2012). Godlewski and Kline (2012) defined mental toughness as

control—acting in an influential manner, commitment—the tendency to engage in

situations rather than remain apart, challenge—the belief that change is a normal part of

life, and confidence—the belief in one’s ability to achieve success. In their model, these

three preentry factors predicted work attitudes, including initial adjustment and

organizational commitment (Godlewski & Kline, 2012). Organizational commitment

(both normative and affective) then predicted turnover intentions and actual employee

turnover (Godlewski & Kline, 2012).

In a similar study, Holtom et al. (2014) explored job attitudes and job

embeddedness for their utility in predicting turnover in at the U.S. Air Force Academy.

They defined person-organization fit as the compatibility between an individual and an

organization, and they operationalized it through survey questions to ascertain value and

goal congruence between them (Holtom et al., 2014). When compared to job satisfaction,

organizational commitment, and job embeddedness, person-organization fit was the most

powerful predictor of turnover (r = -.13, p < .01; Holtom et al., 2014). The relative

weight of person-organization fit in explaining the variance was 45.03%, followed by job

embeddedness (19.69%) and job satisfaction (14.81%; Holtom et al., 2014). Although

person-organization fit may develop over time rather than exist preentry, there is some

similarity between the concepts of expectations about the military and person-

organization fit, and both have negative relationships to turnover.

The wide range of individual factors in the study of military employee turnover

clearly outlines the interest in this topic and the complexity of this issue. Although this

research did not take account of several factors including mental health, general health,

physical fitness, country of birth/ethnicity, children, preentry commitment, expectations,

attitudes and intentions, vocational aspirations, desire for a military career, mental

toughness, job satisfaction, organizational commitment, job embeddedness, and person-

organization fit, they may be good candidates for future study in relation to cognitive fit.

The researcher included many of the other factors under discussion here in this research:

military service (Navy) and occupation, cognitive ability, gender, and length of service.

Employee Fit

Interactional psychology described in simple terms is the relationship between a

person and his or her environment (Kristof-Brown & Guay, 2011). The definition of the

concept of employee fit, fundamentally based in interactional psychology (Kristof-Brown

& Guay, 2011), is the compatibility between an individual and his or her work

environment, otherwise stated as person-environment fit (Billsberry et al., 2012; Duffy,

Autin, & Bott, 2015; Kristof-Brown & Billsberry, 2012; Kristof-Brown & Guay, 2011;

Maynard & Parfyonova, 2013; Thompson et al., 2015). The concept of person-

environment fit is prevalent in industrial and organizational psychology and in the human

resources management literature (Kristof-Brown & Guay, 2011).

Person-environment fit. Many different personal attributes and environmental

factors may be relevant to person-environment fit (Kristof-Brown & Guay, 2011). From

the broad definition of person-environment fit as the compatibility between an individual

and a work environment, several different dimensions have emerged, including person-

vocation, person-job, person-organization, person-group/team, and person-individual fit

(Kristof-Brown & Billsberry, 2012; Kristof-Brown & Guay, 2011). Learning fit is

another new conceptualization with demonstrated benefits in job satisfaction (Felstead,

Gallie, Green, & Inanc, 2015). Scholars have further operationalized each of these types

of fit to facilitate measurement (see Figure 2). As Figure 2 shows, the concept of fit is

popular and it has resulted in the outgrowth of many different conceptualizations of the

relevant factors for person-environment fit and its various dimensions (Kristof-Brown &

Guay, 2011). This upsurge of fit conceptualizations and dimensions has led to a call for

more precise definitions and constructs (Kristof-Brown & Guay, 2011). It has also caused

a discussion of the frame of reference; whether to compare the person to the environment

or the environment to the person (Hardin & Donaldson, 2014; Kristof-Brown & Guay,

2011). Most prior research has measured the extent to which a person fits in a work

environment, but recent developments have indicated that either the person or the

environment has utility as the frame of reference (i.e., the extent a person matches the

environment or the environment matches the person; Hardin & Donaldson, 2014).

Figure 2. Employee fit—Types and relationships. Person-environment (PE) fit is the

relationship between many individual and organizational attributes. This diagram shows

the conceptualizations of fit in the literature on PE fit, the relationship between them, and

the types pertinent to the proposed research. Developed from information in Kristof-

Brown and Guay (2011).

Person-vocation fit. Although not a primary focus of this research, person-

vocation fit bears mentioning because of its historical underpinnings and relevance to the

Navy’s job placement process. The history of vocational choice theories is long and

includes seminal works such as Frank Parson’s guidance on choosing a career in the early

1900s, Donald Super’s life-span approach proposing growth, exploration, establishment,

maintenance, and disengagement career stages in the mid-1900s, and John Holland’s

RIASEC model using six occupational types (realistic, investigative, artistic, social,

enterprising, and conventional; Kristof-Brown et al., 2005). These vocational choice

theories represent the origin of employee fit and the person-vocation fit dimension of

person-environment fit (Kristof-Brown & Guay, 2011; Marcus & Wagner, 2015).

Person-vocation fit is also relevant to the Navy’s job placement process because

one can argue either that the military/Navy is a vocation or that individual career fields

within the military/Navy are separate vocations. On one hand, there are several aspects of

military/Navy life that are similar regardless of individual career fields, so one could

define the military as a vocation. On the other hand, work in individual career fields may

vary widely, from administrative work in an office environment to mechanical work on a

flight line, so one could define each career field as a vocation. Additionally, in the Navy

one must transfer from job to job within a career field, and the jobs one can choose vary

in specifics including location, job tasks, and experience level. However, although

person-vocation fit lends itself to research on individual military/Navy career fields, this

research instead focuses on the concept of person-job fit, and specifically, the dimension

demands-abilities fit, based on its applicability to the cognitive testing of all military

applicants for the explicit process of selecting and placing applicants into military jobs.

Person-job fit. The category of fit that is most relevant to this research is person-

job fit. Person-job fit is the relationship between the requirements of a job and the

characteristics of an employee (Boon et al., 2011; C. Chen et al., 2014; Gabriel et al.,

2014; Kristof-Brown & Guay, 2011). Person-job fit is a concept hiring officials use

because it has solid legal support for use in making selection decisions (Sekiguchi &

Huber, 2011). In much of the prior research, researchers have measured person-job fit

subjectively using a survey and asking individuals if they perceive their skills and

abilities are a good match for the requirements of their job (Boon et al., 2011; Freund &

Kasten, 2012).

Empirical research demonstrates a significant relationship between person-job fit

and several positive employment outcomes in many settings. In a meta-analysis that

included 62 studies and 225 effect sizes, Kristof-Brown et al. (2005) found a strong

correlation between person-job fit and job satisfaction (p = .56), organizational

commitment (p = .47), and intent to quit (p = -.46). Quratulain and Khan (2015) also

demonstrated that person-job fit has a positive effect on job satisfaction (β = .43, p < .01),

although that effect was weaker if the employee perceived high work pressure (β = -.17, p

< .01). Y. Peng and Mao (2015) obtained a similar result where person-job fit positively

correlated with job satisfaction (r = -0.443, p < .01). Han, Chiang, McConville, and

Chiang (2015) found that person-job fit correlated positively with psychological

ownership (β = .52, p < .01), which they defined as a feeling of ownership about their

jobs, and that psychological ownership had a positive correlation with contextual

performance (β = .44, p < .01), which includes organizational citizenship behaviors.

Farzaneh, Farashah, and Kazemi (2014) found that person-job fit positively influenced

organizational commitment (β = 0.14, p < .01) and that organizational commitment

significantly affected organizational citizenship behaviors (β = .51, p < .01). In addition

to this indirect relationship, person-job fit also related directly to organizational

citizenship behaviors (β = .06, p < .05; Farzaneh et al., 2014). Person-job fit related

positively to performance (β = .675, p < .001) and sense of well-being (β =.809, p < .001;

Lin, Yu &Yi, 2014). Another study demonstrated a significant relationship between

person-job fit and innovative work behavior (γ = .23, p < .05; Afsar, Badir, & Khan,

2015). Finally, person-job fit related negatively to employee burnout (β = -17, t = -3.34)

and turnover intentions (β = -.46, t = -12.91; Babakus, Yavas, & Ashill, 2011). These

results highlight the benefits of strong person-job fit because key employment outcomes

such as job satisfaction, psychological ownership, organizational commitment,

organizational citizenship, and innovative work behavior are likely to result in

performance and retention. Just as important, person-job fit also relates negatively to

intent to quit and burnout, which may lead to employee turnover.

Some of the research on person-job fit has focused on its association to personal

influencers such as general self-efficacy and vocational interest in making career choices.

General self-efficacy an individual’s self-perception of his or her ability to perform in a

wide-range of situations, or in other words, his or her self-confidence in his or her coping

skills (Song & Chon, 2012). General self-efficacy is a core component of self-evaluation,

and it relates directly to person-job fit (β = .426, 95% bias-corrected bootstrap confidence

interval of .294-.546, SE = .064, p = .001) and indirectly to career choice through person-

job fit and vocational interests (βstandardized = .371, 95% bias-corrected bootstrap

confidence interval of .239-.519, SE = .072, p = .000; Song & Chon, 2012).

In the context of employee well-being, Warr and Inceoglu (2012) examined the

associations between person-job fit and both job engagement and job satisfaction. The

method they used compared wanted job features to actual job features to measure person-

job fit, where job features included a supportive environment, competition and financial

focus, personal influence, challenging workload, ethical principles, career progress,

amount of social contact, and status (Warr & Inceoglu, 2012). A poor fit between wanted

and actual job features resulted in a significant negative association with job satisfaction

(r = -.14) and a positive relationship to job engagement (r = .27; Warr & Inceoglu, 2012).

Other types of fit may interact with person-job fit in employment decisions. J.

Peng et al. (2014) examined the interaction between person-job fit and person-

organization fit and theorized that a person with high person-organization fit, but low

person-job fit, may be more likely to leave, while a person with high person-organization

fit and high person-job fit may be more likely to stay. As they expected, person-

organization fit had a significant negative relationship with turnover intentions (β = -.273,

p < .001; J. Peng et al., 2014). The interaction between person-job fit and person-

organization fit related significantly to turnover intentions (β = -.154, p < .01) and was

stronger when person-job fit was high than when person-job fit was low (J. Peng et al.,

2014). On the other hand, Christensen and Wright (2011) researched the influence of

person-organization fit on job choice, attempting to isolate the effects of person-

organization fit and person-job fit. They found, after controlling for person-job fit, that

person-organization fit (operationalized as public service motivation) did not increase the

likelihood of choosing a public-service job, implying that person-job fit may play a more

important role in job choice than person-organization fit (Christensen & Wright, 2011).

Workplace or self-modification strategies may improve person-job fit over time

(Hinami, Whelan, Miller, Wolosin, & Wetterneck, 2013). In a population of hospitalists,

job-switching early in a career improved person-job fit (median fit was slightly but

statistically significantly higher for individuals who made one job change; 4.4 v. 4.0 on a

5-point Likert-type scale), indicating that individuals recognize and act to improve fit—

often with positive results (Hinami et al., 2013). Job modification strategies, such as

adjusting work hours or workload, were effective in improving person-job fit for

established employees (Hinami et al., 2013). Hinami et al. (2013) also demonstrated that

employees gradually increased person-job fit over time, likely through experiential

learning and socialization/value sharing (Spearman coefficient r = .149; p < .001; Hinami

et al., 2013).

Researchers have conceptualized person-job fit with two dimensions; demands-

abilities fit, and needs-supplies or supplies-values fit (C. Chen et al., 2014; Kristof-Brown

et al., 2005). Needs-supplies or supplies-values fit measures the match between the

individual’s needs, preferences, and desires, and what the job provides (C. Chen et al.,

2014; Kristof-Brown et al., 2005). Demand-abilities fit is the congruence between a

person’s knowledge, skills, and abilities, and job tasks (Kristof-Brown et al., 2005), and

employers typically measure it through the employee’s perception of this match (Bogler

& Nir, 2015; Kristof-Brown & Billsberry, 2012; Melvin, Hale, & Foster, 2013).

Demands-abilities fit. Demands-abilities fit is the match between the demands of

a job, and an individual’s abilities (Park, Beehr, Han, & Grebner, 2012). The basis of the

concept of demands-abilities fit is traditional hiring practices in which employers select

and hire an individual for a job based on a comparison of his or her abilities with the

requirements of the job (Kristof-Brown & Guay, 2011). Two other aspects of demands-

abilities fit that are important to the relationship between an individual and a specific job

are time and energy (Park et al., 2012).

A recent study described the content dimensions of demands-abilities as

quantitative workload and job complexity, defining job complexity as the level of skill

utilization compared to a job’s mental requirements (Park et al., 2012), which is a

concept similar to cognitive fit. Park et al. (2012) used the difference between demands

and abilities to measure fit, and showed it had a positive relationship to psychological

strain, both anxiety (r = .23, p < .01) and depression (r = .18, p < .01), indicating that

those with greater abilities than needed on the job experienced less strain, and those with

greater demands than abilities experienced more strain (Park et al., 2012). Optimism,

internal locus of control, and self-efficacy all weakly moderated this relationship (Park et

al., 2012). Of note, the use of fit difference in Park et al.’s study is akin to the

researcher’s method of measuring cognitive fit.

There is evidence that demand-abilities fit is relevant to several key employment

outcomes. Research on demands-abilities fit further supported results on person-job fit,

showing that an employee’s perception of the fit between his or her abilities and job

demands predicted both job commitment and job satisfaction (Bogler & Nir, 2015;

Kristof-Brown et al., 2005; McKee-Ryan, & Harvey, 2011). Demand-abilities fit relates

to job meaningfulness, which includes three elements: work that is meaningful, has

meaningful consequences, and has a positive impact on others (Tims et al., 2016).

Theorizing that organizational effectiveness relates to job commitment and job

satisfaction, Bogler and Nir (2015) were interested in finding factors that predicted these

organizational outcomes in elementary school teachers. The results of their study

indicated that a teacher’s perceived fit between demands and abilities was the single

variable that affected all four outcomes tested: organizational commitment (R2 adjusted

=.165; p < .001), professional commitment (R2 adjusted = .222; p < .001), intrinsic

satisfaction (R2 adjusted = .336; p < .001), and extrinsic satisfaction (R2 adjusted = .224;

p < .001; Bogler & Nir, 2015). Gabriel et al. (2014) explored the causal relationship

between person-job-fit and job satisfaction and found that the perception of person-job fit

predicted job satisfaction (γ = .03, p < .05).

In a recent study about turnover intentions, demand-ability fit negatively

correlated with turnover intentions (r = -.16, p < .01; J. Peng et al., 2014). In the same

study, demand-ability fit had a positive correlation with work engagement (r = .44, p <

.01) and there was a significant positive correlation between work engagement and

turnover intentions (r = -.51, p < .01; Peng et al., 2014). Additionally, recent discussions

about fit have highlighted the need to include time as a variable; fit may be dynamic

because both individuals and organizations change (Gabriel et al., 2014). These results

support the theoretical foundations of this research to explore the relationship between

cognitive fit and employee turnover.

Cognitive Ability

The definition of cognitive ability is an individual’s ability to learn (Ones &

Viswesvaran, 2011). Cognitive ability as another term for general intelligence, namely

knowledge, recall of knowledge, and ability to work with knowledge (Mumford et al.,

2015) or as the capacity to problem-solve, plan ahead, and learn from experience (Oh et

al., 2014). Performance is arguably the most important construct for measuring employee

value to an organization (Maltarich et al., 2010; Ones & Viswesvaran, 2011) and when

one is selecting people to hire, many regard general cognitive ability as the most

powerful predictor of job performance (Ones & Viswesvaran, 2011).

Another study of interest, based on the similarity of the test population and

cognitive ability testing method, reported that cognitive ability predicted task

performance at β = .54 (Oh et al., 2014). This research population was South Korean

military officers, and the method of measuring cognitive ability was the Korean Police

Officers Aptitude Battery (Oh et al., 2014). The results showed the relative weight for

predicting task performance of cognitive ability was 58.93%, followed by

conscientiousness (33.22%), and openness to experience (3.30%; Oh et al., 2014). The

other predictors Oh et al. (2014) tested included emotionality, extraversion,

agreeableness, and honesty-humility, all of which had a relative weight of less than 3%.

In addition to its predictive value for employee performance, Maltarich et al. (2011) have

recognized cognitive ability as an important objective measurement of job qualification.

It provides a more precise measure of an individual’s on-the-job mental challenge than

other measures of job skill such as education or experience (Fine & Nevo, 2008). Of note,

there may be some potential for adverse impacts when conducting cognitive testing

(Klein, Dilchert, Ones, & Dages, 2015), and differences in perceptions about cognitive

ability between older and younger employees (Truxillo, McCune, Bertolino, &

Fraccaroli, 2012).

Some studies have claimed that lower cognitive ability is better for environments

characterized by time pressure and unpredictable task changes (Beier & Oswald, 2012).

A recent review of this literature using the resource theories of cognitive processing and

skill acquisition as a theoretical framework did not support this claim. Beier and Oswald

(2012) proposed that the direction for future research should include broadening the

range of skills and abilities examined.

Research on the longer term value of cognitive ability in relation to civilian

employee turnover is scarce (Maltarich et al., 2010; Ryan & Ployhart, 2014; Zaccaro et

al., 2015) and the results are mixed (Boudreau et al., 2001). Studies on the relationship

between general cognitive ability and turnover in the civilian sector have only shown a

correlation coefficient of 0.02, meaning that as cognitive ability increases, the likelihood

of turnover slightly increases (Allen et al., 2010). So, employees who demonstrate high

cognitive ability, indicating they are likely to be top performers, may be the same people

who will leave the organization voluntarily (Maltarich et al., 2010). However, Maltarich

et al. (2010) demonstrated that the cognitive demands of a job are pertinent to employee

turnover decisions for jobs with high cognitive demands using job satisfaction as a partial

mediator.

The study conducted by Maltarich et al. (2010) was the first of its kind to examine

the relationship between voluntary turnover and the alignment between an individual’s

cognitive ability and the cognitive demands of a job. The observations for their study

came from the National Longitudinal Survey of Youth, 1979 Cohort, the data for which

included respondent’s results from the ASVAB (Maltarich et al., 2010). To determine the

cognitive demands of particular jobs, Maltarich et al. collected average levels of ability

from the Occupational Information Network webpage. They designed their research

using a product of coefficients method to relate cognitive ability to job satisfaction and

job satisfaction to predict voluntary turnover (Maltarich et al., 2010).

For jobs with high cognitive demands, both coefficients were statistically

significant (β = -.03, one-tailed p < .05; loge (HR) = -0.48, one-tailed p < .001), leading to

a statistically significant product of the coefficients (z = 1.70, one-tailed p < .05;

Maltarich et al., 2010). While the results for jobs with low or medium cognitive demands

did not show a significant relationship between job satisfaction and cognitive ability and

voluntary turnover, the results for jobs with high cognitive demands suggested that the fit

between demands of the job and the abilities of the individual (demand-ability cognitive

fit) may be important for understanding turnover outcomes. Additionally, the results of

this study could indicate that some individuals with high cognitive ability may be

intentionally choosing jobs that have low cognitive demands and that they may retain

well (Erdogan et al., 2011a; Erdogan, Bauer, Peiro, & Truxillo, 2011b; Maltarich et al.,

2010; Thompson et al., 2013). This conjecture could refute past claims that hiring

overqualified applicants could result in increased employee turnover (Maltarich et al.,

2010). Overall, Maltarich et al.’s (2010) study offers evidence that there may be an

important relationship between cognitive fit and employee turnover that needs additional

examination.

Cognitive Testing in the U.S. Military

The U.S. military has been a leader in using testing as a selection screening tool

since 1917, when the Army developed the Alpha and Beta tests (Rumsey, 2012; Rumsey

& Arabian, 2014a). Competition between the branches of the military for the ablest

recruits led to the Selective Service Act of 1948 to improve the equitable distribution of

human talent (Held, Hezlett, et al., 2014). Implemented in 1950, the AFQT was a single

test for all branches of the military and it included a minimum score for service entry

(Held, Hezlett, et al., 2014). During this time, the United States still had a conscripted

military, and it used the AFQT categories as the basis for equally distributing both the

most highly qualified and the lowest qualified individuals into the services (Held, Hezlett,

et al., 2014). In 1974, based on the shift to an all-volunteer force, and the success of the

ASVAB in the Air Force and Marine Corps, the Department of Defense directed the use

of a single test battery for both selection and classification in all branches of the U.S.

military (Watson, 2010). In response, the military implemented the Armed Services

Vocational Aptitude Battery (ASVAB) service-wide in January 1976 (Held, Hezlett, et

al., 2014). In 1996-97, the military updated to the Computerized Adaptive Test (CAT-

ASVAB) which includes nine sub-tests, itemized in Table 2 (Held, Hezlett, et al., 2014).

Table 2

Armed Services Vocational Aptitude Battery (ASVAB) Sub-Tests

ABBREVIATION SUBTEST

AR Arithmetic Reasoning

WK Word Knowledge

PC Paragraph Comprehension

MK Mathematics Knowledge

GS General Science

EI Electronics Information

AS Auto and Shop Information

MC Mechanical Comprehension

AO Assembling Objects

Note. The source for the ASVAB sub-tests is the ASVAB website at http://official-

asvab.com/docs/asvab_fact_sheet.pdf.

All of the U.S. Armed Services use the ASVAB as a cognitive screening tool to

determine eligibility for service based on minimum qualification. They do not use

ASVAB subtest scores separately; they combine them into various composites (Held,

Hezlett, et al., 2014). The AFQT score, which is a composite score that includes the

arithmetic reasoning, word knowledge, paragraph comprehension, and mathematics

knowledge sub-tests, is how the services measure general cognitive ability (Arkes &

Cunha, 2015; Held, Hezlett, et al., 2014). The other services develop and use other

composites individually (Grant et al., 2012). For example, an Army composite called

skilled technical or ST is a composite of GS + MK + MC + VE (Grant et al., 2012). This

composite has proven an accurate predictor of training success for the Army’s Operating

Room Specialist course (p < 0.0001), with a 5-time improvement in the odds of first

attempt completion for each increase of 10 points in the ST composite score (Grant et al.,

2012).

The ASVAB test is not static—the subtests have changed over time (Rumsey,

2012; Rumsey & Arabian, 2014b). There is ongoing research focused on adding two

additional sub-tests based on the importance of skill in receiving, transmitting, and

interpreting computerized information, and integrating the Assembling Objects subtest

into the AFQT (Held & Carretta 2013; Held, Carretta, & Rumsey, 2014; Rumsey &

Arabian, 2014a, 2104b; Trippe, Moriarty, Russell, Carretta, & Beatty, 2014). Another

study found that a composite of ASVAB test scores had utility in predicting skill in

multi-tasking (Hambrick et al., 2011).

For classification into a career field, each of the services uses different

combinations of subtests as composite scores based on recruit training success (Held,

Hezlett, et al., 2014). The Navy uses training success to define cognitive requirements by

career field and gender (Watson, 2010). Each rating-gender combination has a cognitive

requirement, which the Navy developed by tracking the ASVAB line scores of sailors

who successfully complete entry-level technical training without setbacks, which it calls

first-pass pipeline success (Watson, 2010). The Navy then aggregates these line scores to

develop minimum requirements.

The Navy administers the ASVAB to over one million individuals annually (Held,

Hezlett, et al., 2014). With nearly 40 years of use, study sample sizes are large and they

have inspired multiple studies (Held, Hezlett, et al., 2014). Of note, in a review of the

literature on attrition from the military services, higher AFQT scores had a weak

association with lower employee turnover in 26 prior research studies (Knapik et al.,

2004).

Navy’s Algorithm for Cognitive Fit

In 2009, the U.S. Navy redesigned its job placement process to improve initial

training success (Watson, 2010) and it began using a decision support system using

cognitive fit called the Rating Identification Engine (RIDE) to match individuals to

available jobs (Rumsey, 2012). Before implementing RIDE, the Navy only used ASVAB

scores to limit the placement of sailors into ratings where they would face challenges

(Watson, 2010). As long as an applicant met the minimum requirements, he or she was

eligible for the rating. RIDE improves upon this process by recognizing that classifying

sailors in ratings where they are overqualified, and therefore underchallenged, may be

just as bad as placing them in ratings where they are overchallenged.

The Navy developed its RIDE algorithm to improve personnel utilization by

increasing job satisfaction, reducing attrition, and promoting retention (Watson, 2010).

The Navy used the Yerkes-Dodson law as the theoretical framework for RIDE (Watson,

2010; Yerkes & Dodson, 1908). Yerkes and Dodson (1908) found that moderate levels of

electrical stimulus were the most effective in rapid habit formation. This visualization of

this relationship is an inverted U, and subsequent research has developed it further to

apply to the relationship between human performance and cognitive arousal (Watson,

2010). Using this framework, individuals who are underchallenged or overchallenged in

the context of cognitive ability are less likely to perform well than individuals who are

appropriately challenged. This concept is similar to work engagement, meaning an

individual’s physical, cognitive, and emotional involvement in the workplace (Bakker,

2011; Venz & Sonnentag, 2015).

The RIDE algorithm works to place individuals in ratings where their cognitive

ability closely matches that of other successful sailors assigned to the rating (Watson,

2010). RIDE S-score and Q-score utility curves for each rating using 75,000 Navy

recruiting and training records from 1996-1998 (Watson, 2010), and subsequently

updated with 60,000 records from 2011-2013. Watson (2010) used a comparison of

actual ASVAB scores of sailors with successful completion of the training pipeline

(without repeating any portion) for each Navy enlisted rating to build the S-score utility

curve. The purpose of the Q-score utility curve is to measure overqualification by

comparing an individual’s cognitive ability, using his or her AFQT score as a measure of

his or her overall general cognitive ability, to other applicants who go to the rating

(Watson, 2010).

The Navy compares sailors’ test scores to the utility curves to give an S-score and

a Q-score for each rating, and the Navy uses the composite of these two utility scores as a

measure of cognitive fit for each rating (Watson, 2010). RIDE identifies all of the Navy

jobs for which an individual is qualified, rank orders them according to this cognitive fit,

and then searches for job availability (Held, Carrera, & Rumsey, 2014). A review of the

impact of this new process showed that sailors with high cognitive fit were more likely to

complete their initial training, more likely to receive promotion, and less-likely to leave

(Department of the Navy, 2012). However, because job placement operates on a first-

come, first-served basis, job availability limits the process. This process constraint

reduces the ability of RIDE to optimize cognitive fit, and in some cases inevitably results

in sailors ending up in jobs where they are over- or underchallenged.

Based on the initial review of RIDE results (Department of the Navy, 2012), it

seems reasonable that sailors who are cognitively overqualified or underqualified (i.e.,

low demands-abilities fit) will have a higher turnover rate. However, as previously noted,

prior research results are mixed (Boudreau et al., 2001; Maltarich et al., 2010). In fact,

some research has identified a subset of workers with high cognitive ability who

purposefully choose jobs with low cognitive demands and do not leave (Maltarich et al.,

2010). These results signal a complex relationship between cognitive fit and employee

turnover and a need for additional research.

Summary

It is usual to consider employee selection and turnover separately (Maltarich et

al., 2010). Current research on employee turnover primarily focuses on the situational

antecedents to turnover events (Boudreau et al., 2001; Hom et al., 2012) rather than

individual factors that could act as predictors of future turnover when hiring an employee.

Since employers often use cognitive ability in hiring decisions, it is a measurable prehire

attribute that may be relevant in predicting future employee retention. Cognitive ability is

as an individual’s ability to learn (Ones & Viswesvaran, 2011) and there is wide

acceptance of its generalizability as a predictor of job performance (Schmidt, 2014), but

the majority of research on cognitive ability focuses on selection, hiring, and performance

without exploring their relationships to employee retention and turnover decisions. When

researchers have studied general cognitive ability in relation to turnover, they have only

found a small relationship with an effect size of 0.02, meaning that as cognitive ability

increases, the likelihood of turnover slightly increases (Allen et al., 2010). However,

Maltarich et al. (2010) demonstrated that the cognitive demands of a job are pertinent to

employee turnover decisions, which suggests that the cognitive fit between the demands

of the job and the abilities of the individual may offer a more relevant predictor of future

turnover outcomes than general cognitive ability.

The U.S. military has used cognitive ability testing to prescribe minimum

requirements for new recruits since World War II. Subsequent research on cognitive

ability has produced mixed results when correlated with retention, but traditionally, the

basis of these studies has been general cognitive ability rather than the match between the

cognitive demands of a specific career field and an individual’s cognitive ability (Allen et

al., 2010; Knapik et al., 2004). On the other hand, past research on person-job fit in the

civilian sector, and more specifically demands-abilities fit, has shown strong correlation

to job commitment, job satisfaction, and intent to quit (Kristof-Brown et al., 2005). These

results provide support for further research on the relationship between cognitive fit and

employee turnover. In addition, most recent research on demands-abilities fit used

subjective measurements of an individual’s perceived fit via survey response (Bogler &

Nir, 2015; Freund & Kasten, 2012). While subjective measurements of fit may be useful

in examining outcomes such as job satisfaction or dissatisfaction, an objective measure of

fit may provide a more useful measure for hiring new employees (Fine & Nevo, 2011),

and for predicting employee turnover.

The U.S. Navy’s hiring process offers an opportunity to examine the relationship

between cognitive ability and employee turnover using the concept of demands-abilities

fit in a new way. This quantitative study contributes to the body of knowledge by

examining the relationship between demands-abilities fit and employee turnover, using

cognitive ability as an objective measurement as recommended by Maltarich, Reilly, and

Nyberg (2011) and Lu, Wang, Lu, Du, and Bakker (2014). The research design for this

study used individual ASVAB scores and the RIDE algorithm to measure cognitive

ability against the Navy’s ASVAB standards for individual occupations to determine an

objective measurement of cognitive fit. The results of this research may lead to

improvements in selection and placement of new employees based on the ability to

predict future performance and retention using cognitive fit.

Chapter 3: Research Method

Employee turnover is a prime concern the U.S. Navy (Pinelis & Huff, 2014).

Failure to retain high-performing sailors in the U.S. Navy increases recruitment and

reenlistment costs, and results in the promotion of lower quality and less experienced

Navy personnel. The Navy uses monetary bonuses (with an average cost of $47,948.00

per enlisted sailor offered a bonus) as an incentive to encourage sailors to stay based on

their skill set and manning level, training costs, or criticality to the mission (Coughlan et

al., 2014; Pinelis & Huff, 2014). When not enough sailors remain, the Navy recruits and

trains additional sailors; however, it only hires them at entry level—leaving an

experience gap. Additionally, the Navy promotes sailors according to vacancies at the

next higher paygrade (Arkes & Cunha, 2015; Kumazawa, 2010). The Navy orders sailors

in a competitive group based on several factors including advancement exam scores,

performance evaluations, education, and awards to determine their relative quality

(Kumazawa, 2010). However, this only results in the best quality sailors gaining

promotion if there are fewer vacancies than sailors eligible to promote because, if the

number of vacancies is higher than the number eligible to promote, the entire competitive

group will receive promotion to fill the Navy’s requirements, regardless of the sailors’

quality or experience. These undesirable outcomes highlight retention as fundamental to

workforce quality in an entry-level hiring system. As a potential strategy for the U.S.

Navy to reduce personnel costs and maintain a high-quality workforce, this study

determined the extent to which a measurement of cognitive fit, calculated using a sailor’s

cognitive ability measured by his or her ASVAB test results compared to the cognitive

ability requirements for the job he or she receives, may predict employee turnover.

The purpose of this non-experimental, quantitative study was to examine the

relationship between cognitive fit and employee turnover in the U.S. Navy. The U.S.

Navy measures cognitive ability through the ASVAB and uses the results in the hiring

process for those desiring to enlist. The researcher used secondary case-file data from the

U.S. Navy’s Career Waypoints personnel database for all enlisted sailor retention

decisions that occurred in 2014. The data include S- and Q-scores, the two measurements

of cognitive fit the Navy uses, computed using ASVAB test scores, employee turnover

outcomes, and control variables: gender and length of service. The researcher used

logistic regression to examine the relationship between cognitive fit and U.S. Navy

enlisted sailor turnover decisions. The goal of this research was to determine whether

employee turnover decreases when cognitive fit increases.

As a potential strategy for the U.S. Navy to reduce personnel costs and maintain a

high-quality workforce, the researcher designed the study to address the following

research question: To what extent do cognitive fit, gender, and length of service predict

employee turnover amongst U.S. Navy enlisted sailors? The hypothesis for this research

question, in null and alternative form, is as follows:

H10. Cognitive fit, gender, and length of service do not predict employee

turnover amongst U.S. Navy enlisted sailors.

H1a. Cognitive fit, gender, and length of service significantly predict employee

turnover amongst U.S. Navy enlisted sailors.

The focus of this section is on the quantitative research method, and it begins with

a description of the design the researcher used to investigate the probability of employee

turnover based on cognitive fit, while controlling for gender and length of service. The

explanation of the chosen research method also includes details about the population and

sample, and particulars about the secondary data the researcher used from the U.S. Navy,

including source, processing, and analysis. This section concludes with a description of

assumptions, limitations, and ethical standards.

Research Methods and Design

The researcher chose a quantitative design study, using multinomial logistic

regression to explore whether cognitive fit predicts employee turnover. The purpose of

this study was to determine if cognitive fit, gender, and length of service (as the predictor

variables) have a relationship to employee turnover (as the criterion variable) in a way

that is measurable and significant.

The study sample included all active U.S. Navy enlisted sailors, paygrades E1

thru E6, with up to 14 years of service who made a retention decision in 2014. Archival

data for this research came from the U.S. Navy, and the data included measurements of

cognitive fit, similar to the approach used in prior research to compute cognitive fit by

comparing ASVAB test scores to the average level of ability required by occupation

computed using data from the Occupational Information Network website (Maltarich et

al., 2010). In this research, the researcher calculated cognitive fit using the sailor’s

cognitive ability, measured by ASVAB test results, compared to two factors: training

school success and a comparison to the AFQT scores of the rating population. Training

school success (S-score) is a function of ASVAB scores for successful training

completion with no set-backs, and has a value that ranges from -100 to 0, where -100 is

not qualified, and 0 is perfectly qualified. The AFQT of the rating population (Q-score)

has a value that ranges from 0 to 100, where 0 is perfectly qualified and 100 is

significantly over-qualified. The researcher added Q-score and S-score together to

provide a numerical value for cognitive fit that ranges from -100 to 100, with optimum fit

at 0.

The data the Navy provided also included employee turnover outcomes, including

decisions to reenlist or separate from Naval service voluntarily or involuntarily. The

push-pull model establishes a basis for operationalizing employee turnover using

voluntary and involuntary separation in a way that can link to cognitive fit. Functional

turnover is the removal of the lowest performers and is beneficial to an organization

(Becker & Cropanzano, 2011). The U.S. Navy initiates functional turnover actions by

involuntarily separating sailors who are lower performers than their peers, or who are not

eligible for reenlistment. On the other hand, individuals across the performance spectrum

may self-initiate voluntary turnover. It may be in the organization’s best interest for them

to leave if they are poor performers, but when top performers voluntarily leave, it can

negatively affect organizational performance (Becker & Cropanzano, 2011). Identifying a

predictive relationship between cognitive fit and employee turnover, including the three

types of retention outcomes (reenlistment, voluntary separation, or involuntary

separation), may signify an opportunity to improve retention by improving cognitive fit.

First, the researcher conducted a descriptive analysis of the data set (Field, 2009).

The researcher inspected a histogram of cognitive fit for data concerns, divided the

dataset into three groups based on turnover outcome categories (involuntary separation,

voluntary separation, and reenlistment), and compared them based on the independent

variables cognitive fit, gender, and length of service.

Logistic regression was the method the researcher used to answer the

research question, “to what extent does cognitive fit predict employee turnover

amongst U.S. Navy enlisted sailors while controlling for gender and length of

service?” since it calls for analysis about a predictive relationship with a

categorical outcome (employee turnover; Field, 2009; C. Peng et al., 2002). Based

on previous research (Hoglin & Barton, 2013), The researcher included gender as

a categorical factor, and included length of service (measured in months from

initial active duty service data to turnover outcome approval month) as a

covariate. Additionally, the researcher added interaction terms between cognitive

fit, gender, and length of service to examine the combined effect of these

variables (Field, 2009). To address the assumption of linearity and account for the

expected curvilinear relationship between cognitive fit and retention, the

researcher tested both cognitive fit and the square of cognitive (Field, 2009).

Based on the size of the dataset, which included 56,847 cases, to validate the

value of the statistical tests, the researcher also tested 1% and 10% subsets of the

data (Ertas, 2015). Additionally, based on the uneven distribution of outcomes

(64.5% reenlistment, 30.6% voluntarily separated, and 4.9% involuntarily

separated), a randomly selected stratified subset with 200 cases for each outcome

was also tested to validate the results. The researcher performed the same analysis

on each of these four datasets using a multinomial model to distinguish the type of

separation using polytomous employee turnover outcomes: involuntary

separation, voluntary separation, or reenlistment.

Population

The study population was the active component of the U.S. Navy for 2014, which

was approximately 327,000 personnel, and it included individuals the Navy recruited

from across the United States and who serve worldwide. This population is both useful

and appropriate because it represents a cross-section of the workforce and it includes

sailors from all demographics and ratings from initial entry through mid-career.

Sample

The study sample was all active U.S. Navy enlisted sailors, paygrades E1 thru E6,

with up to 14 years of service who made a retention decision in 2014. In other words, the

sample included all sailors who separated from naval service or reenlisted to continue

their naval service in 2014. The sample included sailors from initial entry because the

military separates up to 17.8% of new recruits during their initial training (Gibson,

Hackenbracht, & Tremble, 2014), through 14 years of service, which is a limitation based

on the Navy’s Career Waypoints system. The observations for this study come from

secondary data, which came from the U.S. Navy’s Career Waypoint system. Permission

from the U.S. Navy to use this data is in Appendix A. The scope of the Navy’s

reenlistment policy and processes which only requires this subset of sailors (E1-E6 with

up to 14 years of service) to utilize the Career Waypoint system limited the selection

criteria for the research sample. The Navy does not keep the same kind of data on sailors

in paygrades E-7 through E-9, or on officers, in the Career Waypoints system, which is

why the researcher did not include them in the study sample.

Career Waypoints is a Navy decision-support information technology

system that sailors in paygrades E1 thru E6 with up to 14 years of service use to

notify the Navy of their intention to separate from naval service, or request

permission to reenlist and continue to serve. The Navy originally collected some

of the data available in the system from individual applications for enlisted

service in the Navy, and they include cognitive testing results, which the Navy

uses for eligibility and occupational placement. Since only this subset of

employees uses Career Waypoints, the sample had the same constraints. The data

included sailor demographics, ASVAB scores the Navy originally collected

during the application process for naval service and used for eligibility and

occupational placement, and subsequent requests and outcomes for reenlistment.

The U.S. Navy deidentified the data prior to providing it to the researcher to

protect the identity of the test subjects.

There were 56,847 total U.S. Navy enlisted sailor retention decisions in

2014; 36,650 reenlisted, 17,509 voluntarily separated, and 2,653 separated

involuntarily. The researcher discarded 35 cases that were missing outcome data.

Of the total, 11,272 (20.6%) were female and 45,120 (79.4%) were male. Most of

the sailors in the sample were in paygrades E4 or E5 (Table 3). All U.S. Navy

ratings were in the data set (Table 4).

Table 3

Paygrade Composition

E2 3 .0%

E3 8,742 15.4%

E4 21,780 38.3%

E5 20,205 35.5%

E6 6,117 10.8%

Total 56,847 100.0%

Table 4

Rating Composition

HM 6,262 11.0% ABF 568 1.0% MN 203 .4%

MA 2,635 4.6% SO 562 1.0% CTM 202 .4%

IT 2,472 4.3% PS 535 1.0% MMS(SS-W) 198 .3%

AT 1,958 3.4% FC(AEGIS) 531 .9% CSS 195 .3%

LS 1,952 3.4% AS 528 .9% UT 191 .3%

ET(OTH) 1,717 3.0% CTI 525 .9% MM(NUC-TR) 186 .3%

AM 1,603 2.8% GSM 510 .9% SB 185 .3%

OS 1,551 2.7% QM 509 .9% AWO 182 .3%

MM(OTH) 1,524 2.7% STS 501 .9% RP 178 .3%

AO 1,458 2.6% EM(SS-N) 481 .9% EOD 161 .3%

CS 1,384 2.4% IC 479 .8% GSE 160 .3%

BM 1,348 2.4% HT 475 .8% MR 142 .2%

ABH 1,242 2.2% BU 475 .8% AWV 139 .2%

AD 1,177 2.1% MMS(SS-AX) 394 .8% SW 135 .2%

AE 1,106 1.9% AME 383 .7% MU 134 .2%

YN 997 1.8% CM 361 .7% EM(NUC-TR) 133 .2%

GM 974 1.7% ET(SS-NV) 360 .6% AWF 120 .2%

FC 878 1.5% PR 358 .6% AWR 119 .2%

MM(SW-N) 821 1.4% EM(SW-N) 349 .6% YNS 110 .2%

MM(SS-N) 806 1.4% EO 337 .6% ET(NUC-TR) 99 .2%

CTR 795 1.4% ET(SS-RF) 319 .6% LSS 90 .2%

EM(OTH) 756 1.3% FT 297 .6% ITS 63 .1%

CTT 712 1.3% ET(SS-N) 290 .5% LN 62 .1%

EN 684 1.2% ND 271 .5% EA 43 .1%

IS 671 1.2% CTN 270 .5% NC(C) 25 .0%

DC 663 1.2% ET(SW-N) 269 .5% SN 15 .0%

AZ 644 1.1% MC 265 .5% NC(CRF) 11 .0%

AC 636 1.1% AWS 237 .5% AN 9 .0%

SH 619 1.1% AG 231 .4% FN 6 .0%

ABE 595 1.0% CE 229 .4%

STG 584 1.0% MT 228 .4%

When designing a research study, researchers perform a statistical power

calculation to ensure the dataset includes a large enough sample to achieve at least

an 80% chance of detecting an effect if it exists in the population (Field, 2009).

However, there is a lack of consensus on the best statistical power calculation

method for logistic regression; options include the likelihood ratio, the Wald test,

proportion tests, or various approximations for research with multivariates

(Demidenko, 2007). Additionally, statistical power calculations provide the

minimum number of cases required to obtain the desired probability of detecting

an effect. Researchers sometimes use large datasets in similar studies on

employee turnover such as Weaver’s (2015) research on why federal employees

leave, which included 263,475 participants, and Ertas’ (2015) research on

turnover intentions of millennial federal employees, which included 266,000

participants. Although both of these studies used full data sets, Ertas accounted

for the potential of large sample sizes enlarging the value of statistical tests by

validating the model through 1% and 10% subsets of the larger sample.

In this case, the U.S. Navy provided a large dataset relevant for examining

the research question. Although the method the researcher used to draw this

sample was to provide all cases that included a retention result in 2014—which

arguably is not random—it does include nearly 20% of the Navy’s full active duty

population, a broad range of career fields and paygrades for Navy employees, and

it also closely mirrors the gender composition of the Navy. Based on these

observations, the researcher utilized the full data set the Navy provided for

statistical testing, with validation using 1% and 10% randomly selected subsets in

the method utilized by Ertas (2015). Additionally, based on the uneven

distribution of outcomes (64.5% reenlistment, 30.6% voluntarily separated, and

4.9% involuntarily separated), a randomly selected stratified subset with 200

cases for each outcome was also tested to validate the results. The researcher

performed the same analysis on each of these four datasets.

Materials/Instruments

The data for this study included demographics (gender, length of service) and

retention outcomes (reenlisted, voluntarily separated, or involuntarily separated) of

sailors whose enlistment contracts ended in 2014 and cognitive fit, calculated using the

sailor’s cognitive ability measured by his or her ASVAB test results from his or her

initial recruitment compared to training success and the general cognitive ability of other

applicants in the same career field. These data came from the U.S. Navy’s Career

Waypoint system. The U.S. Navy collected these data for personnel management

purposes.

For initial enlistment, the U.S. Navy uses the ASVAB. All U.S. military services

utilize this test battery, which has nine subtests (listed in Table 2) to screen applicants for

military service cognitively, and if qualified, to place new recruits into occupations

(Held, Hezlett, et al., 2014). The Navy administers it either as a paper-and-pencil (P&P)

test, which takes three hours, or as a CAT, which reduces the test time to approximately

one and a half hours (Held, Hezlett, et al., 2014). For basic eligibility, all of the military

services use a composite score of two math and two verbal subtests called the AFQT

(Held, Hezlett, et al., 2014). In addition, the Navy also uses different combinations of

ASVAB subtest results tailored to particular occupations (Held, Hezlett, et al., 2014). The

Navy has 85 enlisted occupational fields, called ratings, which have different training

requirements, training times, and other requirements; they therefore require different

ASVAB subtest combinations to place individuals dependably in occupations for which

they are cognitively suited (Held, Hezlett, et al., 2014).

The Navy is the only service that routinely revalidates ASVAB requirements by

rating (Held, Hezlett, et al., 2014). Events that trigger revalidation include an increase in

academic failures in occupation training courses, major changes to training requirements,

reductions in training time allowed, new or redefined occupational fields, and changes in

the recruiting environment affecting average recruit ASVAB scores (Held, Hezlett, et al.,

2014). ASVAB scores can predict successful completion of initial training requirements

(Held, Hezlett, et al., 2014). Based on multiple ASVAB standards and validation studies,

the predictive validity of the Navy’s occupational ASVAB coefficients averages 0.55,

with a range of 0.25 to 0.85 depending on the rating (Held, Hezlett, et al., 2014). Current

information about ASVAB reliability is in Table 5 and is on the ASVAB website at

http://official-asvab.com/reliability. ASAB reliability ranges from 0.85 to 0.97 depending

on the version of the test: P&P or CAT, and the subtest of interest.

Table 5

Reliability for Armed Forces Qualification Test Composite and Armed Services

Vocational Aptitude Battery Sub-Tests

AFQT AR WK PC MK

P&P CAT P&P CAT P&P CAT P&P CAT P&P CAT

0.94 0.97 0.87 0.92 0.88 0.93 0.75 0.85 0.85 0.93

GS EI AS MC AO

P&P CAT P&P CAT P&P CAT P&P CAT P&P CAT

0.80 0.87 0.79 0.87 0.81 n/a 0.79 0.85 0.84 0.82

Note. The source of the ASVAB reliability data is the ASVAB website at http://official-

asvab.com/reliability.

Operational Definition of Variables

Cognitive fit. Cognitive fit is the match between an individual and a job based on

cognitive ability (Maltarich et al., 2010). Cognitive fit is an independent continuous

variable from archival data that is the result of comparing a sailor’s cognitive test results

with two factors: training school success (S-score) and AFQT scores for other sailors in

the same rating (Q-score; Watson, 2010). The Navy combines these two scores to

approximate Yerkes-Dodson’s law (Watson, 2010). For this research, the researcher

added Q-score and S-score together to provide a numerical value for cognitive fit The

value of cognitive fit is zero if the sailor is a perfect fit for his or her assigned rating, a

positive value (from 0 to 100) if the sailor is overqualified for the rating, and a negative

value (from 0 to -100) if the sailor is underqualified for the rating.

Employee turnover outcome. Employee turnover outcome is a dependent

categorical variable from archival data. There are three possible outcomes for an

employee turnover event: voluntary separation or turnover, involuntary separation or

turnover, and reenlistment to continue service.

Involuntary separation or turnover. For the purpose of this study, this action

includes sailors who the Navy did not permit to reenlist based on performance after

comparing them to their peers, or who were ineligible to reenlist because they no longer

met the enlistment criteria for their ratings. The Navy initiated these turnover actions. For

this research, the researcher included sailors who were ineligible to reenlist and who did

not receive approval for reenlistment. The Navy codes used for this category were: forced

separation (FSP), ineligible separation (ESP), denied final in-rate (DFI), ineligible (IEG),

and voluntary separation (VSP) cases where a sailor was not approved to reenlist in rate

and there were no other options to convert into another rating.

Voluntary separation or turnover. For the purpose of this study, this action

includes sailors who chose not to reenlist. The individual initiates these turnover actions.

For this research, the researcher included sailors who requested to separate and/or join the

Navy Reserve. The Navy codes used for this category were: voluntary separation (VSP),

requested Navy Reserve (RQR), and intends to separate (ITS).

Reenlistment. For the purpose of this study, this action includes sailors who chose

to reenlist. The individual initiated these turnover actions and the Navy approved them.

For this research, the researcher included sailors who received approval to reenlist in their

current rating, or in a new rating. The Navy codes used for this category were approved

in rate (AIR) and approved conversion (ACV). Appendix B describes all of these

variables.

Gender. Gender is a categorical variable from archival data that signifies if an

individual is male or female.

Length of service. Length of service is an interval variable from archival data

that measures the amount of time a sailor served in the Navy, calculated in months from

when an individual initially entered the U.S. Navy to the time when the Navy approved

his or her employee turnover outcome.

Data Collection, Processing, and Analysis

The U.S. Navy has granted permission for use of Career Waypoints data

(Appendix A). The data on the specified research population of U.S. Navy enlisted sailors

in paygrades E1-E6 with up to 14 years of service are resident in the Career Waypoint

System. The dataset includes a Navy-assigned record identifier that is not personally

identifiable, rating, gender, S-score and Q-score for cognitive fit, and turnover outcome

and approval date. Appendix B includes a complete listing of all the variables and a

description of the type of data and how the researcher coded and computed the data.

Although this research used data on U.S. Navy sailors, it did not meet the definition of

research involving human subjects (National Defense, 32 C.F.R. § 219.102(f), 2014;

Department of Defense, 2011b) under Exemption Category 4, as determined by Naval

Sea Systems Command’s Human Research Protection Official (Appendix C).

The first statistical procedure in this research study was the use of simple

descriptive statistics to characterize the population. Gender composition, average length

of service, and paygrade for U.S. Navy sailors who separated prior to 14 years of service,

are of interest, as is average cognitive fit. The next statistical was multinomial logistic

regression analysis using cognitive fit as the predictor variable and turnover outcome

(voluntary separation, involuntary separation, or reenlistment) as the categorical variable.

Assumptions

Assumptions are a key underpinning of research, representing the researcher’s

perspective and intertwined with his or her logic and the suppositions and hypotheses

presented in the study design (Farquhar, 2012). In this study, the researcher’s ontological

stance, or view of the world, is nomothetic, meaning that the phenomenon cognitive fit

exists independently of social perceptions (Farquhar, 2012). In alignment with this

perspective, the researcher’s epistemology is positivist—leading to a research design

utilizing real, measurable phenomena (Farquhar, 2012). Based on the researcher’s

underlying ontological, epistemological, and axiomatic standpoint, there are two key

assumptions in this research design. To provide a basis for observation, one assumption is

that sailor’s ASVAB test scores are an accurate reflection of their cognitive ability, taken

personally, independently, and to the best of their ability. Although it is clear that some

people attempt to gain entry into the U.S. military by cheating on the ASVAB test, this

assumption is reasonable given the relatively small occurrence of this behavior and the

large sample size the researcher used in this research.

The second assumption is an epistemological consideration, because it could lead

to false positive or false negative results. As a way to discriminate between voluntary and

involuntary separations, the research plan includes the assumption that all sailors who

requested reenlistment wanted to continue their naval service. This assumption is more

problematic, because is it likely that some sailors have not made a final decision about

staying in the Navy when they reach the timeline for requesting permission to reenlist.

However, the fact that these sailors wanted to maintain their option to remain in the Navy

is an indicator they were somewhat interested in remaining in the service.

Limitations

Items outside of the researcher’s control often limit the applicability and

usefulness of research. There are three limitations in this proposed research design:

generalizability to the general population, use of cross-sectional rather than longitudinal

design, and the omission of potentially significant moderating or extraneous variables

from the data collection and analysis—including alternative job availability, leadership,

and command climate, which all affect employee turnover. Although U.S. Navy sailors

join the service from communities across the United States with demographics that

closely reflect the U.S. population, some of the employment processes are unique to the

U.S. Navy, potentially limiting generalizability. Additionally, this study utilized an

instrument designed and used exclusively by the U.S. military. The lack of a tool to

measure cognitive fit in the civilian workforce reduces the potential generalizability of

the findings of this research study.

Second, this study is non-experimental and cross-sectional. A cross-sectional

design is necessary for the type of event (turnover outcomes) under consideration, since

two of the three (voluntary and involuntary separation) result in termination of

employment. However, observations of the third (reenlistment) result in continued

service that will have a termination outcome sometime in the future. Since this study is

non-experimental and cross-sectional, it will not be able to address causality.

The third limitation is the omission of potentially significant moderating or

extraneous variables beyond the control variables from data collection and analysis. The

potential exists that variables the researcher has not included in the proposed research

design might have a greater impact on turnover outcomes than cognitive fit. One variable

the researcher has not included in the plan that has produced an effect on employee

turnover is alternate job availability, which researchers often operationalize using the

Bureau of Labor’s unemployment rate (Pinelis & Huff, 2014). Based on the research

design which only includes employment outcomes for one year, variability in the

unemployment rate is not significant enough to include in this study. Other variables with

demonstrated relevance to retention are job satisfaction and organizational commitment

(Lytell & Drasgow, 2009), leadership, and command climate—however these variables

are situational rather than individual antecedents and they are not relevant to developing

pre-hire selection criteria.

Delimitations

Delimitations denote the scope of a research project. In the case of this research,

the population defines the scope of the project. The researcher delimited the population to

U.S. Navy enlisted sailors in paygrades E-1 through E-6 with up to 14 years of service.

This population omits Naval officers and enlisted sailors in paygrades E-7 through E-9

with greater than 14 years of service. These delimitations are necessary based on the data

available in the Navy’s Career Waypoint database. Since the Navy selects officers for

entry and places them in jobs using different criteria than enlisted sailors, the impact of

excluding this group from the population should not be significant. Additionally, when an

enlisted sailor receives promotion to E-7 and above, he or she becomes a careerist,

because he or she will rarely separate before becoming retirement eligible. So, the subset

of sailors that reenlisted will be smaller than it would be if it included them, but there

should still be enough data to determine discriminant factors between groups. Future

research on cognitive fit and its relevance to performance outcomes could consider

whether cognitive fit predicts promotion to E-7 through E-9.

Ethical Assurances

In the conduct of responsible research, researchers have an obligation to

themselves, study participants, their colleagues, and society to make choices based on

ethical values. Two of the primary risks of research to study participants are a violation of

privacy and a breach of confidentiality. In this research design, the use of a different,

discrete record identifier that does not include personally identifiable information

mitigates this risk. In this way, the data the researcher collected from the Navy’s Career

Waypoint database does not include information that might identify a particular

individual. The researcher protected the sailor data using password access, and will

dispose of it after seven years by deleting the database from the hard drive of the

computer in use for the research. Additionally, to mitigate risk to participants, the

researcher submitted this research proposal to a Navy Human Research Protections

official (Appendix C) and the Northcentral University Institutional Review Board to

ensure the safety of participants and the protection of their rights.

Scientists must communicate research procedures clearly and to report results

accurately to prevent the waste of time and resources, and to sustain the construct of

scientific research which builds upon previous results. This research design is

substantially different than previously conducted research, and the researcher has built it

to be straightforward and repeatable. The responsible conduct of research is an issue of

public concern because scientific results may influence decisions that affect society. In

this study, the U.S. Navy might use the results to change recruiting and job placement

procedures. Careful use of statistical methods, accurate reporting, and conclusions and

recommendations drawn from careful analysis of the results are necessary.

Summary

The potential utility of cognitive fit as a predictor of both performance and

retention for hiring decisions is a new twist on an often-used attribute. In the past,

researchers have mainly used cognitive ability as a predictor of performance in employee

selection and hiring decisions (Maltarich et al., 2010). Previous research on the

relationship between cognitive ability and employee turnover has used general cognitive

ability rather than a more focused measure of the cognitive match between an

individual’s skills and the demands of the job (Boudreau et al., 2001; Knapik et al.,

2004). Comparing voluntary turnover to the level of cognitive match between an

employee and his or her job is a recent development (Maltarich et al., 2010). Although

the results demonstrate that the cognitive demands of a position make a difference in

employee turnover decisions, the relationship is complex and in need of further study

(Maltarich et al., 2010).

The researcher designed this quantitative research study to determine whether

there is a significant and measurable relationship between cognitive fit and employee

turnover in the U.S. Navy. The Navy offers a unique opportunity to study this

phenomenon because of its hiring process, which uses cognitive ability testing and

validated cognitive requirements for each Navy career field to place sailors into jobs.

This research adds to the body of knowledge on human resources management practices

by determining if cognitive fit is a predictor of future retention that the Navy can use to

select and hire top talent.

Chapter 4: Findings

High-performing employees are important to organizational success (Crook et al.,

2011) and the competition for talent is on the rise (Maltarich et al., 2010). The loss of top

talent to employee turnover represents a significant loss of organizational effort and

financial resources (Godlewski & Kline, 2012). The purpose of this non-experimental,

quantitative study was to examine the relationship between cognitive fit and employee

turnover in the U.S. Navy, with the goal to determine if employee turnover decreases

when cognitive fit increases. If cognitive fit predicts employee turnover, the Navy may

alter hiring and placement processes to increase the likelihood of retaining talented

employees in the future.

The researcher chose a quantitative design study, using multinomial logistic

regression to explore whether cognitive fit predicts employee turnover. The researcher

included gender and length of service as covariants in the study based on previous

research by Hoglin and Barton (2013) on employee turnover, and the relevance of these

variables based on the kaleidoscope career model (Mainiero & Sullivan, 2005; Sullivan

& Mainiero, 2007). This chapter begins with an overview of the data the U.S. Navy

provided. A statement of the research question follows the description of the data set.

Then, the chapter presents statistical analysis of the data to answer the research question

with an explanation of each of the steps the researcher used to conduct the statistical

analysis.

Results

The primary goal for using the data provided by the U.S. Navy in this study was

to determine if a predictive relationship exists between cognitive fit, gender, length of

service, and employee turnover. There were 56,812 cases in the sample; 36,650 reenlisted

(64.5%), 17,391 voluntarily separated (30.6%), and 2,771 (4.9%) separated involuntarily.

The researcher discarded 35 (.1%) cases that were missing employee turnover outcome

data. Of the sample, 11,725 (20.6%) were female, and 45,087 (79.4%) were male.

In the study, the researcher used a multinomial model and polytomous employee

turnover outcomes: involuntary separation, voluntary separation, or reenlistment. The

researcher tested both cognitive fit and the square of cognitive fit to account for the

expected curvilinear relationship between cognitive fit and employee turnover (Field,

2009). Finally, to validate the findings, the researcher conducted the statistical tests on

the full dataset, a 1% subset, and a 10% subset of the data (Ertas, 2015). The researcher

also had concerns about the difference in proportions between the three outcomes, so she

also tested a stratified subset of the data by randomly selecting 200 cases from each

outcome.

Descriptive statistics about cognitive fit for the full sample reveal a mean of -

28.13, with a median of -34.48 and standard deviation of 38.35. The mode for this data is

0, with 4,639 cases. The mean for length of service is 62.02 months, with a median of 54

months, and standard deviation of 31.79. Sailor gender in this sample is 79.8% male, and

20.2% female. A comparison of the independent covariants cognitive fit, gender, and

length of service by turnover outcome group for the full dataset is in Table 6.

Table 6

Descriptive Data for Predictor Variables by Turnover Outcome

Voluntary Involuntary Reenlistment Total

Mean Cognitive Fit

(-100 to 100)

-25.9984 -27.0097 -29.2362 -28.1305

Gender

(percentage M/F)

77.9/22.1 80.9/19.1 80.6/19.4 79.8/20.2

Length of Service

(months)

61.4 63.22 62.22 62.02

The primary goal of this study was to determine if cognitive fit predicted

employee turnover, while controlling for gender and length of service. The researcher

utilized U.S. Navy retention data for enlisted sailors in paygrades E1-E6 with up to 14

years of service to explore the relationship between these variables. The mean values for

cognitive fit overall and in all three turnover outcome groups are below the optimum

value of zero, indicating that sailors in this dataset have less than optimum cognitive fit

for their assigned career fields. Of note, all of these scores are very close together, even

though the scale ranges from -100 to 100. Mean cognitive fit for sailors who separated is

slightly better than for those who reenlisted, which is opposite of what one might expect.

The percentages of females compared to males is slightly higher than overall 2014

enlisted gender demographics in the Navy, which was 18% female and 82% male

(Department of Defense, 2014). This was not surprising, and the gradual increase of

female accessions in the Navy over time explains it. In the sample, more males than

females reenlisted or separated involuntarily, while more females voluntarily separated.

The average number of months of service for sailors making turnover decisions is 57.77

for females and 64.19 for males, and equal approximately five years in the Navy. This

makes sense since the initial obligation for new sailors is four to six years, and as the base

of the pyramid, first-term enlisted sailors are the largest subset in the Navy, but the

difference between males and females, especially since there are so many fewer females

in the Navy, is noteworthy.

A histogram of the variable cognitive fit reveals some anomalies (Figure 3). The

data includes 2,234 cases where sailors were not qualified, and therefore have a cognitive

fit of -100. Sailors who were not qualified for their ratings did not meet the minimum

requirements and received placement waivers. Since these cases were outliers, the

researcher removed them from the dataset. There are also 4,639 cases of sailors who were

a perfect fit (cognitive fit equals zero). This anomaly may indicate that some sailors

chose their “best fit” rating when officers explained it to them and offered it as a choice.

Of note, the percentage of females perfectly qualified is significantly higher than the

proportion one might expect based on the dataset: 1,938 (41.7%) were female and 2,705

(58.3%) were male, compared to the percentages of males (79.8%) and females (20.2%)

in the full data set. Since normal distribution is not an assumption for logistic regression

and it is such a large data set, reducing the impact of individual data points, the researcher

retained these cases.

The researcher computed z scores for all three main variables, cognitive fit,

gender, and length of service to test for additional outliers in the dataset, and the three

data subsets. For both cognitive fit and length of service, there were some cases that were

more than three standard deviations from the mean, so the researcher did not include

them in her computations in order to reduce Type I and Type II error rates and possible

distortion of the results. After these alterations, the full dataset included 54,333 total

cases, the 10% data subset included 5,462 cases, the 1% data subset included 555 cases,

and the stratified data subset included 574 cases.

Figure 3. Cognitive fit by gender.

The researcher used logistic regression to investigate the research question, which

asked, “to what extent does cognitive fit predict employee turnover amongst U.S. Navy

enlisted sailors, while controlling for gender and length of service?” The null hypothesis

was that cognitive fit, gender, and length of service do not predict employee turnover

amongst U.S. Navy enlisted sailors. The alternative hypothesis was that cognitive fit,

gender, and length of service significantly predict employee turnover amongst U.S. Navy

enlisted sailors.

Logistic regression is similar to multiple regression, but researchers use it when

the outcome variable is categorical and the predictor variables are continuous or

categorical (Field, 2009). In this case, the outcome variable was employee turnover.

There were three possible outcomes: voluntary separation, involuntary separation, and

reenlistment. The predictor variables were gender, which is also categorical, and length

of service, which is continuous. In the regression models, the researcher added interaction

terms between cognitive fit, gender, and length of service to examine the combined effect

of these variables.

There are three assumptions for logistic regression: multicollinearity,

independence of errors, and linearity. The data met the assumption of multicollinearity

since the predictor variables are not similar, and they met the assumption of

independence of errors because there is no overlap of cases in the data due to any

multiple inclusion of individuals. However, for the assumption of linearity there was a

potentially an issue, since the expected relationship between cognitive fit and employee

turnover was curvilinear. To overcome this concern, the researcher conducted the logistic

regression testing the predictor cognitive fit in its standard form, and the predictor

cognitive fit squared to account for a possible quadratic relationship.

In order to account for the size of the data set, and the disproportionate number of

cases by retention outcome (64.4% reenlisted, 30.7% voluntarily separated, and 4.9%

involuntarily separated), the researcher did additional tests to validate the model. To

determine if the large sample size affected the value of the statistical tests (Ertas, 2015),

the researcher drew and modeled random subsets of 10% and 1%. As a final step, the

researcher drew a stratified random subset with 200 cases from each of the three

categorical retention outcome groups, for a total of 600 cases that were proportional

based on turnover outcomes.

Multinomial logistic regression tests the relationships between variables, and it is

necessary with a categorical dependent variable with more than two categories. The data

provided by the U.S. Navy included information about the type of separation outcome,

whether it was voluntary or involuntary. The researcher used a multinomial model to

distinguish the type of separation using polytomous employee turnover outcomes:

involuntary separation, voluntary separation, or reenlistment. The reference category for

this multinomial logistic regression model was sailors who reenlisted. Logistic regression

calculates effect through odds ratios. An odds ratio higher than one means the

independent variable for a sailor voluntarily or involuntary separating was higher than for

a sailor reenlisting. An odds ratio of less than one means the chance of a sailor

voluntarily or involuntarily separating was lower than for a sailor reenlisting. The

multinomial logistic regression results are in Tables 7-10, with the Table 7 depicting

analysis of the full dataset, Table 8 using the 10% data subset, Table 9 using the 1% data

subset, and Table 10 using the stratified subset.

Table 7

Multinomial Logistic Regression Results: Full Dataset

Full Dataset

95% CI for Odds Ratio

Turnover

Outcome

Variable B(SE) Lower Odds

Ratio

Upper

Voluntary Intercept -.800(.032)***

Gender(F) .130(.065)* 1.003 1.138 1.292 Service .000(.000) .999 1.000 1.001

Fit .001(.001) .999 1.001 1.002

Fit2 .000(.000)*** 1.000 1.000 1.000 Fit*Gender(F) .002(.001) .999 1.002 1.004

Fit2*Gender(F) .000(.000) 1.000 1.000 1.000

Fit*Service .000(.000)* 1.000 1.000 1.000 Fit2*Service .000(.000)** 1.000 1.000 1.000 Gender(F)*Service .001(.001) .999 1.001 1.003

Fit*Gender(F)*Service .000(.000) 1.000 1.000 1.000

Fit2*Gender(F)*Service .000(.000) 1.000 1.000 1.000

Involuntary Intercept -

2.653(.067)***

Gender(F) -.004(.142) .754 .996 1.316

Service .002(.001) 1.000 1.002 1.004

Fit .003(.001) 1.000 1.003 1.005

Fit2 .000(.000) 1.000 1.000 1.000

Fit*Gender(F) .002(.003) .995 1.001 1.006

Fit2*Gender(F) .000(.000) .996 1.002 1.009

Fit*Service .000(.000) 1.000 1.000 1.000

Fit2*Service .000(.000)* 1.000 1.000 1.000 Gender(F)*Service .000(.002) .996 1.000 1.004

Fit*Gender(F)*Service .000(.000) 1.000 1.000 1.000

Fit2*Gender(F)*Service .000(.000)* 1.000 1.000 1.000

Overall Model Cox and Snell R2 = .003

Nagelkerke R2 = .004

McFadden R2 = .002

Goodness of Fit Deviance: Chi-square = 64929.125,

df = 69412, Sig. = 1.000

Pearson: Chi-square = 75816.219,

df = 69412, Sig. = .000***

Note: Correlation significance * p < .05, ** p < .01, *** p < .001

Table 8

Multinomial Logistic Regression Results: 10% Dataset

10% Dataset

95% CI for Odds Ratio

Turnover

Outcome

Variable B(SE) Lower Odds

Ratio

Upper

Voluntary Intercept -.709(.098)***

Gender(F) .297(.206)** .898 1.346 2.017 Service -.001(.001) .996 .999 1.002

Fit .003(.002) .999 1.003 1.007

Fit2 .000(.000) 1.000 1.000 1.000

Fit*Gender(F) .003(.005) .994 1.003 1.013

Fit2*Gender(F) .000(.000) 1.000 1.000 1.000

Fit*Service .000(.000) 1.000 1.000 1.000

Fit2*Service .000(.000) 1.000 1.000 1.000

Gender(F)*Service -.001(.003) .993 .999 1.005

Fit*Gender(F)*Service .000(.000) 1.000 1.000 1.000

Fit2*Gender(F)*Service .000(.000) 1.000 1.000 1.000

Involuntary Intercept -

2.587(.215)***

Gender(F) -.032(.471) .385 .968 2,435

Service .000(.003) .994 1.000 1.006

Fit .002(.005) .992 1.002 1.011

Fit2 .000(.000) 1.000 1.000 1.000

Fit*Gender(F) -.001(.011) .977 .999 1.021

Fit2*Gender(F) .000(.000) .999 1.000 1.000

Fit*Service .000(.000) 1.000 1.000 1.000

Fit2*Service .000(.000) 1.000 1.000 1.000

Gender(F)*Service .004(.007) .991 1.004 1.018

Fit*Gender(F)*Service .000(.000) 1.000 1.000 1.000

Fit2*Gender(F)*Service .000(.000) 1.000 1.000 1.000

Overall Model Cox and Snell R2 = .005

Nagelkerke R2 = .006

McFadden R2 = .003

Goodness of Fit Deviance: Chi-square = 7971.567,

df = 9540, Sig. = 1.000

Pearson: Chi-square = 9887.257,

df = 9540, Sig. = .006**

Note: Correlation significance * p < .05, ** p < .01, *** p < .001

Table 9

Multinomial Logistic Regression Results: 1% Dataset

1% Dataset

95% CI for Odds Ratio

Turnover

Outcome

Variable B(SE) Lower Odds

Ratio

Upper

Voluntary Intercept 1.886(.707)**

Gender(F) -1.417(2.788) .001 .242 57.215

Service .009(.011) .988 1.010 1.032

Fit -.014(.016) .956 .986 1.017

Fit2 .000(.000) .999 1.000 1.001

Fit*Gender(F) -.108(.062) .795 .898 1.015

Fit2*Gender(F) -.001(.001) .996 .999 1.002

Fit*Service .000(.000) 1.000 1.000 1.001

Fit2*Service .000(.000) 1.000 1.000 1.000

Gender(F)*Service .096(.069) .963 1.101 1.260

Fit*Gender(F)*Service .004(.002) .999 1.004 1.008

Fit2*Gender(F)*Service .000(.000) 1.000 1.000 1.000

Involuntary Intercept 1.318(.736)

Gender(F) -1.500(2.823) .001 .223 56.472

Service .004(.012) .981 1.004 1.027

Fit -.010(.016) .959 .990 1.022

Fit2 .000(.000) .999 1.000 1.001

Fit*Gender(F) -.115(.063) .788 .891 1.008

Fit2*Gender(F) -.001(.002) .996 .999 1.002

Fit*Service .000(.000) .999 1.000 1.000

Fit2*Service .000(.000) 1.000 1.000 1.000

Gender(F)*Service .101(.069) .966 1.106 1.266

Fit*Gender(F)*Service .004(.002) .999 1.004 1.008

Fit2*Gender(F)*Service .000(.000) 1.000 1.000 1.000

Overall Model Cox and Snell R2 = .047

Nagelkerke R2 = .060

McFadden R2 = .031

Goodness of Fit Deviance: Chi-square = 814.486,

df = 1058, Sig. = 1.000

Pearson: Chi-square = 1084.828,

df = 1058, Sig. = .277

Note: Correlation significance * p < .05, ** p < .01, *** p < .001

Table 10

Multinomial Logistic Regression Results: Stratified Dataset

Stratified Dataset

95% CI for Odds Ratio

Turnover

Outcome

Variable B(SE) Lower Odds

Ratio

Upper

Voluntary Intercept .448(.337)

Gender(F) -.369(.860) .230 .692 2.082

Service -.007(.005) .987 .993 1.000

Fit -.007(.008) .983 .993 1.004

Fit2 .000(.000) 1.000 1.000 1.000

Fit*Gender(F) .045(.021)* 1.018 1.046 1.074 Fit2*Gender(F) .000(.000) 1.000 1.000 1.001

Fit*Service .000(.000) 1.000 1.000 1.000

Fit2*Service .000(.000) 1.000 1.000 1.000

Gender(F)*Service .004(.014) .986 1.004 1.021

Fit*Gender(F)*Service -.001(.000) .999 .999 1.000

Fit2*Gender(F)*Service .000(.000) 1.000 1.000 1.000

Involuntary Intercept .199(.336)

Gender(F) 1.269(.914) 1.102 3.558 11.483

Service -.004(.005) .990 .996 1.003

Fit .000(.008) .990 1.000 1.011

Fit2 .000(.000) 1.000 1.000 1.000

Fit*Gender(F) .013(.018) .990 1.013 1.037

Fit2*Gender(F) -.001(.000) .999 .999 1.000

Fit*Service .000(.000) 1.000 1.000 1.000

Fit2*Service .000(.000) 1.000 1.000 1.000

Gender(F)*Service -.022(.015) .959 .978 .998

Fit*Gender(F)*Service .000(.000) 1.000 1.000 1.000

Fit2*Gender(F)*Service .000(.000) 1.000 1.000 1.000

Overall Model Cox and Snell R2 = .040

Nagelkerke R2 = .045

McFadden R2 = .018

Goodness of Fit Deviance: Chi-square = 1205.441,

df = 1082, Sig. = .005**

Pearson: Chi-square = 1109.505,

df = 1082, Sig. = .274

Note: Correlation significance * p < .05, ** p < .01, *** p < .001

The results include Cox and Snell’s, Nagelkerke’s, and McFadden’s overall

model assessments measuring R2, along with deviance and Pearson’s goodness-of-fit

measurements. One concern about these model assessments is the difference in

significance between deviance and Pearson’s measurements. The researcher tested for the

possibility of overdispersion in the full set of data with standard cognitive fit (Pearson =

1.09, and deviance = 0.94; Field, 2009). Since neither of these values was particularly

high, and both were close to 1, the researcher did not find cause for concern that the data

were overdispersed.

In the full dataset, gender, cognitive fit squared, the interaction of cognitive fit

with length of service, and the interaction of cognitive fit squared with length of service

were all statistically significant for voluntary turnover. In the same dataset, the interaction

of cognitive fit squared with length of service, and the three-way interaction of cognitive

fit squared with both gender and length of service were statistically significant for

involuntary turnover. Of note, cognitive fit is not statistically significant, but cognitive fit

squared is statistically significant, indicating that the relationship between voluntary

turnover and cognitive fit is curvilinear as expected. However, there is no significant

relationship between cognitive fit and involuntary turnover. The measures of R2 for the

full dataset are similar, and represent very small effects, meaning the model is weak and

only explains .002 to .004% of the turnover outcomes. For voluntary turnover, the odds

ratio for gender is the only statistically significant result (p < .05) that indicates a

measurable impact, specifically that females are 1.138 times more likely to separate

voluntarily than to reenlist. All of the odds ratios for the other statistically significant

results are 1.000, indicating that there is a positive effect, but the size of the effect is less

than .001.

The researcher used 10% and 1% random subsets of the data to test the value of

these statistically significant findings. R2 increased, but only to .06%. In the 10% subset,

gender remained statistically significant for voluntary turnover, with an odds ratio of

1.346, p < .01, but it was not significant for involuntary turnover or in the 1% subset.

None of the other factors were significant. The lack of findings in these subsets refutes

the value of the statistically significant relationships noted in the full dataset.

In the final test using stratified data with equal numbers of cases for each turnover

outcome, R2 was only .04%. In addition, the only statistically significant result using this

subset was the interaction of cognitive fit and gender for voluntary turnover, and it was a

linear instead of a curvilinear relationship. The odds ratio for this interaction was 1.046, p

< .05, meaning the change in odds of voluntary turnover for females with lower cognitive

fit was 1.046.

Evaluation of Findings

There is limited prior research on general cognitive ability as a predictor of

employee turnover (Maltarich et al., 2010; Ryan & Ployhart, 2014; Zaccaro et al., 2015),

and it has had mixed results (Boudreau et al., 2001). In 2010, Maltarich et al. conducted a

study that, for the first time, examined the relationship between voluntary turnover and

cognitive fit instead of general cognitive ability. Maltarich et al. used ASVAB results

from the National Longitudinal Survey of Youth and determined the cognitive demands

of particular jobs by collecting average levels of ability from the Occupational

Information Network webpage. Maltarich et al.’s research design used the product of

coefficients method to relate cognitive ability to job satisfaction and job satisfaction to

predict voluntary turnover, and the results identified a significant relationship between

job satisfaction, cognitive ability, and voluntary turnover for jobs with high cognitive

demands, but no relationship for jobs with low or medium cognitive demands.

In the current research model, cognitive fit explained less than 1% of employee

turnover outcomes. When examining the mean values of cognitive fit by turnover

outcome, it was apparent that the three groups were very similar, and mean cognitive fit

for sailors who separated was slightly higher than for those who reenlisted. Also worth

noting, the average cognitive fit for the dataset was -28.1305, significantly below the

optimum value of zero, indicating that most sailors are underqualified for their jobs and

potentially affecting the results. Although statistically significant relationships emerged

when the researcher tested the model on the full dataset, the effects were very small (less

than 0.001), which was even smaller than past research on general cognitive ability,

which measured an effect size of 0.02 (Allen et al., 2010). These results indicate that

cognitive fit is not an important predictor of future employee turnover, further validated

through testing of the 10%, 1%, and stratified subsets.

The researcher’s model used an objective measurement of fit, and the results are

similar to prior research on objective measures of general cognitive ability and cognitive

fit (Maltarich et al., 2010; Ryan & Ployhart, 2014; Zaccaro et al., 2015). However, unlike

Maltarich et al.’s (2010) design, the researcher used the Navy’s RIDE algorithm to

determine cognitive fit, and did not group jobs into low, medium, and high categories

since the algorithm includes precise job demand measurements for each Navy rating.

Furthermore, the algorithm included voluntary and involuntary separation as separate

turnover outcomes, precipitating the use of multinomial logistic regression for statistical

analysis.

In addition, the results differ from previous Navy research on the RIDE algorithm,

which indicated that sailors with high cognitive fit were less likely to separate

(Department of the Navy, 2012). The outcome of this research may be different based on

methodology; the Navy’s research defined high cognitive fit as placement in any of the

top 25 best-fit ratings (Department of the Navy, 2012) whereas this research assigned

each person-job match a numerical value for cognitive fit. The results also differ from

previous research that used a broader definition and a more subjective measurement of

fit. J. Peng et al. (2014) found a significant relationship between the interaction between

person-job fit and person-organization fit and to turnover intentions (β = -.154, p < .01).

In the same study, demand-ability fit had a positive correlation with work engagement (r

= .44, p < .01) and there was a significant positive correlation between work engagement

and turnover intentions (r = -.51, p < .01; Peng et al., 2014). Other research on demands-

abilities fit has shown that an employee’s perception of his or her fit predicted both job

commitment and job satisfaction (Bogler & Nir, 2015; Kristof-Brown et al., 2005;

McKee-Ryan & Harvey, 2011), and Gabriel et al. (2014) found that the perception of

person-job fit predicted job satisfaction (γ = .03, p < .05). Although not all of these

studies address employee turnover, they are about related concepts, and they may indicate

the importance of employee perception in the relationship between cognitive fit and

turnover.

Prior research on the Kaleidoscope Career Model identified career trends based

on gender (Sullivan & Mainiero, 2007). In addition, Hoglin and Barton’s (2013) research

noted gender as an attribute related to military retention. As expected based on these

previous results, gender was a statistically significant predictor of voluntary employee

turnover as an individual factor, or through the interaction with other factors, in three of

the four tests. This result supports the KCM finding that females and males enact their

careers differently (Sullivan & Mainiero, 2007). In addition, the difference between

males and females in average number of months of service (57.77 for females and 64.19

for males) also corroborates this claim since sailors enlist at the entry level, and Sullivan

and Mainiero (2007) found that while males and females both desired challenge at the

outset of their careers, females more frequently chose balance than men at mid-career. As

a final note, length of service as an independent variable was not statistically significant,

which differs from prior research by Hoglin and Barton (2013). However, length of

service was statistically significant when interacting with cognitive fit terms in both

voluntary and involuntary employee turnovers when using the full dataset. In addition, all

of the results were positive, indicating that these interactions grew slightly over time.

Summary

The loss of top talent to employee turnover negatively impacts organizational

success, both from the perspective of human capital and from a financial standpoint

(Godlewski & Kline, 2012). Failure to retain high-performing employees is a problem

because it increases recruitment and reenlistment costs, and it can result in the promotion

of lower quality and less experienced personnel. The goal of this study was to examine

cognitive fit as a predictor of employee turnover of U.S. Navy enlisted sailors using a

quantitative research design and multinomial logistic regression. Although the square of

cognitive fit and some of the other interactions between variables were statistically

significant for voluntary and involuntary turnover in the full dataset, the effect sizes were

very small, and further testing of 10%, 1%, and stratified subsets of the data refuted the

value of these findings. These results indicate that cognitive fit is not an important

predictor of future employee turnover.

Chapter 5: Implications, Recommendations, and Conclusions

Retaining top-performing sailors is one of the U.S. Navy’s top priorities.

Employee turnover is a key concern for the U.S. Navy because high turnover results in

increased personnel costs and a lower quality and less experienced workforce (Pinelis &

Huff, 2014). The purpose of this non-experimental, quantitative study was to examine the

relationship between cognitive fit and employee turnover in the U.S. Navy. The research

question that guided this research project was, to what extent does cognitive fit, gender,

and length of service predict employee turnover amongst U.S. Navy enlisted sailors?

The Navy collects data on sailors when it recruits them and at their retention

decision points in the Career Waypoints system. The Navy provided those data for sailors

who made a retention decision in 2014. The Navy measures cognitive ability using the

ASVAB, and it uses the results in the hiring process for those desiring to enlist. The data

the researcher used in this study were secondary case-file data from the U.S. Navy’s

Career Waypoints personnel database for all enlisted sailor retention decisions that

occurred in 2014. The data included ASVAB test scores and employee turnover

outcomes, as well as gender, paygrade, and length of service.

To conduct this research, the researcher used a quantitative design, using

multinomial logistic regression to explore whether cognitive fit predicts employee

turnover. The researcher used polytomous employee turnover outcomes: involuntary

separation, voluntary separation, or reenlistment, and included gender and length of

service as covariants in the study based on previous research by Hoglin and Barton

(2013) on employee turnover, and the relevance of these variables based on the

kaleidoscope career model (Mainiero & Sullivan, 2005; Sullivan & Mainiero, 2007). The

researcher tested both cognitive fit and the square of cognitive to account for the expected

curvilinear relationship between cognitive fit and employee turnover (Field, 2009).

Finally, to validate the findings, the researcher conducted the statistical tests on the full

dataset, a 1% subset, and a 10% subset of the data (Ertas, 2015).

In the full dataset, gender, cognitive fit squared, and the interactions of cognitive

fit and the square of cognitive fit with length of service were statistically significant for

voluntary turnover. The interaction of cognitive fit squared with length of service, and the

three-way interaction of cognitive fit squared with gender and length of service were

statistically significant for involuntary turnover. However, the results revealed that the

proposed model explained less than 1% of employee turnover. In addition, only the odds

ratio for gender indicated a measurable impact; the rest of the odds ratios showed a very

small effect size. Finally, the results of the same analysis on 10%, 1%, and stratified

random subsets of the data provided evidence that cognitive fit is not an important

predictor of employee turnover.

There are three limitations of this study: generalizability to the general population,

sample techniques including the use of a cross-sectional design, and the omission of

potentially significant moderating or extraneous variables. First, although the personnel

of the U.S. Navy closely reflect the general U.S. population, the terms of employment for

enlisted sailors differ from those of other citizens. The Navy contracts sailors for terms of

enlistment usually lasting two to four years; during their contract, they have very limited

options to separate voluntarily from service. Based on that difference, other settings may

not replicate the behavior the researcher observed in the Navy dataset. Additionally, the

lack of a tool to measure cognitive fit in the civilian workforce reduces the potential

generalizability of the findings of this research study. Second, the use of a cross-sectional

design may also present a limitation, especially since length of service was not a

predictor as the researcher expected (Hoglin & Barton, 2013). Also, selecting the study

sample based on retention actions, where every case included a retention decision, limited

analysis options. Other similar retention studies reported results in terms of hazard ratios

(Maltarich et al., 2010), which are the relative likelihood of employee turnover occurring

based on one standard deviation difference in cognitive fit. An added limitation stems

from the omission of other potentially relevant variables, such as compensation and job

availability.

Although existing theory and empirical research do not directly explain the

relationship between cognitive ability and employee turnover (Maltarich et al., 2010), the

theory of employee fit and its key construct, demands-abilities fit, provided a basis for

considering why cognitive ability might employee turnover. Employee fit is the

alignment between an individual and his or her work environment (Billsberry et al., 2012;

Kristof-Brown & Billsberry, 2012; Kristof-Brown & Guay, 2011; Maynard &

Parfyonova, 2013; Thompson et al., 2015). Person-job fit is one of the dimensions of

employee fit that has gained recognition, and it includes the key concept of demands-

abilities fit (Kristof-Brown & Guay, 2011). In 2010, Maltarich et al. conducted a study

that, for the first time, examined the relationship between voluntary turnover and

cognitive fit instead of general cognitive ability, and found a curvilinear relationship

between cognitive fit and voluntary turnover for jobs with high cognitive demands. Based

on this theoretical framework and Maltarich et al.’s findings, the researcher expected

cognitive fit to have a curvilinear relationship with employee turnover, with

overqualification leading to a higher incidence of voluntary turnover, and

underqualification leading to a higher incidence of involuntary turnover. The researcher

chose two other factors as possible covariates, gender and length of service, based on

their proven relevance to employee turnover in previous research.

There were two key aspects of this research that are different than past studies.

First, the Navy’s RIDE algorithm offers a more precise way to measure cognitive fit

objectively than ever before. Second, in this model, the researcher included voluntary and

involuntary separation as separate turnover outcomes, and suggested that over- and

underqualification may predict not only employee turnover, but also whether it will be

voluntary or involuntary.

The researcher’s analysis of the full dataset produced significant results, finding

that there is a curvilinear relationship between cognitive fit, the interaction of fit and

length of service, and voluntary turnover. The researcher also found a curvilinear

relationship between the interaction of fit, length of service, and involuntary turnover.

However, the effect size was very small, and in tests of smaller subsets of the data, these

results did not hold up. From these results, the researcher concluded that cognitive fit and

interactions with gender and length of service are not important predictors of employee

turnover. Of note, since most of the sailors in the dataset were underqualified, with mean

cognitive fit -28.1305, it is possible this had an effect on the result, masking a stronger

relationship.

This chapter continues with a discussion of the implication of this research for job

placement, for predicting future retention, and for future research. Several

recommendations for the U.S. Navy and future researchers follow this discussion. The

chapter ends with several conclusions about cognitive fit as a useful construct for

predicting positive employment outcomes.

Implications

Implications for job placement. From this dataset, mean cognitive fit was -

28.1305, which implies that the Navy was not optimally placing sailors into jobs where

they had the best cognitive fit. Although this research did not identify cognitive fit as an

important predictor of future employee turnover, the Navy previously found that

cognitive fit predicts positive employment outcomes such as training completion,

promotion, and retention (Department of the Navy, 2012), so this situation may be

affecting training costs and promotion results.

Implications for predicting future retention. The results of this study differ

from previous research on employee fit and its many conceptualizations (Kristof-Brown

& Guay, 2011). As reported by Kristof-Brown and Guay (2011), researchers have

proposed many factors as meaningful to the alignment between an employee and a job,

such as demands, abilities, values, climate, goals, personality, and ethics. Past research on

person-job fit in the civilian sector, and more specifically demands-abilities fit, has

shown strong correlation to job commitment, job satisfaction, and intent to quit (Kristof-

Brown et al., 2005), all concepts related to employee turnover. The prevalence of

previous research has measured employee fit subjectively by asking research participants

to rate alignment.

One of the distinctions of this research was the use of an objective measurement

of fit. Maltarich et al.’s (2010) study was the only other attempt to measure fit

objectively. The conjecture was that an objective measure of fit may provide a more

useful measure for hiring new employees (Fine & Nevo, 2011) and for predicting

employee turnover. However, in this research, cognitive fit was not an important factor in

predicting future retention. This implies that an objective measurement of cognitive fit,

without some subjectivity on the part of the employee, may not be adequate for

predicting future employee retention during the hiring process.

Implications for future research. There are several implications from this study

for future research. First, cognitive fit is a relatively new conceptualization of employee

fit, and methods to measure it objectively are still under development. The Navy uses the

ASVAB test and a unique algorithm to conduct this measurement, and this research

utilized that construct. Other methods of measuring cognitive fit may be worth

developing to test this concept further.

Next, the researcher considered all Navy sailors across the spectrum, and did not

group their jobs based on cognitive demands. Maltarich et al. (2010) found a relationship

between employee turnover and jobs with high cognitive demands, but not the other

groups. These results imply the need to consider more specificity in the study group to

understand the utility of cognitive fit on employment outcomes fully.

For future research on the Navy, a reexamination of how the researcher

categorized cases into the three employee turnover outcomes (reenlistment, voluntary

separation, and involuntary separation) may be helpful. Additionally, in this study the

researcher combined the Navy’s S-score for training success and the Q-score for rating

norms into one cognitive fit scale. Future research on these individual factors, and the

manner in which they work together may be beneficial.

Finally, future research on cognitive fit should expand the aperture to consider

other positive employment outcomes, such as training success and promotion. Other

outcomes that prior researchers have considered on employee fit, including job

satisfaction, and job commitment, may also be of interest.

Recommendations

Recommendations for job placement. Although the Navy has been using the

RIDE algorithm since 2009, there is no requirement to provide the information on best fit

jobs to the applicant or to utilize it in the placement process. Job placement operates on a

first come, first served basis, so job availability limits the process. This process constraint

reduces the Navy’s ability to optimize cognitive fit, and in some cases, inevitably results

in the Navy placing sailors in jobs where they are over- or underchallenged. The Navy

could use cognitive fit to limit career field choices for applicants, delaying recruitment

until best fit career fields are available.

Recommendations for predicting future retention. Since the results of this

research found that cognitive fit was not an important predictor of future retention, it may

prove valuable to consider the interaction between cognitive fit and career interests on

retention. The U.S. Navy utilizes a career interests inventory called Jobs in the Navy to

augment placement options based on cognitive ability. It is currently offering this

questionnaire on a voluntary basis, but it is adapting it for use with all new Navy recruits.

The interaction between interests and cognitive fit could have strong future application in

hiring and placement practices.

Recommendations for future research. Future research could focus on methods

of measuring cognitive fit, especially for civilian organizations who do not use a tool like

the ASVAB to test cognitive ability, and who hire new personnel throughout their

organizations, not just at entry level.

Future research with greater specificity in the study group may identify greater

effects that the Navy can put into practice. For example, Maltarich et al. (2010) detected a

predictive relationship between cognitive ability and jobs with high cognitive demands,

but did not identify a relationship for jobs with low or medium cognitive demands.

Options for future research should include grouping jobs by cognitive demands, and, for

the Navy in particular, by rating. One may also want to study a subset of sailors with only

high or low cognitive fit.

For future research on cognitive fit in the Navy, there are several

recommendations to consider. First, the categorization of outcomes is a subject that needs

more attention. In this study, the researcher grouped sailors who received approval for

reenlistment in their current rating together with sailors who received approval to convert

to a new rating as a part of their reenlistment. Since the Navy’s measurement of cognitive

fit is based on the match between a sailor and a rating, and the sailor’s rating changes if

he or she converts for reenlistment, one could argue that cognitive fit in the original

rating is not relevant and that these cases should not be in the reenlistment category.

Additionally, sailors who were ineligible to reenlist may have a different behavior pattern

than sailors who requested to separate. Studying these two groups separately may yield

important differences in the results.

One final recommendation about categorizing outcomes in future research is to

examine voluntary separation outcomes more closely using all retention requests, rather

than only the final request. In Career Waypoints, sailors can submit reenlistment

applications monthly, starting when there are fifteen months remaining on their

enlistment contracts, until there are only three months remaining on the contract. These

monthly applications offer additional information about the nature of the final outcome.

For example, if the Navy denies a sailor reenlistment and there are more than six months

left on his or her enlistment contract, he or she can reapply each of the remaining months

either to reenlist in rate or to convert to a new rate. Based on the process in the Career

Waypoints system, once a sailor reaches the six-month point, if he or she has not received

approval for reenlistment in his or her current rate, he or she can still apply—but only to

convert to a new rating. In this research, the researcher manually included sailors who did

not receive approval for reenlistment in rate, and who choose not to apply for conversion,

in the involuntary separation category. Others could use the history of retention requests

to detect these cases or other anomalies, and to categorize them appropriately. Finally,

since the Navy uses two measurements to determine cognitive fit based on training

success and rating norms, future research could study S-score and Q-score values

separately to measure their predictive value for employee turnover.

Researchers should not limit the usefulness of cognitive fit as a construct for

predicting employment outcomes to predicting future employee turnover. Cognitive fit

may be relevant to other positive outcomes such as training success and promotion. The

researcher’s final recommendation is for future research to consider other applications for

cognitive ability as a measurement of employee fit.

Conclusions

This research offers evidence that cognitive fit and interactions with gender and

length of service are not important predictors of employee turnover. This finding is

contradictory to Maltarich et al.’s (2010) research. This research also adds to the

literature on human resources by proposing a new construct for measuring cognitive fit,

and examining voluntary and involuntary employee turnover as separate outcomes. Other

researchers may extrapolate the findings of this study to other organizations, and most

directly to the other U.S. military services, and offer new ideas for future research on

cognitive fit. The results of this study benefit the U.S. Navy, and other military services

and organizations, by exploring ways to improve the hiring process, and optimizing

placement, utilization, and retention of personnel.

References

Afsar, B., Badir, Y., & Khan, M. M. (2015). Person-job fit, person-organization fit and

innovative work behavior: The mediating role of innovation trust. Journal of High

Technology Management Research, 26, 105-116.

doi:10.1016/j.hitech.2015.09.001

Al-Emadi, A. Q., Schwabenland, C., & Qi, W. (2015). The vital role of employee

retention in human resource management: A literature review. IUP Journal of

Organizational Behavior, 15(3), 7-32. doi:10.1016/j.hrmr.2005.11.003

Allen, D., Bryant, P., & Vardaman, J. (2010). Retaining talent: Replacing misconceptions

with evidence-based strategies. Academy of Management Perspectives, 24(2), 48-

64. doi:10.5465/amp.2010.51827775

Arkes, J., & Cunha, J. M. (2015). Workplace goals and output quality: Evidence from

time-constrained recruiting goals in the US navy. Defence & Peace Economics,

26, 491-515. doi:10.1080/10242694.2014.891352

Arkes, J., & Mehay, S. (2014). The impact of the unemployment rate on attrition of first-

term enlistees. Defence and Peace Economics, 25(2), 125-138.

doi:10.1080/10242694.2013.752244

Babakus, E., Yavas, U., & Ashill, N. J. (2011). Service worker burnout and turnover

intentions: Roles of person-job fit, servant leadership, and customer orientation.

Services Marketing Quarterly, 32(1), 17-31. doi:10.1080/15332969.2011.533091

Bakker, A. (2011). An evidence-based model of work engagement. Current Directions in

Psychological Science, 20(4), 265-269. doi:10.1177/0963721411414534

Becker, W. J., & Cropanzano, R. (2011). Dynamic aspects of voluntary turnover: An

integrated approach to curvilinearity in the performance-turnover relationship.

Journal of Applied Psychology, 96(1), 233-246. doi:10.1037/a0021223

Beier, M., & Oswald, F. (2012). Is cognitive ability a liability? A critique and future

research agenda on skilled performance. Journal of Experimental Psychology-

Applied, 18, 331-345. doi:10.1037/a0030869

Bernerth, J. B., & Aguinis, H. (2016). A critical review and best-practice

recommendations for control variable usage. Personnel Psychology, 69(1), 229-

283. doi:10.1111/peps.12103

Billsberry, J., Talbot, D. L., & Ambrosini, V. (2012). Mapping fit: Maximizing

idiographic and nomothetic benefits. In J. Billsberry & A. L. Kristof-Brown

(Eds.), Organizational fit: Key issues and new directions (pp. 124-141).

Chichester, UK: Wiley-Blackwell. doi:10.1002/9781118320853.ch6

Bogler, R., & Nir, A. E. (2015). The contribution of perceived fit between job demands

and abilities to teachers’ commitment and job satisfaction. Educational

Management Administration & Leadership, 43, 541-560.

doi:10.1177/1741143214535736

Boon, C., den Hartog, D. N., Boselie, P., & Paauwe, J. (2011). The relationship between

perceptions of HR practices and employee outcomes: Examining the role of

person-organisation and person-job fit. International Journal of Human Resource

Management, 22(1), 138-162. doi:10.1080/09585192.2011.538978

Boudreau, J., Boswell, W., Judge, T., & Bretz, R., Jr. (2001). Personality and cognitive

ability as predictors of job search among employed managers. Personnel

Psychology, 54(1), 25-50. doi:10.1111/j.1744-6570.2001.tb00084.x

Breaugh, J. (2014). Predicting voluntary turnover from job applicant biodata and other

applicant information. International Journal of Selection and Assessment, 22,

321-332. doi:10.1111/ijsa.12080

Cabrera, E. F. (2009). Protean organizations: Reshaping work and careers to retain

female talent. Career Development International, 14(2), 186-201.

Carette, B., Anseel, F., & Lievens, F. (2013). Does career timing of challenging job

assignments influence the relationship with in-role job performance? Journal of

Vocational Behavior, 83(1), 61-67. doi:10.1016/j.jvb.2013.03.001

Center for Naval Analysis. (2014). Attrition and reenlistment of first-term sailors;

Update through end of FY14. Washington, DC.

Chen, C., Yen, C., & Tsai, F. C. (2014). Job crafting and job engagement: The mediating

role of person-job fit. International Journal of Hospitality Management, 37(1),

21-28. doi:10.1016/j.ijhm.2013.10.006

Chen, G., & Ployhart, R. E. (2006). An interactionalist analysis of soldier retention

across career stages and time (Report no. ADA448543). Texas A and M

University: U.S. Army Research Institute for the Behavioral and Social Sciences.

doi:10.1037/e500422012-001

Christensen, R. K., & Wright, B. E. (2011). The effects of public service motivation on

job choice decisions: Disentangling the contributions of person-organization fit

and person-job fit. Journal of Public Administration Research & Theory, 21, 723-

743. doi:10.1093/jopart/muq085

Coughlan, P. J., Gates, W. R., & Myung, N. (2014). One size does not fit all:

Personalized incentives in military compensation. Defense & Security Analysis,

30, 360. doi:10.1080/14751798.2014.948283

Crook, T., Todd, S. Y., Combs, J. G., Woehr, D. J., & Ketchen, D. R. (2011). Does

human capital matter? A meta-analysis of the relationship between human capital

and firm performance. Journal of Applied Psychology, 96, 443-456.

doi:10.1037/a0022147

Demidenko, E. (2007). Sample size determination for logistic regression revisited.

Statistics in Medicine, 26, 3385-3397. doi:10.1002/sim.2771

Department of Defense. (2011a). Population report. Retrieved from

http://prhome.defense.gov/portals/52/Documents/POPREP/poprep2011/appendix

d/d_32.html

Department of Defense. (2011b). Protection of human subjects and adherence to ethical

standards in DoD-supported research (Instruction 3216.02). Retrieved from

http://dtic.mil/whs/directives/corres/pdf/321602p.pdf

Department of Defense. (2014). Population representation in the military services.

Retrieved from http://prhome.defense.gov/RFM/MPP/AP/POPREP.aspx

Department of the Navy. (2012). RIDE measures of effectiveness (OPNAV N132).

Washington DC: Hewlett Packard

Duffy, R. D., Autin, K. L., & Bott, E. M. (2015). Work volition and job satisfaction:

Examining the role of work meaning and person-environment fit. Career

Development Quarterly, 63(2), 126-140. doi:10.1002/cdq.12009

Erdogan, B., Bauer, T., Peiro, J., & Truxillo, D. (2011a). Overqualified employees:

Making the best of a potentially bad situation for individuals and organizations.

Industrial and Organizational Psychology-Perspectives on Science and Practice,

4(2), 215-232. doi:10.1111/j.1754-9434.2011.01330.x

Erdogan, B., Bauer, T., Peiro, J., & Truxillo, D. (2011b). Overqualification theory,

research, and practice: Things that matter. Industrial and Organizational

Psychology-Perspectives on Science and Practice, 4(2), 260-267.

doi:10.1111/j.1754-9434.2011.01339.x

Ertas, N. (2015). Turnover intentions and work motivations of millennial employees in

federal service. Public Personnel Management, 44, 401-423.

doi:10.1177/0091026015588193

Farquhar, J. D. (2012). Philosophical assumptions of case study research. In Case study

research for business (pp. 15-30). London: Sage. doi:10.4135/9781446287910.n3

Farzaneh, J., Farashah, A., & Kazemi, M. (2014). The impact of person-job fit and

person-organization fit on OCB: The mediating and moderating effects of

organizational commitment and psychological empowerment. Personnel Review,

43, 672-691. doi:10.1108/pr-07-2013-0118

100

Feldman, D., & Maynard, D. (2011). A labor economic perspective on overqualification.

Industrial and Organizational Psychology-Perspectives on Science and Practice,

4(2), 233-235. doi:10.1111/j.1754-9434.2011.01331.x

Felstead, A., Gallie, D., Green, F., & Inanc, H. (2015). Fits, misfits and interactions:

Learning at work, job satisfaction and job-related well-being. Human Resource

Management Journal, 25, 294-310. doi:10.1111/1748-8583.12071

Field, A. (2009). Discovering statistics using SPSS. London, England: Sage.

Fine, S., & Nevo, B. (2008). Too smart for their own good? A study of perceived

cognitive overqualification in the workforce. International Journal of Human

Resource Management, 19, 346-355. doi:10.1080/09585190701799937

Fine, S., & Nevo, B. (2011). Overqualified job applicants: We still need predictive

models. Industrial and Organizational Psychology: Perspectives on Science and

Practice, 4(2), 240-242. doi:10.1111/j.1754-9434.2011.01333.x

Ford, M. T., Gibson, J. L., DeCesare, A. L., Marsh, S. M., & Griepentrog, B. K. (2013).

Pre-entry expectations, attitudes, and intentions to join predict military tenure.

Military Psychology, 25(1), 36-45. doi:10.1037/h0094755

Freund, P. A., & Kasten, N. (2012). How smart do you think you are? A meta-analysis on

the validity of self-estimates of cognitive ability. Psychological Bulletin, 138,

296-321. doi:10.1037/a0026556

Gabriel, A. S., Diefendorff, J. M., Chandler, M. M., Moran, C. M., & Greguras, G. J.

(2014). The dynamic relationships of work affect and job satisfaction with

perceptions of fit. Personnel Psychology, 67, 389-420. doi:10.1111/peps.12042

George, C. (2015). Retaining professional workers: What makes them stay? Employee

Relations, 37(1), 102. doi:10.1108/ER-10-2013-0151

Gibson, J. L., Hackenbracht, J., & Tremble, T. R. (2014). An event history analysis of

first-term soldier attrition. Military Psychology, 26(1), 55-66.

doi:10.1037/mil0000030

Godlewski, R., & Kline, T. (2012). A model of voluntary turnover in male Canadian

Forces recruits. Military Psychology, 24(3), 251-269.

doi:10.1080/08995605.2012.678229

Grant, J., Vargas, A. L., Holcek, R. A., Watson, C. H., Grant, J. A., & Kim, F. S. (2012).

Is the ASVAB ST composite score a reliable predictor of first-attempt graduation

for the U.S. Army operating room specialist course. Military Medicine, 177,

1352-1358. doi:10.7205/milmed-d-12-00189

Hambrick, D. Z., Rench, T. A., Poposki, E. M., Darowski, E. S., Roland, D., Bearden, R.

M., & Brou, R. (2011). The relationship between the ASVAB and multitasking in

101

Navy sailors: A process-specific approach. Military Psychology, 23, 365-380.

doi:10.1080/08995605.2011.589323

Han, T., Chiang, H., McConville, D., & Chiang, C. (2015). A longitudinal investigation

of person-organization fit, person-job fit, and contextual performance: The

mediating role of psychological ownership. Human Performance, 28, 425-439.

doi:10.1080/08959285.2015.1021048

Hardin, E. E., & Donaldson, J. I. (2014). Predicting job satisfaction: A new perspective

on person-environment fit. Journal of Counseling Psychology, 61, 634-640.

doi:10.1037/cou0000039

Held, J. D., & Carretta, T. R. (2013). Evaluation of tests of processing speed, spatial

ability, and working memory for use in military occupational classification (No.

NPRST-TR-14-1). Millington TN: Navy Personnel Research Studies and

Technology.

Held, J. D., Carretta, T. R., & Rumsey, M. G. (2014). Evaluation of tests of perceptual

speed/accuracy and spatial ability for use in military occupational classification.

Military Psychology, 26(3), 199-220. doi:10.1037/mil0000043

Held, J. D., Hezlett, S. A., Johnson, J. W., McCloy, R. A., Drasgow, F., & Salas, E.

(2014). Introductory guide for conducting ASVAB validation/standards studies in

the U.S. Navy (No. NPRST-TR-15-1). Millington TN: Navy Personnel Research

Studies and Technology.

Hinami, K., Whelan, C., Miller, J., Wolosin, R., & Wetterneck, T. (2013). Person-job fit:

An exploratory cross-sectional analysis of hospitalists. Journal of Hospital

Medicine, 8(2), 96-101. doi:10.1002/jhm.1995

Hoglin, P. J., & Barton, N. (2013). First-term attrition of military personnel in the

Australian Defence Force. Armed Forces & Society, 41(1), 43-68.

doi:10.1177/0095327X13494743

Holland, J. (1959). A theory of vocational choice. Journal of Counseling Psychology,

6(1), 35-45. doi:10.1037/h0040767

Holtom, B. C., Smith, D. R., Lindsay, D. R., & Burton, J. P. (2014). The relative strength

of job attitudes and job embeddedness in predicting turnover in a U.S. military

academy. Military Psychology, 26, 397-408. doi:10.1037/mil0000055

Hom, P., Mitchell, T., Lee, T., & Griffeth, R. (2012). Reviewing employee turnover:

Focusing on proximal withdrawal states and an expanded criterion. Psychological

Bulletin, 138, 831-858. doi:10.1037/a0027983

Hong, E. N. C., Hao, L. Z., Kumar, R., Ramendran, C., & Kadiresan, V. (2012).

Effectiveness of human resource management practices on employee retention in

institute of higher learning: A regression analysis. International Journal of

102

Business Research and Management, 3(2), 60-79. Retrieved from

http://www.cscjournals.org/manuscript/Journals/IJBRM/Volume3/Issue2/IJBRM-

81.pdf

Hu, J., Erdogan, B., Bauer, T. N., Jiang, K., Liu, S., & Li, Y. (2015). There are lots of big

fish in this pond: The role of peer overqualification on task significance,

perceived fit, and performance for overqualified employees. Journal of Applied

Psychology, 100, 1228-1238. doi:10.1037/apl0000008

Jackofsky, E. F. (1984). Turnover and job performance: An integrated process model.

Academy of Management Review, 9, 74-83. doi:10.2307/258234

Klein, R. M., Dilchert, S., Ones, D. S., & Dages, K. D. (2015). Cognitive predictors and

age-based adverse impact among business executives. Journal of Applied

Psychology, 100, 1497-1510. doi:10.1037/a0038991

Knapik, J. J., Jones, B. H., Hauret, K., Darakjy, S., & Piskator, E. (2004). A review of the

literature on attrition from the military services: Risk factors for attrition and

strategies to reduce attrition (No. USACHPPM-12-HF-01Q9A-04). Aberdeen

Proving Ground, MD: Army Center for Health Promotion and Preventive

Medicine.

Kumazawa, R. (2010). Promotion speed and its effect on attrition of Navy-enlisted

personnel: Addressing heterogeneity in high school credentials. Applied

Economics, 42, 2563-2576. doi:10.1080/00036840801964450

Kristof-Brown, A. L., & Billsberry, J. (2012). Fit for the future. In J. Billsberry & A. L.

Kristof-Brown (Eds.), Organizational fit: Key issues and new directions (pp. 1-

19). Chichester, UK: Wiley-Blackwell. doi:10.1002/9781118320853.ch1

Kristof-Brown, A. L., & Guay, R. P. (2011). Person-environment fit. In S. Zedeck (Ed.)

APA handbook of industrial and organizational psychology: Maintaining,

expanding, and contracting the organization (Vol. 3, pp. 3-50). Washington, DC:

American Psychological Association. doi:10.1037/12171-001

Kristof-Brown, A. L., Zimmerman, R. D., & Johnson, E. C. (2005). Consequences of

individuals’ fit at work: A meta-analysis of person-job, person-organization,

person-group, and person-supervisor fit. Personnel Psychology, 58(2), 281.

doi:10.1111/j.1744-6570.2005.00672.x

Kulkarni, M., Lengnick-Hall, M. L., & Martinez, P. G. (2015). Overqualification,

mismatched qualification, and hiring decisions. Personnel Review, 44, 529-549.

doi:10.1108/PR-11-2013-0204

Li, N., Barrick, M. R., Zimmerman, R. D., & Chiaburu, D. S. (2014). Retaining the

productive employee: The role of personality. Academy of Management Annals,

8(1), 347-395. doi:10.1080/19416520.2014.890368

103

Lin, Y., Yu, C., & Yi, C. (2014). The effects of positive affect, person-job fit, and well-

being on job performance. Social Behavior & Personality: An International

Journal, 42, 1537-1547. doi:10.2224/sbp.2014.42.9.1537

Liu, S., Luksyte, A., Zhou, L., Shi, J., & Wang, M. (2015). Overqualification and

counterproductive work behaviors: Examining a moderated mediation model.

Journal of Organizational Behavior, 36(2), 250-271. doi:10.1002/job.1979

Lobene, E., & Meade, A. (2013). The effects of career calling and perceived

overqualification on work outcomes for primary and secondary school teachers.

Journal of Career Development, 40, 508-530. doi:10.1177/0894845313495512

Lu, C., Wang, H., Lu, J., Du, D., & Bakker, A. B. (2014). Does work engagement

increase person-job fit? The role of job crafting and job insecurity. Journal of

Vocational Behavior, 84(2), 142-152. doi:10.1016/j.jvb.2013.12.004

Lytell, M. C., & Drasgow, F. (2009). “Timely” methods: Examining turnover rates in the

U.S. Military. Military Psychology, 21, 334-350.

doi:10.1080/08995600902914693

Mafini, C., & Dubihlela, J. (2013). Determinants of military turnover of technical air-

force specialists: An empirical case analysis. Mediterranean Journal of Social

Sciences, 4, 523. doi:10.5901/mjss.2013.v4n3p523

Mainiero, L., & Sullivan, S. (2005). Kaleidoscope careers: An alternate explanation for

the opt-out revolution. Academy of Management Executive, 19(1), 106-123.

doi:10.5465/ame.2005.15841962

Maltarich, M. A., Nyberg, A. J., & Reilly, G. (2010). A conceptual and empirical analysis

of the cognitive ability-voluntary turnover relationship. Journal of Applied

Psychology, 95, 1058-1070. doi:10.1037/a0020331

Maltarich, M. A., Reilly, G., & Nyberg, A. J. (2011). Objective and subjective

overqualification: Distinctions, relationships, and a place for each in the literature.

Industrial & Organizational Psychology, 4(2), 236-239. doi:10.1111/j.1754-

9434.2011.01332.x

Marcus, B., & Wagner, U. (2015). What do you want to be? Criterion-related validity of

attained vocational aspirations versus inventoried person-vocation fit. Journal of

Business and Psychology, 30(1), 51-62. doi:10.1007/s10869-013-9330-9

Maynard, D., & Parfyonova, N. (2013). Perceived overqualification and withdrawal

behaviours: Examining the roles of job attitudes and work values. Journal of

Occupational and Organizational Psychology, 86, 435-455.

doi:10.1111/joop.12006

104

McKee-Ryan, F. M., & Harvey, J. (2011). “I have a job, but …”: A review of

underemployment. Journal of Management, 37, 962-996.

doi:10.1177/0149206311398134

Melvin, B., Hale, R., & Foster, M. (2013). The importance and challenge of ability

assessment. Career Planning & Adult Development Journal, 29(4), 98-113.

Mumford, M. D., Watts, L. L., & Partlow, P. J. (2015). Leader cognition: Approaches

and findings. Leadership Quarterly, 26, 301-306.

doi:10.1016/j.leaqua.2015.03.005

Nyberg, A. J. (2010). Retaining your high performers: Moderators of the performance–

job satisfaction–voluntary turnover relationship. Journal of Applied Psychology,

95, 440–453.

Oh, I., Le, H., Whitman, D. S., Kim, K., Yoo, T., Hwang, J., & Kim, C. (2014). The

incremental validity of honesty-humility over cognitive ability and the big five

personality traits. Human Performance, 27(3), 206-224.

doi:10.1080/08959285.2014.913594

Ones, D. S., & Viswesvaran, C. (2011). Individual differences at work. In T. Chamorro-

Premuzic, S. von Stumm, & A. Furnham (Eds.), The Wiley-Blackwell handbook

of individual differences (pp. 379-407). Wiley-Blackwell.

doi:10.1002/9781444343120

Park, H. I., Beehr, T. A., Han, K., & Grebner, S. I. (2012). Demands-abilities fit and

psychological strain: Moderating effects of personality. International Journal of

Stress Management, 19(1), 1-33. doi:10.1037/a0026852

Peng, C., Lee, K., & Ingersoll, G. (2002). An introduction to logistic regression analysis

and reporting. Journal of Educational Research, 96(1), 3-14.

Peng, J., Lee, Y., & Tseng, M. (2014). Person-organization fit and turnover intention:

Exploring the mediating effect of work engagement and the moderating effect of

demand-ability fit. Journal of Nursing Research, 22(1), 1-11.

doi:10.1097/jnr.0000000000000019

Peng, Y., & Mao, C. (2015). The impact of person-job fit on job satisfaction: The

mediator role of self-efficacy. Social Indicators Research, 121, 805-813.

doi:10.1007/s11205-014-0659-x

Pinelis, J. K., & Huff, J. M. (2014). The economy and enlisted retention in the Navy.

(Report No. DRM-2014-U-007301-Final). Washington DC: Center for Naval

Analysis.

Quratulain, S., & Khan, A. K. (2015). How does employees’ public service motivation

get affected? A conditional process analysis of the effects of person-job fit and

105

work pressure. Public Personnel Management, 44(2), 266-289.

doi:10.1177/0091026014568461

Rainayee, R. A. (2013). Employee turnover intentions: Job stress or perceived alternative

external opportunities. Business and Management, 5(1), 48-59. Retrieved from

http://search.proquest.com/openview/07867f1bfab6239e9e522f1ef1453f0d/1?pq-

origsite=gscholar

Rumsey, M. G. (2012). Military selection and classification in the United States. In J. H.

Laurence & M. D. Matthews (Eds.), The Oxford handbook of military psychology

(pp. 129-147). New York, NY: Oxford University Press.

doi:10.1093/oxford/9780195399325.013.0054

Rumsey, M. G., & Arabian, J. M. (2014a). Introduction to the special issue on selected

new developments in military enlistment testing. Military Psychology, 26(3), 131-

137. doi:10.1037/mil0000041

Rumsey, M. G., & Arabian, J. M. (2014b). Military enlistment selection and

classification: Moving forward. Military Psychology, 26(3), 221-251.

doi:10.1037/mil0000040

Ryan, A., & Ployhart, R. E. (2014). A century of selection. Annual Review of Psychology,

65(1), 693-717. doi:10.1146/annurev-psych-010213-115134

Schmidt, N. (2014). Personality and cognitive ability as predictors of effective

performance at work. Annual Review of Organizational Psychology and

Organizational Behavior, 1(1), 45-65. doi:10.1146/annurev-orgpsych-031413-

091255

Sekiguchi, T., & Huber, V. L. (2011). The use of person-organization fit and person-job

fit information in making selection decisions. Organizational Behavior and

Human Decision Processes, 116, 203-216. doi:10.1016/j.obhdp.2011.04.001

Shaw, J. D., Park, T., & Kim, E. (2013). A resource-based perspective on human capital

losses, HRM investments, and organizational performance. Strategic Management

Journal, 34, 572-589. doi:10.1002/smj.2025

Smith, D. R., Holtom, B. C., & Mitchell, T. R. (2011). Enhancing precision in the

prediction of voluntary turnover and retirement. Journal of Vocational Behavior,

79(1), 290-302. doi:10.1016/j.jvb.2010.11.003

Song, Z., & Chon, K. (2012). General self-efficacy’s effect on career choice goals via

vocational interests and person-job fit: A mediation model. International Journal

of Hospitality Management, 31, 798-808. doi:10.1016/j.ijhm.2011.09.016

Sullivan, S. & Mainiero, L. (2007). The changing nature of gender roles, alpha/beta

careers and work-life issues: Theory-driven implications for human resource

106

management. Career Development International, 12(3), 238-263.

doi:10.1108/13620430710745881

Super, D. (1953). A theory of vocational development. American Psychologist, 8(1), 185-

190. doi:10.1037/h0056046

Thompson, K. W., Shea, T. H., Sikora, D. M., Perrewé, P. L., & Ferris, G. R. (2013).

Rethinking underemployment and overqualification in organizations: The not so

ugly truth. Business Horizons, 56(1), 113-121. doi:10.1016/j.bushor.2012.09.009

Thompson, K. W., Sikora, D. M., Perrewé, P. L., & Ferris, G. R. (2015). Employment

qualifications, person-job fit, underemployment attributions, and hiring

recommendations: A three-study investigation. International Journal of Selection

and Assessment, 23(3), 247-262. doi:10.1111/ijsa.12112

Tims, M., Derks, D., & Bakker, A. B. (2016). Job crafting and its relationships with

person-job fit and meaningfulness: A three-wave study. Journal of Vocational

Behavior, 92(1), 44-53. doi:10.1016/j.jvb.2015.11.007

National Defense, 32 C.F.R. § 219.102(f) (2014). Retrieved from

http://www.ecfr.gov/cgi-bin/text-idx?tpl=/ecfrbrowse/Title32/32tab_02.tpl

Trippe, D. M., Moriarty, K. O., Russell, T. L., Carretta, T. R., & Beatty, A. S. (2014).

Development of a cyber/information technology knowledge test for military

enlisted technical training qualification. Military Psychology, 26(3), 182-198.

doi:10.1037/mil0000042

Truxillo, D. M., McCune, E. A., Bertolino, M., & Fraccaroli, F. (2012). Perceptions of

older versus younger workers in terms of big five facets, proactive personality,

cognitive ability, and job performance. Journal of Applied Social Psychology, 42,

2607-2639. doi:10.1111/j.1559-1816.2012.00954.x

Tzafrir, S. S., Gur, A. B., & Blumen, O. (2015). Employee social environment (ESE) as a

tool to decrease intention to leave. Scandinavian Journal of Management, 31(1),

136-146. doi:10.1016/j.scaman.2014.08.004

van Iddekinge, C. H., Roth, P. L., Putka, D. J., & Lanivich, S. E. (2011). Are you

interested? A meta-analysis of relations between vocational interests and

employee performance and turnover. Journal of Applied Psychology, 96, 1167-

1194. doi:10.1037/a0024343

Venz, L., & Sonnentag, S. (2015). Being engaged when resources are low: A multi-

source study of selective optimization with compensation at work. Journal of

Vocational Behavior, 91(1), 97-105. doi:10.1016/j.jvb.2015.09.008

Warr, P., & Inceoglu, I. (2012). Job engagement, job satisfaction, and contrasting

associations with person-job fit. Journal of Occupational Health Psychology,

17(2), 129-138. doi:10.1037/a0026859

107

Watson, S. (2010). Testing, validating, and applying an empirical model of human

performance in a high-performance organization. In P. E. O’Connor & J. V. Cohn

(Eds.), Human performance enhancement in high-risk environments: Insights,

developments, and future directions from military research (pp. 16-36). Santa

Barbara, CA: ABC-CLIO.

Weaver, T. L. (2015). Intent to exit: Why do US federal employees leave? International

Journal of Public Administration, 38, 442. doi:10.1080/01900692.2014.949739

White, L. A., Rumsey, M. G., Mullins, H. M., Nye, C. D., & LaPort, K. A. (2014).

Toward a new attrition screening paradigm: Latest Army advances. Military

Psychology, 26(3), 138-152. doi:10.1037/mil0000047

Yerkes, R. M., & Dodson, J. D. (1908). The relation of strength of stimulus to rapidity of

habit-formation. Journal of Comparative Neurology & Psychology, 18, 459.

doi:10.1002/cne.920180503

Zaccaro, S. J., Connelly, S., Repchick, K. M., Daza, A. I., Young, M. C., Kilcullen, R.

N., & Bartholomew, L. N. (2015). The influence of higher order cognitive

capacities on leader organizational continuance and retention: The mediating role

of developmental experiences. Leadership Quarterly, 26, 342-358.

doi:10.1016/j.leaqua.2015.03.007

108

Appendixes

109

Appendix A: Research Request and Approval

April 1, 2016

From: CAPT Renee J. Squier, USN

To: Director, Navy Personnel Plans and Policy (OPNAV N13) Subj:

REQUEST TO CONDUCT RESEARCH

Ref: (a) DoD1 3216.02

Encl: (l) Research method

(2) Human Research Determination

l. Respectfully request permission to conduct a quantitative, correlational study using secondary data from

the Career Waypoints system to determine if cognitive fit predicts employee turnover.

2. The research plan is to compute cognitive fit for each sailor by comparing U.S. Navy enlisted sailor

Armed Services Vocational Aptitude Battery (ASVAB) test scores to the cognitive demands for their career

fields in the Navy. This measurement of cognitive fit will then be compared to retention to determine if it is

related to employee turnover. The full research method is provided in enclosure (l). The results of this study

could benefit the U.S. Navy, and other military services and organizations by providing a measurable pre-

hire predictor to improve hiring processes and better match individuals with jobs—optimizing placement,

utilization, and retention of personnel.

3. Per reference (a), although this research will use data on U.S. Navy sailors, the Human Research

Determination included as enclosure (2) deems the study exempt.

4. The proposed study sample is active duty U.S. Navy enlisted sailors, paygrades El thru E6 with up to 14

years of service who reenlisted or separated in calendar year 2014. No personally identifiable information

(name or social security number) will be utilized. The Career Waypoints data elements listed below are

requested to conduct the proposed research:

Cognitive Fit Turnover Outcomes

Rating Gender

Race/Ethnicity Marital Status

Age Length of Service

Paygrade Educational Level

R. J. SQUIER

110

DEPARTMENT OF THE NAVY OFFICE OF THE CHIEF OF NAVAL OPERATIONS

2000 NAVY PENTAGON

WASHINGTON, D.C. 20350-2000

1040

SerN13/ 072

8 Apr 16

From: Director, Military Personnel Plans and Policy (N 13)

To: CAPT Renee J. Squier, USN subj:

REQUEST TO CONDUCT RESEARCH

Ref: (a) Request to Conduct Research

1. Your research request (reference (a)) to conduct a quantitative, correlational study using secondary data from the Career Waypoints to determine if cognitive fit predicts

employee turnover system is approved.

2. Please share the results of your research with us when it is complete.

U.S. Navy

Copy to:

N132

111

Appendix B: Research Variables

Variable

name Type

Level of

Measurement Description

Cognitive

Fit

Independent

Continuous Participant test score compared to rating

norms (Q-score) and training success (S-

score). These scores are added to obtain a

value for cognitive fit.

Turnover

Outcome

Dependent

Categorical

Voluntary Separation: requested to

separate or transition to the Navy Reserve

(NES Codes: VSP, RQR, and ITS)

Involuntary Separation: not selected for

retention or ineligible to reenlist (NES

Code: FSP, ESP, DFI, IEG, and VSP cases

where a sailor was not approved to reenlist

in-rate, and there were no options to

convert to another rating, as noted in the

application type reason)

Reenlistment: approved for reenlistment in

the current rating or to convert to another

rating (NES Code: AIR, ACV)

Rating Independent Categorical All U.S. Navy enlisted career fields

Gender Independent Binary 0 for Male

1 for Female

Length of

Service

Independent Interval Number of months of service since initial

enlistment

Paygrade Independent Categorical E1 through E6

112

Appendix C: Human Subjects Research Determination

March 28, 2016

From: CAPT Renee J. Squier, USN

To: Mr. Daniel Wallace, NAVSEA HRPO

Subj: REQUEST HUMAN SUBJECTS RESEARCH DETERMINATION

Ref: (a) DoDI 3216.02

Encl: (a) Research method

1. Per ref (a), request review and Human Subjects Research determination on the

proposed quantitative, correlational study to determine if cognitive fit predicts employee

turnover using secondary data from the Navy’s Career Waypoints system. The results of

this study could benefit the U.S. Navy, and other military services and organizations by

providing a measurable pre-hire predictor that could improve hiring processes to better

match individuals with jobs, optimizing placement, utilization and retention of personnel.

2. Although this research will use data on U.S. Navy sailors, the data will not be

obtained through intervention or interaction with the individual or in a context where an

individual would have a reasonable expectation of privacy, nor will it include personally

identifiable information. The proposed study sample is active duty U.S. Navy enlisted

sailors, paygrades E1 thru E6 with up to 14 years of service who reenlisted or separated

in calendar year 2014. Personally identifiable information including name and social

security number will be removed prior to data transfer. The data elements listed below

are planned for use:

Cognitive Fit Turnover Outcome

Rating Gender

Race/Ethnicity Marital Status

Age Length of Service

Paygrade Educational Level

3. The research plan is to compute cognitive fit for each sailor by comparing U.S.

Navy enlisted sailor Armed Services Vocational Aptitude Battery (ASVAB) test scores to

the cognitive demands for their career fields in the Navy. This measurement of cognitive

fit will then be compared to retention to determine if it is related to employee turnover. A

complete description of the research method is provided in enclosure (a).

R. J. SQUIER

113

5000

Ser HRPO/048

31 Mar 2016

MEMORANDUM

From: NAVSEA HQ Human Research Protection Official (HRPO)

To: CAPT Renee J. Squier, USN, NAVSEA 00

Subj: NAVSEA HQ HRPO DETERMINATION OF HUMAN SUBJECT RESEARCH

FOR PROTOCOL “Determination of cognitive fit predictors for employee

turnover using the Navy’s Career Waypoints system”

Ref: (a) DoDI 3216.02

(b) SECNAVINST 3900.39D

(d) OPNAV ltr 3900 Ser N093/15U0075 of 19 Aug 2015

Encl: (1) NAVSEA HQ Human Subject Research Determination Checklist

1. References (a) and (b) require performers engaged in research that may involve human

subjects supported by a Federal agency to submit pertinent documentation for a

determination of research prior to commencement of such research. CAPT Squier, the

performing entity, submitted the following documentation: Research Method (for

Cognitive Fit Predictors study). In accordance with reference (b), a review of the protocol

and exemption determination has been completed by the HRPO.

2. Based on my review of the submitted documentation, I have determined that the

research activity is “Exempt research involving human subjects” under exemption

category 4 of 32 CFR 219.101(b).

3. By references (c) and (d), as NAVSEA HQ HRPO, I determine that the protocol and

the exemption determination appear to be in compliance with the DoD policies based

upon the review documented in Enclosure (1) and of the performer-provided

documentation. You are authorized to commence research. As principal investigator you

are informed that significant modifications to the research protocol or research materials

must be reported to the NAVSEA HQ HRPO.

4. Refer questions to Daniel F. Wallace, NAVSEA HQ Human Research Protection

Official, by phone at 540-653-8097 or by email at [email protected].

Daniel F. Wallace, PhD

Copy to:

SEA 05H - Gray

SEA 05H – Markiewicz