METHODOLOGY ON HUMAN CAPITAL PRACTICES
BUSI 240-Organizational Behavior 1
Liberty University
2022
Research Design
A research design is used to structure the research to address the research questions. Cooper and
Schindler (2014) explained that research design is the strategy for a study and the plan by which
the strategy is to be carried out. It specifies the methods and procedures for the collection,
measurement and analysis of data. Kothari and Garg (2014) describe a research design as the
arrangement of conditions for collection and analysis of data in a manner that aims to combine
relevance to the research purpose with economy in procedure.
The study adopted descriptive research design. The descriptive research designs describe the
phenomenon. Descriptive studies are designed primarily to document what is going on or what
exists (Trochim et al., 2016). Serekan (2010) explained a good research design has a clearly
defined purpose and has consistency between the research questions and the proposed research
methods. Cooper and Schindler (2014) posits that, if the research is concerned with finding out
who, what where, when or how much, then the study is descriptive.
Research Philosophy
The study adopted positivism research paradigm since the study involves hypotheses testing and
thus seek to obtain the objective truth or reality and also predict what may happen in future. The
study took a viewer/observer approach. This is supported by Aligu, Bello, Kasim & Martin (2014)
who in their study argue that positivism could be regarded as a research strategy and approach
of the viewer and observer. The positivism paradigm of exploring social reality is based on the
philosophical ideas of the French Philosopher, August Comte (Thomas, 2010).
The study is a descriptive research. Walliman (2011) noted that since descriptive research attempts
to examine situations in order to establish what the norm is, that is, what can be predicted to happen
again under the same circumstances, positivism research paradigm is appropriate. Thomas (2010)
also stated that the observation and reason are the best means of understanding human behavior,
true knowledge is based on experience of senses and can be obtained by observation and
experiment.
In addition, Aligue et al., (2014) explains that positivist paradigm emphasizes that genuine, real
and factual happenings could be studied and observed scientifically and empirically and could as
well be elucidated by way of lucid and rational investigation and analysis. The study, therefore,
agrees by the authors views and Vermaa’s (2014) conclusion that, research philosophy is a rich and
multifaceted source for understanding, scientific theories and models and their testing.
Population of the Study
In the study, the focus of analysis was PSI and the unit of observation was the PSGs. Sekeran
(2010) defines population as the entire group of people or things of interest that the researcher
wishes to investigate, and Williman (2011) defines population is a collective term used to describe
the total quantity of cases which are the subject of your study. Therefore, population is the total
number of elements or observations that the study wishes to make some inferences on. Therefore,
the target population was 150,000 PSGs who are estimated to be in Nairobi by 2016 (KNPSWU,
2016).
Sampling Frame
The sampling frame for the study is the merged list of private security industry companies obtained
from KSIA and PSIA. The merged list of members is then arranged in alphabetical order for
ease of identification (See Appendix 4). Sampling frame is the list of elements from which the
sample is actually drawn (Cooper & Schindler, 2014). Gujarati & Porter (2010) define a sample as
a subset of the population.
Sample Size
Williman (2011) argued that no sample will be exactly representative of a population. Williman
(2011) further stated that sample is the small part of a whole population carefully selected to show
what the whole is like. Kothari and Garg (2014), stated that in case of finite population as in the
study, the below stated formula will be applicable to determine the appropriate sample size that will
be a true representative of the population.
z
2
. p.q.N
n .............................................................................................................................................3.1
e
2
(N 1) z
2
. p.q
where p = sample proportion of the characteristics of the population
q = 1 – p;
z = the value of the standard variant at a given confidence level and to be
worked out from table showing area under Normal Curve; for example, z=1.96,
if tested at 95% confidence level
n = size of sample.
e = error margin
N= target population
Z=1.96, if tested at 95% confidence level.
(i) Calculation of the study sample size
The target population is 150 000 security guards. The estimate is within 2 per cent of the true
value with 95 per cent confidence level. In this case, e(error margin) in this case is 0.02; level
of significance is 0.05 %; Z-tabulated value is 1.96.
N = 150,000
e = .02 (since the estimate should be within 2% of true value);
z = 1.96 (as per table of area under normal curve for the given confidence level of 95%).
Let me assume p to be p = .02 (This may be on the basis of my experience or on the basis of
past data).
1.96
2
0.02
1
0.02
150,000
0.02
2
150,000
1
1.96
2
0.02
1
0.02
= 188.004 =189………………………………3.2
In the study, after proportional allocation of samples, oversampling was experienced.
Oversampling was experienced from 189 to 213, which according to the study, the difference
of 24 catered for possible attrition and non-response. Oversampling simply means using a
sampling rate which is greater or generally substantially greater than Nyquist rate. The Nyquist
rate is the sampling theorem/computed sample size determined by the formulae in equation 3.2)
due to experienced imbalanced dataset. This is supported by Hernandez, Ariel & Francisco
(2013) explained that dealing with a problem of imbalanced datasets is by applying some
oversampling or under sampling techniques so as to improve the accuracy of instance selection
methods on imbalance datasets. See Appendix 4 for the sampling frame.
(ii) Calculation of sample size of the companies whose employees were
interviewed.
The total number of companies is 112 companies. The estimate is within 2 per cent of the true
value with 95 per cent confidence level. In this case, e(error margin) in this case is 0.02; level
of significance is 0.05 %; Z-tabulated value is 1.96.
N = 112
e = .02 (since the estimate should be within 2% of true value);
z = 1.96 (as per table of area under normal curve for the given confidence level of 95%).
Let me assume p to be p = .02 (This may be on the basis of my experience or on the basis of
past data).
1.96
2
0.02
1
0.02
112
0.02
2
112
1
1.96
2
0.02
1
0.02
= 71 companies.......................................................3.3
Systematic random sampling was used to obtain the sample from the population. This is because
systematic random sampling procedure provides an unbiased and efficient sampling technique, and
it is preferred when multiple sampling technique are implemented like in the case of the study. The
following procedure was followed in systematic random sampling: First, all PSIA and KSIA
companies were merged and arranged in alphabetical order. Secondly, random sampling was done
through lottery method to which number 48 was obtained from random sampling with replacement.
Number 48 is the integer that was selected. Third, the interval was determined = 112/71= 1.577 =
2. Meaning that for every 2nd company after 48, was selected. Lastly, from the list of 112
companies, 71 companies were selected systematically.
Sampling Techniques
The study adopted systematic random sampling and purposive sampling techniques. Systematic
random sampling is employed to identify the private security firms whose employees were
interviewed. According to Baran and Jones (2016), in systematic random sampling, there is an
equal chance (probability) of selecting each unit from within the population when creating the
sample frame. The study adopted systematic random sampling since it is easy to select and reduces
the potential for human bias in the selection of cases.
Baran and Jones (2016) explained that to create systematic random sampling, there are seven steps,
namely: Defining the population; Choosing sample size; Listing of the population; Assigning
numbers to cases; Calculating the sampling fraction by dividing sample (s) with the total
population size (N); Selecting the first unit from the random number table and Selecting sample.
The study adopted purposive sampling also known as judgmental, selection or subjective sampling
to focus on particular characteristics of a population that are of interest, which could best answer
the research questions since it is based on the random means (Cooper & Schindler, 2014). It is
suitable for the study since it is flexible, enables the researcher to select a sample based on purpose
of the study, the company whose employees are interviewed are known also based on knowledge of
the population. With purposive sampling it is easy to get the sample of subjects with specific
characteristics (Williman, 2011).
Data Collection Instruments
The study adopted triangulation technique in data collection since the study sought to obtain a
comprehensive understanding of the existing challenges of performance of the PSGs in PSI. Carter,
Byant, DiCenso, Blythe, and Nevile (2014), defined data triangulation as the use of multiple
methods of data sources in qualitative research to develop a comprehensive understanding of the
phenomenon. Hence, the study adopted both structured and unstructured questionnaires,
participant interview guide and key informant interview guide provided in Appendix 2 and 3
as data collection tools. Primary data was collected, cross sectional in nature.
According to Williman (2011), questionnaires are particularly suitable tool for collecting data and
enable the researcher to organize the questions and receive replies without actually having to talk to
every respondent. The study also adopted qualitative interview guide as a tool to help get in depth
information. Baker and Edwards (2012) argued that successful field research depends on the
investigator’s trained abilities to look at people, listen to them, think and feel with them and talk to
them. The study also used key informant interview guide as a data collection tool. This is
supported by Fabeil (2013), that key informant interview guide serves as s suitable way to gather
relevant information needed to address the knowledge gaps, since it provides information from
knowledgeable people in the sector.
Pilot Testing
Cooper and Schindler (2014) explained that a pilot test is conducted to detect weakness in design
and instrumentation to provide proxy data for selection of probability sample. Therefore, pilot test
draw subjects from the target population and stimulate the procedures and protocols that have been
designated for data collection (Mugenda & Mugenda, 2010). For example, if the study is a survey
to be executed by mail, the pilot should be mailed. Cooper and Schindler (2014) noted that for
most student questionnaires, the minimum number for pilot is 10 responses, although for large
surveys between 100 and 200 responses is usual. Hence this study recognizes pilot testing an
important process of carrying out a preliminary study or assessment, going through the entire
research process with a recommended small sample size.
Simon (2011) also posited that a sample size of 10-20% of the sample size for the actual study is
reasonable number of participants to consider enrolling in a pilot. The study ensured thorough pre-
test of the data collection instruments in the study is done to avoid non-responses bias by detecting
the existing questionnaire problem, revising the questionnaire, allocating maximum data collection
period, sending reminders to potential respondents and also maintaining confidentiality. Pre-test is
the small-scale trial of the data collection tools/instruments. The study considered 5
questionnaires for pre-test, which was 20% of the 22 questionnaires which was considered for the
pilot test.
Reliability of the Instrument
According to Mugenda and Mugenda (2010), reliability is using the internal consistency technique
where data is determined from scores obtained from a single test administered. Cronbach’s
Coefficient Alpha was computed to determine how items correlate among themselves. The
interpretation of the result is that a high coefficient will imply that items correlate highly among
themselves. According to Sekaran, (2010) a value of at least 0.7 is recommended. Cronbach’s
Alpha is a general form of The Kunder-Richardson (K-R)
Validity of the Instrument
According to Cooper and Schindler (2014), there are two types of validity: internal validity and
external validity. Internal validity seeks to explain if the conclusions drawn about a demonstrated
experimental relationship truly imply cause. Validity determines whether the findings are really
about what they appear to be about. Yong & Pearce (2013) stated that various techniques of
determining validity include: Expert opinion, Bartlett’s Sphericity test, exploratory and
confirmatory factor analysis. The study adopted Bartlett’s sphericity test and expert opinion. The
goal of factor analysis is to find the latent structure of the dataset by uncovering common factors
(Hooper, 2012). Table 4.3 shows the validity test results.
Pilot tests was done to help find out if the sampling methodology was right, testing technical
appropriateness of the instruments, for example, detect any questionnaire problem and establish
logistics. The study adopted Bartlett's Test to test if the items used in the structured questionnaire
to measure the various domains. The questionnaire was revised after pilot test to avoid any
questionnaire error, during the main data collection.
