Research and Statistical Method Assignment - Analysis and Design

This assignment critically examines a research study, focusing on the crucial aspects of research methodology and statistical analysis. The analysis begins by addressing the importance of sample size determination, considering factors such as population heterogeneity and desired accuracy levels, and evaluating the chosen sample size of 15,000 employees from a population of 69,000 across 63 banks. The study then investigates the simple random sampling method employed, highlighting its advantages like convenience and minimal knowledge requirements, while also acknowledging potential disadvantages like non-representativeness in heterogeneous populations. Furthermore, the assignment assesses the reliability and validity of the measurement variables, including quantitative and qualitative job insecurity, and psychological distress, identifying potential validity concerns with certain measures. Finally, the assignment explores the role of control variables like age, gender, and education level in influencing employee priorities and the overall correlational research design, discussing its advantages in data collection and theory building, and its limitations regarding cause-and-effect relationships.

RESEARCH AND STATISTICAL METHOD
STUDENT ID:
[Pick the date]

Question 1
A relevant issue that researcher faces in relation to conducting research is to decide on the
relevant sample size that is requisite for conducting the research. The key condition which a
sample should meet is that it should be representative of the underlying population so that
any conclusions that may be drawn from the same are reliable. Also, in the process of
determining sample size, a trade-off is present which needs to be detailed. On one hand, the
representativeness of the sample can be improved by choosing a large sample size but it has a
potential cost and time factor attached which needs to be taken into cognizance. The
alternative is to go for a smaller sample size which tends to save on costs and data collection
time but may compromise on the accuracy of the sample which may be biased and hence
would have an adverse impact on the results (Eriksson & Kovalainen, 2015).
In wake of this trade off, the researcher needs to consider the priority according to each of the
factors and also ascertain the level of accuracy that would be required. For instance, if the
subject of the study is such that a higher accuracy level is desired, then it makes sense to
choose a higher sample size. Yet another relevant parameter is the level of heterogeneity
present in the population. A largely homogeneous data can be well captured by a small
sample size but for a heterogeneous population, a larger sample size is recommended so that
the vagaries and deviations can be reproduced in the sample. These, factors are indicated in
the formula for minimum sample size which is captured below (Hair et. al., 2015).
The above formula validates that population having higher standard deviation would require a
greater minimum sample size. The accuracy factor is represented by MOE (Margin of Error)
which represents the acceptable tolerance level and needs to be decided by the researcher
(Flick, 2015).
In the wake of the theoretical framework presented above, the relevant details of the current
study need to be critically analysed. For the given study, the relevant population has 69,000
employees which are spread across more than 60 banks. The sample size selected for the
study is 15000 which form 21% of the given population. A key aspect which justifies the
rather large sample is that the population data would be quite heterogeneous considering that
the data corresponds to 63 banks which are divided into different levels along with differing

demographic parameters. In this case, a selection of a lower sample size could lead to a non-
representative sample selection and thus would lead to issues regarding validity and
reliability (Hillier, 2016). Hence, the given sample size seems to be fair in wake of the
heterogeneous nature of the underlying population along with the accuracy desired by the
researcher.
Question 2
The sampling method which has been used for the given study is simple random sampling
since a random sample of 15,000 has been selected from the given population with no
classification has been done for various demographic parameters. A characteristic of this
sampling method is that every element represent in the population has an equal chance of
getting selected. This has been facilitated in the given case through the use of addresses since
randomly 21% employees have been selected by each bank and personalised envelops have
been mailed by the employer to collect the data.
Advantages
1) A key advantage of this method of sampling is the underlying convenience and
requirement of minimal knowledge. This is pivotal considering the scope of the given
study which involves selection of 15,000 employees (Flick, 2015).
2) For each of the 63 banks, there has been no classification based on the key attributes and
hence classification errors would not be witnessed which could be an issue in stratified
random sampling (Hair et. al., 2015).
3) This method is quite commonly used and tends to lead to representative sample selection
owing to samples being randomly selected. This is especially true when the sample size is
large enough (Eriksson & Kovalainen, 2015).
Disadvantages
1) One of the key issues in relation with simple random sampling is that in case of
heterogeneous population, the sample may not be representative. This is prevalent for the
given study also since for neither of the banks there has been any quota for key attributes
such as employment level, gender and education level. In such a situation, it is possible
that a particular gender or employment level may have a representation in the sample

which exceeds the actual representation level in the population. Further, over-
representation of one attribute automatically implies under representation of the other
(Hillier, 2016). For example, it is likely that a particular gender could have an
overwhelmingly high representation in the sample data and hence the data obtained from
such a sample would not be representative of the population as a whole. Ann additional
concern in this sampling could be regarding representation of those attributes which has
limited presence. For instance, there may be only a few employees at the top levels and
hence there is a fair chance of these level of employees being underrepresented leading to
questioning of the validity and reliability of the study. .
2) Owing to the issues in relation to the sample being representative, it is possible that
standard error for such a sample would be higher and hence accuracy would be comprised
to some extent (Flick, 2015).
Question 3
It is pivotal that the given measure should be reliable which implies that similar data can be
obtained if the same study is replicated with different sample data. The validity on the other
hand for measurement variable implies that it should accurately capture the key parameters
under study. In the context of the given research study, the discussion regarding the variables
of measurement and their underlying reliability & validity is carried out below.
 Quantitative job insecurity– The reliability of measure can be captured using
cronbach’s alpha. Considering that the given parameter is quantitative thus, value
higher than 0.8 would be required to ensure satisfactory performance in terms of
reliability. This condition has been fulfilled in the case of given measure. Owing to
the usage of the concerned measure in a previous study based in a similar context, the
validity concerns also do not arise for the given measuring variable (Hair et. al.,
2015).
 Qualitative job insecurity – The reliability of measure can be captured using
cronbach’s alpha. Considering that the given parameter is qualitative thus, value
higher than 0.7 would be required to ensure satisfactory performance in terms of
reliability. This condition has been fulfilled in the case of given measure. Validity
concerns could potentially arise since the researcher has decided to go ahead with 10

selective measures and this selective choice can undermine the validity of the
measurement variable (Hastie, Tibshirani & Friedman, 2011).
 Psychological distress – The reliability of measure can be captured using cronbach’s
alpha. Considering that the given parameter is qualitative thus, value higher than 0.7
would be required to ensure satisfactory performance in terms of reliability. This
condition has been fulfilled in the case of given measure. But, there are validity
concerns in this case taking into consideration the fact that the questionnaire used is
about five decades old and hence may lack validity under the current environment.
Thus, it would be better that any other measurement for psychological distress should
be selected based on some recent study as it would help in alleviation of validity
concerns (Hillier, 2016).
The broad conclusion that can be derived from the discussion carried out above is that for the
measurement variables used, reliability concerns are absent but validity concerns are present
to some extent.
Question 4
The given research study aims to highlight the association level between the qualitative and
quantitative measures in relation to well-being and job security. However, it is noteworthy
that the given research does not focus only on the above variables but considers additional
demographic variables related to age, gender & education in the capacity of control variables.
This is essential as potential changes in the control variables could alter the dependent
variable and thus undermine the validity of the relationship between the independent and
dependent variable. This is because a change in any of the control variables can produce a
corresponding change in the dependent variable but it would not be on the basis of the
independent variable (Hastie, Tibshirani & Friedman, 2011). The researchers intend to
measure the changes observed in dependent variable only on account of the change made in
the independent variables and hence identify the various control variables. By keeping the
control variables identified constant, it becomes possible for the researcher to attribute that
any change in the dependent variable is on account of changes in the independent variables
(Flick, 2015).
The relationship between the various control variables and the measurement variables can be
highlighted. One of the control variables is age. This is because there is difference in job

expectations of the different age groups. For the employees who are aged and have been in
the company since long are more concerned with quantitative measures since they want
stability in the career and would not prefer to switch to another organisation. Also, the
qualitative issues for these employees would cease to be that important considering the
adaptability with organisational processes and culture over time. On the contrary, the younger
employees would be more concerned about qualitative measures since they would be more
ambitious and would not be hesitant to shift jobs for a better work environment. In
comparison, the quantitative factors would be of lesser importance in comparison with their
elder counterparts (Hillier, 2016).
The education level and gender also tend to play a significant role in influencing the priorities
of the employees. It is expected that employees which have higher educational qualifications
would value the qualitative measures more considering they would have high expectations
regarding the work environment and nature of work. Additionally, owing to their better
educational qualifications, they would not worry about quantitative measures as they would
be able to secure job elsewhere. However, for the employees lacking in any higher education
qualification, quantitative factors would be more crucial in comparison to qualitative factors.
This is because these employees would face difficulty in finding an alternate job and hence
they are willing to compromise on the attributes of work (Hair et. al., 2015). Further, gender
is also a crucial variable impacting the preferences of the employees and thus need to be
controlled for this research study.
Question 5
For the purposes of the given research, the design used would be categorised as correlational
design considering the emphasis is on using a particular sample data for understanding
association between qualitative and quantitative measures. This research design has various
advantages and limitations that have been outlined below.
Advantages
1) The given research design leads to higher collection of data as compared to studies
based on experimental design, One potential explanation of this could be the aim of
the correlation design to track the broad trend rather to find precise values and test
given hypotheses (Hastie, Tibshirani & Friedman, 2011)

2) This research design tends to provide leads for a plethora of other researches. Since
the correlational research design is concerned with understanding association and not
underlying causes, hence further descriptive and experimental research design can be
deployed Hillier, 2016).
3) Besides, correlation research design plays a vital role in theory building by providing
initial direction of potential association between two or more variables which then can
be carried forward with additional researches for building theory (Flick, 2015).
Disadvantages
1) One of the key limitations of this research design is that only association between the
variables is indicated and hence a cause and effect relation cannot be confirmed.
Additionally, even if association relationship is formed , the underlying causes for the
same cannot be explored using the given research design and it would require
assistance from other research designs (Hastie, Tibshirani & Friedman, 2011)
2) The number of variables that can be represented in this research design is rather
limited since at a given time association relationship between only two variables can
be captured (Hair et. al., 2015).

References
Eriksson, P. & Kovalainen, A. (2015) Quantitative methods in business research (3rd ed.).
London: Sage Publications.
Flick, U. (2015) Introducing research methodology: A beginner's guide to doing a research
project (4th ed.). New York: Sage Publications.
Hair, J. F., Wolfinbarger, M., Money, A. H., Samouel, P., & Page, M. J. (2015) Essentials of
business research methods (2nd ed.). New York: Routledge.
Hastie, T., Tibshirani, R. & Friedman, J. (2011) The Elements of Statistical Learning (4th
ed.). New York: Springer Publications.
Hillier, F. (2016) Introduction to Operations Research (6th ed.). New York: McGraw Hill
Publications.