Research and Statistical Methods for Business Assignment

RESEARCH AND STATISTICAL METHODS FOR BUSINESS
Assignment
Student Id and Name
[Pick the date]

Question 1
A key problem that researches often face is with regards to determining the sample size
which would be required for conducting the research. It should be noteworthy that the key
objective of the sampling process is that the chosen sample should be representative of the
actual population. As a general rule, it is better that a higher sample size is selected as it
enhances the chance of the sample being representative of the population even though this
depends on the underlying sampling method deployed. However, a higher sample size leads
to a smaller standard error which captures the difference between the population parameter
and sample statistic. But a key problem associated with a large sample is the high amount of
time and resources that would be required in collecting and analysing data collected from
such large sample (Hair, et.al., 2015).
Hence, the above theoretical discussion clearly reflects that there is an inherent trade off
involved with regards to sample size whereby on one side the higher sample size can
potentially deliver higher precision and accuracy but at a higher cost while on the other side a
lower sample size would lower the accuracy but the cost involved would be lesser. In this
trade off, it is imperative that a suitable middle ground be attained. In this regards, the
following formula for minimum sample size is useful (Lind, et.al., 2012).
It is apparent from above that the minimum sample size depends on the level of accuracy
desired coupled with the level of heterogeneity in the population data. The higher the
heterogeneity level, higher would be the minimal sample required. MOE requires the margin
of error and the sample is inversely related to the square of the MOE (Flick, 2015).
For the given case, the population comprises of 69000 employees and out of these about 21%
employees or 15,000 employees have been chosen for the sample. The sample size does not
seem high to me considering that the underlying population seems to be heterogeneous
considering that in total this has employees from 63 banks and additionally different levels.
Hence, considering the number of organisations and levels involved, it makes sense to choose
a higher sample size as lower sample size can potentially lead to misrepresentation of some
banks and underrepresentation of others. Thus, these sampling errors can potentially have an

adverse impact on the reliability of these results. Hence, it would be fair to conclude that such
a large sample size would be necessary for the given study (Lind, et.al., 2012).
Question 2
The current sampling method that has been deployed in the given study is simple random
sampling. This is a sampling method where requisite samples are randomly selected and each
element of the sample has an equal chance of getting selected. Thus in order to conduct
simple random sampling, the various employees can be numbered from 1 to 63000 and then
15,000 random numbers can be generated and the corresponding employees would form the
sample. Like other sampling techniques, simple random sampling method has certain
advantages and disadvantages which have been outlined below (Hair, et.al., 2015).
Advantages of simple random sampling
1) A key advantage of simple random sampling is the ease of use which is involved. This is
especially relevant considering the given sample size of 15000 employees and thereby
convenient implementation is imperative.
2) Another key advantage is that this method is free from errors related to classification since
no classification of data is required to implement the same.
3) It usually leads to the selection of a representative sample especially if the sample size is
large enough. This representative sample through the use of inferential statistical tools can
then be yield to derive useful conclusions about the population (Lieberman, et.al., 2013).
Disadvantages of simple random sampling
1) This method is not very useful with regards to data that is heterogeneous and has
significant attributes attached. Take the example of the current research. In this case, there
are significant attributes that are involved such as bank name, level of employees, gender
etc. which need proportionate representation. However, this seems difficult in random
sampling as it may so happen that a particular bank has large representation in the sample
while other lower. Further, in case there is a bank with only limited employees such as
500, then it might be possible that no representation from the same may be available
despite the larger sample size (Flick, 2015). Thus, in this case, it is better to choose
stratified sampling method which involved sorting the population in accordance with the