A Report on Statistical Analysis: Job Satisfaction & Performance

This report provides a detailed review of the statistical analysis employed in a research paper focusing on the impact of employee job satisfaction on organizational performance within the hotel industry. The original study utilized ANOVA (F-test), Chi-square tests, and t-tests, selected based on the nature of the variables under examination. The report explains the rationale behind using each test, such as Chi-square for nominal data and ANOVA for comparing means across multiple categories. It also discusses the use of Bartlett’s test of sphericity to assess variable associations. Furthermore, the report suggests alternative statistical methods, including the Z test for large sample sizes, the Wilcoxon-Mann-Whitney test for non-normally distributed data, and ad hoc tests like the least significant difference test to refine ANOVA results. Levene’s test is proposed as an alternative to Bartlett’s test. The report concludes by advocating for the use of descriptive statistics like mean and standard deviation to provide a clearer understanding of participant responses.

Running head: Statistics 1
Statistics
by
Course:
Tutor:
University:
Department:
Date:

Statistics 2
Part A
The research by Deshpande, Arekar, Sharma, and Somaiya (2012) was aimed at examining
the effect of employee job satisfaction on the performance of an organization with a specific
focus on the hotel industry. The statistical analysis approaches used include ANOVA (F-test),
Chi-square test, and t-test. The three statistical analysis methods were chosen because of the
multiple differences in the variables under study. For instance, a Chi-square test was to be
undertaken in instances where the data types of both variables were nominal. Analysis of
variance (ANOVA) was to be used in situations where only a single data set was nominal and
the other one is of a ratio in nature i.e. more than two categories exist. Moreover, an
independent t-test was to be carried out if one data set was ratio or interval (on condition that
only two categories exist) in nature and the other one nominal. Additionally, the overall
satisfaction levels were measured using the mean and median values.
The Bartlett’s test of sphericity was used to ascertain the strength of the association between
the study variables by making use of an approximate Chi-square value. More specifically, the
approach was used to test the null hypothesis that there was no correlation in the study
variables.
The analysis of variance was used to test the difference in the means of the six computed
variables for job satisfaction which includes role clarity, financial benefits, work stress, work
environment, workers relations and welfare. An F-test was used to compare equality of two
different variances. The decision to reject or accept the null hypothesis was based on the
computed and statistic value of F. If the computed F value was larger than the F statistic, then
the null hypothesis was rejected and vice versa. More specifically, ANOVA was used to test
the association between experience, salary, and department with job satisfaction.

Statistics 3
An independent sample t-test was conducted to ascertain the overall satisfaction of employees
based on gender. The t-test was to test whether there was a significant variation between the
mean responses of the male and female.
Part B
The Z test could be used to test the hypothesis that overall employee satisfaction was based
on gender. This is because the Z test is more suitable where the sample size is large (n is
greater than 30), the population variance is computable or provided, and the population and
sample variance is assumed to be equal. In this instance, the sample population was 72 and
the standard variation can be calculated. The Wilcoxon-Mann-Whitney test can also be used
as an alternative to the two-sample t-test especially when there is doubt as to whether the
study variables are normally distributed in the two groups. The Wilcoxon-Mann-Whitney test
takes all the observations from the two groups understudy and arranges them on the basis of
their size (without adhering to group membership) and the sums them up.
An ad hoc test such as the least significant difference test can also be conducted in addition to
one way ANOVA. The study has used one way ANOVA to compare two means from two
independent and unrelated groups (such as job satisfaction and salary, job satisfaction and
work experience, and job satisfaction and department). An F distribution has been used to test
whether the two means are equal, and the test returned a significant f-statistic. This however
only shows the difference in the means but does not tell which specific group had a difference
in means. Thus, an ad hoc test such as the least significant difference test can further be
carried out to find out exactly which group had a difference in means.
Alternatively, Levene’s test could be used instead of Bartlett’s test in F-statistic. Levene’s
test also tests whether the variances in the group(s) are equal, and also examines the

Statistics 4
assumption of equality of the variances of the populations from which the samples are
obtained (Corder & Foreman, 2014).
Other descriptive statistical analysis methods such as mean and standard deviation can be
used to show the participants responses on each variable. This would provide insight into the
general view of the participants as pertains to the research variables.

Statistics 5
References
Corder, G. W., & Foreman, D. I. (2014). Nonparametric statistics: A step-by-step approach
(2nd ed.). New Jersey, NJ: John Wiley & Sons.
Deshpande, B., Arekar, K., Sharma, R., & Somaiya, S. (2012, January). Effect of employee
satisfaction on organization performance: An empirical study in hotel industry.
In Ninth AIMS International Conference on Management held at Pune, India, 9(1),
671-623.