Quantitative Methods Assignment 1: Research Methods in HRM

This assignment solution addresses a quantitative methods assignment (7SSMM002) focused on research methods in HRM and organizational analysis. The solution analyzes statistical output, addressing nominal and interval-level data, internal consistency, data distributions, and out-of-range scores. It examines multiple regression, including predictor variables, effect size, and beta coefficients, and also discusses Multivariate Analysis of Co-Variance (MANCOVA), interpreting predictors, outcome variables, and assumptions. Finally, the solution explores Structural Equation Modeling (SEM), evaluating model fit, factor loadings, path coefficients, and indirect relationships. The assignment covers topics like job involvement, organizational commitment, and job satisfaction, offering a comprehensive analysis of statistical techniques commonly used in HRM research. The student has provided answers to all the questions asked in the assignment brief, providing the word count for each section as requested.

7SSMM002 QUANTITATIVE METHODS ASSIGNMENT 1
RESEARCH METHODS IN HRM AND ORGANISATIONAL ANALYSIS7SSMM002
QUANTITATIVE METHODS ASSIGNMENT– WORTH 50% OF YOUR TOTAL MARK FOR THIS MODULE
For this assignment, make reference to the statistical output with which you have been supplied.All
of the information that you will need to complete the sections that follow is contained within the
supplied statistical output.
NOTE WELL:
(A) DO NOT EXCEED THE WORD LIMIT FOR EACH SECTION – YOU WILL BE MARKED DOWN
FOR DOING SO. STATE YOUR WORD COUNT FOR EACH SECTION.
(B) DO NOT PRINT OUT THE STATISTICAL OUTPUT BECAUSE THERE’S TOO MUCH OF IT
(C) YOU CAN INCLUDE REFERENCES. IN-TEXT REFERENCES ADD TO YOUR WORD COUNT.
THE REFERENCE LIST WILL NOT ADD TO YOUR WORD COUNT.
Nominal-Level Data
Before we conduct analyses, it is important to check our data for artifacts relevant to the statistical
tests that we will perform(Field, 2018). This can be achieved in various ways, but one approach is to
peruse frequency tables. Frequency tables are particularly relevant tonominal-level data. We can
also use frequency tables to check interval-level data, but, in this case, we will use different
approaches for our interval or scaled variables, described later.
1. Name each nominal variable in this study and briefly comment on the frequency tables
associated with each (50 words).
The nominal variables are :
Gender- from table it is found that male were 451 and female were 829
Ethnicity- the most employees were Australian that is 441, 398 were New Zealand and 138 was
European
Occupation – the table shows that 709 were full time basis and 344 were part time.
State word count:
Interval-Level Data: Internal Consistency
The scaled variables (i.e., variables that, in practice, we treat as being at the interval level) in this
study relate to variables that are psychological in nature. Oftentimes, we want to get some idea
about evidence in favour of unidimensionality for each construct. With simple measures where

7SSMM002 QUANTITATIVE METHODS ASSIGNMENT 2
items are the only facet1, this is often achieved using exploratory or confirmatory factor analysis. But
in the present case, this work has already been completed for us in previous research. So, as is
routinely assumed in applied research, it will suffice to check the internal consistency for each
measure and to assume the factor structure will hold.
2. Name each interval-level variable in this study and state if the coefficients alpha are at an
acceptable level (50 words).
Here, variables are
Positive and negative affect terms, job involvement items and organizational commitment items
It is found that cronbach’s alpha value is .800 that means that it is at acceptable level that value is
significant.
Also, in job involvement items cornbach alpha value is .905 that means it is at acceptable level.
State word count:
Interval-Level Data: Data Distributions and Out-of-Range Scores
For scaled variables, we need to check about whether our data make sense as we did when we used
frequency tables for our nominal variables. For scaled variables, depending on what is under
scrutiny, it can sometimes be convenient to look at summary statistics for the aggregated scale (i.e.,
the average of the items that make up the scale, for example [item 1 + item 2 + item]/3) as well as
the range of the scale. This will help to ensure that we have no out-of-expected-range data, that our
standard deviations are not too high relative to our mean values, and so that our mean values are
not at some unexpected value. We can also check the N for each of our variables because this can
sometimes vary as some respondents might refuse to answer specific questions. If anything seems
to be amiss, then we can look back at our data set, perhaps at the item level, and find out more
about any problems that arise.
Another consideration is the distribution shape for each of our variables. In an ideal world,
frequency distributions relating to our aggregated scales would be perfectly bell-shaped to fulfil a
common assumption in many statistical tests that variables should be normally distributed. One way
to address this is by looking at histograms of our aggregated scales. We cannot expect perfection
with real world data, but we can at least hope for a distribution that is not cataclysmically different
from normal. Bear in mind that many or most statistical tests are robust against the assumption of
normality.Also some tests assume multivariate normality, for which specific tests have been
1 A facet, in this context, is a source of variability (e.g., items) that does not include the object of measurement
(i.e., the people being measured, see Brennan, 2001).

7SSMM002 QUANTITATIVE METHODS ASSIGNMENT 3
developed (e.g., Mardia’s coefficient). However, it is rare that applied data meet the assumption of
multivariate normality (Byrne, 2010).
3. Comment on the descriptive statistics and distribution shape of the Job Involvement
variableONLY (50 words).
By analyzing the graph it is stated that mean of job involvement is 2.54 and standard deviation
is .803. besides that, minimum value is 1 and maximum is 5. Thus, it means that most employees are
involved as fixed term in organization. The SD is .803 that is high so it indicate that values are
spread.
The distribution shape of job involvement is bell curved which means that it is unimodal
distribution.
State word count:
Multiple Regression
Often our choice of statistical technique is based on how our original variables were measured.
Multiple regression is routinely used when the variables under scrutiny are at the interval-
level(Spicer, 2005). Nominal variables can be included in multiple regression analyses, but these
need to be dummy-coded first. In the present case, all of the variables included in the multiple
regression output are at the interval-level.
4.Name the variables that are included as predictors in the multiple regression output. How are
these predictors arranged into blocks? Name the outcome variable in this analysis. Comment on
the effect size and significance of the models relevant to the multiple regression. Comment on the
beta coefficients associated with specific predictors and what they mean. Comment very briefly
on the residual plots associated with this analysis. (150 words)
In this the predictor variable is job involvement and organization commitment. They are arranged as
positive and negative affect, org commitment and job involvement.
The effect size in regression is standardized mean difference. And significance is constant.

7SSMM002 QUANTITATIVE METHODS ASSIGNMENT 4
The outcome variable in analysis is job satisfaction.
The beta coefficients of positive and negative affect is .258 and of job involvement is .220 and of
organization commitment is .488. it means they all are positive so outcome will increase by beta
coefficient value.
It is found that residual plots is values are scar rated which means that there is problem associated
with regression. It has to be changed so that proper value of analysis is identified. Besides that, it
might be due to predictor variables taken in doing it or missing values of data. this has led to scatter
diagram of plots in graph.
State word count:
Multivariate Analysis of Co-Variance (MANCOVA)
Sometimes researchers are interested in group differences (e.g., differences between people from
different countries, different genders, etc) and how they related to variables measured on a scale.
We might also be interested in controlling for variables that could interfere with the results of the
study (e.g., age, affect, etc). These types of variables are often referred to as covariates.
Multivariate analysis of co-variance (aka MANCOVA or simply multivariate analysis of variance
[MANOVA] with the addition of a co-variate) can help us to address questions that involve data
configurations akin to that described above(Field, 2018).
5. Name the variables that are included as predictors in the MANCOVA output . Name the
outcome variables in this analysis. Name the co-variate in this analysis if there is one. Comment
briefly on the assumptions of this analysis and whether they have been met. Comment briefly on
the multivariate and univariate effect size and significance of the models relevant to the
MANCOVA, the profile plots, and what they mean.(150 words)
In this predictors are intercept, panas, naq bullied, emphasis , naq bullied* emphasis.
The outcome variables are pillai’s trace, wilks lambda , hotelling trace and roy largest root.

7SSMM002 QUANTITATIVE METHODS ASSIGNMENT 5
It has been evaluated that assumptions made in analysis is not met as the significance value
obtained is P= .000 that is less than P= 0.05. so, there is no relationship between independent and
dependent variables.
It is stated that effect size of variable on intercept, panas, etc. is not having relationship with two
variables on numeric scale.
The profile plot state that organization commitment depends on variables that are bullied and not
bullied. Also, on basis of job involvement as well organization commitment changes. In case of
bullied, job involvement is affected even when employee work full time.
State word count:
Structural Equation Modeling
Structural equation modeling (SEM) can allow us some flexibility, often if we’re interested in looking
at relationships between multiple, scaled variables. There are several advantages associated with
this technique, in that, unlike multiple regression, it takes simple measurement models (i.e., usually
internal consistency) into consideration when estimating model parameters and it allows for the
presence of multiple predictors and outcomes. There is a lot of output associated with SEM but we
are only interested in a small portion of it in the current case. Open the output with the file name
“Standardised Coefficients for Structural Model (Graphic)” . This provides a graphic of the model of
interest along with standardized factor loadings and standardized path coefficients. Depicted is a
proposed partially mediated model(Finch & French, 2015), where there is a hypothesized direct
effect between organisational commitment (OrgCt) and job satisfaction (JobSat). But there is also a
hypothesized indirect effect between OrgCt and JobSat through job involvement (JobInv). The
graphic doesn’t tell us about the indirect effect, but the Amos output does. Note that we are
assuming here that confirmatory factor analyses have been run for each construct independently
and the outcomes were admissible. Also, we could test other models (e.g., direct links only, fully
mediated model) but we won’t in the interests of brevity.Please do NOT print all of the Amos
output! There are 24 pages of it and most of it we won’t even use. But do refer to it in response to
the guidance below.
6. Describe the fit of the model and whether there is any indication of Heywood cases. Make a
general comment on the factor loadings associated with the model and comment on any relevant
path coefficients and their statistical significance. State briefly if there is any evidence for an
indirect relationship between OrgCt and JobSat in the Amos output. (150 words)

7SSMM002 QUANTITATIVE METHODS ASSIGNMENT 6
It is evaluated that the model is fit in this because it will show is testing hypothesis between casual
variables. This means organization commitment and job satisfaction relationship will be analysed
from it.
The model is beneficial as allow in considering all factors such as organization involvement. With
that it can be identified whether there is relationship with job satisfaction or not. Besides that, how
it impact on job involvement. This will help in deriving statistically significance value of all variables
associated in it.
Moreover, it is found that there is indirect relationship between organization commitment and job
satisfaction. It is because job satisfaction is dependent on organization commitment. So, job
satisfaction will be impacted by job involvement. It will impact due to decrease in job involvement.
State word count:
References
Brennan, R. L. (2001). Generalizability theory. New York: Springer Verlag.
Byrne, B. M. (2010). Structural equation modeling with AMOS: Basic concepts, applications, and
programming (2nd ed.). New York: Routledge.
Field, A. (2018). Discovering statistics using IBM SPSS statistics (5th ed.). Thousand Oaks: Sage.
Finch, W. H., Jr., & French, B. F. (2015). Latent variable modeling with R. New York, NY, US:
Routledge/Taylor & Francis Group.
Spicer, J. (2005). Making sense of multivariate data analysis. London: Sage.