CHAPTER ONE

INTRODUCTION

1.1 BACKGROUND OF THE STUDY

Every employee has to miss work once in a while; on occasion, personal things need to be taken
care of during work hours. And while it’s not a requirement (other than for conditions mandated
by the Family and Medical leave act related to family and health, or for jury duty) that
companies provide time off from work (paid or unpaid), most do have policies that grant
employees some time for excused absences.

But when those days off are too frequent, it becomes a problem and is called absenteeism. And
that could be a sign that your organization and work environment need some adjustments before
the activity starts impacting the availability and productivity of your workforce, along with the
company’s profitability.

Reasons for taking time off from work can be varied, but they generally fit into three categories:
approved absences, occasional absences, and chronic absenteeism. And while it would be
difficult to compile a comprehensive list of all the reasons employees miss work, let’s dive a
little deeper into these three categories.

(i) Approved Absences: If an employee asks for and is given permission to be absent
from work, this is an approved absence. Legitimate reasons for this type of absence
include earned vacations, holidays, maternity or paternity leave, long-term medical
leave, jury duty, and anything that needs to be taken care of during work hours that
cannot be scheduled outside of them.
(ii) Occasional Employee Absence: In addition to approved absences from work, there
will be times when an employee needs time off that wasn’t approved in advance. Life
happens, and not everything can be planned for ahead of time. Examples of typical
occasional absences from work can include sick days, childcare issues, bereavement
for a family member or friend, legal issues involving the court, and the age-old
standard of car trouble. These are genuinely occasional in that employees don’t abuse

1
 the availability of this time off and utilize them only when necessary. Companies
(should) plan for workers to need unplanned time off, on occasion.

(iii) Chronic Absenteeism: Chronic employee absenteeism refers to times when an

employee is out regularly without permission from their employer. While the two
categories above are manageable (most of the time) from an employer’s standpoint,
chronic absenteeism is not as it disrupts the business’s day-to-day operations. From
corporate profits to the morale of other employees, having workers who are
constantly MIA creates a headache for everyone.

If an employees is disengaged, call in sick all the time, show up late more often than not, leave
early every day, or take extra-long lunch breaks (to name just a few examples of chronic
absenteeism), then that institution or company has a serious issue on its hand that needs to be
dealt with.

Home and family responsibilities are among the top ten causes of long-term absenteeism and
among the top five causes of short-term absenteeism (Cameron, 2015), but other things usually
cause chronic absenteeism. And while discontent can feed discontent (absenteeism equals more
work for others equals even more absenteeism), there are usually some primary causes of
employee absenteeism in the first place:

(i) Low Employee Engagement: When employees don’t feel valued, it’s hard for them to
feel engaged at work. Why should they care about the company if the company
doesn’t care about them? Treating employees with respect and providing helpful
feedback, when necessary, makes for a healthy work environment.

(ii) Lack of a Flexible work Schedule: While much of the workforce went remote at the
beginning of the Covid-19 pandemic, many companies are starting to bring
employees back into the office. So after nearly two years of learning to schedule their
work/life activities from home, being asked to return to more stringent work hours
(and commuting) could turn employees against the “old way” of doing business. They

2
 were trusted to manage their time and get their job done on their own, and going back
to an inflexible work schedule won’t be boosting morale or engagement anytime
soon.

(iii) Mental Health Issues: The mental well-being of its employees has become a
priority for many companies in the last several years, even more so since the
beginning of the Covid-19 pandemic. According to (MMH, 2021), “51% of
employees reported worse mental health at work since the pandemic began, and 30%
of employees were scared to disclose mental health issues for fear of being fired or
furloughed.” One study demonstrated that depression alone is thought to cost
companies around $44 Billion in lost productivity per year, and about 40% of
employees said their company did not provide adequate policies or procedures to
address health and well-being during the COVID-19 pandemic. Anxiety, depression,
or other mental health illnesses can often lead to an employee feeling unwell enough
to miss work quite often.

(iv) The Equal Employment Opportunity Commission defines harassment as “unwelcome

conduct based on race, colour, sex (including pregnancy), national origin, religion,
age, disability, or genetic information,” and sexual harassment is the most prevalent
form of harassment in the workplace. A 2018 Hiscox Workplace Harassment Study
found that 35% of workers feel they have been harassed at work and that among
women, the figure is even higher at 41%. In addition, between 2010 to 2017,
employers paid out nearly $1 billion to settle harassment charges. No one wants to
work in a hostile work environment. Still, often employees have no option other than
to keep their job due to external circumstances, which can lead to chronic
absenteeism as they try to avoid harassment.

(v) Poor Leadership: An employee could be taking time off work to escape a bad boss,
too. Whether it’s too much micromanaging or not managing at all, poor leadership is

3
 often a cause of absenteeism. Research shows that it takes about 22 months for a
former employee’s stress level to return to a healthy range after a negative
management experience. And according to Gallop (2018), U.S. businesses lose $360
billion each year in lost productivity from employees who are unhappy with their
managers.

Multivariate analysis is based in observation and analysis of more than one statistical outcome
variable at a time. In design and analysis, the technique is used to perform trade studies across
multiple dimensions while taking into account the effects of all variables on the responses of
interest. The development of multivariate methods emerged to analyse large databases and
increasingly complex data. Since the best way to represent the knowledge of reality is the
modelling, we should use multivariate statistical methods. Multivariate methods are designed to
simultaneously analyse data sets, i.e., the analysis of different variables for each person or object
studied. Keep in mind at all times that all variables must be treated accurately reflect the reality
of the problem addressed. There are different types of multivariate analysis and each one should
be employed according to the type of variables to analyse: dependent, interdependence and
structural methods. In conclusion, multivariate methods are ideal for the analysis of large data
sets and to find the cause and affect relationships between variables; there is a wide range of
analysis types that we can use.

There are two determining factors that we have to take into account when doing a multivariate
approach:

(I) the multidimensional nature of the data matrix and

(II) The purpose of trying it, preserving its complex structure. This is based on the belief
that the variables are interrelated, so that only the set of the same test may provide a
better understanding of the studied object obtaining information univariate and
bivariate statistical methods are unable to achieve. The joint treatment of the variables
will faithfully reflect the reality of the problem addressed.

Multivariate methods can be classified based on two types of variables:

(i) According to the methods of dependency: Analysed variables are divided into two
groups: dependent and independent variables. The aim is to determine whether the set

4
 of independent variables affects all dependent variables and how. They develop a
hypothesis that attempts to validate empirically are explanatory or predictive
techniques.

(ii) According to the methods of interdependence: It is based to do a reality approach

without specific hypotheses and try to describe reality by synthesizing the relevant
information; they are descriptive or reductive techniques. They can be classified into
two groups based on whether the data analysed are metric or non-metric.

There are steps to perform a multivariate analysis and it can be summarized in:

(i) State the objectives of the analysis. Define problem in its conceptual terms, objectives
and multivariate techniques that are going to be employed.
(ii) Design analysis. To determine the sample size and estimation techniques those are
going to be employed.
(iii) Decide what to do with the missing data.
(iv) Perform the analysis. Identify outliers and influential observations whose influence
on the estimates and goodness of fit should be analysed.
(v) Interpret the results. These interpretations can lead to redefine the variables or the
model which can return back to steps (III) and (IV).
(vi) Validate the results. At this point, we must establish the validity of the results
obtained by analysing other results obtained within the sample and it is generalized to
the population from which it comes.

1.2 STATEMENT OF THE PROBLEM

Research has showed how people have been using different Multivariate techniques like the
Discriminant Analysis, Hotelling’s T-square etc. to reduce dimensionality in a dataset. But the
problem is most multivariate methods are not suitable for data reduction. The only multivariate
technique that is suitable for dimensionality reduction is Factor Analysis.

Factor analysis is a statistical method used to describe variability among observed, correlated
variables in terms of a potentially lower number of unobserved variables called factors. Which
makes it the perfect Multivariate technique for this dataset.

5
1.3 AIM AND OBJECTIVES OF THE STUDY

The main aim of this research work is to apply Factor Analysis on the factors influencing
Absenteeism to work.

The specific Objectives are

(i) To implement Factor analysis to reduce the total number of variables to a smaller
number of common factors
(ii) To employ Factor Analysis to extract the factors that has eigenvalues greater than one
(iii) To implement Factor Analysis to determine the major factor(s) that influence
Absenteeism to work.

1.4. RESEARCH QUESTIONS

Some research questions that could be addressed through Factor analysis on the factors

influencing absenteeism to work include:

(i) What are the major factors that influence absenteeism to work?

(ii) What interventions or policies have been effective in reducing the rate of absenteeism

to work and how can these interventions be tailored to specific populations or

regions?

1.5 SIGNIFICANCE OF THE STUDY

Attendance at workplace is very important to reaching key milestones in a company, promotion

to a more demanding roles and having access to opportunities. Studies shows that workers who

are chronically absent (miss 18 days or more in a year for any reason) are less likely to be

effective in their workplace. This study will help us identify the factors that influence

6
absenteeism to work, there correlation and possible recommendations on how to tackle these

factors.

1.6. SCOPE OF STUDY

The scope of this study will be from a courier company in Brazil from July 2007 to July 2010

with records of absenteeism at the workplace. Factor analysis will be performed on the data to

enable us to discern underlying factors that describe the data. This underlying factor will lead to

dimension reduction of the 16 (From Transportation, Distance from Residence, Service time,

Age, etc.) variables in the data.

Overall, the scope of study, on the application on the factors that influence absenteeism is wide

and may involve collaboration between experts from different fields, including public health,

transportation, and statistics. The aim of these studies is to identify effective interventions that

can reduce the factors that influence absenteeism to work and improve punctuality to your

workplace.

1.7 DEFINITION OF TERMS

(i) Factor: A Factor is a circumstance, fact, or influence that contributes to a result or a

number or quantity that when multiplied with another produces a given number or
expression.

(ii) Factor Analysis: Factor analysis is a statistical method used to describe variability
among observed, correlated variables in terms of a potentially lower number of
unobserved variables called factors. For example, it is possible that variations in six
observed variables mainly reflect the variations in two unobserved variables.

7
(iii) Dimensionality: This is the quality of having many different features or qualities,

especially in a way that makes something seem real, rather than being too simple.

(iv) Variable: A variable is a factor that can change in quality, quantity, or size, which

you have to take into account in a situation.

(v) Multivariate Analysis: Multivariate analysis is a set of techniques used for analysis of
data sets that contain more than one variable, and the techniques are especially
valuable when working with correlated variables.

(vi) Eigenvalues: In linear algebra, an eigenvector or characteristic vector of a linear

transformation is a nonzero vector that changes at most by a scalar factor when that

linear transformation is applied to it.

(vii) Absenteeism: Absenteeism is any failure to report for or remain at work as scheduled,
regardless of the reason. Absenteeism is usually unplanned, for example, when
someone falls ill, but can also be planned, for example during a strike or wilful
absence.

(viii) Work: Work is something produced or accomplished by effort, exertion, or exercise

of skill.

(ix) Workplace: A workplace is a place (such as an office, shop, or factory) where people
work.

8
(x) Principal component analysis (PCA): This is a statistical procedure that uses an

orthogonal transformation to convert a set of observations of possible variables into a

set of observations of possibly correlated variables into a set of values of linearly

uncorrelated variables, called principal components.

(xi) Level of significance: This is the measurement of the statistical significance. It

defines whether the null hypothesis is assumed to be accepted or rejected. It is

expected to identify if the result is statistically significant for the null hypothesis to be

false or rejected.

(xii) Factor loadings: These are correlation coefficients between observed variables and

latent common factors. Factor loadings can also be viewed as standardized regression

coefficients, or regression weights.

(xiii) Influence: Influence refers to the power to change or affect someone or something-

especially the power to cause changes without directly forcing those changes to

happen. Influence can also refer to a person or thing that affects someone or

something in an important way.

(xiv) Variance: The variance is a measure of variability. It is calculated by taking the

average of squared deviations from the mean. Variance tells you the degree of spread

in your data set.

(xv) Communalities: Communalities indicate the amount of variance in each variable that
is accounted for.

9
 CHAPTER TWO: LITERATURE REVIEW

2.1 Introduction

This chapter discusses published information in multivariate analysis, different weed control
methods and some other information in this particular subject area within a certain time period.
This Chapter will just be a simple summary of the sources, but it will also contain an
organizational pattern that combines both summary and synthesis.

2.2 History of Multivariate Analysis

The origin of Multivariate normal distribution can be traced to the writings of Gauss, Bravais,
Shols, Galton and Edgeworth during the 19th century. But the key figure whose quest for
knowledge on laws of heredity triggered off research on the theory and applications of the
multivariate normal distribution is Francis Galton in a lecture delivered at the Royal
Anthropological institute in 1885, Galton presented his data on heights of parents and adult
children in the form of a bi variate frequency chart.

In 1928, Wishart presented his paper ’The Precise distribution of the sample covariance
matrix of the multivariate normal population’, which is the initiation of MVA. Fischer,
Hotelling, Roy, and Xu et al. (1930) made a lot of fundamental theoretical work on multivariate
analysis. At that time, it was widely used in the fields of psychology, education, and biology. In
the middle of the 1950s, with the appearance and expansion of computers, multivariate analysis
began to play a big role in geological, meteorological. Medical and Social and Science. From
then on, new theories and new methods were proposed and tested constantly by practice and at
the same time, more application fields were exploited. With the aids of modern computers, we
can apply the methodology of multivariate analysis to do rather complex statistical analysis.

10
2.3 Application of Factor Analysis

Charles (2002) pioneered the use of Factor Analysis in the field of psychology and is sometimes
credited with the invention of factor analysis. He discovered that school children's scores on a wide
variety of seemingly unrelated subjects were positively correlated, which led him to postulate that a
general mental ability, or underlies and shapes human cognitive performance. His postulate now
enjoys broad support in the field of intelligence research, where it is known as the g theory.

Stephenson (2005), a student of Spearman, distinguishes R Factor Analysis, oriented toward the
study of inter-individual differences, and Q Factor Analysis oriented toward subjective intra-
individual differences.

Cattell (2001) expanded on Spearman's idea of a two-factor theory of intelligence after performing
his own tests and factor analysis. He used a multi-factor theory to explain intelligence. Cattell's
theory addressed alternative factors in intellectual development, including motivation and
psychology. Cattell also developed several mathematical methods for adjusting psychometric
graphs, such as his "scree" test and similarity coefficients. His research led to the development of his
theory of fluid and crystallized intelligence, as well as his 16 Personality Factors theory of
personality. Cattell was a strong advocate of Factor Analysis and psychometrics. He believed that all
theory should be derived from research, which supports the continued use of empirical observation
and objective testing to study human intelligence.

Factor analysis is used to identify "factors" that explain a variety of results on different tests. For
example, intelligence research found that people who get a high score on a test of verbal ability are
also good on other tests that require verbal abilities. Researchers explained this by using factor
analysis to isolate one factor, often called crystallized intelligence or verbal intelligence, which
represents the degree to which someone is able to solve problems involving verbal skills. Factor
analysis in psychology, is most often associated with intelligence research. However, it also has
been used to find factors in a broad range of domains such as personality, attitudes, beliefs, etc. It is
linked to psychometrics, as it can assess the validity of an instrument by finding if the instrument
indeed measures the postulated factors.
Rayerson (2004) the safest approach to creating a portfolio is to diversify stocks. A
balance of risk levels protects the investment in case adverse conditions alter one market.

11
Investment professionals use Factor Analysis to anticipate movement in a variety of industries. It
may provide clues that would otherwise go unnoticed. If one portfolio holds stocks in both
commodities and technology, a sudden increase in the price of a related variable, oil for example,
may not seem to factor into the equation without proper analysis.

Taylor (2010) many factors influence the staffing of a company. Through statistical interpretation,
human resource professionals can create a balanced environment. A staffer might combine different
variables together to determine if a company can benefit from fewer contractors and more in-house
talent. Testing allows proper screening of employees using Factor Analysis. Market research and
analysis can be the key to getting the best fit in graduates each year.

Kiovisto (2014) Insurance companies rely on actuarial tables and statistics to create policies. Florida
is prone to hurricanes. Certain data may show that drivers between the ages of 25-40 handle stress
and emergencies better than any other age groups. Based on that fact, automotive policyholders in
Florida who fall within that range may get discounts on coverage. Studying variables is the only way
insurance companies can make decisions regarding deductibles, rates and available plans.

Zahari (2013) asserts that even industries that seem less obvious need to focus on market research
and analysis to survive. Restaurants take demographics and target customers into account when
creating menus. The sweet shop next to a university is going to plan differently than the family
restaurant in a tourist area for menu items and advertising. From competitors to the ethnicity
breakdown of a community, data collection allows for cost-effective planning and a successful.

Pohlmann (2004) observed that education uses this technique in decision-making processes. A
school council looks at classroom sizes and testing results to set salary and staffing limits for
teachers. Data analysis goes into determining the curriculum each year for education from grade
school to graduate programs. Even industries that require continuing education will look at different
factors in creating options.

Ocal (2007) explains that Factor analysis plays a role in most industries. Through statistical
planning, companies can make better choices for everything from multi-channel marketing to
inventory control. Data is a powerful tool and factor analysis uses it to get results.

Prasetya (2018) discovered that Factor Analysis has also been widely used in physical sciences such
as geochemistry, hydrochemistry, astrophysics and cosmology, as well as biological sciences such

12
as ecology,molecular biology,neuroscience,and biochemistry,In groundwater
qualitymanagement, it is important to relate the spatial distribution of different chemical
parameters to different possible sources, which have different chemical signatures. For example, a
sulphide mine is likely to be associated with high levels of acidity, dissolved sulphates and transition
metals. These signatures can be identified as factors through R-mode factor analysis, and the
location of possible sources can be suggested by contouring the factor scores. In geochemistry,
different factors can correspond to different mineral associations, and thus to mineralization.

Xu (2013) Factor analysis is used for summarizing high-density oligonucleotide DNA

microarrays data at probe level for Affymetrix Gene Chips. In this case, the latent variable
corresponds to the RNA concentration in a sample.

2.3.1 Rules for Retaining Components

According to Kellow (2006), in the initial extraction process, FA will drive as many components as
the number of measured variables. After the initial components are extracted, the analyst must
decide on how many components should be retained to meaningful represent the original correlation
matrix. The initial component eigen values, percent of variance accounted for and cumulative
variance accounted for. According to Steves, probably the most widely used criterion is that of
Kaiser (1970): retain only those components whose eigenvalues are greater than one (1992). This is
the default option in many statistical packages (e.g., SPSS). Other method for retaining factors,
however maybe more defensible and perhaps meaningful in interpreting the data. Indeed, after
reviewing empirical findings on its utility, (Preacher and McCallum, 2003) report that “the general
conclusion is that there is little justification for using the Kaiser criterion to decide how many factors
to retain”. One reasonable alternative to Kaiser rule is Cattell’s (1966) scree test, which provides a
graphical representation of the eigen values relative to their magnitude (this option is available in
most major statistical packages). The basic idea is to plot eigenvalues on the ordinate (Yaxis) of a
bivariate scatter with order of magnitude represented on the abscissa (X axis). Then, a visual
inspection of the scree plot is under taken to identify a point at which an inflection occurs that
signifies a flattening of this line of best fit. eigenvalues that occur before the first value that signifies
a flattening are then retained (Steven, 1992).

13
A fairly common technique noted in the literature (Kellow, 2004) combines the two approaches.
Eigenvalues greater than one are initially retained and the scree plot test issued subsequently to
assess the tenability of the model. Because eigenvalues represent reproduced variance, this is
equivalent. The second stage, evaluate the parsimony of the solution relative to the contribution of
each component to reproducing the original variance in the data. A potential disadvantage of this
approach is the arbitrary criterion of retaining eigenvalues greater than one in the first stage.
Because FA studies typically rely on sample data. Eigenvalues (Reproduced variance) should be
expected to change (even with large samples) slightly from sample to sample. In addition, the
interpretation of what constitutes a “Meaningful” amount of variance accounted for (which
eigenvalues represent) is inherently subjective (Thompson, 2002).

2.3.2 Rotation Strategies

Once an appropriate number of components have been determined, the analyst charged with the task
of interpreting the components. This process often is facilitated by geometrically rotating the factors
to obtain a sharper conceptual solution. Because the starting point for locating factors in geometric
space is arbitrary, rotating the factors does not change the overall variance explained by the
components, although the eigenvalues associated with the respective components are not necessarily
the same as the unrotated solution (Thompson, 1996). Two methods of rotation are available:

(i) Orthogonal
(ii) Oblique

Orthogonal rotation constrains the obtained solution such that the obtained factors are uncorrelated.
The overwhelming choice of analysts who opt for an orthogonal solution is the varimax procedure,
which is the default option in most popular statistical packages (Kellow, 2004). For various reasons
(Tabachnick and Fidell, 2001) varimax is generally an excellent choice if one prefers an orthogonal
solution, although other options are available.

In contrast to orthogonal solutions, oblique rotation solutions allow factors to be correlated. At

times, the quest for simple structure is inhibited by the assumption of uncorrelated factors.
“Typically this is indicated by variables having coefficients that are large in absolute value on two or
more factors (which is sometimes called multivocal vs univocal) (Thompson, 2004). The use of an

14
oblique solution, such as oblimin or promax (Tabachnick and Fidell, 2001) often best describes the
reality of constructs being investigated. Rarely does one assume that multidimensional constructs,
such as school climate, are composed of dimensions that are completely independent of one another.
Most statistical programs will provide an estimate of the correlation between components when an
oblique rotation is requested. Tabachnick and Fidell recommend performing such an analysis and
examining the correlation for values.

In order to interpret Factor analysis, one must consult the correlations between variables and
components, often referred to as “loadings” as noted by Thompson (1996), these coefficients are
merely “weights” assigned to variables to indicate their importance. However, these obscures are
very important difference between these values when oblique as opposed to orthogonal rotational
strategies are used.

If an orthogonal rotation is used, the correlation between a variable and a component represents the
total contribution of the variable to the respective component (called a structure coefficient). In the
use of orthogonal rotation, the components will be uncorrelated and the structure coefficients and
pattern coefficients will be identical. In contrast, when an oblique rotation is employed, the
correlations coefficient associated with a particular variable and a component indicates the unique
contribution of that variable to the component after partial ling out the variance attributable to the
variables covariance with other components (called a pattern coefficient) (Tabachnick and Fidell,
2001). This is analogous to regression analysis, where the beta ( β ) weights indicate the contribution
of individual predictors in “explaining” the criterion variable. If the predictors are perfectly
uncorrelated, these weights indicate both the total and unique contribution of a predictor variable.
However, when the individual predictor variables are correlated with one another which is usually
the case, the weights indicate the unique contribution of the variable to explaining the criterion in
the presence of other predictors.

As noted by Thompson (2004), “Persons first learning of rotation are often squeamish about the
ethics of this procedure”. It should be stressed however, that component rotation simply expresses
the data in a different dimensional space. The wise analyst would do well to go beyond default
settings by exploring both orthogonal and oblique rotation strategies.

Evaluation analyst are encouraged to explore a variety of option at each of the FA process, and to
allow informed judgement to guide the process rather than strict, arbitrary criteria. There are

15
additional options that are infrequently used because they are not readily available in most packages.
For instance, some have suggested a promising variant of the scree plot in which standard errors are
computed to supplement interpretation of the number of components to retain (Nasser, 2002). The
FA to obtain composite scores is a valuable tool when dealing with correlated variables. In addition,
the use of component scores rather than a large number of individual variables is better given the
fact that, all other things being equal, using fewer predictors (in regression case) makes for a more
powerful analysis.

2.4 ABSENTEEISM

“Absenteeism is the practice or habit of being an absentee and an absentee is one who habitually
stays away from work.” According to Labour Bureau of Shimla: Absenteeism is defined as the total
man shifts lost because of absence as percentage of total number of men shifts scheduled to work in
other words, it signifies the absence of an employee from work when he is scheduled to be at work.
Any employee may stay away from work if he has taken leave to which he is entitled or on ground
of sickness or some accident or without any previous sanction of leave. Thus, absence may be
authorized or unauthorized, wilful or caused by circumstances beyond one’s control. Maybe even
worse than absenteeism, it is obvious that people such as malingerers and those unwilling to play
their part in the workplace can also have a decidedly negative impact. On this problem various
studies and researches have been carried out and some of the prominent researches and analysis are
mentioned here. Akyeampong (2007) wrote a research paper Trends and seasonality in Absenteeism.
In this paper the author focusses on that at which time period the employees are more absent. In this
paper he said that illness-related absences are highly seasonal, reaching a peak during the winter
months (December to February) and a trough during the summer (June to August). The high
incidence in winter is likely related to the prevalence of communicable diseases at that time,
especially colds and influenza. The low incidence during the summer may be partly because many
employees take their vacation during these months. Because of survey design, those who fall ill
during vacation will likely report ‘vacation’ rather than ‘sickness or disability’ as the main reason
for being away from work. It refers to workers absence from their regular task when he is normally
schedule to work. The according to Webster’s dictionary Compared with the annual average, part-
week absences are roughly 30% more prevalent in the winter months and almost 20% less so during

16
the summer months. Seasonality is much less evident in full week absences. Romero and Lee (2012)
also wrote a research paper titled A National Portrait of Chronic Absenteeism in the Early Grades.
In this paper he focused on the following points:

(i) How widespread is the Problem of Early Absenteeism?

(ii) Does Family Incomes Impact Early Absenteeism?
(iii) What is the Impact of Early Absenteeism on Academic Achievement?
Nordberg and Red (2016) wrote a paper on Absenteeism, Health Insurance, and Business
Cycles. In this he wants to evaluate how the economic environment affects worker
absenteeism and he also isolate the causal effects of business cycle developments on
work-resumption prospects for on-going absence spells, by conditioning on the state of
the business cycle at the moment of entry into sickness absence. The author found out
that:
(i) That business cycle improvements yield lower work-resumption rates for persons
who are absent, and higher relapse rates for persons who have already resumed
work.
(ii) That absence sometimes represents a health investment, in the sense that longer
absence ‘now’ reduces the subsequent relapse propensity.
(iii) That the work-resumption rate increases when sickness benefits are exhausted,
but that work-resumptions at this point tend to be short-lived. Having examined
the available literature it was realised that no single theory exists on the subject,
but that theories exist on why people fail to attend work. All information, reports
and statistics on the subject highlight absenteeism as a problem and an area that
greatly interests managers and researchers. Most of the literature is categorised
into two areas:
(i) Factors that cause absenteeism
(ii) Management’s response to the causes. The literature review is comprised of
two sections, the first examining the causes under a number of headings. The fact
the causes are identified in the literature demonstrates the firms do regard
absenteeism as a sufficient problem to warrant analyses being made and records
being kept. The second part of this literature review moves on to critically

17
 examine management’s responses to the problem as outlined in the literature, and
to try and assess the actual effectiveness of these responses.

2.4.1 The Factors that Influence Absenteeism to Work

The causes of absence are unlikely to be explained by any single factor, and current thinking sees its
causes in terms of multiple factors. Graham and Bennett (1995) believe that the factors contributing
to non-attendance include the nature of the job, personal characteristics of the worker and
motivating incentives. Up until the late 1970s, much of the research into absence focused on trying
to find a single factor to explain it If this were possible then employers would have been able to
solve the problem. It is in no way as easy as that, as Nicholson (1977) has identified. He splits
absence into three categories. Firstly, pain avoidance which puts forward the argument of job
dissatisfaction which cannot be seen as a single cause of absence, but without any doubt is one of a
number of factors that influence absenteeism. The second theory put forward is the adjustment to
work. This argues that employees adapt to the situation found in the workplace and that new
employees will observe absence behaviour of their colleagues. This raises many questions about the
culture, management style, even the work conditions and in the workplace.

Another adjustment to work perspective sees absence in terms of an employee’s response to both the
intrinsic and extrinsic rewards found in the workplace, and is associated with the equity and
exchange theory Rhodes and Steers (1990). This argues that individuals expect a fair exchange in
what they bring to their jobs in terms of skill, knowledge and commitment and the rewards or
outcomes they get out of it. One must raise the question of whether these relate to intrinsic factors
such as job satisfaction, or extrinsic factors such as pay and benefits. If either falls short of
employee’s expectations they will go absent? The third theory sees absence as a result of a decision
made on the basis of the cost and benefit associated with absence. If the employee values a day off
more day pay — will they go absent? This does not explain why some employees are motivated to
go to work while others stay away. There has been research to support the view that the provision of
occupational sick pay, which reduces the economic cost of absence, leads to higher absenteeism.
More recent research has tended to emphasise the complex nature of the factors influencing absence,
and is associated in particular with the ideas of Nicholson (1977), Steers and Rhodes (1978,1984)
and Rhodes and Steers (1990). The implications of the earlier research were that absence could be

18
avoided as long as the cause was identified and the appropriate policies applied. Steers and Rhodes
(1984) argue that absence behaviour needs to take into account variations in the personal
characteristics, attitudes, value and backgrounds of individuals and the fact that people do become
genuinely ill and have domestic difficulties from time to time.

CHAPTER THREE

METHODOLOGY

3.1 Introduction

This chapter deals with the specific procedures or techniques used to identify, select,
process, and analyze information about Factor analysis. This methodology section allows the
researcher to critically evaluate this study’s overall validity and reliability. This chapter will
answer two main questions. How was the data collected or generated? How was it analyzed?

3.2 Method of Data Collection

This part explains the underlying need for data collection and how it captures quality
evidence that seeks to answer all questions that have been posed.

To improve the quality of information, it is expedient that data is collected so that the
researcher can draw inferences and make informed decisions on what is considered factual.

3.3 Method of Data Analysis

The data for this study obtained from a courier company in Brazil from the records of
absenteeism at work from July 2007 to July 2010 as presented in Chapter four of this study is
analyzed using SPSS software.

3.4 Technique of Data Analysis

The analysis technique employed in this research work Factor Analysis technique.

19
3.4.1 Multivariate Analysis

Multivariate Analysis is a subdivision of statistics encompassing the simultaneous

observation and analysis of more than one outcome variable.

Multivariate Analysis is of different types, the choice of which type to use in a particular
does require selecting appropriate data transformations and standardizations (Kenkel, 2006). An
appropriate Multivariate Analytical strategy should take into account the statistical relevance,
data structure and objectives of the study. Some of the Multivariate Analysis technique are:

(i) Cluster Analysis

(ii) Factor Analysis
(iii) Canonical Correlation
(iv) Analysis of Variance etc.

The technique that will be used for the analysis of this research is Factor Analysis.

3.5 Statistical Tool

The statistical tools employed for data analysis in this study is the Factor Analysis. Factor
Analysis is used to reduce many individual items into a fewer number of dimensions. For the
purpose of this study, Factor Analysis will be used to reduce the number of variables to the
smallest number of common factors. The factors that are being considered for the data analysis is
as follows:

(i) Transportation Expense

(ii) Distance from Residence to work
(iii) Service time
(iv) Age
(v) Work load
(vi) Hit target
(vii) Disciplinary failure
(viii) Education
(ix) Social drinkers
(x) Pet
20
(xi) Weight
(xii) Height
(xiii) Body mass Index

3.6 Factor Analysis

Factor Analysis is a statistical method used to describe variability among observed,

correlated variables in terms of potentially lower number of unobserved variables called factors.
For example, it is possible that variations in six observed variables mainly reflect the variations
in two unobserved (underlying) variables. Factor analysis searches for such joint variations in
response to unobserved latent variables. The observed variables are modelled as linear
combinations of potential factors, plus “error” terms.

The statistical model for Factor Analysis attempts to explain a set of p observations in
each of n individuals with a set of k common factors ( f i , j ) where there are fewer factors per unit
than observations per unit (k<p). Each individual has k of their own common factors, and these
are related to the observations via factor loading matrix (L € R p X k), for a single observation,
according to

x i ,m −ʯ i=l i ,1 f 1 , m+ …+l i ,k f 1 ,m +ϵ i ,m (1)

Whereby

x i ,m is the value of the ith observation of the mth individual

ʯi is the observation mean for the ith observation

l i , j is the loading for the ith observation of the jth factor,

f j , m is the value of the jth factor of the mth individual, and

ϵ i , m is the (i ,m ¿ th unobserved stochastic error term with mean zero and finite variance.

In matrix notation

X −M =LF+ ϵ (2)

21
Where observation matrix X ϵ R p X n , factor matrix F ϵ R p X n, error term matrix ℇ ϵ R p X n and mean
matrix M ϵ R p X n whereby the ( i ,m ) th element is simply M i ,m =ʯ i.

Also, we will impose the following assumptions on F :

(i) F and ℇ are independent

(ii) E ( F )=0; where E is Expectation
(iii) Cov ( F )=I where Cov is the covariance matrix, to make sure that the factors are
uncorrelated, and I is the identity matrix.

Suppose Cov ( X−M )=Σ . Then

Σ=Cov ( X−M )=Cov (LF +ϵ ), and therefore, from the conditions imposed on F above,

T
Σ=LCov ( F ) L +Cov( ϵ ) , or setting ϕ=Cov (ϵ),

T
Σ=L L +ϕ

Note that for any orthogonal matrix Q , if we set L' =LQ and F ' =QT F , the criteria for being
factors and factors loadings still hold. Hence a set of factors and factor loadings is unique only
up to an orthogonal transformation.

3.6.1 Variance-Covariance Matrix

Variance covariance matrix is a symmetric matrix (i.e. X =X T ) and always positive semi definite
matrix. The diagonal values of the covariance matrix represent the variance of the variables x i,
while the off-diagonal entries represent the covariance matrix between two different variables. A
positive value in covariance matrix means a positive correlation between the two variables, while
a negative value indicates a negative correlation and zero value indicate that the two variables
are uncorrelated or statistically independent.

The variance covariance matrix is given by:

22
 [ ]
Var ( y 1 , y 2 ) Cov( y 1 , y 2) ⋯ Cov ( y 1 , y m )
S= ⋮ ⋱ ⋮ (3)
Cov ( y m , y ) Cov ( y m , y) ⋯ Var ( y , y m)

Where

( y − y)( y i− y )
n n
( y i− y j) ( y j − y j )
Var ( y )=∑ i and Cov ( yi , y j ) =∑
i=1 n−1 i=1 n−1

The correlation matrix could also be used which is equivalent to performing Factor Analysis on
the variance covariance matrix of the standardized variables.

x i−x i
Y i=
Si
(4)

Where x i is the mean and Si is the standard deviation

3.6.2 Eigen Values and Eigen Vectors

The eigen values are obtained by the equation | A−λI |=0(5)

Where A is the variance covariance matrix, λ is the latent value or characteristics roots, and I is
the identity matrix.

The eigen vectors are obtained by substituting each value of the characteristics roots (λ) in
equation (5).

Significance test for equality of remaining roots:

1
Degree of freedom ( DF ) = ( m−k +2 ) ( m−k−1 )=0 for k =0(6)
2

Degree of freedom for χ 2tab

Where m is the number of variables.

23
 2
χ statistics is computed by the formula:

[
χ 2= ( n−1 )− ( 16 )(2 p+ 1+ 2p )][−log|S|+ P log traces
e
p ]
(7)

Where n is number of observations, p is number of variables and S is the variance covariance

matrix.

If χ 2com is greater than χ 2tab, the null hypothesis is rejected at a certain level of significance and
conclude that all λ i ' s are equal otherwise, the null hypothesis is not rejected.

3.6.3 Trace

The Trace is defined as the sum of the eigen values. The determinant of S is the product of the
eigen values (Knill, 2012).

m
Trace S = ∑ λi (8)
i=1

Where λ i is the eigen values i=1 , … , m

The formula for χ 2 statistics is different when 0< k < p−1 or when the correlation matrix is used.

3.6.4 Loadings

Factor loadings are part of the outcome from factor analysis, which serves as a data reduction
method designed to explain the correlations between observed variables using smaller number of
factors.

Formula

When the Principal Components method is used, the matrix of estimated factor loadings, L is
given by:

'
L =¿

When the maximum likelihood method is used, the matrix factor loadings is obtained through an
iterative process.

24
Notation

Term Description

^
( λ ¿ ¿ 1 ¿ e^1) ¿ ¿ eigenvalue-eigenvector pairs

3.6.5 Communalities

This is the proportion of each variable’s variance that can be explained by the factor’s (e.g., the
underlying latent continua). It is also noted as h2 and can be defined as the sum of squared factor
loadings for the variables

Formula

2 2 2 2
hi =Li 1 + Li 2 +…+ Lℑ (10)

Where i=1 , 2 ,… p

Notation

Term Description

L Matrix of Factor loadings

3.6.6 Variance

Variability in the data explained by each factor. Variance equals the eigenvalue if you use
Principal Components to extract factors and do not rotate the loadings.

Formula

When a correlation matrix is used, the proportion of variance explained by the j th factor is
calculated as follows:

^L1 j2+ L^ 2 j2+ …+ ^L pj 2 λj

= (11)
tr (R) tr (R)

25
When a covariance matrix is used, the proportion of variance explained by the j th factor is
calculated as follows:

^L1 j2+ L^ 2 j2+ …+ ^L pj 2 λj

= (12)
tr (S) tr (S)

Notation

Term Description

L Matrix of factor loadings

λj th
j eigenvalue

tr (R)trace of correlation matrix

tr (S ) trace of covariance matrix

26
 CHAPTER FOUR: ANALYSIS OF DATA/IMPLEMENATION, CONCLUSION AND
RECCOMENDATION

4.1 Introduction

This chapter discusses data presentation, analysis of the data presented and discussion of the
result obtained from the analysis.

4.2 Results and Discussion

The methodologies presented in the previous chapter will be applied to the absenteeism
to work data obtained from a database that was created with the purpose of taking records from a
courier company in Brazil from July 2007 to July 2010. The data collected was primary in nature
and included factors like Transportation expense, distance from residence, Age, work load, hit
target etc. The data is then subjected to Factor Analysis.

Table 1: Eigen Values and Vectors Decomposition

Total Variance Explained

27
Component Initial Eigenvalues Extraction Sums of Squared Loadings R

Total % of Variance Cumulative % Total % of Variance Cumulative % Total

1 1.760 19.552 19.552 1.760 19.552 19.552 1.

2 1.625 18.051 37.603 1.625 18.051 37.603 1.
3 1.264 14.039 51.642 1.264 14.039 51.642 1.
4 1.068 11.864 63.505 1.068 11.864 63.505 1.
5 .956 10.621 74.127
6 .854 9.491 83.617
7 .632 7.023 90.640
8 .430 4.780 95.420
9 .412 4.580 100.000

Extraction Method: Principal Component Analysis.

From Table 1, retention of four components is inevitable. This is because they have eigen
values of at least one. The four components retained accounted for approximately 63.51% (from
table 1) of the variation in the dataset. We can conclude that it is great representation by the
selected components. The retention of the four components is further corroborated by the scree
plot in figure 1. The plot clearly indicates that the four components should be retained by a close
observation of the upper part of the line graph

28
Figure 1: Scree Plot

Table 2: Summary of Important Components in the Data Set

Component Matrixa

Component

1 2 3 4

29
ICD .780 -.285 .111 -.053
Transportation_expense .047 .607 -.488 .219
Distance_from_residence .517 .648 -.002 .143
Age -.262 -.028 .849 .034
Work_load -.237 .023 -.224 -.541
Hit_target .192 -.349 -.062 .610
Disciplinary_failure -.708 .421 .024 .094
Social_drinker .300 .672 .488 -.034
Social_smoker -.359 -.057 .028 .566

Extraction Method: Principal Component Analysis.

a. 4 components extracted.

Table 2 above shows the important components in the dataset. The four components that have
eigenvalues greater than one and are retained constitutes of about 63.51% of the variation in the
dataset. This is sufficient enough to guarantee an adequate representation by the retained
components.

Table 3: Rotated Component Matrix

Rotated Component Matrixa

30
 Component

1 2 3 4

ICD -.820 .131 .026 .115

Transportation_expense .291 .377 -.655 .040
Distance_from_residence -.135 .772 -.300 .059
Age .192 .162 .849 .085
Work_load .110 -.214 -.099 -.576
Hit_target -.189 -.167 -.079 .681
Disciplinary_failure .822 .063 .043 -.084
Social_drinker .010 .850 .217 -.097
Social_smoker .407 -.134 .044 .517

Extraction Method: Principal Component Analysis.

Rotation Method: Varimax with Kaiser Normalization.
a. Rotation converged in 5 iterations.

Table 3; concentrated on the four PC’s that explains 63.51% of the total variability of the dataset
retained.

Component 1 is loaded highly (strong positive relationship) with disciplinary failure

Component 2 identifies social drinker with a strong positive relationship.

While Component 3 identifies Age with a strong positive relationship.

And finally, Component 4 doesn’t have a strong positive relationship with any of the factors.

Table 4: Communality table

31
 Communalities

Initial Extraction

ICD 1.000 .704

Transportation_expense 1.000 .657
Distance_from_residence 1.000 .708
Age 1.000 .791
Work_load 1.000 .400
Hit_target 1.000 .535
Disciplinary_failure 1.000 .688
Social_drinker 1.000 .780
Social_smoker 1.000 .453

Extraction Method: Principal Component Analysis.

From the table above, we can see the results from the communalities of each factor extracted.
This table helps to identify which of the factor(s) is/are retained from the dataset. The factors
with high values are the component retained. Therefore, the factors retained are ICD, Distance
from residence to work, Age and Social drinker.

32
Figure 2: Biplot of the Components

We can see from Figure 2 above the component plot, we can see how the retained components
are influenced by the factors considered and how it influences the PC because of how further
apart it is from the origin.

33
 4.3 Conclusion

Conclusively, this study is carried out to determine the factors influencing absenteeism to work
using Factor Analysis methodology where all the determinants with eigen values greater than 1
were extracted and retained. The data obtained for this study was obtained from a database that
was created with the purpose of taking records from a courier company in Brazil from July 2007
to July 2010 to aid the investigation of the researcher. The results of the analysis showed that
Factor analysis was able to reduce the dataset to four components. The results showed that the
factors retained are ICD, Distance from residence to work, Age and Social drinker as the major
factors influencing absenteeism to work.

4.4 Recommendation

From the result of this study, the following recommendations are made:

(i) One primary key to reducing absenteeism is ensuring that the attendance policy is
clear and understood by all current employees and new hires. Employers can also
discourage employees from being absent by taking proactive steps like: rewarding
good attendance, providing emotional and health support.
(ii) Implement flexible work environments which can result to higher engagement levels,
increased job satisfaction.
(iii) Organizations should also try to establish the specific reasons for absences and what
can be done to improve the situation.
(iv) Providing personal time off, childcare support, and wellness program can go a long
way in improving employee engagement and reducing absenteeism rates. Managers
should also identify the root causes of absenteeism.

REFERENCES

34
Bandalos, Deborah L. (2017). Measurement Theory and Applications for the Social Sciences.
The Guilford Press.

^ Harman, Harry H. (1976). Modern Factor Analysis. University of Chicago Press. pp. 175,
176. ISBN 978-0-226-31652-9.

^ Polit DF Beck CT (2012). Nursing Research: Generating and Assessing Evidence for Nursing
Practice, 9th ed. Philadelphia, USA: Wolters Klower Health, Lippincott Williams & Wilkins.

^ Meng, J. (2011). "Uncover cooperative gene regulations by microRNAs and transcription

factors in glioblastoma using a nonnegative hybrid factor model". International Conference on
Acoustics, Speech and Signal Processing. Archived from the original on 2011-11-23.

^ Liou, C.-Y.; Musicus, B.R. (2008). "Cross Entropy Approximation of Structured Gaussian
Covariance Matrices". IEEE Transactions on Signal Processing. 56 (7): 3362–
3367. Bibcode:2008ITSP...56.3362L. doi:10.1109/TSP.2008.917878. S2CID 15255630.

^ Zwick, William R.; Velicer, Wayne F. (1986). "Comparison of five rules for determining the
number of components to retain". Psychological Bulletin. 99 (3): 432–442. doi:10.1037//0033-
2909.99.3.432.

^ Horn, John L. (June 1965). "A rationale and test for the number of factors in factor
analysis". Psychometrika. 30 (2): 179–
185. doi:10.1007/BF02289447. PMID 14306381. S2CID 19663974.

^ Dobriban, Edgar (2017-10-02). "Permutation methods for factor analysis and

PCA". arXiv:1710.00479v2 [[Link]].

^ * Ledesma, R.D.; Valero-Mora, P. (2007). "Determining the Number of Factors to Retain in

EFA: An easy-to-use computer program for carrying out Parallel Analysis". Practical
Assessment Research & Evaluation. 12 (2): 1–11.

35
^ Tran, U. S., & Formann, A. K. (2009). Performance of parallel analysis in retrieving
unidimensionality in the presence of binary data. Educational and Psychological Measurement,
69, 50-61.

b
^ Velicer, W.F. (1976). "Determining the number of components from the matrix of partial
correlations". Psychometrika. 41 (3): 321–327. doi:10.1007/bf02293557. S2CID 122907389.

^ Courtney, M. G. R. (2013). Determining the number of factors to retain in EFA: Using the
SPSS R-Menu v2.0 to make more judicious estimations. Practical Assessment, Research and
Evaluation, 18(8). Available online: [Link]

^ Warne, R. T.; Larsen, R. (2014). "Evaluating a proposed modification of the Guttman rule for
determining the number of factors in an exploratory factor analysis". Psychological Test and
Assessment Modeling. 56: 104–123.

^ Ruscio, John; Roche, B. (2012). "Determining the number of factors to retain in an

exploratory factor analysis using comparison data of known factorial structure". Psychological
Assessment. 24 (2): 282–292. doi:10.1037/a0025697. PMID 21966933.

^ Garrido, L. E., & Abad, F. J., & Ponsoda, V. (2012). A new look at Horn's parallel analysis
with ordinal variables. Psychological Methods. Advance online
publication. doi:10.1037/a0030005

^ Revelle, William (2007). "Determining the number of factors: the example of the NEO-PI-
R" (PDF).

^ Revelle, William (8 January 2020). "psych: Procedures for Psychological, Psychometric, and
PersonalityResearch".

^ Kaiser, Henry F. (April 1960). "The Application of Electronic Computers to Factor

Analysis". Educational and Psychological Measurement. 20 (1): 141–
151. doi:10.1177/001316446002000116. S2CID 146138712.

36
^ Bandalos, D.L.; Boehm-Kaufman, M.R. (2008). "Four common misconceptions in exploratory
factor analysis". In Lance, Charles E.; Vandenberg, Robert J. (eds.). Statistical and
Methodological Myths and Urban Legends: Doctrine, Verity and Fable in the Organizational
and Social Sciences. Taylor & Francis. pp. 61–87. ISBN 978-0-8058-6237-9.

^ Larsen, R.; Warne, R. T. (2010). "Estimating confidence intervals for eigenvalues in

exploratory factor analysis". Behavior Research Methods. 42 (3): 871–
876. doi:10.3758/BRM.42.3.871. PMID 20805609.

^ Cattell, Raymond (1966). "The scree test for the number of factors". Multivariate Behavioral
Research. 1 (2): 245–76. doi:10.1207/s15327906mbr0102_10. PMID 26828106.

^ Alpaydin (2020). Introduction to Machine Learning (5th ed.). pp. 528–9.

^ Russell, D.W. (December 2002). "In search of underlying dimensions: The use (and abuse) of
factor analysis in Personality and Social Psychology Bulletin". Personality and Social
Psychology Bulletin. 28 (12): 1629–46. doi:10.1177/014616702237645. S2CID 143687603.

^ Mulaik, Stanley A (2010). Foundations of Factor Analysis. Second Edition. Boca Raton,
Florida: CRC Press. p. 6. ISBN 978-1-4200-9961-4.

^ Spearman, Charles (1904). "General intelligence objectively determined and

measured". American Journal of Psychology. 15 (2): 201–
293. doi:10.2307/1412107. JSTOR 1412107.

37
38