An exploration of the relationship
between emotions and self-reported
productivity over time
Moranda Lam
A research project submitted in partial fulfilment of the
requirements for the degree of MA by coursework and Research
Report in the field of Industrial Psychology in the Faculty of
Humanities, University of the Witwatersrand, Johannesburg.
15 February 2010
Declaration
I hereby declare that this research report/thesis is my own, unaided work. It is being
submitted in partial fulfilment for the degree of Master of Arts by Coursework and
Research Report in the Field of Industrial Psychology at the University of the
Witwatersrand, Johannesburg. It has not been submitted before for the purpose of any
other degree or examination at any other University
Moranda Lam
15 February 2010
Word count: 23 540
University of the Witwatersrand, Johannesburg
SENATE PLAGIARISM POLICY
Declaration by Students
TI ___________________________ (Student number: _________________) am a
student registered for ______________________ in the year __________. I hereby
declare the following:
? TI am aware that plagiarism (the use of someone else?s work without their
permission and/or without acknowledging the original source) is wrong.
? TI confirm that the research report submitted is my own unaided work except
where I have explicitly indicated otherwise.
? TI have followed the required conventions in referencing the thoughts and ideas
of others.
? TI understand that the University of the Witwatersrand may take disciplinary
action against me if there is a belief that this is not my own unaided work or
that I have failed to acknowledge the source of the ideas or words in my
writing.
TSignature: _________________________ Date: ________________________
Abstract
For many years, the importance of emotions has been underestimated in the
workplace. This is because the workplace was believed to be a space that did not
accommodate the expression of emotions. However, towards the end of the twentieth
century, researchers became more interested in the role of emotions in the workplace,
since it is said that people do not always work in an objective manner based on cold
cognitive conditions. As a result, it has led to the development of various models and
theories, one of which is the happy-productive worker hypothesis. The current study
is based on this model. However, the approach to understanding this hypothesis in the
current study is slightly different from how it has been traditionally assessed. The
current study expanded the happiness construct to explore whether there is a
relationship between arousal, pleasantness and self-reported productivity over time.
Results from the current study were found to support the happy productive worker
hypothesis, as a significant relationship was found between pleasantness and self-
reported productivity. However, this relationship was only significant in the absence
of the arousal dimension. This, therefore, indicates that arousal plays an important
role in understanding emotions in relation to self-reported productivity in the
workplace.
Furthermore, a repeated measures approach was used to observe within
subject effects to assess for potential patterns. The relationship between emotions and
self-reported productivity was only significant at specifically 10H00 and 12H00 and
not at 16H00 and 19H00. This may be due to the low response rate for the 16H00 and
19H00 questionnaires. In addition, only slight changes were found in the change of
emotions and self-reported productivity as separate constructs over time. It is also
important to note that the data used in the study was somewhat skewed due to the
biased age and cultural groups of the sample. Thus, this violates the assumption of
normality. Consequently, these effects may have impacted on the findings and
applicability of the results to alternative contexts. Thus, more research in this field is
required.
Acknowledgements
I wish to express my sincere thanks and gratitude to the following people:
? My supervisor, Karen Milner, for your patience, advice and powerful
intellectual presence. Thank you for believing in me and supporting this study.
Your suggestions have been most valuable.
? Mike Greyling, for your support, belief and mentorship. You are an
inspirational academic. Your principled dedication and knowledge to the field
of psychology is truly humbling.
You have both helped to make this study possible.
Contents
Page
Chapter 1. Introduction 1
Chapter 2. Literature review 2
2.1. Introduction 2
2.2. Emotions 2
2.2.1. Definition: affect, moods and emotions 3
2.2.2. Common theories on emotions 4
2.2.2.1. Basic emotions 4
2.2.3. Emotions in the workplace
2.2.3.1. Employee positive emotion and favourable outcome model 8
2.2.3.2. AET (Affective events theory) model 9
2.3. Productivity 12
2.3.1. Definition 13
2.3.2. Triple P model 13
2.3.3. Studies on productivity 16
2.4. Emotions and self-reported productivity 17
2.4.1. Happy productive worker hypothesis issues 17
2.5. Implications of previous research 18
2.6. Summary 19
2.7. Research Questions 20
Chapter 3. Methodology 21
3.1. Introduction 21
3.2. Research design 21
3.3. Procedure 21
3.3.1. Stage one 23
3.3.2. Stage two 24
3.4. Sample 25
Page
3.5. Instruments 27
3.6. Analysis 28
3.7. Ethics 30
3.8. Summary 3
Chapter 4. Results 32
4.1. Introduction 32
4.2. Demographics in relation to emotions and self-reported productivity 32
4.3. Is there a relationship between reported events and emotions? 33
4.4. Emotion and self-reported productivity 34
4.4.1. Is there a relationship between arousal and self-reported productivity? 34
4.4.2. Is there a relationship between pleasantness and self-reported productivity? 35
4.4.3. Is there a relationship between the interaction between arousal and
pleasantness and self-reported productivity? 35
4.4.4. Test for multicollinearity. 35
4.5. Longitudinal results 36
4.5.1. Patterns in emotions and self-reported productivity over time 36
4.5.2. Is there a pattern in the relationship between emotions and self-reported
productivity over time? 40
4.5.3. Re-examining the relationship between emotions and self-reported productivity at
10H00 and 12H00 42
4.6. Summary 43
Chapter 5. Discussion 45
5.1. Introduction 45
5.2. Demographics in relation to emotions and self-reported productivity 45
5.3. Is there a relationship between reported events and emotions? 45
5.4. Emotions and self-reported productivity 46
5.4.1. Relationship between emotions (arousal and pleasantness) and self-reported
productivity 47
Page
5.5. Longitudinal results 50
5.5.1. Patterns in emotions and self-reported productivity over time 50
5.5.2. Is there a pattern in the relationship between emotions and self-reported
productivity over time? 52
5.5.3. Re-examining the relationship between emotions and self-reported productivity
at 10H00 and 12H00 54
5.6. Summary 55
Chapter 6. Limitations and implications for future research 56
Chapter 7. Conclusion 58
Chapter 8. Reference list 60
Appendices
Appendix A: Affect circumplex 66
Appendix B: Affect grid 67
Appendix C: Affective events theory model 68
Appendix D: Triple P model 69
Appendix E: Biographical questionnaire 70
Appendix F: Questionnaire 1 72
Appendix G: Questionnaire 2 73
Appendix H: Questionnaire 3 74
Appendix I: Questionnaire 4 76
Appendix J: Data collection schedule 78
Appendix K: Participant information sheet 79
Appendix L: Example of events in relation to reported emotions (first 100 responses) 81
Appendix M: Ethics clearance certificate 85
Chapter 1
Introduction
For many years, the importance of emotions was underestimated in the workplace because
the workplace was believed to be a space that did not tolerate the expression of emotions.
However, it has since been acknowledged that individuals do bring their affective states,
traits and emotions to work (Clements-Croome, 2000). This is because organisational
members seldom carry out their work in an objective manner based on cold cognitive
conditions. Organisational researchers have thus begun to take an interest in the effects of
emotions in the workplace (Clements-Croome, 2000).
Since research on emotions became more popular and prominent in the twentieth
century, researchers began to look at the relationship between emotions and various
behavioural outcomes such as decision making processes (Isen & Means, 1983) and
performance (Bell & Staw, 1989). One of the more popular studies on emotions is the
investigation of the happy-productive worker hypothesis (Staw & Barsade, 1993). However,
mixed results have been found to support or disapprove this hypothesis (Iffaldano &
Muchinsky, 1985; George & Brief, 1992). Furthermore, the theoretical content underlying
the two constructs are also unclear. Firstly, happiness is a rather confining term that is used to
describe how an individual is feeling emotionally. It only depicts one particular emotional
state. Therefore, using a broader construct may be more suitable to better understand the
influence of emotions on the behavioural outcome, productivity. Secondly, productivity in
the past has been very poorly defined and researchers have used various types of
organisational outcomes to broadly define and measure the construct. Thus, the current study
will use a more precise measure by means of a subjective evaluation of productivity.
Furthermore, many of the studies which have examined the relationship between emotions
and productivity have only assessed this relationship based on one moment. Since emotions
are said to be highly susceptible to change, it was therefore important to assess the happy-
productive worker hypothesis longitudinally to examine whether there are any patterns in this
relationship over time.
1
Chapter 2
Literature review
2.1. Introduction
The literature review will explore various discrepancies and potential areas of development
within the research on emotions and productivity in the workplace. This will include a
general definitional clarification of each construct followed by a discussion of their
respective roles specific to the workplace environment. Thereafter, various models and
theories for both constructs will be discussed.
2.2. Emotions
Although the concept of emotions was first addressed by William James in 1884 (Gross &
Munoz, 1995), research emphasising emotions only made substantial advances towards the
end of the twentieth century. This progress provided a better understanding of the structure
and roles of emotions in human behaviour (Lord & Kanfer, 2002).
Since William James first questioned ?What is an emotion?? in 1884, until today,
more then a century later, there does not seem to be a satisfactory answer. This is because
emotion was seen to be a ?fuzzy category? with no clear boundaries of what can or cannot be
considered to be an emotion and whether there are particular conditions from which an
emotion can be experienced. Unlike terms such as ?bachelor? which may be precisely defined
as an unmarried man, there is no one true definition of an emotion that would enable one to
decide what is and what is not an emotion (Gross & Munoz, 1995). Attempts to clarify the
underlying definition or characteristic of an emotion to better understand this concept have
included explorations of the role of emotions within various models, theories of emotions
and also differentiating emotions from other similar terms which are often used
interchangeably. The following sections will consider some of the models and theories;
however it is important to first differentiate the term ?emotions? from other similar terms.
2
2.2.1. Definition: affect, moods and emotions
The terms which are often confused and used interchangeably with emotions are affect and
mood. Both moods and emotions fall under the general term of affect. Affect is a
psychological state with the fundamental characteristic of possessing change (Weiss &
Kurek, 2003). If one envisions a continuous stream of experiences, changes of some kind
may be observed between the particular moments. These changes may be seen as an
individual changing from one state to another. These states have the potential to vary in
length of time. For instance, the pain from dropping a book on one?s foot may last for a
moment and anger may last for a few hours. As a result, regardless of how long it takes for
people to change from one state to another and how long they maintain a particular state,
states are nevertheless time bound constructs possessing the potential to change at any
moment (Weiss & Kurek, 2003). Thus, if affect is defined as a state, the features that
describe the nature of a state can therefore also be used to characterise affect. Consequently,
moods and emotions are also considered to be states as they constitute the two forms of
affect.
Taking into consideration the lack of definitional clarity between emotions and affect
in past research, the studies on emotions and affect which will be referred to throughout this
paper will therefore be considered as comparable or alike due to the structural complexity
and characterisation of the affect construct. It was, however, important for the purpose of the
current study to carefully define and distinguish between emotions and moods, as they are
considered subcategories of affect and thus carry a different meaning.
There have been several definitions which have attempted to differentiate moods
from emotions. However, there are two differentiations that are most appropriate to highlight
for the context of this study. Weiss and Kurek (2003) suggest that moods are generally
considered to be more dispersing affective states, deficient of a defining object and disjointed
from specific objects or state of affairs. Emotions on the contrary, are believed to be affective
states that are connected to particular objects or circumstances. In other words emotions are
related or attached to some person, event or object.
A similar differentiation was found in the work of Izard (1993). Although Izard
(1993) considers emotions to be a complex phenomenon, he emphasises the practical
3
component of emotions which has the ability to manifest in the form of a physical action,
biasing of perceptions, or a feeling state. Emotions are said to be associated with a specific
stimulus and occupy a shorter period in time while moods on the other hand are more
enduring, diffused and less related to specific stimuli (Weiss & Kurek, 2003). Staw and
Barsade (1993) further support this differentiation by recognising that emotions generally
denote a strong reaction to a specific object or cause; in contrast, moods usually refer to a
milder, more diffuse affective state that may not be directed towards a single attitudinal
object. Weiss and Kurek (2003) provide examples that distinguish between the language of
emotions and moods. For instance, ?I?m feeling down? and ?I?m feeling kind of perky? are
considered statements which form part of the language of mood. On the other hand, ?I?m
feeling angry with my boss? and ?I?m feeling guilty about missing my son?s soccer game?
are statements in the language of emotions. From these examples it can be seen that the
reported emotions are linked to some sort of event, person or object that are absent from the
mood statements. Consequently, when looking at emotions it is necessary to relate them to
some specific person, object, behaviour or event.
2.2.2. Common theories on emotions
Over the years, there have been a number of theories and models developed in an attempt to
gain a better understanding of the role of emotions in various contexts. Initially, it would be
useful to consider theories that address the basic role and meaning of emotions as well as the
fundamental units of the construct.
2.2.2.1. Basic emotions
One of the more prominent models within the field of emotions is the basic emotions
structure. Over the years, researchers were struggling to decide which emotions to label as a
basic emotion. This is because researchers tend to question the basis upon which one is
willing to call an emotion ?basic? (Lord & Kanfer, 2002). Izard (1992) for example, proposes
three criteria: basic emotions have innate and unique neural substrates, unique facial
expressions and unique feeling states, while Ekman (1994) suggests further criteria that
incorporate primate and automatic appraisal and distinct physiology. Thus, after thorough
discussions and negotiations, researchers eventually came to a consensus that the range of
emotions disperses from a universal set of nucleus emotions (Lord & Kanfer, 2002).
4
Having considered the different criteria to assess those emotions that qualify as a
basic emotion, six emotions are repeatedly identified within this category. These emotions
are happiness, surprise, fear, sadness, anger and disgust. Although many researchers agree
that these six emotions are sufficient to form the basic emotion structure, the prototype
theory, which is developed by Shaver, Schwartz, Kirson and O? Conner (1987), has critiqued
that ?surprise? cannot be considered as a basic emotion. Shaver et al. (1987) felt that surprise
does not clearly demonstrate a particular kind of feeling due to the varying applicability of
this feeling to sudden situations. Despite this critique, however, these six basic emotions have
since formed the fundamental units of the basic emotions structure.
After the establishment of the basic emotions, subsequent research used this structure
as the core foundation to further build upon the concept of emotions. Of relevance to the
current study is the prototype theory, which has been previously mentioned. According to
Shaver et al. (1987), the basic emotions can be further subdivided into twenty-five
subordinate categories. Each subcategory is distinguishable based on the intensity of the
emotion, antecedents and/or context for which the emotion is expressed. Alternately,
emotions can also simply be categorised as positive and negative emotions. These two
categories would then form the two superordinate levels of emotions (Shaver et al., 1987).
This classification into two superordinate levels is based on the proposition that affective
states, despite the differences between moods and emotions, appear to fall along a bipolar
dimension ranging from positive to negative. This continuum is referred to by terms such as
?valence?, ?hedonic tone? or ?pleasantness?. Further examination has found that these affect
states, whether positive or negative states, are sometimes aroused or quiescent. For that
reason, an emotional arousal or activation dimension is perceived as relevant to complement
the valence dimension (Lord & Kanfer, 2002). In contrast to this view, Bradburn (1969)
found that positive and negative affective states do not lie on a bipolar dimension but rather
stand as two separate dimensions. It is therefore important to clarify whether these
dimensions should be treated as bipolar opposites or as two separate dimensions in order to
understand and assess them correctly.
Thus, noting these two different views, attempts which have been made to clarify this
incongruity will be discussed next. Firstly, this led to the development of the Positive Affect
Negative Affect Schedule (PANAS) by Tellegen, Watson and Clark (1999). Further
5
developed from the PANAS, a geometric scale called the affect circumplex (Refer to
Appendix A) can be used as an attempt to understand whether affective states comprise two
separate dimensions or a single bipolar construct (Tellegan, et al., 1999). This affect
circumplex was developed by Larsen and Diener (1992) and has the affect terms
circumscribed in a particular order. The order and distance between the terms are arranged
along the circumference according to the similarity in meaning between the affect terms. In
such a structure, the location of the affect terms are determined by the main longitude and
latitude which are at right angles to each other, each representing a dimension with a
particular meaning. The latitude dimension represents the continuum between pleasant and
unpleasant affect and the longitude dimension characterises activated versus inactivated
affects.
Hence, based on the structure of the affect circumplex model by Larsen and Diener
(1992), it is suggested that although a range of emotions can be experienced, they however
need to be categorised on a circular continuum. This circular continuum nevertheless still
relies on two main emotion dimensions, namely arousal and pleasantness. These two
dimensions are measured on a bipolar scale. The extremes on the bipolar scales of each of the
dimensions represent the positive and negative aspect of the particular emotion.
Consequently, the affect circumplex supports the notion that positive and negative emotions
are bipolar in nature.
Another illustration to show that positive and negative emotions fall on a bipolar
continuum is demonstrated by one of the dimensions on the affect grid (Refer to Appendix
B), which is pleasantness. This model uses the broader affect construct to understand
emotions and is constructed by Russell, Weiss and Mendelsohn (1989). The affect grid was
designed to record judgements about single instances of affect and is based on a similar
concept to the affect circumplex. The feeling presented by a single facial gesture or a single
word constitutes the judgements expressed. According to this model, the affective state that a
person can experience is proposed to lie within two dimensions, which are represented on a
perpendicular set of axes. The horizontal axes from left to right represent the pleasantness-
unpleasantness dimension, where the vertical axes from top to bottom represent the arousal-
sleepiness (Russell et al., 1989). Thus, this model comprises four quadrants. Each quadrant
represents an inclination towards different emotions. Therefore, someone with high arousal
and high pleasantness would tend towards a feeling of excitement, a person with high arousal
6
and low pleasantness would tend to be calm, individuals with low arousal and high
pleasantness would lead to anxiety, and lastly, someone with low arousal and low
pleasantness would tend towards depression.
When relating the affect grid to positive and negative emotions, pleasantness can be
more closely associated to the positive and negative emotions. Emotional arousal, on the
other hand has less clear boundaries to show whether high arousal is a positive or negative
emotion. It would be dependent on the context or environment in which the person finds
themselves. For example, an excited person can feel just as emotionally aroused as an angry
person. But the excited person is more likely to be experiencing more positive emotions in
that particular circumstance. A more blurred example is a person who is feeling eager. Whilst
this emotion may involve high arousal, it would be dependent on the situation to determine
whether it is a positive or negative emotion. For example, a person who is eager to find out
whether they have been retrenched is likely to interpret this emotion as more negative than
someone who is eager to find out about his or her salary increase.
As a result, pleasantness and arousal can be described as two dimensions of emotional
affect. The current study will therefore look at emotions in line with the way emotions was
conceptualised by Shaver et al. (1987) and Russel et al. (1989). Furthermore, it is
acknowledged that pleasantness and arousal are not the only dimensions that can be used to
describe and understand emotions yet is sufficient to only use these two in the current study.
In addition it is necessary to note that according to Russel et al. (1989), these two dimensions
are conceptually separate even if they happen to be correlated positively or negatively in
certain situations. Yet, in the study of Estes and Adelman (2008), they found a consistent
relationship between the arousal and pleasantness. This may be an indication that this
relationship may be contextually and operationally influenced. Nevertheless, the affect grid is
potentially suitable for a study that requires any judgments about affect or emotions of either
a descriptive or subjective kind (Russell et al., 1989).
2.2.3. Emotions in the workplace
Despite the general theories on emotions, it was not until the end of the twentieth century that
research on emotions gained its prominence in organisational literature. Only during the last
decade of the twentieth century did researchers begin to pay attention to this topic. According
7
to Flam (1993), a serious analytical interest in emotion in general and in negative emotions in
particular is almost missing in the studies of work and organisational life. Likewise, Ashforth
and Humphrey (1995) commented that everyday emotions that are part of organisational
routines have been surprisingly neglected, given that they saturate and are an inherent part of
the workplace.
Domagalski (1999) suggested that this disregard for the emotional subtext of
organisations was largely due to the longstanding emphasis on rationality and more
deliberate modes of performance in organisations. Yet, even when attention was directed to
the socio-emotional needs of employees, as was the case with the Hawthorne studies (Brief
& Weiss, 2002), superiors were advised to maintain a controlled posture in order to
effectively deal with the ?irrational? emotional episodes of the members of the workforce. As
a result, employee emotions were considered negative and irrational. Emotions were
considered something that had a need to be managed and controlled so that they would not
interfere with rational organisational functioning (Stearns & Stearns, 1986). However, later
investigations acknowledged that employees seldom carry out their tasks in an objective
manner based on cold cognitive conditions. Consequently, organisational researchers have
since begun to take an interest in the effect and role of emotions in the workplace (Clements-
Croome, 2000). A few models have therefore been developed to explore at the role of
emotions in the workplace.
2.2.3.1. Employee positive emotion and favourable outcome model
The employee positive emotion and favourable outcome model was not specifically
developed to look at emotions in the workplace. However, the constructs used in this model
can be applied to an organisational context and are relevant to the current study. The
employee positive emotion and favourable outcomes model was developed by Staw, Sutton
and Pelled (1994). Similar to the previous models discussed, this model also categorises
emotions into the two broad positive and negative terms as previously discussed. This model
suggests that positive emotions bring about favourable outcomes on the job through three
possible intervening processes (Staw et al., 1994). The first process suggests that positive
emotions have desirable effects on an individual independent of the person?s relationship
with others, including greater task activity, persistence, and enhanced cognitive functioning.
The second process proposes that people with positive rather than negative emotions benefit
8
from more favourable responses by others. People with positive emotions are more
successful at influencing others. They are more likable, and a halo effect may occur when
warm or satisfied employees are rated favourably on other desirable attributes. The last
process suggests that people with positive feelings react more favourably to others, which is
reflected in greater altruism and cooperation with others (Staw et al., 1994). In essence, Staw
et al. (1994) suggests that employees who travel through organisational life with a positive
emotion will reap more favourable outcomes from their workplace in comparison to their
more negative counterparts.
2.2.3.2. AET (Affective events theory) model
Having considered a few theories and models on emotions, the Affective Events Theory
(AET) (Refer to appendix C) appears to be the most all-encompassing theory on emotions in
the context of this current study (Weiss and Cropanzano, 1996). The AET is useful as it
represents a fuller understanding of emotions in the workplace. Although Weiss and
Cropanzano (1996) initially presented their model as an untested theory, and research was
still in the early stages, results to date have been strongly supportive of the core ideas in the
theory.
This theory is represented by a model which was developed by Weiss and
Cropanzano (1996). In the AET, Weiss and Cropanzano (1996) argue that aspects of the
work environment including environmental conditions, roles and job design initiate emotions
in organisational settings. AET focuses on the structure, causes and consequences of
affective experiences at work by examining the antecedents of employee? experiences of
affective work events and the affective, attitudinal and behavioural reaction to these events.
The antecedents of work constitute the affective events which are described colloquially as
?hassles and uplifts? that act systematically to determine affective states. In turn, these states
lead to attitudinal and behavioural outcomes (Weiss & Cropanzano, 1996).
One of the more important outcomes of AET research is attempting to understand job
satisfaction from a new perspective. For many years, researchers have been puzzled by
consistent findings suggesting that job performance bears no relation to workers? feelings of
satisfaction (Judge, Thoreson, Bono & Patton, 2001). Various organisational studies in the
past have studied affective experiences not as affect itself but rather as related constructs
9
such as job satisfaction, psychological strain and burnout (Weiss & Kurek, 2003). Fewer
studies have looked at affect in terms of emotions and moods in the workplace.
Although the regularly used constructs to measure to affect may potentially be
influenced by affective experiences and are related conceptually, they are not the same as
actual affective experiences. Job satisfaction, which is the most frequently used construct to
measure affect is in fact often categorised as an attitude and not as an affective state (Staw &
Barsade, 1993). As an attitude, job satisfaction is best conceptualised as an evaluation of
one?s job which is influenced in part by affective events that have occurred at work, as a
source of error, and the mood whilst making the evaluation (Brief, Butcher & Roberson,
1995). Thus, job satisfaction generally encompasses a large cognitive component.
A further noted difference between job satisfaction and affective experiences is that
job satisfaction is a fairly stable construct as compared to affect (Weiss & Kurek, 2003).
Additionally, the AET makes it clear that a range of emotions are transient states rather than
aspects of work life that remain constant for long periods of time. Employees at work may, in
one moment, be overjoyed with a successful outcome and in the next moment, be
disappointed and perhaps even angry when their boss does not appear to share their
enthusiasm. Thus, job satisfaction is not a good measure of affect since affect is highly
susceptible to change. Researchers today have argued that job satisfaction constitutes a set of
attitudes towards work that do not necessarily include affective feelings (Weiss & Kurek,
2003). As a result, research that specifically addresses affective states such as emotions
rather than substitutes for the outcome of these states is required.
Noting that emotions are susceptible to change on a moment to moment basis, the
study conducted by Ashkanasy, Zerbe and Hartel (2002) suggested that it may be worth
examining emotions as well as the relationship between emotions and productivity through
inspecting ?within subjects? effects rather than only considering the differences and/or
patterns between individuals. In agreement with this view, Stone, Shiffman and De Vries
(1999) have looked at within subject effects of emotions. In their study, Stone et al. (1999)
captured momentary affective states among employees using an ecological momentary
assessment. Participants were signalled at random points on various days and asked to
indicate their emotional sates. These states were measured using the pleasant versus
10
unpleasant and aroused versus sleepy dimensions as on the affect grid. These immediate
affective states were collected over a two-week period.
In addition, the AET also describes how employee behaviours and attitudes are
constantly changing as they encounter everyday hassles and uplifts in their workplace. In the
AET, events related to emotions are seen to be a central aspect of understanding employee
attitudes and behaviour. Ashkanasy et al. (2002) argued that it is the little things building up
that determine how employees think and behave. They suggest that uplifting events, such as a
complimentary comment by a superior, or a friendly and supportive act by a colleague
following a negative occurrence such as a lost sale, can reverse the negative behavioural
consequence that would normally be expected to emerge from the event (Ashkanasy et al.,
2002).
Results from this study showed that individual affective experiences examined over
time have distinctive patterns (Stone et al. 1999). Although patterns were observed, no
distinct cycles were detected. Similar results were found in the study of Dickman (2002) and
Egloff, Tausch, Kohlmann and Krohne (1995). Dickman (2002) found that arousal tends to
increase throughout the day in the presence of some sort of manipulation such as smoking or
caffeine intake. Egloff et al. (1995), furthermore, found that the positive moods increases
from morning to evening. However, Stone et al. (1999) found that the average levels of
positive and negative affects were found to differ among individual participants. The
variability of the affect levels fluctuate in the intensity of the peaks and valleys and the time
span for movement from one state to another (Stone et al., 1999). Fluctuations are, however,
predictable and explainable by the mere aspect that individuals are exposed to different
experiences. This suggests that emotions, being sensitive to any sort of stimuli, have a high
susceptibility to change and fluctuation. The conclusion from their study was that affective
experiences are state constructs that vary in direction, intensity, subjective experience and
length of time.
In summary, emotions are often misinterpreted and used interchangeably with other
constructs such as mood. However, the distinguishing factor, according to various
researchers as previously discussed, is that emotions refer to the feelings that are experienced
in conjunction with events that individuals experience (Weiss & Cropanzano, 1996). These
feelings are also said to be susceptible to change on a moment to moment basis (Stone et al.,
11
1999). Thus, given the transitory nature of emotions, the construct can be observed and
studied over time to trace potential patterns and within subject effects which may emerge.
Emotions have also in the past been understood and measured using various related
construct. However, it has often been conceptualised and described as two dimensions
namely pleasantness and arousal (Russel et al., 1989). Thus, the current study will also
approach the concept of emotions from this point of view.
Furthermore, according to Weiss and Cropanzano?s (1996) AET model, emotions or
aspects of affect in theory can be related to various behavioural outcomes such as decision
making, performance, psychological stress, judgements and interpersonal behaviour (Lucas
& Diener, 2003). Yet, the behavioural outcome, which the current study is most interested in
testing, is productivity. The AET model also specifically suggests that there is some
association between productivity with emotions (Weiss & Cropanzano, 1996). Thus various
behavioural outcomes could have been chosen, but based on the proposed relationship by the
AET, and with productivity being one of the more important behavioural outcomes in the
workplace, the current study will therefore inspect this suggested relationship.
2.3. Productivity
Since the 1930s there has been a great deal of interest in the relationship between employee
well-being and productivity (Zelenski, Murphy & Jenkins, 2008). Hersey (1932) reported a
positive relationship between daily emotions and performance, whereas Kornhauser and
Sharp (1932) reported that worker attitudes (more cognitive assessments of emotions) were
not related to productivity. Establishing whether or not emotions in the workplace promote
productivity has important implications for management and strategies for workplace
improvements. Thus, understanding the role of productivity in the workplace is important and
the current study will look at the relationship between productivity and emotions. As a result,
to get a better understanding of this construct, the definition of productivity will first be
considered. This will be followed by an examination of the Triple P model and an
explanation of how it can be assessed. Thereafter, there will be a brief look at previous
studies on workplace productivity.
12
2.3.1. Definition
Productivity, like emotions, is a complex construct which is often poorly differentiated from
its neighbouring terms. The earliest appearance of the term ?productivity? was used by
Quesnay (1766) (as cited in Tangen, 2004). Subsequent to its first appearance two centuries
ago, the term productivity has since been applied and explored in many different
circumstances and at various levels of aggregation. However, similar to emotions,
productivity is also often confused with its neighbouring terms. There is therefore a need to
clarify the theoretical and operational definition of the construct. This can be achieved
through the use of the triple P model (Tangen, 2004) which will be discussed next.
2.3.2. Triple P model
An effort to clarify this definitional confusion has been found in the triple P model (Tangen,
2004) (Refer to appendix D). The schematic representation of the triple P model can be used
to distinguish between the following commonly confused terms: productivity, performance,
effectiveness, efficiency and profitability (Tangen, 2004). This model allows the main
differences between the terms to be clearly discernable.
Briefly looking at the triple P model, effectiveness and efficiency seem to shadow the
whole model. In essence, effectiveness can be defined as doing the right things, whereas
efficiency is described as doing things right (Sink & Tuttle, 1989). However, neither of these
constructs is appropriate for the current study.
Looking more specifically at the heart of the model, the outer perimeter forms the
performance factor. Performance is the umbrella term for excellence and includes
profitability and productivity as well as other non-cost factors such as quality, speed,
dependability, delivery and flexibility. In relation to performance, productivity and
profitability are more specific concepts related to the input-output ratio. Performance
encompasses a broader term, which covers both overall economic and operational aspects.
Slack, Chambers and Johnston (2001) nevertheless suggest that the performances that each
organisation strives to achieve are very case specific. Due to the overly generalised definition
of performance, measurements of performance may be unclear or unspecific as to which
13
aspect of performance one is measuring, even though at times only one facet is of particular
interest.
Forming the middle perimeter of the triple P model is the construct of profitability. It
cannot be denied that productivity and profitability are interdependent; however, they do not
always go hand in hand (West, 1999). In general terms, profitability appears to be the goal
for the success and growth for any given business, which can strictly be defined as the ratio
between cost and revenue or assets versus liabilities (West, 1999). Profitability focuses on
the monetary input-output relationship which influences the price factors such as price
recovery. Consequently, profitability does not address the direct issues of interest in this
study and thus will not be considered further here.
Lastly, productivity forms the central part of the triple P-model. Tangen (2004)
suggests that productivity is a multi-dimensional term, for which the definition may vary
depending on the context in which it is used. Despite the range of definitions used to describe
productivity, there appears to be common underlying characteristics that are embraced by the
term. The features tend to follow the classic definition of productivity, which is described as
the ratio of the quantity and quality of units produced to the labour per unit of time (Bolton &
Drew, 1991).
The triple P model clarifies the definition of productivity and allows for a better
understanding of the differences and similarities with its confounding neighbouring terms. It
is, however, difficult to measure productivity using the method explained in the model,
especially due to the wide range of occupations which will be incorporated into the current
study. According to Koss and Lewis (1993), the way productivity is defined and interpreted
depends on the context in which it is used. As a result, when comparing work outputs using
objective evaluations, it is necessary to consider the occupation as well as the level in the
organisation in which the employee is working. This can then raise questions about the
meaningfulness of productivity assessments.
For instance, jobs that have productivity levels that can be defined using only the
quantity of service or products delivered on a daily basis, assessments of productivity can be
easily evaluated by a third person such as a superior or colleague. However, in many other
occupations, results of work input are often only observed on a longer term basis. In those
14
circumstances, work output may be results of inputs accumulated over many days rather than
a one day input-output routine (Kupers, 1998). For a salesman, the number of
products/services sold daily may predict their daily productivity which can be easily assessed
by a third person. On the other hand, occupations such as architects, auditors and attorneys
operate on longer term input-output system where the fruits of their input are usually only
observed over a few or many days, if not longer. Furthermore, their outputs may not be
tangible.
Therefore, to objectively assess the daily productivity levels of those who operate on
the longer term input-output system would be inappropriate (Kupers, 1998). Thus for the
purpose of the current study, self-reported productivity measures were considered to be more
useful as daily productivity was measured over a period of time. It is important to note that
previous research has mostly utilised superior or management evaluations in terms of
performance and productivity levels. Thus, it was interesting to change the perspective to
consider subjective evaluations. This is because an individual may have different views from
their superiors or colleagues about how productive they have been. This is especially true in
occupations where results of their work cannot be discretely counted. Furthermore, it is
assumed that employees themselves are clearer about what they have and have not been
doing, which contributes to how productive they see themselves (Vora, 1992).
According to Broman?s (2004) definition, productivity can also be examined by
considering the relationship between the actual and potential output of a process. For the
purpose of the current study, the application of this concept seems reasonable, since the gap
between actual and potential productivity levels can be evaluated or judged by the individual
as well as by others. The current study will be using self evaluations. Therefore, the higher
the reported actual productivity level, the closer the individual feels to achieving their
optimum potential. The potential or optimum productivity level that the individual relates to
their actual productivity level is, however, only in relation to that given instance. This is
because the individual is only reporting their actual productivity level in relation to the
potential optimum level they are working at in the given or immediate environment. They
may feel that their potential to work at an even higher productivity level could increase if all
the necessary resources are available or if they were working in a different environment. For
example, an employee such as a receptionist working in an organisation on an open day may
be very distracted by the visitors walking in and out and constantly asking questions. On that
15
particular day, the receptionist may not be as productive as every other average day at work
but she may still be working as productively as she could, given the circumstances. As a
result, this concept allows for contextual flexibility.
2.3.3. Studies on productivity
Having considered the definition of productivity, this section will now look at some previous
studies which on productivity. The links between productivity and human well being are of
interest to many kinds of social contexts. Argyle (2001) points out that there are mixed or
inconsistent evidence to show that cognitive states affect worker productivity. For instance,
work by Wright and Staw (1999) examines connections between worker affect and
supervisors? ratings of workers. Depending on the affect measure, the authors find interesting
but mixed results. Amabile, Barsade, Meuller and Staw (2005) uncovers evidence that
happiness appears to provoke greater creativity. Lastly, Paterson, Warr and West (2004)
detect influences of emotion and affect upon productivity while Sanna, Turley and Mark
(1996) suggests that those individuals who are experiencing a negative emotion put forth the
most effort.
There is an analytical literature by economists that it examines the interconnections
between psychological forces (in particular, biased perception) and human productivity. The
research done by Benabou and Tirole (2003) focuses on the interactions between self-
deception, malleability of memory, and ability and effort. The authors consider the possibility
that self-confidence enhances the motivation to act, which in essence can be related to the
idea that there can be a connection between affect and productivity.
Therefore, there are various psychological traits that can be combined in very
complex ways to predict productivity. However, across the various traits typically required of
employees, emotions may be one of the more important factors in understanding productivity
as employees are said to seldom carry out work activities in an objective and cold cognitive
conditions (Clements-Croome, 2000).
16
2.4. Emotions and self-reported productivity
Having gained a better understanding of both the constructs, this section will now delve more
deeply into the proposed relationship between emotions and productivity as suggested by the
AET. One of the more popular theory which illustrate this suggested relationship by the AET
is the happy productive worker hypothesis (Staw & Barsade, 1993).
2.4.1. Happy productive worker hypothesis issues
Managers and researchers have long believed that the happy worker is a productive one (Staw
& Barsade, 1993). Various welfare and human resource programmes have been based on the
happy productive worker hypothesis. These programmes were attempts made by management
to reduce labour conflict and improve production rates (Wright & Staw, 1999). Despite
considerable research on the subject, uncertainty remains today as to whether happier
workers are indeed more productive (Wright & Cropanzano, 2004). Researchers have
suggested that inconsistent findings linking happiness and productivity may be due to
inconsistent measurement. Most often in studies of happiness and productivity, happiness has
also been operationalised as job satisfaction (Brief & Weiss, 2002). However, job satisfaction
may not be an effective proxy for happiness as previously discussed.
In recent years, researchers began to conduct more careful analyses of existing
studies. They developed new paradigms for testing the relationship and modified the original
job satisfaction-productivity hypothesis. For example, Judge et al. (2001) suggested that the
meta-analysis of Iffaldano and Muchinsky (1985), which found an average correlation of .17
between satisfaction and performance, underestimated the true correlation. In their updated
meta-analysis, Judge et al. (2001), found a higher correlation of .30. Both these correlations
are, however, weak and on the low side of average. Other researchers have suggested that
positive affect and positive emotions are more likely to make workers more productive than
job satisfaction (George & Brief, 1992).
In relation to the happy productive worker model, emphasis is only placed upon the
effect of happiness on productivity. This may be problematic as it only emphasises one of
many possible emotions. This is not to say that it is not a good predictor, as Oswald, Proto
and Sgroi (2008) found a positive relationship between happiness and productivity.
17
However, it is seen to be limiting to explain productivity in relation to other emotions, such
as anger. It is only able to relate emotions to productivity in terms of how happy a person is
feeling and not with any other emotion. As a result, the explanation on the emotional side of
the hypothesis is rather narrow and confined. This is supported by Lucas and Diener (2003)
who state that after a careful analysis of the nature of emotions, it is seen that emotions can
play a much more complicated role. It depends on the way people approach the world and the
specific actions people take in response to the events. Frijda (1999), for example, views
emotions as more than just a feeling of pleasantness or pain combined with an appraisal of an
object or event as good or bad. He argues that emotions have three additional components,
namely action readiness (readiness for changes in behaviour towards the environment),
autonomic arousal and cognitive activity changes. Although each of these components is
elicited in reaction to some stimulus, they also prepare people to deal with the stimulus in a
specific way. Fear and anger, for example, have distinct patterns of action, autonomic
arousal, and cognitive activity changes. People will act very differently towards an
unpleasant stimulus depending on whether they feel fear or anger. In essence, Frijda (1999)
suggest that human behaviour is related not only to how pleasant people feel but also by their
level of emotional arousal.
When applying the emotions theory, as discussed in the previous section to the happy
productive worker hypothesis, happiness can only be considered in relation to high
pleasantness. This is because high or low emotional arousal does not necessarily give a clear
indication of whether the person is happy. Yet, a study conducted by Dickman (2002) found
that peak levels of productivity were largely motoric in nature, indicating that arousal was
positively correlated to productivity. Therefore, both arousal and pleasantness is suggested to
be positively related to productivity.
2.5. Implications of previous research
A constraint identified in many emotions and productivity research is that participants are
often limited to those in managerial positions. This is because researchers felt that the work
of managers is relatively unstructured and is highly subjected to influence by the person
occupying the role (Staw & Barsade, 1993). Though this may be true, not all employees at
levels lower then management are nowadays working in what use to be called ?machine
paced? or highly routine roles. This is especially true due to the vast technological
18
advancement in the workplace. In many occupations such as a draftsman in an architectural
firm or a trainee accountant in an accounting firm, employees are working in positions which
are lower than managerial positions, but have the flexibility to complete their work in a given
time. In addition to their flexibility they also have access to resources such as the internet to
help them complete their tasks on time and more efficiently. Consequently, this study did not
limit participants to only employees from specific organisational levels. All levels of
employees were invited provided they fulfilled certain criteria which will be discussed in the
next chapter.
Lastly, it is also worth noting that the majority of the research in this area has been
conducted in other countries. In South Africa, this is a considerably under researched area.
Thus, it would be interesting to explore whether the relationship can be influenced by the
context or environment one is in.
2.6. Summary
In summary, this research is relevant on a number of levels. The purpose of this study lies
further than only reintroducing the importance of emotions into the workplace. The main
focus of the current study is therefore to re-examine the happy productive worker theory by
broadening the happiness construct and look at the relationship between emotions and self-
reported productivity. The need to re-examine this relationship is also supported by the
previous misunderstanding and/or of constructs as well as the employee positive emotion
favourable outcome model, which in the current study would be increased self-reported
productivity (Staw et al., 1994). Furthermore, due to contextual factors, it made sense to
assess the association between emotions and productivity using subjective measures, where
productivity was rated in relation to the optimum level.
In addition, Ashkanasy et al. (2002) highlighted the need to consider within subject
effects. Thus, the current study examined the patterns of emotions and self-reported
productivity over time through a repeated measures design. This allowed for the assessment
of within subject effects in terms of both emotion and productivity constructs. The
relationship between emotions and self-reported productivity was also examined over time.
By this it is meant that this relationship will be examined cross-sectionally at different times
of the day over a period of time. However, this is not to suggest that it was anticipated that
19
there would be a difference in this relationship at the various times of the day, but rather to
use it as a means to trace patterns of this relationship as well as emotions and self-reported
productivity as their own constructs over time.
Furthermore, it has been heavily emphasised that emotions are related to the events in
which individuals are exposed to. As suggested by the AET model, emotions change in
relation to the hassles and uplifts a person is exposed to (Ashkanasy et al., 2002). Thus, it
makes sense to also examine the relationship between emotions and their immediate events
to understand the context in which emotions are experienced.
2.7. Research Questions
The following research questions stem from the rationale explained in this section and
will be addressed in the results and discussion chapters.
o Is there a relationship between reported events and emotions?
o Is there a relationship between emotions and self-reported productivity?
? Is there a relationship between arousal and self-reported productivity?
? Is there a relationship between pleasantness and self-reported
productivity?
? Is there a relationship between the interaction of arousal and
pleasantness with self-reported productivity?
o Are there any changes or patterns of emotions over time?
o Are there any changes or patterns of self-reported productivity over time?
o Is there a pattern in the relationship between emotions and self-reported
productivity over time?
20
Chapter 3
Methodology
3.1. Introduction
The previous chapter discussed the literature on emotions and productivity. In this chapter
the methodology of the research will be discussed. This will include an explanation of the
research design, the procedure that was undertaken in order to carry out the research, the
participants that formed part of the research, how the data was collected, how the data was
analysed and finally, various ethical considerations pertaining to the research.
3.2. Research design
This research was carried out quantitatively and took the form of a longitudinal non-
experimental research design.
3.3. Procedure
The current study formed part of a larger project where three other researchers also examined
the role of emotions in the workplace. However, each researcher explored emotions in
relation to a different construct. In light of this, each researcher nevertheless recruited
participants that would be pooled together to form a larger sample. Therefore, all participants
were required to complete questionnaires designed to measure their emotional states as well
as the four additional constructs that are specific to each researcher.
The criterion to decide whether individuals could participate in the study was that had
to have access to their e-mail. This is because participants were required to complete some
questionnaires during the day and some in the evening through accessing their email. This
will be further explained in the next section. As a result, for the purpose of this study, the
sample was limited to working adults.
Since the current study forms part of a larger research project, an integration of the
questionnaires from each researcher was first compiled. Necessary questions from every
researcher of the larger research were put together for each of the particular questionnaires.
The integrated questionnaires were constructed in such a way as to create a better flow to
21
facilitate a good understanding of the questions and to ensure that there were no repetitive
questions. As a result, five separate questionnaires were compiled. The first questionnaire
(Refer to appendix E) compiled was the biographical questionnaire used to collect
demographic information. The other four questionnaires (Refer to appendix F-I) consisted of
specific questions that were intended to answer the main research questions of interest to the
current study as well as those of the other researchers. All four of these questionnaires were
self-report surveys.
After the assembly of the final questionnaires, a distribution schedule (Refer to
appendix J) was constructed. The distribution schedule incorporated the formulation of the
particular days and times on which the particular questionnaires were to be sent out to
participants. Participants were expected to complete questionnaires as soon as possible. It
was decided to carry out the study over a three-week period. During these three weeks only
two or three days of the five working days throughout each week were used to collect data.
The selected days were patterned in this way to evenly spread the days of data collection
between weeks. Times were set according to an estimation of the general working hours
across organisations.
As such, the first questionnaire (Refer to appendix F) was set for 10H00. This is
because a time lapse was required for the participants to be able to answer various questions
in the first questionnaire. For example, since the study partially aims to assess the level of
productivity of the participant as well as the particular event that may be attributed to the
reported productivity and emotional levels, these can only be evaluated provided participants
have already spent some time at work, for experiences or events to take place. This is based
on the assumption that by 10H00, participants have already been at work for some time.
The second questionnaire (Refer to appendix G) was set to be around lunchtime and
the third (Refer to appendix H) just before day end. These two questionnaires were set at
these times in order to provide sufficient time for potential variations in responses from the
previous questionnaires. The last questionnaire (Refer to appendix I) was set for 19H00 to
examine the emotions and self-reported productivity levels of participants who brought work
home.
22
Once the questionnaires and schedule were put together, questionnaires were posted
onto a web-based data collection database, Survey Monkey. From here onwards, the study
can be said to be divided into two stages. The first stage refers to the sampling process as
well as to the session where demographic data was gathered together with the collection of
participant?s consent. The second stage is the period whereby the core data of the study was
collected. Further details of each stage will be discussed next.
3.3.1. Stage one
After the compilation of the questionnaires, the sample was recruited via a snowballing
technique. The recruitment began with the use of personal contacts. Given that this study
forms part of a bigger project, each of the researchers involved in the study aimed to recruit
approximately 15 participants based on the criteria that they had to have access to their email
throughout the day. Thereafter, each researcher from the larger project met individually with
each of their recruited participants to explain the purpose and requirements for administering
the questionnaires. Information such as the longitudinal nature of the study and the expected
participation was explained. Furthermore, a participant information sheet (Refer to appendix
K) and the distribution schedule were also provided to participants during this session, where
the details of the study were more vividly explained and consent to participation in the study
was requested.
Once permission was granted, participants were sent an email that contained a link
which would redirect them to Survey Monkey where they would be able to complete the
initial survey (Biographical questionnaire). Incorporated in this survey was also an example
of the questionnaires that the participants could expect to be receiving in the following
weeks. This trial allowed the participants to familiarise themselves with how to answer the
questions; any queries that they may have had in that regard may be clarified before the
actual data collection begins. Based on this trial, it was seen that the self-report
questionnaires took no more than five minutes for participants to complete. Participants were
also requested to respond as soon as possible, from when the emails were received.
Responses to this initial survey were recorded and returned electronically to the
Survey Monkey database. Upon completion of this survey, the email addresses of the
participants were captured onto Survey Monkey and each participant was then allocated a
23
reference number. This allowed the researcher to ensure anonymity of the participants as the
personal details of the participants from this set of data was deleted by an independent
person. The same procedure was carried out for future questionnaires. This reference number
was then used for all further correspondence. Once all participants completed the initial
survey, it was then possible to progress to stage two of the study.
3.3.2. Stage two
Stage two of the study is defined by a three-week period where the four self-reported
questionnaires were sent out according to the above discussed distribution schedule. These
questionnaires were repeatedly distributed to the participants at the specified times
throughout the three weeks. Each questionnaire was completed after automatically generated
emails were sent out to each participant. These automatically generated emails served as
notifications to remind the participants when to complete the questionnaires. Also, the email
contained a link or web address to redirect the participant to the site to complete the
questionnaire. All data collected in the three weeks, as previously stated, were recorded on
the Survey Monkey database with anonymity guaranteed. Thereafter, email was the primary
source of communication with participants throughout this period.
Recorded responses for all the participants that were gathered separately by each
researcher were then combined and used for data analysis and conclusions were made. When
all the participants were put together the total sample came to 64 participants. Though 64
participants were recruited, due to the longitudinal nature of the study, not all participants
completed all the questionnaires that were scheduled or sent through. Consequently only
participants who completed four or more questionnaires were included in the study. As a
result, the final sample amounted to 49. Although it is not ideal to include those participants
who only completed four questionnaires, this amount was sufficient to look at the variability
of the constructs over time (at least a day), given the difficulties experienced in the sampling
and data collection process. Thus, this option appeared to have been the most pragmatic
solution.
24
3.4. Sample
Looking at table 1, the final sample consisted of 49 participants of which 21 were females
and 28 were males. On average, 14 observations were recorded for each participant.
Participants who were included in the sample had to have completed a minimum of two
questionnaires. The maximum questionnaires that could have been completed are 32
questionnaires.
The mean age of the sample was 27.14 with a standard deviation of S=7.94. The
participants? ages ranged between 21 and 62, which was relatively young, with 63 percent of
them younger than the mean age. The sample was therefore skewed and does not fulfil the
assumption of normality.
In terms of race, 22 participants were White, seven were Black, two were Coloured
and 19 were Asian. Thus, the majority of the sample were White and Asian participants. It
should also be noted that 76 percent of the sample was English speaking and as such,
language barriers shouldn?t have posed any problems to the results of the current study.
Although not everyone had English as their first language, each participant was still literate
enough to understand the language and subsequently answer the questionnaires.
Twenty five percent of participants had completed secondary education while 74
percent also had some tertiary education. Only one participant (2%) held an education level
below matric. Twelve participants (24.5%) matriculated and did not further their education.
73.5 percent of the participants went on to further obtain a diploma (12.3%), degree (59.2%t)
or some other form of higher education (2%).
Lastly, the majority (86 %) of the participants were single (n=43). This can be related
back to the age range of the sample. Seeing as the sample was relatively young, it is
reasonable to find that majority were single or not yet married.
25
Table 1. Descriptive statistics of the sample
Variable Category Frequency
Cumulative
Frequency Percentage
Cumulative
Percentage
Gender Female 21 21 42.86 42.86
Male 28 49 57.14 100
Race White 22 22 44.90 44.90
Black 7 29 14.29 59.19
Coloured 2 31 4.08 63.27
Asian 19 49 38.78 100
Age 21 1 1 2.04 2.04
22 5 6 10.2 12.24
23 5 11 10.2 22.45
24 9 20 18.37 40.82
25 12 32 24.49 65.31
27 6 38 12.24 77.55
28 3 41 6.12 83.67
29 1 42 2.04 85.71
30 1 43 2.04 87.76
31 1 44 2.04 89.8
33 1 45 2.04 91.84
34 1 46 2.04 93.88
39 1 47 2.04 95.92
59 1 48 2.04 97.96
62 1 49 2.04 100
Marital Single 43 43 87.76 87.76
Status In a relationship 2 45 4.08 91.84
Married 4 49 8.16 100
26
3.5. Instruments
Demographic details were collected in order to summarise the sample. A short biographical
blank was distributed to each participant to collect this information. The kinds of questions
used to summarise the data were, ?What gender are you?? and ?How old are you??. These
questions are sufficiently general so as to ensure the anonymity of participants.
Emotions were assessed by means of the affect grid (Refer to appendix B), developed
by Russel et al. (1989). The affect grid is composed of two dimensions. The horizontal axis
represents the pleasant-unpleasant dimension and the vertical axis measures the level of
arousal. Each of the above dimensions was measured on a nine point scale. This would then
form a grid containing 81 blocks. Participants were required to report their emotional state by
rating how pleasant they feel on the x-axis and their level of arousal on the y-axis on the grid.
Due to the structure of this scale, item analysis and assessment of internal consistency
are, of course, impossible with a single-item scale. Assessment of the reliability of the affect
grid was therefore determined indirectly through an interjudge reliability assessment by the
developers of the scale. Validity of the scale was therefore determined by the upper bound of
the reliability value. It is difficult to calculate the reliability of the affect grid, especially for
the self-reporting purposes, because firstly the assessment of interjudge reliability would
require that a second judge have access to the subject's mental state, which is impossible.
Secondly, a test-retest estimate of reliability, although possible, would be inappropriate for
any scale, including the affect grid, that measures states that can change quickly. As a result,
the interjudge reliability assessment revealed an estimated reliability of .91 and .81 for the
pleasantness and arousal dimensions respectively (Russel et al., 1989).
Productivity was measured using a single item, self-reported questionnaire. This one
item was derived from a study conducted by Wyon, Tham, Croxford and Oreszczyn (2000).
The original item was used to rate how well employees felt that they have worked during the
day in relation to their full capacity on a scale of 0-100 percent. The context in which this
item was developed examined the effects of health on self-estimated productivity. Although
this item was developed for a somewhat different purpose it could still be used for the current
study, as the daily productivity level of employees was also compared to their full capacity.
27
The main difference was that instead of assessing productivity in relation to health-related
factors, the current study examined self-reported productivity in relation to emotions.
In addition to the emotion and self-report productivity scales that were used, other
related questions were also asked. These questions involved exploring the possible factors
that may contribute to the reported emotional state and self-reported productivity. The
purpose of these questions were to investigate whether there is a relationship between
reported events and emotions as the theories suggests. Thus the question, ?Please specify a
particular event, person, object or any other aspect of work that may be attributed to the
reported emotion or vice versa? was asked in each questionnaire.
3.6. Analysis
To examine the relationship between emotions and self-reported productivity, a linear model,
which is able to accommodate the longitudinal design, was required. In particular, the design
had a repeated measures structure in that there were multiple measurements for each
individual. However, as the time period did not reflect interventions or treatment effects, the
model assumed an autoregressive correlational structure. This assumes that observations for
the same person that are closer together in time will be more similar then those further apart.
The analysis for the current study was done through an extension of a regression
equation. This extension is called the Hierarchical Linear Model (HLM) which was designed
by Bryk and Raudenbush (1992). The HLM provides a conceptual and statistical mechanism
for investigating and drawing conclusions regarding the influence of phenomena at different
levels of analysis (Hofmann, 1997). The HLM was designed to alleviate the inferential
problems associated with using statistical techniques that view hierarchical data on a single
level (Kreft, De Leeuw & Van der Leeden, 1994). One of the primary advantages of
hierarchical models is that they allow one to simultaneously investigate relationships within a
particular hierarchical level, as well as relationships between or across hierarchical levels. In
order to model both within and between level relationships, one needs to simultaneously
estimate two models: one that models relationships within each of the lower level units, and a
second that models how these relationships within units vary between units. This type of
two-level modelling approach strictly defines the hierarchical linear model.
28
This model was chosen over the ordinary least squares (OLS) regression because of
the issues around independent observations. According to Osborn (2000), hierarchical or
nested data tend to be more homogonous than when data for each person are independently
measured from a larger selection. In the current study, homogeneity is evident particularly
because observations are collected repeatedly with the same individuals, and are thus not
fully independent.
Most analytic techniques, like the OLS, hold independence of observations as a
primary assumption (Osborn, 2000). Osborn (2000) suggests that since this assumption is not
met in hierarchical data sets, using an analysis like the Ordinary Least Squares regression,
would result in standard errors that are far too small. As a result, this in turn will result in a
greater probability of rejecting the null hypothesis. Since the HLM does not hold
independence of observations as an assumption, it would therefore serve as the best statistical
technique to use.
There are a number of properties of the current study that motivate the use of the
HLM. As previously mentioned, the current study uses a repeated-measures approach. Since
the current study is repeatedly collecting information from the same subjects, the information
will therefore become hierarchical in nature. Consequently, the analysis for this type of data
would be commonly approached through the HLM and Repeated Measures ANOVA.
However, unlike a typical Repeated Measures ANOVA, the current study is not interested in
repeated measures in different conditions. Rather, the primary focus of the current study is on
multiple readings of the same information. Thus, the HLM appears to be more appropriate
than a Repeated Measures ANOVA.
Due to the repeated measures structure the current study relates variables at different
levels. At one level, differences are observed between people. At the second level,
differences are observed within subjects by means of the multiple observations collected over
time. The purpose of the HLM is to analyse hierarchical data on two levels. And since the
current study aims to analyse emotions on the levels of people and observations, the HLM
would be more statistically appropriate.
Furthermore, the random effects are also estimated for each subject in the HLM
(Verbeke & Molenberghs, 2000). This was therefore a suitable analysis to use because it did
29
not limit a set interval between the observations in a within subjects design over time; this is
characteristic of this study especially with the delay questionnaire distributions experienced
during data collection.
From the above, it is therefore evident that the HLM was the most suitable analytical
technique to fulfil the purpose of the current study. It should be noted, however, that the
content and processes of the model are complex. Thus, assistance was obtained in conducting
and interpreting the analysis of the study.
Over and above the HLM, additional analyses were also conducted. Simple
correlation and t-tests were done to consider the relevance and potential effects of the
demographic information. The HLM as well as the other analyses were completed using a
computerised statistical program called SAS (version 9.2). SAS (version 9.2) analysed the
data using the Proc Mixed procedure which essentially is SAS? (version 9.2) implementation
of the HLM.
3.7. Ethics
The presence of any unethical issues in a research process was considered unprincipled to
continue with the study until all the issues had been resolved. All actions taken in the current
study were carefully monitored to ensure that they were within ethical boundaries. For ethical
reasons, extra care was taken to ensure confidentiality. After the completion of the research
study, a summary of the results was given to the participants. For ethical reasons,
confidentiality was assured, as no individual results were reported. Due to the use of the
snowballing technique, permission was also obtained from the snowballed participants for all
the other researchers from the larger project to access their information from the
questionnaires. Furthermore, data from the study will be kept for two years before it will be
destroyed.
In addition, minimal risk was posed towards participants; research was only carried
through with the participants who provided informed consent and no deceptive techniques
were used. Also, an information sheet was provided to participants before their decision to
participate was requested. The information sheet indicated the exact intentions of the
30
research, as well as all the other necessary details, to the participants. Lastly, participants
were also informed of their right to withdraw from the study at any point.
3.8. Summary
The current study involved a complex quantitative methodology utilising a repeated measures
design which involved data collection across different times of day and different days of the
week over a three-week time period. The sample was collected through a snowballing
technique via the four researchers from the larger project. Various ethical issues that were
relevant to the research were addressed.
31
Chapter 4
Results
4.1. Introduction
In order to assess whether emotions are related to self-reported productivity over time, the
primary analysis used was a time series analysis. Due to the repeated measures design of the
current study, the data was analysed using the HLM and repeated measures ANOVA. These
analyses allowed for the measures the differences ?within? subjects through the multiple
observations collected over time. In addition, t-tests, correlations and nonparametric tests
were also performed. Summarised results for these tests are illustrated below.
4.2. Demographics in relation to emotions and self-reported productivity
It is useful to firstly examine whether any of the demographical characteristics may in some
ways be related to the main variables in the current study. In view of the fact that the
majority of the demographic variables are nominal data, one way ANOVAs were conducted
to assess whether there is a difference in arousal, pleasantness and self-reported productivity
levels between the different demographical groups. Results for the analysis are shown in
table 2.
Table 2. One way ANOVA: difference between the demographic groups and their level
of arousal, pleasantness and self-reported productivity
Variable Effect DF F-value P-value
Gender Arousal 1 .71 .40
Race Arousal 4 3.31 .02*
Marital Status Arousal 2 10.93 <.0001*
Gender Pleasantness 1 .06 .80
Race Pleasantness 4 11.01 <.0001*
Marital Status Pleasantness 2 5.44 .004*
Gender Self-reported Productivity 1 8.32 .004*
Race Self-reported Productivity 4 9.66 <.0001*
Marital Status Self-reported Productivity 2 3.02 .05*
*p < 0.05
32
From table 2 it can be seen that there is a significant difference in the level of arousal
and pleasantness across the racial groups and between those who do and don?t have partners
or spouses. In terms of self-reported productivity, there appears to be a significant difference
in the self-reported productivity level between gender, racial groups as well as different
marital status groups.
4.3. Is there a relationship between reported events and emotions?
With regards to the main variables of the study as previously mentioned in the literature
review, the nature of emotions is characterised by its relation to some event. This is what
distinguishes it from moods. Consequently, the current study tested this assumption by
categorising the reported events into either some or no event attributed by the reported
emotions according to the participant?s subjective view.
Table 3. Two sample t-test for emotions and self-reported productivity level between the
presence or absence of a significant event.
Variable DF T-value P-value
Self-reported Productivity 1124.80 -.80 .42
Arousal 1126 .71 .48
Pleasantness 1126 1.64 .10
*p < .05
Due to the diverse responses to this question, it was decided to categorise the events
in terms of some or no event as opposed to person and object related events as discussed
previously (Staw & Barsade, 1993). A two sample t-test was therefore used to test whether
there was a difference between those who reported some and those who reported no event
that can be related to their emotion or self-reported productivity level. However, as shown in
table 3, there is no significant difference in the emotional and self-reported productivity
levels between those who do and don?t relate these factors to some event.
33
4.4. Emotion and self-reported productivity
To understand the relationship between emotions and self-reported productivity, the analysis
was divided into three parts. Firstly, the relationship between the two emotion constructs,
arousal and pleasantness was looked at individually in relation to self-reported productivity;
then, the relationship between the interaction of both the emotions constructs and self-
reported productivity was considered. Thereafter, the patterns for the changes in emotions
and self-reported productivity as well as the relationship between the two constructs over
time will be considered. The following research questions will be addressed in this section:
o Is there a relationship between emotions and self-reported productivity?
? Is there a relationship between arousal and self-reported productivity?
? Is there a relationship between pleasantness and self-reported productivity?
? Is there a relationship between the interaction of arousal and pleasantness
with self-reported productivity?
4.4.1. Is there a relationship between arousal and self-reported productivity?
A simple correlation was performed to assess whether there was an overall relationship
between emotions and self-reported productivity. Results shown in table 4 indicate that there
was a significant relationship between arousal and self-reported productivity. The
relationship found between arousal and self-reported productivity is positive but weak. An
almost significant relationship was found between pleasantness and self-reported
productivity.
Table 4. Correlation between emotions and self-reported productivity
Variable DF
Corr.
Coeff. (r) F-value P-value
Arousal 1075 .18 14.34 .0002*
Pleasantness 1075 .06 3.76 .05
*p < .05
34
4.4.2. Is there a relationship between pleasantness and self-reported productivity?
Furthermore, in table 4, it is seen that there is a positive but weak relationship (r = .06) was
found between pleasantness and self-reported productivity.
4.4.3. Is there a relationship between the interaction between arousal and pleasantness
and self-reported productivity?
Though the emotions constructs were measured separately, it would nevertheless be
interesting to test the effect of emotions on self-reported productivity as one construct. The
interaction between arousal and pleasantness is key to the understanding of emotions in this
study. Therefore, the effect of the interaction between arousal and affect on self-reported
productivity was assessed by means of a multiple. The results are presented in table 5. As
such, a simple correlation was performed to assess the relationship between self-reported
productivity and the interaction between arousal and pleasantness. As a result, it is found that
there is an almost significant relationship between self-reported productivity and the
interaction between arousal and pleasantness.
Table 5. Multiple regression to assess relationship between interaction of arousal and
pleasantness with self-reported productivity
Variable DF F-value P-value
Arousal *Pleasantness 1073 3.66 .06
*p < .05
4.4.4. Test for multicollinearity.
It was unclear why no significant relationship was found in the overall relationship between
pleasantness and self-reported productivity, as there was a significant relationship between
arousal and not with the interaction between arousal and pleasantness. It was therefore useful
to test whether there was an occurrence of multicollinearity where arousal may be related to
pleasantness by using a simple correlation. A correlation was therefore used to test for this.
35
Table 6. Effects of multicollinearity
Variable Estimate F-value P-value
Arousal*Pleasantness .32 123.73 *<.0001
Pleasantness*Self-reported Productivity .62 9.78 *<.002
*p < .05
Results from the correlation shown in table 6 indeed indicate that there was a
significant positive relationship between arousal and pleasantness. Thus, when excluding
arousal from the analysis, a significantly strong positive relationship between pleasantness
and self-reported productivity was found.
4.5. Longitudinal results
Having examined the cross-sectional relationship between emotions and self-reported
productivity, the following sections will now consider the within subject effects and the
patterns of each construct over time. To introduce the longitudinal design of this part of the
study, the change in emotions and self-reported productivity over time will first be
considered followed by a more detailed exploration of relationship between constructs at the
different times of the day. Thus, this section will address the following three research
questions:
o Are there any changes or patterns of emotion over time?
o Are there any changes or patterns of self-reported productivity over time?
o Is there a pattern in the relationship between emotions and self-reported productivity
over time?
4.5.1. Patterns in emotions and self-reported productivity over time
To assess this, the average arousal, pleasantness and self-reported productivity levels were
graphed to examine the potential for any type of pattern. The graphs for each variable are
illustrated in figures 1 to 6. The change in each variable was examined from a broad weekly
overview to more specific times of the day.
36
Figure 1. Change in emotions (arousal and pleasantness) over three weeks
Figure 2. Change in self-reported productivity level over three weeks
Figure 1 and figure 2 show that there is little change in both emotions constructs as
well as self-reported productivity over the three weeks. All three constructs virtually stay at a
consistent level with little change throughout the weeks.
37
Figure 3. Change in emotions (arousal and pleasantness) on a daily basis
Figure 4. Change in self-reported productivity level on a daily basis
Figure 3 and figure 4 also show that arousal, pleasantness and self-reported
productivity hardly change on a daily basis. Although there appears to be some change in
arousal and pleasantness in the last few days as seen in figure 3, the difference between the
38
highest and lowest level is less than one interval apart. Consequently, self-reported
productivity, arousal and pleasantness appear to remain somewhat steady over a weekly and
daily basis.
Figure 5. Change in emotions (arousal and pleasantness) throughout the day
* T1 = 10H00, T2 = 12H00, T3 = 16H00, T4 = 19H00
Figure 6. Change in self-reported productivity level throughout the day
* T1 = 10H00, T2 = 12H00, T3 = 16H00, T4 = 19H00
39
Figure 5 shows that arousal and pleasantness levels appear to stay somewhat the same
throughout the day. Similar to Figure 3, even though arousal appears to decrease slightly
towards the end of the day, the difference is also within a small range. The same can be said
with pleasantness where a slight increase is observed as seen on the daily graph. The self-
reported productivity level on the other hand appears to significantly decrease towards the
end of the day as shown in figure 6. But this measure was taken after working hours and is
therefore to be expected.
4.5.2. Is there a pattern in the relationship between emotions and self-reported
productivity over time?
After considering the relationships between emotions and self-reported productivity, this
relationship was re-examined longitudinally or over time as an attempt to identify patterns.
Thus, correlations were used to measure the relationship at the different times of the day
separately rather than looking at the overall relationship which was done in the earlier
section. Considering that there was a more visible change throughout the day, as shown in
figures 5 and 6, this relationship is therefore only re-examined using this time frame. Results
for this analysis are shown in table 7.
Table 7 shows how the relationship between self-reported productivity and emotions
change throughout the day. This analysis is not the same as that in figure 4 and figure 5.
Results from figures 4 and 5 show how each of the arousal, pleasantness and self-reported
productivity constructs individually change throughout the day. Table 7, on the other hand,
shows how the relationship between arousal and self-reported productivity and pleasantness
and self-reported productivity change throughout the day.
40
Table 7. Relationship between self-reported productivity and arousal and self-reported
productivity and pleasantness throughout the day.
*p<.05
Variable Time Estimate DF
F-
value P-value
Arousal*Time 1072 4.24 .006*
Arousal 10H00 .03 274 -430 <.0001*
Arousal 12H00 .03 48 19.70 <.0001*
Arousal 16H00 .004 204 -0.41 .69
Arousal 19H00 .001 170 .20 .85
Pleasantness*Time 1072 5.05 .002*
Pleasantness 10H00 .02 274 3.04 .003*
Pleasantness 12H00 .02 291 2.45 .02*
Pleasantness 16H00 .008 204 .80 .42
Pleasantness 19H00 -.02 170 -1.61 .11
As tabulated, the relationship between both emotion constructs and self-reported
productivity significantly changes throughout the day. The strength of the relationship
between arousal and self-reported productivity decreases as the day proceeds. Further, the
relationship between arousal and self-reported productivity was only significant at 10H00
and 12H00.
In terms of the relationship between pleasantness and self-reported productivity,
similar results are found. The relationship between pleasantness and self-reported
productivity also changes throughout the day. From 10H00 to 16H00, there is a decrease in
the strength of the relationship while at 19H00 there appears once again to be an increase but
in the opposite direction. Again, only a significant relationship was found at 10H00 and
12H00 between pleasantness and self-reported productivity.
_________
It is important to note that the degrees of freedom differ for each finding because not all participants filled in all questionnaires, thus the
response rate at each time is different.
41
Thus, similar findings were found between both emotions constructs and self-reported
productivity. Both relationships significantly changed throughout the day. However, only at
10H00 and 12H00 was arousal and self-reported productivity and pleasantness and self-
reported productivity the strongest and significantly related to one another.
4.5.3. Re-examining the relationship between emotions and self-reported productivity
at 10H00 and 12H00
Subsequent to only finding significant relationships between the two emotion constructs and
self-reported productivity at 10H00 and 12H00, it makes sense to re-examine these
relationships within the context of these two times. By doing so, effects at 16H00 and 19H00
would be excluded from the interaction. Consequently, the overall relationship between
arousal and self-reported productivity as well as pleasantness and self-reported productivity
was tested at 10H00 and 12H00. The interaction between time and arousal and time and
pleasantness was also tested in relation to self-reported productivity. Results for these re-
examined relationships can be found in table 8.
Table 8. Relationship between emotions and self-reported productivity at 10H00 and 12H00
only
Variable Estimate DF F-value P-value
Arousal .03 614 5.54 *<.0001
Pleasantness .02 614 4.03 *<.0001
Arousal*Time .001 613 .14 .89
Pleasantness*Time .003 613 .33 .74
*p < 0.05
As seen in table 8, a significant relationship can still be found between arousal and
self-reported productivity (F614= 5.54; p<.0001). Here, a significant relationship is also found
between pleasantness and self-reported productivity, which was previously found to be
insignificant (F1075= 3.76; p=.05) when 16H00 and 19H00 questionnaires were included.
42
4.6. Summary
The following results were found in this chapter:
Demographics in relation to emotions and self-reported productivity
o Individuals from different racial and marital status groups show a significant
difference in levels of arousal, pleasantness and self-reported productivity.
o In addition, there is a significant difference in self-reported productivity levels
between males and females.
Emotions in relation to reported events
o There is no significant difference in the emotional and self-reported productivity
levels between those who do and don?t relate these factors to some event.
Emotions in relation to self-reported productivity
o The significant relationship found between arousal and self-reported productivity is
positive but weak.
o A positive but weak relationship was found between pleasantness and self-reported
productivity.
o There is an almost significant relationship between self-reported productivity and the
interaction between arousal and pleasantness.
o There was a significant positive relationship between arousal and pleasantness. Thus,
when excluding arousal from the analysis, a significantly strong positive relationship
between pleasantness and self-reported productivity was found.
Emotions and self-reported productivity over time
o There is little change in both emotions constructs as well as self-reported productivity
over the three weeks.
o Arousal, pleasantness and self-reported productivity hardly change on a daily basis.
o Arousal and pleasantness levels appear to stay somewhat the same throughout the day
o Self-reported productivity level on the other hand appears to significantly decrease
towards the end of the day
43
o The strength of the relationship between arousal and self-reported productivity
decreases as the day proceeds. The relationship between arousal and self-reported
productivity was only significant at 10H00 and 12H00.
o The relationship between pleasantness and self-reported productivity also changes
throughout the day. Again, only a significant relationship was found at 10H00 and
12H00 between pleasantness and self-reported productivity and not at 16H00 and
19H00.
o Both relationships between arousal and self-reported productivity and pleasantness
and self-reported significantly changed throughout the day. However, only at 10H00
and 12H00 was arousal and self-reported productivity and pleasantness and self-
reported productivity the strongest and significantly related to one another.
o When re-examining the relationship at 10H00 and 12H00 excluding the results at
16H00 and 19H00, a significant relationship can still be found between arousal and
self-reported productivity. Here, a significant relationship is also found between
pleasantness and self-reported productivity, which was previously found to be
insignificant when 16H00 and 19H00 questionnaires were included.
44
Chapter 5
Discussion
5.1. Introduction
The previous chapters have addressed the literature on the topic of emotions and productivity
in the workplace, the methodology used and a presentation of the results from the study. This
chapter will discuss the results found in the study in relation to the literature on emotions and
productivity. The discussion of the results will be presented under the research questions
listed in the first chapter. Thereafter limitations of the study and recommendations for future
research will be discussed.
5.2. Demographics in relation to emotions and self-reported productivity
There have been various significant results found in relation to arousal, pleasantness as well
as self-reported productivity such as with age, race, marital status and different gender
groups. Significant differences in levels of arousal, pleasantness and self-reported
productivity was found between the different racial and marital status groups. In addition,
there is also a difference in self-reported productivity between males and females. These
findings may play an important role in understanding emotions and self-reported productivity
in the workplace. However, these findings diverge from the main focus of the study yet are
important for future researchers to expand upon.
5.3. Is there a relationship between reported events and emotions?
Moving onto the core of the research, in past studies researchers have strongly argued that
what differentiates emotions from moods is the relationship to some sort of event (Staw &
Barsade, 1993). Results from the current study have, however, found no relationship between
emotions and the described events that may have attributed to the reported emotions as
suggested by the model. Table 6 suggests that there is no difference in arousal, pleasantness
and self-reported productivity level between those who reported a significant event and those
who didn?t. Consequently, results from the current study do not support the suggested
association, by the AET model (Weiss & Cropanzano, 1996) as well as other research which
45
supports this hypothesis, between emotions and the proposed events which are attached to
them (Ashkanasy et al., 2002; Staw & Barsade, 1993).
From one point of view, this result can be explained by there being no particularly
interesting event reported on the days of data collection. Associated events (see appendix L)
were nothing out of the ordinary. Examples of reported events include ?I made some tea?,
?Just got out of a meeting? and, ?Developing concepts for ad campaigns?. In general, every
day appeared to be a typical day at work. As such, because there is nothing particularly
interesting happening at work, people may have found it difficult to associate the typical
work events with their reported emotion and productivity levels.
Alternately, this finding may suggest that there is something unusual about the data.
This is because many studies in the past have established and supported this relationship
(Lord & Kanfer, 2002; Ashkanasy et al., 2002). A possible reason for this finding is because
participants used diverse ways to describe the events and these events were in no way
categorised. There may be various ways in which to categorise these events, one of which
may to establish whether these events positively or negatively impacted on their reported
emotion. It may have assisted the researcher to find out whether positive events are related to
positive emotions and vice versa, as proposed by Lord and Kanfer (2002). This is also
supported by the AET model which suggests that events are categorised according to hassles
or uplifting events (Ashkanasy et al., 2002). This essentially denotes a positive or negative
event.
5.4. Emotions and self-reported productivity
The main research questions of the current study will now be discussed. As mentioned in the
results section, the examination of the relationship between emotions and self-reported
productivity was divided into three parts. Thus, this chapter will follow a similar format
discussing the results found in relation to existing literature.
46
5.4.1. Relationship between emotions (arousal and pleasantness) and self-reported
productivity
Research questions to be addressed in this section:
o Is there a relationship between arousal and self-reported productivity?
o Is there a relationship between pleasantness and self-reported productivity?
o Is there a relationship between the interaction between arousal and pleasantness and
self-reported productivity?
Results in the current study found a significant and positive relationship between
arousal and self-reported productivity as well as between pleasantness and self-reported
productivity (Table 3). However, the relationship between pleasantness and self-reported
productivity was only found significant after a test for multicollinearity between arousal and
pleasantness. Thus, arousal and pleasantness was found to be significantly related to each
other in the current study.
According to the affective events theory (Weiss & Cropanzano, 1996) as well as the
favourable outcome model, positive emotions lead to some increased behavioural or
favourable outcome (Staw et al., 1994), such as productivity. This study therefore does
support this hypothesis. Although a high and low score on the arousal dimension is not
indicative of a positive emotion, a high pleasantness rating on the other hand can be
interpreted as a positive emotion. As a result, the more pleasant a person feels the more
productive they tend to be. Thus, a person who experiences more positive emotions could
potentially lead to an increased level of the favourable outcome, which is productivity (Staw
et al., 1994).
In relation to the happy-productive worker hypothesis, as previously mentioned, due
to the confined representation of emotions in the traditional happy-productive hypothesis,
pleasantness and arousal were used as measures of emotions rather than just happiness to test
the hypothesis. Consequently, the current study considered the relationship between emotions
(arousal and pleasantness) and self-reported productivity. However, only the relationship
between pleasantness and self-reported productivity can specifically be used to compare and
be applied to the suggested relationship in the model. This is because pleasantness is also
more closely and/or directly associated with happiness than arousal. Consequently, with
47
pleasantness having a significant and positive relationship with self-reported productivity, the
results from the current study can be seen to support the findings of Oswald et al. (2008) as
well as the happy-productive worker hypothesis (Staw & Barsade, 1993) indicating that an
increased level of happiness can raise productivity levels.
In addition, the correlation found between pleasantness and self-reported productivity
was .62. This value is higher than the correlation value of .30 found in the study of Judge et
al. (2001), which used job satisfaction and performance as the measuring constructs to test
the happy productive worker hypothesis. This therefore indicates that, like many researchers
have suggested, it is important to carefully clarify the meanings of concepts and constructs
before using them as measures for constructs other than that of its own (George & Brief,
1992).
In terms of arousal, it has been found that a person tends to be more productive when
they are feeling more emotionally aroused. This, therefore, supports the findings of Dickman
(2002), even though there was no manipulation, where they also found a positive relationship
between arousal and productivity. Dickman (2002) found that peak levels of productivity are
largely motoric in nature. This finding also makes sense as the bipolar opposite of arousal is
sleepiness. This finding seems reasonable because when a person is sleepy he/she is less
likely to have the energy to do as much work as when they were more aroused. Again, it
needs to be emphasised that neither high nor low emotional arousal is discretely indicative of
a positive emotion. This is because the interpretation of a reported level of arousal is
contextually influenced. For example, a person who is enthusiastic and a person who is
furious can be equally aroused but the emotion of the enthusiastic individual would be
considered as a positive emotion (Russel et al., 1989).
Looking at the relationship between self-reported productivity and the interaction
between the two emotions constructs, emotions is not found to be related to self-reported
productivity. Thus, when considering the emotion construct as a whole it does not support
the findings of the AET model (Weiss & Cropanzano, 1996) as well as the positive emotion
favourable outcome model (Staw et al., 1994). However, this can be explained by the
significant relationship found between arousal and pleasantness. This is because when
looking more specifically at each of the emotions constructs individually, pleasantness and
arousal were both found to be significantly related to self-reported productivity which would
48
then support the AET (Weiss & Cropanzano, 1996), favourable outcome model (Staw et al.,
1994) as well as the happy productive worker hypothesis (Staw & Barsade, 1993).
Thus, combining these results this can be seen as an indication that though
pleasantness is important in understanding the relationship between emotions and self-
reported productivity, arousal plays a more important role. This therefore also supports the
view of Frijda (1999) who found that human behaviour is related not only to how pleasant
people feel but also by their level of emotional arousal. In addition, it also supports Lucas
and Diener?s (2003) recommendation that emotions do play a more complicated role than
only looking at happiness in relation to self-reported productivity. Russel et al. (1989) also
suggested that arousal and pleasantness may be related in certain circumstances. Russel et al.
(1989) states that although arousal and pleasantness are considered conceptually separate,
they may at times be positively or negatively correlated depending on the circumstance in
which it is used.
In addition, although arousal and pleasantness are theoretically different constructs,
they are also in some ways related. Firstly, they both fall under the broader constructs of
emotions and affect. Secondly, whether this relationship is positive or negative it is
determined by the environment and task a person is exposed to. As previously mentioned, a
person who is highly aroused does not necessarily indicate a sense of pleasantness or
happiness. However, a positive relationship found here may be explained by the majority of
responses in the current study reporting a high arousal level as well as a high pleasantness
level and vice versa. Therefore, people would tend to either feel close to excited or depressed
rather than calm or anxious as described by the affect grid (Russel et al., 1989).
In reality, however, it is important to note that although a positive relationship
between arousal and pleasantness was found, it may not always necessarily be this accurate;
sometimes they may be negatively related as well. This may also be the reason why there
was an almost significant relationship found between pleasantness and self-reported
productivity in the initial analysis. It may be because pleasantness and arousal are correlated.
As such, people who feel emotionally aroused tend to feel more pleasant than those who are
less aroused or tired and vice versa. Therefore, this somewhat supports Estes and Adelman?s
(2008) findings that emotional arousal and valence or pleasantness is consistently related as a
significant relationship between arousal and pleasantness was found throughout the day.
49
5.5. Longitudinal results
Following the format in the results chapter, this section will start off discussing the
longitudinal results by looking at the patterns of emotions and self-reported productivity over
time followed by a more detailed discussion of the patterns in relationship between the two
over time in terms of the HLM.
Research questions to be addressed in this section
o Are there any changes or patterns of emotion over time?
o Are there any changes or patterns of self-reported productivity over time?
o Is there a pattern in the relationship between emotions and self-reported productivity
over time?
5.5.1. Patterns in emotions and self-reported productivity over time
As previously mentioned, it was Ashkanasy et al. (2002) that suggested the need to examine
within subject effects. Thus, the change in emotions and self-reported productivity, as
separate constructs, over time was also examined. Yet, it is worth noting that Ashkanasy et al.
(2002) observed the within subject effects for emotions across days and not throughout the
day. The current study, however, considered the change in emotions and self-reported
productivity across weeks, days and throughout the day. The purpose of tracing these patterns
over the different time periods is to get a broader picture of the trends and hone in to get a
closer view of exactly where or when the variations took place.
From figures 1 to 4, it can be seen that there is little change in arousal, pleasantness
and self-reported productivity across the weeks and days at work. Firstly, to look at the trend
of how emotions and self-reported productivity level changes over a weekly basis, there
appears to be no significant difference in arousal, pleasantness and self-reported productivity
levels. In fact, all three constructs appear to be quite steady forming a somewhat straight line.
This may be because the activities the employees are involved in appear to be somewhat
routine. This can be supported by the events described in relation to each reported emotion
being more general activities such as making tea or attending meetings (Refer to appendix
L). A similar pattern was found across the different days of collection.
50
In terms of the change throughout the day, there were only slight changes - a slight
decrease in level of arousal and self-reported productivity and also a slight fluctuation in
pleasantness. Though this change is visible, the difference between the highest and lowest
score is small. Furthermore, the difference between the times that are closer together is also
small, e.g. the difference from 10H00 and 12H00 is just as small as the difference from
12H00 to 16H00 and from 16H00 to 19H00. This finding cannot give a clear indication of
whether to support the results in the study of Ashkanasy et al. (2002) and Stone et al. (1999)
where they found that emotions are susceptible to change on a moment to moment basis
Even though arousal and pleasantness showed little change throughout the day, it does not
imply that they are not susceptible to change. The slight deviation may be a result of a lack of
stimulating or major events occurring.
The slow decrease found with arousal and self-reported productivity throughout the
working day (Refer to figure 5 and 6) is also a reasonable finding as one can expect a person
to become more tired after a days work and thus feel sleepier or tired towards the end of the
day and also less productive. However, this decrease in arousal contradicts the findings of
Dickman (2002) who found an increase in arousal throughout the day. This may be due to the
absence of manipulation which was used in their study. On the other hand, the change in self-
reported productivity level somewhat supports the findings of Dickman (2002). Dickman
(2002) found that there was a ?post lunch dip? in productivity levels. Though there was a
decrease in self-reported productivity levels in the current study, it was only a slight change
and cannot be classified as a ?dip? or drastic change. This may be attributed to the last
questionnaire being sent through at 19H00, when most people have already left work or the
office and therefore feel more relaxed and less productive in terms of the tasks they have to
complete.
Furthermore, when comparing the results found in the work of Stone et al. (1999) to
the current study, much difference is found. It was expected that some fluctuation would
occur across and throughout the days of data collection as found in the results of Stone et
al.?s (1999). A possible reason for the difference in results between Stone et al. (1999) and
the current study is that in their study, their sample was restricted to participants who were at
management level. At management level, employees tend to become more involved with
tasks that require flexible problem solving abilities. As such, they may often find themselves
in stimulating yet challenging and more complex tasks. Results from participating in such
51
activities may motivate opportunities for arousal and self-reported productivity to change.
The sample from the current study, however, consists of employees from various occupations
and organisational levels, most of which are below management level. As such, they may be
involved in less stimulating jobs and perhaps even somewhat routine tasks. Consequent to
this, they may be less likely to experience an as dramatic change in emotions across and
throughout the day(s).
On the other hand, the pleasantness dimension is found to decrease towards midday
and then slightly increase again towards the end of the day (Refer to figure 5). Seeing that
high pleasantness can be classified as a positive affective state or emotion, this finding
therefore contradicts the results in the study of Egloff et al. (1995). Egloff et al. (1995) found
that positive moods increase from morning to evening. Although results from the current
study show that levels of pleasantness slightly decreases during the day and then increases
again towards the end of the day, the pleasantness level at the end of the day is still lower
than the level in the morning. As a result, a tentative explanation for this may be that
employee starts off the day feeling moderately pleasant and as the day goes by s/he becomes
more involved with their work and therefore becomes less concerned about how pleasant
they are feeling. This may be because people generally take less notice of when things are
going smoothly and are more attentive when something goes wrong or when something
extraordinary occurs.
Towards the end of the day, employees may be looking forward to the end of the day,
causing the rise. This anticipation may have led to the increased level of pleasantness.
Though there is a slight variation in pleasantness experienced throughout the day, the change
is also a small difference. Consequently, the above discussion with regards to Stone et al.?s
(1999) finding is also applicable to the pleasantness dimension. People may tend to feel more
pleasant when they are involved in more stimulating activities than those who are working
with more routine tasks. This would however require further investigation in future research.
5.5.2. Is there a pattern in the relationship between emotions and self-reported
productivity over time?
Looking at the AET model but in the work context, Ashkanasy et al. (2002) proposed the
need to look at within people effects. This is because previous research has found a
52
relationship between emotions and productivity on a moment to moment basis but not over a
period of time (Ashkanasy et al., 2002). This method also provides a unique approach to
reconsider the happy-productive worker hypothesis. The happy-productive worker model
was used as the basis for the current study. As previously mentioned, due to the confined
representation of emotions in the traditional happy-productive hypothesis (Judge et al.,
2001), pleasantness and arousal were used as measures of emotions rather than merely
measures of happiness to test the hypothesis. Consequently, the current study considered the
relationship between emotions (arousal and pleasantness) and self-reported productivity over
time.
When looking at the happy-productive worker hypothesis longitudinally in terms of
using emotional arousal and pleasantness, there appears to be a somewhat stable relationship
across the three weeks and days. Yet when looking more closely, there appears to be some
change in the relationship throughout the day. More specifically, the relationship between
arousal and self-reported productivity and pleasantness and self-reported productivity
becomes weaker towards the end of the day (Table 4). Results from table 4 show that only at
10H00 and 12H00 was the relationship between arousal and self-reported productivity and
between pleasantness and self-reported productivity significantly related to each other. The
relationship becomes insignificant at 16H00 and 19H00 for both relationships throughout the
day. This finding is somewhat different to the results in the study of Ashkanasy et al. (2002),
as they found that emotions and productivity was related on a moment to moment basis. This
is because the relationship found at 10H00 and 12H00 links to the findings of Ashkanasy et
al. (2002). However, the insignificant results at 16H00 and 19H00 do not support this
finding. The reason for this may be that the response rates during these two times were lower
than that at 10H00 and 12H00. Thus, it may have affected the sample size and in essence the
significance of the results.
This confounding finding suggests that there may be certain conditions or aspects in
the work environment that results in a significant relationship between the two emotion
constructs and productivity during the morning. This particular facet is however missing
during the afternoon and evening resulting in the insignificant relationship. Looking at some
of the potential reasons for this finding, firstly, there was a generally low response rate for
the last two questionnaires at 16H00 and 19H00. As a result, this would have affected the
statistical power of the results found at these two times. The low response rate may be
53
attributed to the participants possibly having already left work or being ready to leave. As a
result, how they are feeling emotionally would have little to do with how productive they are
because they may not be doing any work, as it has already passed working hours. Secondly,
employees may feel fresh in the morning which may assist them to work more productively
(hence the positive relationship). As the day goes by, people may generally become tired yet
there may still be a certain amount of work which need to be completed before they can leave
the office. Therefore, although employees may be working productively they may not
necessarily be feeling emotionally aroused. Alternately, towards the end of the day, though
people are feeling tired they may also be feeling excited to go home. On the other hand, they
may not be working productively because they are not necessarily concentrating as hard as
earlier in the day as they are now looking forward to going home.
This result, however, is only based on the data from the current study. It is important
to acknowledge that due to the unstable participation rate in the current study throughout the
days of data collection, the strength and sufficiency of the data used. Therefore, if there was a
greater participation rate especially at 16H00 and 19H00, a significant relationship may have
been found and the results of the study would have been more compelling.
5.5.3. Re-examining the relationship between emotions and self-reported productivity
at 10H00 and 12H00
Since the strongest relationship was found at 10H00 and 12H00, the relationship between
emotions and self-reported productivity was re-examined at these times only. A significant
and also slightly stronger relationship was found between arousal and pleasantness with self-
reported productivity when excluding questionnaires completed at 16H00 and 19H00 (Table
7). This indicates that the relationship between emotions and self-reported productivity may
have been significant should there be a higher response rate. This would therefore indicate
that arousal and pleasantness can be used as predictors of self-reported productivity despite
the insignificant relationship found in the afternoon and evening. Consequently, when only
observing effects at 10H00 and 12H00, results support the happy-productive worker
hypothesis because here, pleasantness is significantly related to self-reported productivity.
54
5.6. Summary
In summary, the main results in the current study can be said to somewhat support previous
findings of the happy-productive worker hypothesis where a happier worker can be described
as more productive (Staw & Barsade, 1993). However, due to low response rates at 16H00
and 19H00, the relationship between emotions and self-reported productivity was only
significant at 10H00 and 12H00. Nevertheless, this indicates that the relationship does
change throughout the day. However, it is important to note that pleasantness was only found
to be significant in the absence of emotional arousal. Yet, arousal was significantly related to
self-reported productivity alone as well as in the interaction with pleasantness. This indicates
that arousal plays a key role in determining perceptions of productivity.
55
Chapter 6
Limitations and implications for future research
As with any research there are always limitations. Acknowledging these limitations will
allow for future researchers who are interested in the field to develop on these areas and
possibly make improvements. Even though some of the problems were beyond the control of
the researcher, future researchers should make provision for the potential of a reoccurrence of
the same problems. Furthermore, they can plan on how to deal with them more efficiently
when they arise.
? The first limitation of the study is that the sample is skewed and biased in terms of the
various demographic variables such as age and racial groups. In any given study, it is
ideal to have a normal data set. As such, results achieved from this data have limited
the generalisability to represent the population.
? Secondly, the sample size is rather small and skewed. This would also impact on the
power of the statistic or results and its applicability to other contexts. It would
therefore be beneficial for the quality of future studies to gather a larger and more
representative sample. This may revive the assumption of normality of the data and
also allow the results of the study to be more widely applicable to other contexts,
which in essence will enhance the ecological validity of the study. Having said that,
the results found in the study may be slightly weak but nevertheless still usable.
? In terms of data collection, various problems were unfortunately encountered during
this period, mainly due to technical difficulties that were experienced with the Survey
Monkey server. Some of these difficulties included questionnaires that didn?t go
through, arriving late at the recipient and excluding participants from email list.
Despite the attempts to compromise the problems experienced, response rates
nevertheless fluctuated throughout the period of data collection.
To compensate for the problems experienced, various other measures were used in an
attempt to prevent potentially more serious issues from occurring. Some of the efforts
made to overcome these barriers include extending the days of data collection,
56
keeping a closer eye to see if scheduled questionnaires were sent out punctually and
frequently following up with participants on their progress and seeing into the
problems that they had. Despite the attempts to make up for problems, some of the
issues were beyond the control of the researcher. Thus, future researchers who wish
to replicate this study should be aware of these difficulties which can potentially
occur and make precautions to deal with such issues should they occur.
? Since questionnaires were sent out at 16H00 and 19H00, this appeared to problematic
in terms of response rate. This is because some participants were likely to have left
work already. Lowered response rates also would have decreased the power of the
results as well as the patterns found. Therefore, future research should be more
cautious about when to schedule questionnaires to minimise the risk of a decrease in
responses.
? Lastly, the affect grid and productivity measure may have been slightly less reliable
than a multiple item questionnaire. Although these measures were selected to
accommodate the longitudinal nature of the study, the reliability of the measure is still
questionable. This is however not to say that the measures used did not measure what
was intended to measure. It is rather a suggestion for future researchers to consider a
development of a multiple item measure that will also be suitable to use for a
longitudinal study.
57
Chapter 7
Conclusion
Though the importance of emotions in the workplace has in the past been underestimated and
has only gained prominence in the late twentieth century, the current study has again
supported the notion that emotions cannot be isolated from an organisational environment.
As in many theories and models, emotions have been suggested to be related to various
organisational behavioural outcomes. Yet in the current study, self-reported productivity has
specifically been considered. The choice to assess this behavioural outcome was inspired by
the AET and happy productive worker hypothesis. However, only using happiness to
understand an individual?s self-reported productivity level appeared to be rather limiting.
Thus, the current study aimed to better understand this relationship by expanding the
emotional construct through the use of emotional arousal and pleasantness as measures of
this construct.
A significant relationship was found between emotional arousal and self-reported
productivity as well as between emotional pleasantness and self-reported productivity.
However, when considering the relationship between arousal and self-reported productivity
and pleasantness and self-reported productivity simultaneously, only a significant
relationship was found between arousal and self-reported productivity. Yet, when assessed
individually, a significant relationship was also found between the pleasantness and self-
reported productivity. However, the relationship between pleasantness is related to self-
reported productivity is only significant in the absence of arousal. Thus, this is an indication
that arousal is key to understanding emotions in relation to self-reported productivity in the
workplace. Nevertheless, it can be concluded that the more aroused an individual is feeling
the more productive s/he tends to be. Similarly, the more pleasant a person feels the more
likely s/he is to be productive. Therefore, the current study can be seen to support the happy
productive worker hypothesis, as happiness is more closely related to pleasantness (Staw &
Barsade, 1993). This may also suggest that people who experience positive emotions tend to
be more productive than those who experience negative emotions, which supports the
employee positive emotion favourable outcome model (Staw et al., 1994). However, this is
only based on the significant relationship between pleasantness and self-reported productivity
as high or low arousal is not specifically indicative for positive or negative emotions because
it is highly contextually determined.
58
Furthermore, a unique aspect of the current study was to assess potential patterns in the
relationship between emotions and self-reported productivity through the use of a repeated
measures design. The relationship did change depending on the time of day with only the
10H00 and 12H00 time periods producing significant results.
The patterns for the two emotional constructs as well as self-reported productivity
were also measured and it was found that there was little change or deviation in the level of
emotional arousal and pleasantness. This may be attributed to the lack of stimulating events
which occurred in the workplace. Although no significant relationship was found between
reported events and emotions, a large amount of studies in the past have suggested otherwise.
Nevertheless, referring back to the patterns in change of emotional arousal, pleasantness and
self-reported productivity, arousal and self-reported productivity slightly decreased towards
the end of the day, which can be expected as individuals become more tired towards the end
of a days work thus resulting in a decreased level of arousal and self-reported productivity.
On the other hand, pleasantness was seen to slightly increase which may be explained by
individuals looking forward to going home.
Consequently, the current study has built upon existing literature on the role of emotions in
the workplace by examining the happy productive worker hypothesis through the expansion
of the happiness construct. However, there were various methodological shortcomings were
encountered. These can be used to prompt future researchers of such issues and to assist them
to provide better insight into this field of research.
59
Reference List
Amabile, T. M., Barsade, S. G., Mueller, J. S., & Staw, B. M. (2005). Affect and creativity at
work. Administrative Science Quarterly, 50, 367-403.
Argyle, M., (2001). The psychology of happiness (2nd ed.). London: Routledge
Ashkanasy, N., Zerbe, W., & Hartel, C. (2002). Managing emotions in the workplace. New
York: Sharpe.
Ashforth, B. E. & Humphrey, R. H. (1995). Emotion in the workplace: a reappraisal. Human
Relations, 48, 97- 12.
Bell, N. E., & Staw, B. M. (1989). People as sculptors versus sculptor: the roles of
personality and personal control in organizations. In Arthur, M. B., Hall, D. T., &
Lawrence, B. S. (Eds.), Handbook of Career Theory (pp. 232-251). New York:
Cambridge University Press.
Benabou, R., & Tirole, J. (2003). Intrinsic and extrinsic motivation. Review of Economic
Studies, 70, 489-520.
Bolton, R. & Drew, J. (1991). A multi-stage model of consumers? assessment of service
productivity and value. Journal of Consumer Research, 17, 375-84.
Bradburn, N.M. (1969). The structure of psychological well-being. Chicago: Aldine.
Brief, A. P., Butcher, A., & Roberson, L. (1995). Cookies, disposition, and job attitudes: The
effects of positive mood inducing events and negative affectivity on job satisfaction
in a field experiment. Organisational Behaviour and Human Decision Processes, 62,
55-62.
Brief, A. P., & Weiss, H. M. (2002). Organizational behaviour: affect in the workplace.
Annual Review of Psychology, 53, 279-307.
60
Broman, M. (2004). Assessing productivity in assembly systems. Stockholm: The Royal
Institute of Technology.
Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models. Newbury Park, CA:
Sage.
Clements-Croome, D. (2000). Creating the productive workplace. London: E & FN Spoon.
Dickman, S. J. (2002). Dimensions of arousal: wakefulness and vigor. The Journal of Human
Factors and Ergonomics Society, 44 (3), 429-442.
Domagalski, T. A. (1999). Emotions in organisations: main currents. Human Relations, 52
(6), 833-852.
Egloff, B., Tausch, A., Kohlmann, C., & Krohne, H. W. (1995). Relationship between time
of day, day of week and positive mood: exploring the role of the mood measure.
Motivation and Emotion, 19 (2), 99-110.
Ekman, P. (1994). Strong evidence in the universals in facial expressions: a reply to Russel?s
mistaken critique. Psychological Bulletin, 115, 268-287.
Estes, Z., & Adelmann, J. S. (2008). Automatic vigilance for negative words is categorical
and general. Emotions, 1, 453-457.
Flam, H. (1993). Fear, loyalty and greedy organizations. In S. Fineman (Ed.), Emotion in
organizations (pp. 58- 75). London, U.K.: Sage Publications.
Frijda, N. H. (1999). Emotions and hedonic experience. In D. Kahneman, E. Diener & N.
Schwartz (Eds.), Well-being: The foundation of hedonic psychology (pp. 190-210).
New York: Russel Sage Foundation.
George, J. M., & Brief, A. P. (1992). Feeling good-doing good: a conceptual analysis of the
mood at work organization spontaneity relationship. Psychological Bulletin, 112,
310-329.
61
Gross, J. J., & Munoz, R. F. (1995). Emotion regulation and mental health. Clinical
Psychological Science Practices, 2 (2), 151-164.
Hersey, R. B. (1932). Workers? emotions in shop and home: a study of individual workers
from the psychological and physiological standpoint-short title only in data base.
Philadelphia: University of Pennsylvania Press.
Hofmann, D. A. (1997). An overview of the logic and rationale of hierarchical linear models.
Journal of Management, 23, 723-744.
Iffaldano, M. T., &. Muchinsky, P. M. (1985). Job satisfaction and job performance: a meta
analysis. Psychological Bulletin, 97, 251-273.
Isen, A. M., & Means, B. (1983). The influence of positive affect on decision-making
strategy. Social Cognition, 2, 18-31.
Izard, C. E. (1992). Basic emotions, relations among emotions, and emotion cognition
relations. Psychological review, 99, 561-565.
Izard, C. E. (1993). Four systems for emotion activation: Cognitive and noncognitive
processes. Psychological Review, 100, 68-90.
Judge, T. A., Thoreson, C. J., Bono, J. E., & Patton, G. K. (2001). The job satisfaction-job
performance relationship: a qualitative and quantitative review. Psychological review,
127, 376-407.
Kornhauser, A. W., & Sharp, A. A. (1932). Employee attitudes: suggestions from a study in a
factory. Personnel Journal, 10, 393?404.
Koss, E., & Lewis, D. A. (1993). Productivity or efficiency - measuring what we really want.
National Productivity Review, 12 (2), 273-295.
62
Kreft, I. G. G., de Leeuw, J., & van der Leeden, R. (1994). Review of five multilevel analysis
programs: BMDP-SV, GENMOD, HLM, ML3, and VARCL. The American
Statistician, 48, 324-335.
Kupers, W. (1998). Phenomenology of embodied productivity in services. International
Journal of Service Industry Management, 9 (4), 337-358.
Larsen, R. J., & Diener, E. (1992). Problems and promises with the affect circumplex model
of emotion. Review of personality and social psychology, 13, 25-59.
Lord, R. G., & Kanfer, R. (2002). Emotions and organisational behaviour. In Lord, R. G.,
Klimoski, R. J., & Kanfer, R. (Eds.), Emotions in the workplace (pp. 5-19). San
Francisco: Jossey- Bass.
Lucas, R. E., & Diener, E. (2003). The happy worker: hypotheses about the role of positive
affect in worker productivity. In M.R. Barrick & A.M. Ryan (Eds.), Personality and
work: reconsidering the role of personality in organisations (pp. 30-59). San
Francisco: John Wiley & Sons.
Osborne, J. W. (2000). Advantages of hierarchical linear modelling. Practical Assessment,
Research & Evaluation, 7 (1), 1-3,
Oswald, A. J., Proto, E., & Sgroi, D. (2008). Happiness and productivity. Unpublished thesis,
The University of Warwick, United Kingdom.
Patterson, M., Warr, P., & West, M. (2004). Organizational climate and company
productivity: the role of employee affect and employee level. Journal of Occupational
and Organizational Psychology, 77, 193-216.
Russel, J. A., Weiss, A., & Mendelsohn, G. A. (1989). Affect grid: a single-item scale of
pleasantness and arousal. Journal of Personality and Social Psychology, 57 (3), 493-
502.
63
Sanna, L. J., Turley, K. J., & Mark, M. M. (1996). Expected evaluation, goals, and
performance: mood as input. Personality and Social Psychology Bulletin, 22, 323-
325.
Shaver, P., Schwartz, J., Kirson, D., & O?Conner, C. (1987). Emotion knowledge ? further
exploration of a prototype approach. Journal of Personality and Social Psychology,
52, 1061-1086.
Sink, D. S., & Tuttle, T. C. (1989). Planning and measurement in your organisation of the
future. Norcross: Industrial Engineering and Management Press.
Slack, N., Chambers, S., & Johnston, R. (2001). Operations management (3rd ed.). Harlow:
Pearson Education Limited.
Staw, B. M., & Barsade, S. G. (1993). Affect and managerial performance: a test of the
sadder-but-wiser vs. happier-and-smarter hypotheses. Administrative Science
Quarterly, 38 (2), 304-331.
Staw, B. M., Sutton, R. I., & Pelled, L. H. (1994). Employee positive emotion and favourable
outcomes at the workplace. Organization Science, 5 (1), 51-71.
Stearns, C. Z., & Stearns, P. N. (1986). Anger: the struggle for emotional control in
America?s history. Chicago: The University of Chicago Press.
Stone, A. A., Shiffman, S. S., & De Vries, M. W. (1999). Ecological momentary assessment.
In D. Kahneman, E. Diener & N. Schwarz (Eds.), Well being: the formation of
hedonic psychology (pp. 26-39). New York. U.S.A.: Russel Sage Foundation.
Tangen, S., (2004). Demystifying productivity and performance. International Journal of
Productivity and Performance Management, 54 (1), 34-36.
Tellegen, A., Watson, D., & Clark, L. A. (1999). On the dimensional and hierarchical
structure of affect. Psychological Science, 10, 297-303.
64
Verbeke, G., & Molenberghs, G. (2000). Linear mixed models for longitudinal data. New
York: Springer-Verlag.
Vora, J. A. (1992). Productivity and performance measures: who uses them? Production and
Inventory Management Journal, 33 (1), 46-49.
Weiss, H., & Cropanzano, R. (1996). Affective events theory: A theoretical discussion of the
structure, causes, and consequences of affective experiences at work. Research in
Organizational Behaviour, 18, 1-74.
Weiss, H. M., & Kurek, K. E. (2003). Dispositional influences on affective experiences at
work. In M.R. Barrick & A.M. Ryan (Eds.), Personality at work: reconsidering the
role of personality in organisations (pp. 121-149). San Francisco: Jossey- Bass.
West, M. (1999). Essays on productivity, flexibility, and manufacturing networks.
Linko?ping: Linko?ping University.
Wright, A., & Staw, B. M. (1999). Further thoughts on the happy-productive worker. Journal
of Organisational Behaviour, 20 (1), 31-34.
Wright, T. A., & Cropanzano, R. (2004). The role of psychological well-being in job
performance: a fresh look at an age old quest. Organizational Dynamics, 33, 338?351.
Wright, T. A., & Staw, B. A. (1999). Affect and favorable work outcomes: two longitudinal
tests of the happy-productive worker thesis. Journal of Organizational Behavior, 20,
1-23.
Wyon, D., Tham, K., Croxford, B., & Oreszczyn, T. (2000). The effects on health and self-
estimated productivity of two experimental interventions which reduced airborne dust
levels in office premises. Proceedings of Healthy Buildings, 1, 641 - 646.
Zelenski, J. M., Murphy, S. A., & Jenkins, D. A. (2008). The happy-productive worker thesis
revisited. Job satisfaction Studies, 9, 521-537.
65
Appendices
Appendix A: Affect circumplex
Larsen and Diener, (1992).
66
Appendix B: Affect grid
Extremely high arousal
0 1 2 3 4 5 6 7 8 9
0
1
2
3
4
5
6
7
8
9
Extremely unpleasant
emotion
Extremely pleasant
emotion
Extreme sleepiness
Russel, Weiss and Mendelsohn, (1989).
67
Appendix C: Affective events theory model
Weiss and Cropanzano, (1996).
68
Appendix D: Triple P model
Tangen, (2004).
69
Appendix E: Biographical questionnaire
By completing this questionnaire I consent to participating in a study of emotions in relation
to self-reported productivity by Moranda Lam. I understand that:
? Participation is voluntary.
? No information that may identify me will be included in the research report, and my
responses will remain confidential.
PLEASE ANSWER ALL THE QUESTIONS: Details are required for descriptive purposes
only and anonymity is guaranteed.
1. Age
2. Gender male/ female
3. What race are you?
____________________________________________________________
4. What is your home language?
____________________________________________________
5. What is your current marital status?
______________________________________________
6. Do you have children? If Yes, how many?
__________________________________________
7. What is your highest level of education?
Grade 10- Grade 11 Matric Diploma Degree
Other (specify):
__________________________________________________________
70
8. What is your occupation?
________________________________________________________
9. At what position are you working at?
_______________________________________________
71
Appendix F: Questionnaire 1
Please fill in the following questionnaire and return responses as soon as possible.
Date: ____________________
Time: ____________________
1) Are you at work
Yes/ No
2) How many hours sleep did you get last night?
< 8 hours 6-8 hours 4-6 hours < 6 hours
3) How are you feeling emotionally?
Using the Affect Grid, please mark relevant grid.
4) Has anything happened in the last 24 hours that may have affected how you are feeling?
Yes/No
5) If yes, please specify briefly
_____________________________________________________________________
_____________________________________________________________________
72
Appendix G: Questionnaire 2
Please fill in the following questionnaire and return responses as soon as possible.
Date: ____________________
Time: ____________________
1) How are you feeling emotionally?
Using the Affect Grid, please mark relevant emotion.
2) Was there a particular event, person or object that contributed to the reported emotion?
__________________________________________________________________
3) Has anything happened since the last time you completed a questionnaire that may have
affected your emotion?
Yes/ No
If yes, please specify:
_________________________________________________________________
4) How tired are you?
extremely fairly not tired energetic
5) Please describe briefly what you were doing just before you completed the questionnaire?
_____________________________________________________________________
73
Appendix H: Questionnaire 3
Please fill in the following questionnaire and return responses as soon as possible.
Date: ____________________
Time: __________________
1) Are you working?
Yes/ No
2) How tired are you?
extremely fairly not tired energetic
3) Has anything happened since the last time you completed a questionnaire that may have
affected your emotion?
Yes/ No
If yes, please specify:
_________________________________________________________________
4) How are you feeling emotionally?
Using the Affect Grid, please mark relevant emotion.
5) Was there a particular event, person, object or any other aspect of work that contributed
to the reported emotion?
_________________________________________________________________
74
6) On a scale of 0-100 percent rate how well you have been working today in relation to
your full capacity.-
________________________________________________________________
7) Was there a particular event, person, object or any other aspect of work that contributed
to the reported productivity level?
________________________________________________________________
_________________________________________________________________
75
Appendix I: Questionnaire 4
Please fill in the following questionnaire and return responses as soon as possible.
1) Are you working now? Yes/ No
2) How are you feeling emotionally now?
Using the Affect Grid, please mark relevant emotion.
3) Please specify briefly a particular event, person, object that may be attributed to the
reported emotion since the last time you completed the questionnaire?
________________________________________________________________________
________________________________________________________________________
4) Are you displaying this emotion to others? Yes/No
If not, please specify the relevant emotions you are exhibiting using the second affect grid.
5) How busy are you in relation to the work you bought home?
Not busy Fairly busy Busy Extremely busy
Overloaded
6) On a scale of 0-100 percent, rate how well you have been working in relation to the work
you bought home?
________________________________________________________________
76
7) Please specify a particular event, person or object that may be attributed to the reported
productivity level?
_________________________________________________________________
_________________________________________________________________
8) On a scale of 0-100 percent, rate how well you have been working today in relation to
your full capacity?
________________________________________________________________
77
Appendix J: Data collection schedule
Day Time 1 Time 2 Time 3 Time 4
Week 1 Mon 10H00 12H00 16H00 19H00
Wed 10H00 12H00 16H00 19H00
Fri 10H00 12H00 16H00 19H00
Week 2 Tues 10H00 12H00 16H00 19H00
Thurs 10H00 12H00 16H00 19H00
Week 3 Mon 10H00 12H00 16H00 19H00
Wed 10H00 12H00 16H00 19H00
Fri 10H00 12H00 16H00 19H00
78
Appendix K: Participant information sheet
SCHOOL OF HUMAN & COMMUNITY DEVELOPMENT
UNIVERSITY OF THE WITWATERSRAND
Private Bag 3, WITS, 2050
Tel: (011) 717 4500 Fax: (011) 717 4559
Dear Sir/Madam
My name is Moranda Lam, and I am conducting research for the purposes of obtaining a
Masters degree at the University of the Witwatersrand. The focus of my study is on the
relationship between emotions and self-reported productivity.
In the past the importance of emotions has been underestimated in the workplace. However
towards the end of the twentieth century, researcher became increasingly interested in its
role. As a result more and more theorists began to establish relationships between aspects of
emotions and behavioural outcomes. One of the more important outcomes they looked at was
productivity, which is also of interest in this study. There have been studies which suggest
that a happy worker is a productive one as well as a productive worker being a happy
employee. However this research, does not aim to prove these findings, instead the objective
of this study is to examine whether emotions and productivity have an impact on each other
over time.
This study is a great opportunity to understand processes that we all experience in our daily
lives that most people may not be aware of. Given this, I would like to invite you to
participate in this research study to examine the role of emotions in the workplace. Criteria
for participants are working adults who have regular access to e-mail.
Participation in this research will require you to have immediate access to email. A number
of questionnaires will be sent through to each participant at specified times. Responses will
need to be as immediate as possible because the aim is to measure the emotion being felt at
the moment. Each questionnaire will take no longer than five minutes to complete. The study
will be conducted over a period of three weeks, at various intervals. Participation in this
study is completely voluntary. No one will be advantaged or disadvantaged in any way
79
should they choose to complete the questionnaire or not. No identifying information such as
the name, ID number or place of work is required, hence anonymity is guaranteed. This
research is not intended to investigate any specific individuals or organisations, but rather to
establish general patterns of emotions at work. In addition, the completed questionnaires will
only be seen by me and my supervisor, and responses will be kept confidential.
If you fulfil the criteria for participation in this study and are willing to participate please
complete the following questionnaire. Completion of the questionnaire will be regarded as
consent to participate in the study. Participants will be allocated a reference number on return
of the first set of questionnaires. Reference numbers will need to be stated on questionnaires.
All identification will be through reference numbers to ensure anonymity. Completed
questionnaires will be placed in a box in your organization. Feedback on the study will be
made available for any participant that requests it, in the form of a one page summary.
Your participation in this study would be greatly appreciated. Should you have any queries,
please do not hesitate to contact either myself, or my supervisor, Karen Milner.
Kind Regards Research Supervisor
Moranda Lam (Industrial Psychology Masters) Karen Milner
milanda_hm@yahoo.co.uk Karen.Milner@wits.ac.za
80
Appendix L: Example of events in relation to reported emotions (first 100 responses)
? I have been running around solving and assisting teachers? problems with their internet.
? I was assisting a teacher with internet connection onto our server for security.
? I took an old computer apart.
? The teachers and reception desk is capturing marks for the term end reports.
? Been monitoring the teachers printing the marks.
? I just been installing software on the music computer.
? Was suppose to have a meeting with the security company to discuss the cctv system but they
did not arrive.
? I was taking this big server cabinet apart and helping a lady with her computer.
? Forgot my student card at home and need it today. When I think about it, it dampens my mood.
? Nothing much.
? Discover its better doing something yourself, instead of depending on others.
? Nothing in particular.
? None.
? Access card to office was blocked.
? None.
? Fire drill at 10 this morning.
? None.
? Bank rejected my credit card application.
? Nothing in particular.
? Mainly the same client, briefing all their campaigns late. It?s just not acceptable.
? Had a meeting with client to discuss a possible radio sponsorship as well as finalise a print
campaign.
? It's just been one of those days. Don't feel very productive at this stage although I have lots to
do.
? It was quite a morning. Had to take public transport as my car is being repaired
? The shit has hit the fan. Client is unhappy about certain things and she has now copied my boss
in on the email.
? Nothing significant. Slow morning.
? I attempted to find a different route to work. It didn't work.
81
? Stuck in traffic for longer than usual.
? No work what so ever. Relaxing.
? Lots of work. Boring work.
? Took a long time to wake up. Little jobs coming that I don't need with deadlines.
? I am not feeling 100% well today. No major significant events occurred.
? No significant events occurred.
? None.
? None.
? None.
? I heard that there is a production problem that my colleague could not attend to, so I had to
help.
? I had to take another route to work as the construction on the road caused too much traffic.
? Breakfast with my colleagues was cancelled because my pregnant colleague was having pains.
? Had a meeting where I got a bit of recognition for work I was doing. Had morning tea.
? Took a different route to work because of the traffic lights were out at Wemmer Pan.
? Had a meeting without my team mates so I had to talk on their behalf. Had morning tea.
? Traffic to work was awful. Had a daily progress meeting.
? Had a progress meeting. Been getting a lot of queries.
? I feel frustrated and irritated with my co-workers.
? I had an argument with my boyfriend.
? It was my boyfriend?s birthday. I woke up and made him breakfast in bed at 4 am, he was not
very enthusiastic.
? Have to finish drawings that weren?t done yesterday.
? Had to go out to site and check some things.
? Did not go to work. Took the day off to go to Decorex.
? Not anything I can recall. It was a normal day, everything went well and timing was perfect
? Nothing significant happened this morning. A usual morning.
? We went to watch a video that informed us about the latest happening in the company and it
was very boring.
? I was late for work, overslept so I have lots of work to do and e-mails to respond to.
? Haven?t been at work for the pass two days and not feeling well so I?m tired and sleepy.
? Not much happening.
82
? Not much. I was just late for work.
? Was called by a music/arts promoter for a Heritage Day gig for my band; called a Departments
of Arts
? Just woke up.
? Scheduled a meeting.
? I'm at a new job so trying to organize catch phrases and names for a new product.
? We are moving to a new building at the place I'm working at, on Friday morning.
? I've come to work, I arrived later than usual. It's a quieter start to the day, which is a really good
thing.
? I am busy proof reading my company's new book.
? Had a chat with friends and were sharing some work experiences.
? I was conducting research, and made a few phone calls for personal follow-up.
? I am finishing up my on-line research.
? None. Was completing some work.
? None. I was working.
? I was concentrated on my work.
? Completing my last bit of work.
? My supervisor had requested me to follow up on certain issues.
? Meeting with Nashua.
? Went to meeting.
? I worked till 9. Just came back from a client visit. It went alright.
? I am still searching for positions that I need this weekend and have come up short thus far.
? I ate breakfast!:) Work, going alright at the moment, got a steady flow of what I should be
doing.
? Nothing much. I came to work and worked.
? Nothing much. Had an interview.
? I have had an interview. And I just found out that this is the last week of the Survey! Haha.
Jokes.
? I took my car to get it serviced.
? Woke up on time...
? Nothing to speak of.
? Dispute with co-worker over something trivial.
83
? Work as usual.
? Nothing special.
? Received a sad mail regarding the death of an old friend.
? I interviewed two excellent candidates and invoiced a client for a really nice placement.
? Interviewed a good calibre candidate and then had a good client visit.
? A client has come back to me with a good offer and my candidate has verbally accepted.
? Another offer pending for a secretarial placement.
? I went to visit a client and had a really good reception.
? Nothing really. I have been phoning clients and following up on information I sent them.
? Nothing significant really. I am calling candidates and following up on their interviews and
CVs.
? Very stressed as the manger and seniors are coming to review.
? Started working earlier and accomplished a lot.
? None.
? None.
? None.
? The deadline is today.
? Water out so had to shower at gym. Had to stop by new house to get painting going
84
Appendix M: Ethics clearance certificate
85