THE ANTECEDENTS AND CONSEQUENCES OF THE VARIABILITY IN JOB SATISFACTION by LINDSEY M. KOTRBA DISSERTATION Submitted to the Graduate School of Wayne State University, Detroit, Michigan in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY 2007 MAJOR: PSYCHOLOGY (Industrial/Organizational Psychology) Approved by: Advisor Date ______________________________
125
Embed
The antecedents and consequences of the variability in job satisfaction
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Job satisfaction is one of the most highly researched constructs in
industrial/organizational psychology. In 1992, Cranny, Smith, and Stone estimated
there to be more than 5,000 published articles and dissertations that have, in some
way, examined job satisfaction. In the fifteen years since, job satisfaction has
remained one of the most enthusiastically studied constructs in the organizational
sciences.Why is the topic of job satisfaction so ardently studied? As Locke (1976)
suggests, there are two main reasons why researchers have been systematically
studying the nature and causes of job satisfaction since the 1930s: 1) job satisfaction
can be viewed as an end in itself, since happiness is a goal of life and 2) it
contributes (or is expected to contribute) to other attitudes and outcomes.
Additionally, through the years there have been many different conceptualizations of
job satisfaction. These different conceptualizations have led to different research
methods and many mixed findings, thus furthering the study of job satisfaction as
researchers attempted to definitively define and explain the construct. For example,
in recent years the debate inherent in most job satisfaction literature has centered on
the extent to which job satisfaction is rooted in individual dispositions or in situationalfactors. This debate has lead different researchers in very different directions. Years
of inconclusive research led many researchers to describe the construct of job
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Job satisfaction researchers have varied greatly in their conceptualizations of the
construct. The most inherent difference between these conceptualizations lies in the
hypothesized causal antecedents of job satisfaction, as will be subsequently
discussed.
Weiss and Cropanzano (1996) suggested that, at that time, all specific
theoretical positions on job satisfaction were variants of three general approaches:
The cognitive judgment approach, the social influences approach and the
dispositional approach. The cognitive judgment approach is rooted in equity theory.
As described by Weiss and Cropanzano (1996), the general structure existing for allcognitive judgment theories is that the work environment is represented as a
concrete or abstract set of features (e.g., pay levels, promotion opportunities, etc.)
and that these features are perceived and compared to some set of standards held
by the job incumbents. In these theories, the degree of match between perceptions
and standards can then be thought of as one’s level of job satisfaction. In contrast,
the social influences approach is much like the cognitive judgment approach in that
job satisfaction is seen as a result of one’s perceptions of some desired standards.
The only difference is that in this latter theory, social information is the source of
input for one’s perceptions and standards (Weiss & Cropanzano, 1996). Finally, the
dispositional approach to job satisfaction is based on the idea that to some degree,
job satisfaction results from one’s general tendencies to feel good or bad, and that
this tendency is unaffected by the specific nature of the job (Weiss & Cropanzano,
1996).
8/11/2019 The antecedents and consequences of the variability in job satisfaction
As previously mentioned, these three theoretical conceptualizations differ
primarily in the antecedents hypothesized to cause job satisfaction. The cognitive
judgment approach suggests that job satisfaction is a result of individual perceptions
of fulfilled expectations, standards or needs. The social influence approach
hypothesizes that job satisfaction is caused primarily by cues from the social
environment, while the dispositional approach suggests that job satisfaction is a
result of individual tendencies to naturally feel good or bad. More simply, these
theories differ in the extent to which they suggest job satisfaction to result from
individual differences or from the situation. As a result, research on the origin andnature of job satisfaction naturally shifted focus toward determining the extent to
which it is rooted in situations (i.e., reactions to workplace factors) or in dispositions
inherent to individuals. In other words, job satisfaction researchers have spent much
time caught up in a person-situation debate.
The Person-Situation Debate
The Dispositional Approach: There has been much research supporting the
that differences in personality and trait affectivity predispose individuals to be
differentially satisfied with their jobs (Brief, Butcher, George, & Link, 1993).
Many researchers interested in the dispositional nature of job satisfaction
have focused on identifying the personality traits that may be responsible for
determining an individual’s level of job satisfaction. Judge and Larsen (2001) argued
that neuroticism, extraversion, positive affectivity (PA) and negative affectivity (NA)
are the traits that are best suited to predicting job satisfaction. A meta-analysis
conducted by Judge, Heller, and Mount (2002) lends support to this argument.
Additionally, Judge et al. (2002) found neuroticism to be a consistent correlate of jobsatisfaction, and further found the relationships between job satisfaction with
neuroticism and extraversion to generalize across studies.
There is an abundance of research additionally supporting a relationship
between affective dispositions and job satisfaction (Agho, Mueller & Price, 1993;
individuals characterized by distress, unpleasurable engagement, and nervousness
(i.e., individuals high in NA) are likely to have low levels of job satisfaction. Past
research has also supported a significantly positive relationship between PA and jobsatisfaction (Agho, Mueller, & Price, 1993; Watson, & Slack; 1993). Thus individuals
characterized by high energy, enthusiasm, and pleasurable engagement (i.e.,
individuals high in PA) are likely to have high levels of job satisfaction. A meta-
analysis conducted by Connolly and Viswesvaran (2000) found estimated true score
correlations (i.e., corrected correlations) of PA and NA with job satisfaction of .52 and
-.33 respectively. Thus, there is ample evidence to suggest relationships between
both PA and NA and job satisfaction. While other explanations for these relationships
exist (e.g., NA causes individuals to choose less favorable jobs subsequently leading
to lower satisfaction), much of the past research has suggested evidence of the
relationship between affect and job satisfaction to support the assertion that affect
predisposes individuals to be chronically satisfied or dissatisfied.
Judge, Locke and Durham (1997) further used a combination of several traits,
or what they describe as individuals’ core self-evaluations, to explain the
dispositional nature of job satisfaction. Core self-evaluations are described as a
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Noe, Moeller, & Fitzgerald, 1985). The JCM has been influential to researchers’development of the job satisfaction construct through its emphasis on situational
components.
Also influential, a study conducted in 1987 by Gerhart has been frequently
cited in support of the situational approach. Utilizing a longitudinal field sample of
over 800 individuals, Gerhart found that pay, status, and job complexity added
explanatory power to an equation predicting job satisfaction. In contrast, results did
not support the importance of traits as determinants of job satisfaction. This suggests
that situational factors are not only important predictors, but that they may be more
important than individual traits.
The Staw and Ross (1985) study, which is often cited in support of the
dispositional approach, also provides evidence for situational influences on job
satisfaction. Utilizing a longitudinal data base of over 5,000 individuals, Staw and
Ross obtained a correlation of .44 between combined 1966 and 1969 assessments
of job satisfaction and job satisfaction measured in 1971; leading them to conclude
8/11/2019 The antecedents and consequences of the variability in job satisfaction
that measures of job satisfaction are stable over a 5-year period. However, the
highest correlation between job satisfaction measures over the five year period was
obtained for those individuals who did not change employer or occupation. For those
individuals whose job situations remained constant, the correlations between 1966
and 1971 job satisfaction was .37. For those individuals whose employer and
occupation changed the correlation between 1966 and 1971 job satisfaction was
reduced to .19. Again demonstrating situational aspects to be important. Additionally,
they also found that beyond the effects of attitudinal stability, there was residual
variance in satisfaction that was related to situational aspects. Thus, these resultsmore accurately suggest the presence of both dispositional and situational influences
on job satisfaction.
Situational factors have also been found to mediate the relationship between
dispositions and job satisfaction. More specifically, Judge et al. (2000) found core
self-evaluations (i.e. dispositions) measured in childhood to predict job satisfaction
later in life; however it was also found that job complexity (i.e. a situational effect)
partially mediated this relationship. Thus again, the importance of both dispositional
and situational factors is suggested.
As should be apparent through this review of the literature, the dispositional
approach and the situational approach both have received empirical support. It is
likely that neither of these approaches alone sufficiently explain an individual’s job
satisfaction. Instead of researchers taking a top-down or a bottom-up approach to
job satisfaction, the focus of research should be toward understanding how individual
dispositions and situational characteristics together create an individual's level of job
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Partially because of the conviction with which researchers approached this
person-situation debate, and largely as pointed out by Ilies and Judge (2002),
because of the typical cross-sectional, between-subjects designs that are most
frequently used, it has been particularly difficult to further research on the factors that
influence individuals’ job satisfaction. Additionally, when longitudinal studies were
conducted in the past they were done so over long time frames with very few data
points (e.g., Heller, et al., 2002; Newton & Keenen, 1991; Staw & Ross, 1985; Steel
& Rentsch, 1997; Weaver, 1980). As reviewed above, while we have learned a greatdeal from this past research, there also seems to be a need to move beyond these
traditional approaches in an attempt to better define the causes of job satisfaction.
As a result, several researchers have begun to recognize the importance of
investigating the within- person variation in job satisfaction that may occur over
shorter time frames.
Within Subject Variability in Job Satisfaction: The New Frontier
In traditional approaches to job satisfaction research, time was not considered
an important parameter to consider. However, as suggested by several researchers
(e.g., Ilies & Judge, 2002; Weiss & Cropanzano, 1996), this did not encourage the
investigation of within-individual variation in job satisfaction that may result from
meaningful changes in feelings toward the job. Cross-sectional measurement
assumes that variation around average levels of a construct is randomly distributed
error, which may not be the case. As Kahneman (1999) suggests, the study of
happiness (or satisfaction) can be greatly furthered through obtaining multiple real-
8/11/2019 The antecedents and consequences of the variability in job satisfaction
time measures of the construct as the assessments occur in the individual’s
environment and are not influenced by memory.
One of the most influential and highly cited definitions of job satisfaction was
provided by Locke (1976). Locke described job satisfaction as an emotional reaction
that “results from the perception that one’s job fulfills or allows the fulfillment of one’s
important job values, providing and to the degree that those values are congruent
with one’s needs” (p. 1307). As suggested by Weiss and Cropanzano (1996), though
this is a popular and accepted definition of job satisfaction, research on job
satisfaction to date had not been conducted in a manner consistent with thisdefinition. Locke’s definition is affective, job satisfaction is viewed as an emotional
reaction to the workplace. Affect levels have been shown to fluctuate over time and
to influence immediate perceptions of job stress (e.g., van Eck, Nicolson, & Berkof,
1998) which likely then influence job satisfaction as well (Weiss & Cropanzano,
1996). Thus, fluctuation is expected and meaningful. However, cross-sectional and
two-time point longitudinal studies have not allowed for the identification of such
fluctuations. Given that this approach suggests individual levels of job satisfaction to
fluctuate as one reacts to the job environment, Weiss and Cropanzano seem to be
employing a situational approach. However, Weiss and Cropanzano also suggest
that these patterns of fluctuation can be easily predicted by personality traits. In
general, it seems to be important to utilize momentary assessments of job
satisfaction in the short-term to be able to capture these fluctuation patterns, but also
different individuals are likely to have different reactive patterns which can be
predicted using personality.
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Using a different logic but leading to very similar conclusions, Shoda, Mischel
and Wright (1994) developed a complex theory of personality which is helpful in
understanding not only the importance of investigating the intraindividual variability of
job satisfaction, but also for understanding how both individual dispositions and
characteristics of the situation may interact. Shoda et al. (1994) described how
relatively enduring person variables within the individual interact with situational
characteristics to generate patterns of behavior. The structure of the personality
system can remain stable across situations, but the personality state changes readily
when the situational features that are active change (Mischel & Shoda, 1995). Morespecifically, these authors suggest that an individual’s personality can be described
in terms of stable if…then situation-behavior profiles and argue that variations in
intra-individual behavior should be investigated, as they are not “error” but are
instead meaningful reflections of enduring personality processes (Shoda et al.,
1994). Applying this to job satisfaction research, in support of the situational
approach, there is no reason to expect cross-situational consistency in job
satisfaction because different situations evoke different cognitive-affective reactions,
thus there are likely if-then situation-satisfaction profiles. However, these patterns
are enduring and reflect the influence of stable personality traits.
Job satisfaction researchers have also recently supplied evidence suggesting
the importance of examining the intraindividual variability in job satisfaction. Ilies and
Judge (2002) obtained measures of job satisfaction from 27 individuals at four time
points during the day for 19 working days. Using this procedure, it was found that
36% of the variance in job satisfaction was due to variation within individuals, thus
8/11/2019 The antecedents and consequences of the variability in job satisfaction
suggesting that there is significant within person variability in job satisfaction over
short periods of time. Judge and Ilies (2004) also found state mood to influence state
job satisfaction, and that this effect decays rapidly. An additional recent study sought
to specifically identify the time frame over which individuals significantly vary in their
levels of job satisfaction (Young & Baltes, 2006). Assessing five different constructs
of job satisfaction, Young and Baltes found that while individuals significantly
fluctuated in their levels of job satisfaction between days, between weeks, and
between two week time periods, individuals fluctuate most within the same day. In
addition, utilizing Weiss and Cropanzano’s (1996) Affective Events Theory (AET),Fuller, Stanton, Fisher, Spitzmuller, Russell, & Smith (2003) suggest the relationship
between events and job satisfaction to be mediated by strain, and that this
mechanism likely unfolds during a shorter time frame than one day. In sum, this
previous research supports the assertion that short-term variation in job satisfaction
occurs and that this likely results from reactions to workplace events.
Thus, individuals do seem to meaningfully fluctuate in their levels of job
satisfaction over shorter time frames. Following the logic of Mischel et al. (1994) and
Weiss and Corpanzano (1996), this variability likely results from the appraisal of
different situational events, and these patterns of variability can likely be predicted by
between-subject personality traits. Job satisfaction researchers have begun to
explore the relationships between intraindividual variability and personality. In
particular, Ilies and Judge (2002) found neuroticism to relate to the intraindividual
variability in job satisfaction; additionally, trait affectivity has been suggested as an
important variable to consider when measuring job satisfaction momentarily (Judge &
8/11/2019 The antecedents and consequences of the variability in job satisfaction
research does not address whether this variation predicts important organizational
outcomes (e.g., performance, turnover intentions) above and beyond one’s mean
level of job satisfaction; an additional limitation which the current investigation also
seeks to address. If intraindividual variability in job satisfaction does not predict
important organizational outcomes above and beyond cross-sectional measurements
of job satisfaction, then cross-sectional measurement is likely sufficient for predictive
purposes. Thus, this is an important question to address.
The Present Study
As outlined above, the current study is designed to further explain the role ofpersonality in predicting the within-person variation in job satisfaction, the role of
situational change in predicting the within-person variation in job satisfaction, as well
as to determine whether the variability in job satisfaction predicts organizational
outcomes above and beyond traditional cross-sectional assessments. Further, these
relationships are explored for five different facets of job satisfaction as well as for
multiple conceptualizations of variability.
Within-individual variation can be defined in terms of both frequency and
amplitude changes. In other words, some individuals may change very little in their
levels of satisfaction but these small changes may happen frequently, while other
individuals may experience changes in their levels of job satisfaction that are large in
magnitude, but infrequent. Frequency and amplitude are different conceptualizations
of “variability” and they may have different antecedents and different outcomes.
However, no researcher to date has conceptualized job satisfaction variability in
terms of both frequency and amplitude. Thus, the current study will explore all
8/11/2019 The antecedents and consequences of the variability in job satisfaction
relationships hypothesized for both frequency and amplitude variability in job
satisfaction.
Additionally, research exploring the predictors and outcomes of job
satisfaction variability has focused on measuring the intraindividual variability of
overall job satisfaction. However, there is evidence that job satisfaction is
multifaceted. Facet measures of job satisfaction assess an individual’s satisfaction
with several different aspects of their jobs (e.g., supervision, coworkers, pay,
benefits). Several researchers using facet measures of job satisfaction have found
the facets to differentially relate to both predictor and criterion variables (e.g., Kinicki,Mckee-Ryan, Scriesheim & Carson, 2002; Iaffaldano & Muchinsky, 1985). Given the
previously reviewed research and theory, it has been suggested that the
intraindividual variability in job satisfaction reflects emotional reactions to the working
environment. Thus it is possible that this variability is different for different facets as
these facets represent different aspects of an individual’s working environment.
More specifically, in the present study, five facets of job satisfaction are
investigated. These facets include satisfaction with supervision, coworkers, pay,
promotion, and the nature of work. These facets of job satisfaction are commonly
investigated in the literature and are most often measured with the Job Descriptive
Index (JDI) (Rain, Lane, & Steiner, 1991). In conducting a meta-analysis on the
construct validity of the JDI, Kinicki, et al. (2002) found test-retest reliability
coefficients to be smaller for the facets of supervision and coworker satisfaction (.56
and .59 respectively) than for the facets of pay, promotion, and nature of work
satisfaction (.65, .63 and .67 respectively). The lower test-retest coefficients suggest
8/11/2019 The antecedents and consequences of the variability in job satisfaction
that the facets of supervision and coworker satisfaction may be more variable than
the facets of pay, promotion, and the nature of work. Because an individual’s
interactions with supervisors and coworkers are likely less situationally stable than
either pay, promotion or the nature of one’s work, this assertion is consistent with the
theory of Shoda et al. (1994) in suggesting that an affective state (i.e., job
satisfaction) will vary more frequently as a result of more frequent variation in the
situation. Quarstein, McAfee, and Glassman (1992) suggest situational
characteristics such as pay, promotion opportunities, and working conditions to be
relatively stable over time, while supervisors and coworkers can, at times, behaveerratically and thus may be classified as situational occurrences which are more
transitory. This again suggests the importance of assessing variability for different
facets of job satisfaction, as these facets represent different aspects of the work
environment, aspects which are thought to have different degrees of stability. It
therefore seems important to explore the intraindividual variability in job satisfaction
for different facets of job satisfaction.
However, as already discussed, variability can be conceptualized in multiple
ways. Thus while the facets of supervisor and coworker satisfaction may have
greater frequency variation given that these facets are likely less situationally stable,
this does not mean that these facets have greater amplitude variation. In fact,
logically should an individuals’ pay, or promotion opportunities or the nature of one’s
work change in the situation, it is likely that the affective reaction (i.e., job
satisfaction) to these changes would be large in amplitude. Thus, while these facets
may not frequently vary, they may vary with great amplitude. However, given that
8/11/2019 The antecedents and consequences of the variability in job satisfaction
research has not yet explored frequency vs. amplitude variation in job satisfaction,
no specific hypotheses are made. However, it is loosely expected that the facets of
supervisor and coworker satisfaction will fluctuate with greater frequency, while the
facets of pay, promotion, and nature of work satisfaction will fluctuate with greater
amplitude.
Towards accomplishing the goals of the present study, theoretical evidence is
first reviewed and hypotheses are generated regarding the relationships between the
intraindividual variability of job satisfaction and personality, situational variability, and
organizational variables. These relationships are explored for five different facets of job satisfaction and by defining variability by frequency of change and amplitude of
change.
The Antecedents of Job Satisfaction Variability
Personality: Previous research has suggested that individual dispositions
influence the extent to which individuals are sensitive to situational events (Bowling,
Beehr, Wagner, & Libkum, 2005). Individuals high in neuroticism are described as
having a predisposition to experience anxiety and emotional instability (McCrae &
Costa, 1989). Neuroticism is the personality trait that has been most often suggested
as important to investigate in regard to within-person variability, and past research
has supported its importance to this line of research. For example, by utilizing diary
recordings, Suls, Green, and Hillis (1998) found individuals high in neuroticism to be
more sensitive to stressful events than individuals low in neuroticism while Bolger
and Schilling (1991) found participants high in neuroticism to have more negative
reactions to various daily hassles. Though not specific to job satisfaction, these
8/11/2019 The antecedents and consequences of the variability in job satisfaction
studies demonstrate neuroticism to impact the extent to which individuals react to
situational occurrences. Further as Ilies and Judge (2002) suggest and support,
neuroticism intensifies one’s affective reactions to work-related stimuli, thus resulting
in higher job satisfaction variability. Thus, previous research has supported the idea
of neuroticism relating positively to job satisfaction variability.
Individuals high in extraversion are described as sociable, talkative, assertive
and active (McCrae & Costa, 1989). Extraversion has also been suggested by
previous researchers to be important to investigate when considering within-person
job satisfaction variability (e.g., Bowling et al., 2005). In prior research, extraversionhas been linked to variability in mood or emotional states (e.g., Hepburn, & Eysenck,
1989; Larsen & Ketelaar, 1991; Velting, & Liebert, 1997). More specifically, these
studies all demonstrated extraversion to relate positively to one’s variability in
positive mood. Mood is at least in part situationally determined, as emotion theorists
generally agree that it is an individual’s appraisal of important events/situations that
trigger emotion and mood (e.g., Frijda, 1986; Levenson, 1994). Thus, like
neuroticism, while these studies do not directly reflect job satisfaction, they
demonstrate that extraversion may impact the extent to which individuals react to
situational occurrences. Given that job satisfaction levels are expected to fluctuate as
individuals react to the job environment, extraversion likely also impacts one’s
variability in job satisfaction. However, the one study that has investigated this
relationship did not find extraversion to significantly correlate with variability in job
satisfaction (Ilies & Judge, 2002). In general, there is mixed prior support for the
hypothesis suggesting extraversion to relate to job satisfaction variability.
8/11/2019 The antecedents and consequences of the variability in job satisfaction
The personality trait of openness to experience may be important to
investigate as well. Individuals who are high on openness to experience have an
orientation that is creative, curious, and flexible and further have an affinity for
situations involving novelty, diversity, and change (McCrae & Costa, 1989). In other
words, individuals high in openness to experience have a need for variety. Thus it is
possible that individuals high in openness to experience seek out more variable work
situations, and resultingly may have more variable levels of job satisfaction. Previous
research has also supported openness to experience as relating to within-individual
variability in other contexts. More specifically, Velting and Liebert (1997) foundopenness to experience to positively relate to individuals’ within-day mood
fluctuations. Again, using the same logic as presented above, though not directly
measuring job satisfaction, this study provides evidence to suggest that openness to
experience may relate to how individuals react to their environment, and thus may
impact variability in job satisfaction. While, to date, no study has looked at the
relationship between openness to experience and job satisfaction variability, it is
possible that the two are positively correlated.
Both trait positive and negative affectivity (PA and NA) have additionally been
suggested as important to consider when investigating job satisfaction intraindividual
variability (e.g., Bowling et al., 2005). Reviewing past research lends support to the
importance of NA and PA to job satisfaction variability. In 1990, Parkes found
teachers high in NA to show more symptoms of distress in reaction to a stressful
environment than teachers low in NA, suggesting that high NA individuals react more
strongly to the environment than low NA individuals. Similarly, Marco and Suls (1993)
8/11/2019 The antecedents and consequences of the variability in job satisfaction
found individuals high in NA to be more reactive to negative environmental events.
Also, Brief, Butcher and Roberson (1995) found that the job satisfaction of
individuals high in NA was not as affected by a positive event in the workplace as
was low NA individuals’ job satisfaction. This again demonstrates that trait affect has
an impact on how individuals react to workplace events, but more specifically, the job
satisfaction of high NA individuals is not as strongly affected by positive
environmental change. In fact, research generally suggests individuals high in NA to
more affected by negative events, while individuals high in PA to be more affected by
positive events (Stewart, 1996). In sum, as Weiss and Cropanzano (1996) suggest,affective personality traits seem to act as predispositions, in other words, being high
in PA or NA predisposes individuals to respond to environmental events (either
positive or negative) with more intensity than those low in PA or NA. Given that
previous theory has suggested that different situations evoke different cognitive-
affective reactions, or different situation-satisfaction profiles as well as has
suggested PA and NA to intensify situational reactions, it is likely that NA and PA
relate to within-person job satisfaction variability.
Thus, in sum, a review of the research suggests the following hypotheses.
H1: The personality trait of neuroticism positively relates to intraindividual
variability in job satisfaction.
H2: The personality trait of openness to experience positively relates to
intraindividual variability in job satisfaction.
H3: Dispositional negative affectivity positively relates to intraindividual
variability in job satisfaction.
8/11/2019 The antecedents and consequences of the variability in job satisfaction
defined. Given the absence of research investigating job satisfaction variability in this
way, no direct hypotheses are made. However, the current study explores whether
there are differential relationships between personality and various
conceptualizations of job satisfaction variability. Further, given that variability may be
modeled differently for different facets of satisfaction, the personality-facet variability
relationships are also explored.
Situational Change: Through reviewing the literature, it was suggested that job
satisfaction variability occurs as individuals react to the environment. However, as
already mentioned, no study investigating the within-person variation in jobsatisfaction has attempted to determine whether momentary situational assessments
relate to momentary job satisfaction assessments. Given that immediate perceptions
of job stress influence job satisfaction (Weiss & Cropanzano, 1996) and that different
situations evoke different cognitive-affective reactions (Mischel & Shoda, 1995) and
thus likely different if-then situation-satisfaction profiles, this seems an important
question to directly assess.
Stress researchers often conceptualize stressors as objective external
conditions, or events that have actually occurred, which create stressful demands on
and threats for individuals (Lazarus, 1990). More specifically, job stressors are
aspects of the working environment that create stress for individuals. A great deal of
research suggests the importance of assessing job stressors in regard to their
relation to job satisfaction. This body of research has consistently supported job
stressors as having significant impacts on individuals’ job satisfaction (e.g., Babakus,
Kemey, Mossholder & Bedian 1985). Given that environmental job stressors are so
frequently demonstrated as relating to job satisfaction, they seemed a particularly
important aspect of the job situation to investigate in the current study. While cross-
sectional assessments of job stressors have been related to cross-sectional
assessments of job satisfaction, it is unknown whether momentary assessments of
job stressors will relate to momentary assessments of job satisfaction. However,
given the theory from which this study is based, situational variability is expected to
relate to job satisfaction variability thus the following hypothesis can be made:H5: Intraindividual assessments of job stressor variability are positively related
to intraindividual assessments of job satisfaction.
As with the above hypotheses, this relationship will be investigated for both
frequency variability as well as for amplitude variability and for five different facets of
job satisfaction.
Additionally, in discussing personality as it relates to job satisfaction variability,
personality was described as likely being important to job satisfaction variability
through its impact on how individuals react to situations. In other words, the “reaction
to the environment” which likely determines job satisfaction variability is affected by
personality, suggesting a moderation effect. Given that variability in the situation
likely relates to variability in job satisfaction and that personality likely impacts how
individuals react to situations, the following hypotheses can be made:
H6: The relationship between job stressor variability and job satisfaction
variability is moderated by personality such that this relationship is stronger for
8/11/2019 The antecedents and consequences of the variability in job satisfaction
individuals high in neuroticism than for individuals low in neuroticism.
H7: The relationship between job stressor variability and job satisfaction
variability is moderated by personality such that this relationship is stronger for
individuals high in openness to experience than for individuals low in openness to
experience.
H8: The relationship between job stressor variability and job satisfaction
variability is moderated by personality such that this relationship is stronger for
individuals high in NA than for individuals low in NA.
H9: The relationship between job stressor variability and job satisfactionvariability is moderated by personality such that this relationship is stronger for
individuals high in PA than for individuals low in PA.
This moderation effect is explored for extraversion as well, however given that
the investigation of extraversion is exploratory, a specific hypothesis is not made.
Additionally, these moderating effects are again investigated for both frequency and
amplitude variation and for each facet of job satisfaction.
The Consequences of Job Satisfaction Variability
An additional goal of the present study was to determine whether the
intraindividual variability in job satisfaction relates to organizational variables above
and beyond cross-sectional assessments of job satisfaction. Thus, possible
relationships between the intraindividual variability of job satisfaction and job
performance and turnover intentions are now explored.
Job Performance: Researchers in the organizational sciences have long found
it important to demonstrate a relationship between job satisfaction and job
8/11/2019 The antecedents and consequences of the variability in job satisfaction
performance. In fact, the search for this relationship has been referred to as the “holy
grail” of organizational research (Weiss & Cropanzano, 1996). However, despite all
of the effort researchers have expended, many are still unclear regarding how, or
even if, job satisfaction and job performance relate. For example, Fisher (1980),
viewing job satisfaction as a general attitude, suggests that it is unreasonable to
expect job satisfaction to relate to performance. In fact, research has demonstrated
generally small relationships between job satisfaction and performance (e.g., Katzell,
Thompson, & Guzzo, 1992). However, other researchers have found support for a
relationship between job satisfaction and performance. Podsakoff and Williams(1986) for example, found that by making rewards contingent on productivity, the
connection between satisfaction and performance is high. In general, however,
multiple reviews of the literature seem to suggest a positive, but weak, relationship
between job satisfaction and job performance (Fisher, 2003). As Fisher (2003)
discusses, when assessed at the between-person level the average satisfaction-
performance relationship is generally weak and there is no reason to expect general
satisfaction with the job as a whole to relate to sustained high job performance.
There is evidence however, to suggest that within-person assessments of job
satisfaction relate to job performance. Fisher (2003) found momentary task
satisfaction to relate to momentary task performance, which suggests that individuals
are more satisfied with a particular task at moments they are performing the task
well, and are less satisfied with a particular task at moments they are performing less
well. It is important to point out that is impossible to determine from this study
whether pre-existing satisfaction levels influence subsequent performance, or vice
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Turnover Intentions: Job satisfaction is frequently investigated as an important
predictor of withdrawal behaviors such as turnover. A comprehensive meta-analysis
investigating turnover antecedents found job satisfaction to be one of the strongest
predictors of turnover, = -.19 (Griffeth, Hom, Gaertner, 2000). Thus, when
measured cross-sectionally, the two constructs are negatively related.
Research has additionally explored the satisfaction-turnover relationship from
a longitudinal perspective. Lee and Mtichell (1994) proposed an unfolding model of
turnover, which specifies four basic decision paths that individuals follow whendeciding to quit their jobs. What is particularly important regarding this model is that
research has suggested and supported the voluntary turnover process to unfold over
time, to be precipitated by a “shock” such as being offered a different position, and to
assessed job satisfaction at three different points three months apart. Along the
same vein, Kammeyer-Mueller, Wanberg, Glomb, and Ahlburg (2005) suggest that
because turnover antecedents, like job satisfaction, have been shown to be dynamic,
measuring them in a temporal context should enhance our understanding of the
turnover process. As such, Kammeyer-Mueller et al. (2005) assessed job satisfaction
at 5 different time points 4 months apart and found that when measured over time,
change in the facet of work satisfaction becomes an important predictor of turnover.
Again suggesting job satisfaction change to relate to turnover. While these studies
are important, previous research has demonstrated individuals to meaningfully varyin their levels of job satisfaction over shorter time frames than investigated in the
above mentioned studies. Thus, it is still unknown whether the intraindividual
variability in job satisfaction relates to turnover intentions when assessed
momentarily.
The results of Kammeyer-Mueller (2005) suggest change in job satisfaction to
negatively impact turnover. In other words, individuals who changed in their levels of
satisfaction were more likely to turnover. While this study assessed general change
in satisfaction over many months, it is possible that short term change in job
satisfaction (i.e., variability) also negatively impacts turnover such that individuals
who are more variable in their levels of satisfaction are more likely to intend to
turnover than individuals who vary less in their levels of job satisfaction. Thus the
following hypothesis is made:
H11: Intraindividual variability in job satisfaction is positively related to turnover
intentions.
8/11/2019 The antecedents and consequences of the variability in job satisfaction
In sum, the present study suggests neuroticism, openness to experience, NA
and PA and possibly extraversion to positively relate to job satisfaction variability.
Additionally, it is hypothesized that job stressor variability will relate to the variability
in job satisfaction and that this relationship is moderated by personality, and finally,
the current study suggests job satisfaction variability to relate to job performance and
to turnover intentions, and to predict these variables above and beyond a cross-
sectional assessment of job satisfaction. Further, it is suggested that variability can
be conceptualized in different ways and that variability may be modeled differently for
different facets of job satisfaction. Thus, it was of interest to explore the suggestedrelationships for both frequency and amplitude variation and for different facets of job
satisfaction.
CHAPTER 2
METHOD
Participants
Participants in the present study included full-time staff from a large
Midwestern university. A mass recruitment email (see Appendix A) was sent to all
staff at the university in hopes of recruiting one-hundred and twenty-five individuals
to participate. In regard to assessing intraindividual variability, one-hundred and
twenty-five participants is more than sufficient given the success similar studies have
had in identifying intraindividual variability in job satisfaction using far lessparticipants (e.g., Heller, 2003; Ilies & Judge, 2002; Ilies & Judge, 2004). Further as
Warner (1998) states, in time series studies, representatively sampling time is of
utmost importance, thus many time-series studies contain small numbers of
8/11/2019 The antecedents and consequences of the variability in job satisfaction
participants in order to look at a larger number of time points. In regard to assessing
between subject effects, when =.05, only 44 subjects would be needed to detect a
medium effect (.30, Cohen, 1988) with power = .80 in a regression with five
predictors.
One-hundred and ninety-nine employees responded to the email indicating
their interest in participating. The researcher contacted these interested individuals
via email to explain in greater detail the exact nature of the study and the time
commitments involved with participating. Initially, 132 participants agreed to
participate. However, 31 of these individuals voluntarily withdrew from the studybefore its completion, yielding a final sample of 101 individuals. Thus, power to
detect a medium effect utilizing a regression with five predictors is .99. Participants
received $30.00 compensation for their participation.
Of the 101 individuals, 84.2% are female and 15.8% are male. Regarding
ethnicity, 54.5% of the current sample identified themselves as White/European
American, 33.7% as African American, and 7.9% as Asian. The age of the
participants ranged from 23 to 63, with the average age being 37.91. Additionally,
57.5% percent of the current participants are married or in an exclusive dating
relationship, and 55.4% have children. All 101 participants worked in positions which
allotted them access to the internet throughout each working day (e.g., administrative
assistant, accountant, research coordinator, secretary, housing office coordinator,
office services clerk). All participants worked full-time (ranging from 35-77 hours per
week), with the average number of hours worked per week equaling 41.31. The job
8/11/2019 The antecedents and consequences of the variability in job satisfaction
tenure of the current sample ranged from 2 months to 37 years. A summary of the
demographic characteristics of the sample can be seen in Table 1.
Measures
Job Satisfaction: Two modified versions of Spector’s 1985 Job Satisfaction
Survey (JSS) were used in the present study. The original JSS contains 4 items for
each of 9 facets of job satisfaction (i.e., supervision, coworkers, communication,
contingent rewards, pay, promotion, fringe benefits, operating procedures, and the
nature of work), resulting in a 36-item scale. Respondents indicate their agreement
with each item on a 6-point Likert scale ranging from 1= disagree very much to 6 =agree very much. Higher scores indicate higher levels of satisfaction.
Because participants were asked to complete the JSS at multiple time points
during the same day, the scale was altered slightly in the present study to make
items momentary in nature (see Appendix B). Additionally, because the participants
in the present study were asked to complete the JSS multiple times a day for several
weeks, a shorter version of the scale was desired. Thus, the JSS was reduced to
include only those job satisfaction facets measured in the JDI (i.e., supervision,
coworker, pay, promotion, and the nature of work). Because the Job Descriptive
Index (JDI) is the most frequently used facet measure of job satisfaction (Rain et al.,
1991), it is particularly important to investigate the within person variability of these
To assess the reliability of the 20-item scale, Cronbach’s alpha was calculated
for each facet separately. To ensure that the scale was reliable throughout the
duration of the study, Cronbach’s alpha was calculated for each facet for a random
1/3 of the time points including: Day 1 morning, Day 2 mid-day, Day 3 mid-day, Day
4 mid-day, Day 5 afternoon, and Day 6 mid-day. The reliability information obtained
from these time points was then averaged within each facet to obtain average facet
reliability. This average reliability information can be seen in Table 2.
Additionally, with the assessment of the outcome variables, participants were
asked to complete a reduced (i.e., 20-item), however non-momentary, JSS. This 20-item JSS can be found in Appendix D. Coefficient alphas for the non-momentary
scale in this study are as follows: .811, .830, .880, .702, and .925 for pay, promotion,
supervisor, coworker and work satisfaction respectively.
Personality: Neuroticism, extraversion and openness to experience were
measured using the 50-item IPIP (Goldberg, 1999). Though not included in this
study, the 50-item IPIP also assess conscientiousness and agreeableness.
Respondents indicated the extent to which each item accurately described them on a
scale ranging from 1-very inaccurate to 5-very accurate. An example item is “I am the
life of the party”. Higher scores on the neuroticism subscale indicate higher levels of
emotional stability (i.e. lower levels of neuroticism). Higher scores on the
extraversion and openness to experience subscales indicate higher levels of both of
these constructs. In the present study, coefficient alphas for the neuroticism,
extraversion and openness to experience subscales are .892, .907 and .806
respectively.
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Affectivity: The PANAS (Watson, Clark & Tellegen, 1988) was used to assess
positive and negative affect. When responding to the PANAS, participants are asked
to indicate how the feel in general, or on average. The PANAS consists of 50
adjectives that are rated on a 5-point response scale ranging from 1-very slightly/not
at all to 5-extremly. Coefficient alpha for the PA subscale is .948 in the current study,
and is .953 for the NA subscale. Higher scores indicate higher levels of NA or PA.
Job Stressors: Job stressors were assessed using a twenty-item scale
developed by Frone, Russell, and Cooper (1992). The scale is comprised of three
sub-scales: work pressure (i.e., role overload), lack of autonomy, and role ambiguity.Participants responded on a four-point response scale ranging from 1 (almost
always ) to 4 (almost never / never ). In addition, like job satisfaction, because job
stressors were measured multiple times daily for several days, the questions were
altered to be momentary in nature (see Appendix E). Higher scores indicate higher
job stressors.
Additionally, to ensure that the scale was reliable throughout the duration of
the study, coefficient alpha was calculated for each facet for a random 1/3 of the
time points including: Day 1 morning, Day 2 afternoon, Day 3 mid-day, Day 4 mid-
day, Day 5 afternoon, and Day 6 mid-day. The reliability information was then
averaged and the average reliability is .860.
Performance: In order to assess overall job performance, three questions
were developed. These questions were as follows: 1. “If you had a performance
appraisal or review in the past year, please indicate the overall rating you received
for the quality of your work” with participants indicating their response on a scale
8/11/2019 The antecedents and consequences of the variability in job satisfaction
ranging from 1 = marginal to 5 = above expectations; 2. “Overall what is your usual
performance at work?” with response options ranging from 1 = Consistently below
expectations to 5 = above expectations; and 3. “How do you perform at work relative
to others in your organization (that is, your coworkers)?” with participants responding
on a scale ranging from 1 = well below average to 5 = well above average.
Coefficient alpha for this scale in this study is .728.
Turnover Intentions: A three-item scale developed by Cammann, Fichman,
Jenkins, and Klesh (1979) was used to assess employees' intentions to leave the
organization. Ratings were summated to form an overall score where a highernumber indicates a higher probability of leaving the organization. An example item is
“I will probably look for a new job in the next year” and participants could respond on
a scale ranging from 1-not at all likely to 4-exteremely likely. In the present study,
coefficient alpha for this scale is .902. Higher scores indicate higher intent to
turnover.
Procedure
Upon soliciting participants via email, the researcher sent a follow-up email to
interested individuals. The purpose of this second email was to formally explain the
structure of the study as well as to be sure individuals were fully aware of the time
requirement involved with their agreeing to participate. Individuals were also
informed that compensation would be awarded after the completion of the study.
Further, given that the present study was requesting individuals to indicate their
feelings about their jobs, it was possible that individuals may have felt compelled to
respond in a socially desirable manner in fear that expressing their honest job-
8/11/2019 The antecedents and consequences of the variability in job satisfaction
related feelings may lead to negative repercussions. This was of particular concern
in the present study as participants were being recruited from the same university in
which the study was being conducted. Such participant reactivity to the study may
reduce construct validity (Shadish, Cook, & Campbell, 2002). In order to increase the
accuracy of participant responses, participants were informed that all responses
would remain confidential.
Before beginning the study, individuals needed to formally consent to
participation. The consent form and some general demographic information (e.g.,
gender, ethnicity, job title) were made available online and participants were sent anemail link leading them to the consent form and demographic information. Upon
consenting and completing the demographic information the online survey further
instructed participants to complete the personality measure, and the measure of
positive and negative affectivity. Participants were given one week from the date the
link was sent to complete this initial information.
Once the consent form, demographic information, and initial surveys were
complete, the 20-item JSS and the job stressors scale were also made available
online to consenting participants. Participants were asked to complete the JSS and
the job stressors scale three times daily; once early in the work morning, again in
mid-afternoon, then again in late afternoon. These daily measurements occurred on
6 different days over a three week period. More specifically, participants completed
the JSS and job stressors scale three times daily on Tuesday and Wednesday of
week 1, on Wednesday and Thursday of week 2, and on Tuesday and Wednesday
of week 3. Previous research has demonstrated the majority of within-person job
8/11/2019 The antecedents and consequences of the variability in job satisfaction
satisfaction variability to occur within the same day, however the design of the
present study also allows for between day, between week and between two week
variance to be captured (Young & Baltes, 2006). A reminder email providing a link to
the JSS and job stressors scale was sent out to all participants at the time that they
were expected to fill out the momentary surveys. More specifically, each day of the
study a reminder email was sent at 8:00am, 11:00am and 2:00pm with instructions to
complete the survey by 10:00am, 1:00pm and 4:00pm respectively. Each time a
participant visited the survey website, the job stressors scale was presented followed
by the five facets of job satisfaction presented in random order. The momentaryassessments were collected during the last two weeks of February and the first week
of March.
In addition, the participants were provided with paper copies of the reduced
JSS and of the job stressors scale. They were instructed to complete the paper copy
if they were ever to encounter difficulties accessing the survey online. Participants
were additionally instructed to indicate on the top of the paper survey the date and
exact time of survey completion, and to send any completed paper surveys to the
researcher through campus mail.
Once all momentary measurements were complete, participants were sent
one final email. By clicking on the link included in the email, participants were
directed to a website which asked them to indicate if they experienced any traumatic
life events during the course of the study. If participants indicated that they did
indeed experience a traumatic life event, they were asked to describe the event, and
when it occurred. In addition, participants also completed the non-momentary JSS,
8/11/2019 The antecedents and consequences of the variability in job satisfaction
the turnover intentions scale and the performance scale. Finally, participants also
provided a mailing address, indicating where to send their $30.00 compensation.
CHAPTER 3
RESULTS
Given the extensive time commitment required of the participants in the
current study, several participants did not complete job satisfaction measurements at
every time point. However, no participant was missing more than 3 of the 18 time
points required. Unfortunately, there is as yet no firm guideline for how much missing
data can be tolerated for a sample of a given size (Tabachnick & Fidell, 2001).Again, given the intense nature of the present study, the researcher decided that
completion of 15 of the 18 time points was adequate. Additionally, the amount of
missing data in the current study is less than that found in studies that have utilized
similar measurement procedures. For example, in the Ilies and Judge (2002) study
the maximum number of possible observations across individuals and time periods
was 2052 and data was complete for 1907 or 93% of all observations. In the current
study, the maximum number of observations across individuals and time points was
1818 and data was complete for 1765 or 97% of all observations.
Missing values analyses were conducted in order to further investigate the
nature of the missing data. Little and Rubin (1987) suggested that when data are
missing at random, the chance that a subject’s data are lost is independent of whatthese data would have been had no loss occurred. In order to determine whether
data were missing at random, separate variance t-tests were conducted between the
time points that had 5% or more missing data and job satisfaction scores at all other
8/11/2019 The antecedents and consequences of the variability in job satisfaction
time points. More specifically, a significant t-statistic would indicate that the missing
cases for that time point are significantly correlated with scores at a different time
point, and thus, are not missing at random. These analyses were conducted
separately for each facet of job satisfaction and for job stressors.
In taking a closer look at missing data at each time point for each facet of job
satisfaction, it became apparent that there were no time points for which 5% or more
of the data were missing. In other words job satisfaction data were present for at
least 96 of the 101 participants for each time point in every facet thus there was no
need to conduct t-tests. However, for the job stressors variable, one time point had6.9% missing data. Therefore, separate variance t-tests were conducted between
this time point and every other time point resulting in 15 t-tests. Using an alpha of .01
none of the t-values were significant, thus data was concluded to be missing at
random.
As will be described in greater detail below, periodogram and harmonic
analyses were used to identify and describe variability in this study. Time-series data
that includes missing values can not be analyzed using periodogram or harmonic
analyses, so it was necessary to replace missing values. In the current study, the
mean of the nearest two points was used to replace missing values. Several
researchers suggest that when the amount of missing data is small (i.e., 5% or less),
one’s choice of technique for replacing missing data seems to make little difference
in the results obtained (Raaijmakers, 1999; Roth & Switzer, 1995). Additionally,
several other common methods for replacing missing data were deemed
inappropriate for the present study. As pointed out by Raaijmakers (1999),
8/11/2019 The antecedents and consequences of the variability in job satisfaction
substitution based on the item mean is only a good option when there are low
correlations between the missing variable and the other variables under
investigation. This is unlikely in the current study, given that there are variables in
this study that are the same measure taken at different time points. Raaijmakers also
pointed out that regression and hotdeck procedures are difficult to implement if the
variable having the highest correlation with the missing variable also has missing
data. Given that all time points include missing data for at least one case, regression
and hotdeck procedures do not seem appropriate for the current study. Another
common procedure, using the person mean to replace the missing value, additionallyseemed unsuitable for the present study. The current study suggests that individuals
vary over short periods of time, and the person mean substitution method is
inappropriate for scales characterized by varying means (Raaijmakers, 1999). Thus
the mean of the nearest time points was reasoned to be one’s best guess regarding
the value of the missing time point for the current study. This is a conservative
method for estimating missing values as it reduces the variance of the variable
(Tabachnick & Fidell, 2001). This is of particular importance to note in the present
study, as it is the main purpose of this study to investigate the variability of job
satisfaction. Table 3 displays the means and standard deviations of the job
satisfaction facets, after missing values were replaced.
In order to first identify and describe the within person variability in job
satisfaction, periodogram and harmonic analyses were conducted. In brief,
periodogram analyses are used to identify the periodic components that explain the
largest percentage of the variance in a time series. Harmonic analyses then use this
8/11/2019 The antecedents and consequences of the variability in job satisfaction
identified period to estimate the mean, amplitude and phase that maximize fit to the
observed time series. Both of these procedures will be described in more detail
below.
Before conducting either periodogram or harmonic analyses, it is necessary to
identify and remove any trends in the time series (Warner, 2003). This is important
because any trends that are present will dominant the analyses making it difficult to
detect any cycles that might be present. Regression procedures were used to fit and
remove any linear and/or quadratic trends from the job satisfaction and job stressors
data. Trends were identified and removed separately for every person for every facetof job satisfaction and for the job stressors variable. More specifically, for every
person for every time-series variable, a regression analysis was conducted to predict
the raw time-series data (job satisfaction scores or stressor scores across time) from
the observation number (ranging from 1 – 18). The residuals from this trend analysis
were saved as a new variable and these residuals were used in subsequent
periodogram and harmonic analyses.
As stated before, the main purpose of periodogram analysis is to identify the
periodic components that explain the largest percentage of the variance in a time
series. In periodogram analyses, the overall Sum of Squares (SS) for the time series
is partitioned into a set of N/2, N being the length of the time series, SS components
that correspond to the amount of variance accounted for by different cyclic
components (Warner, 1998). In the current study where the length of the each time
series is 18, the cycle lengths that were fitted to the data include periods of 18/1,
18/2, 18/3, 18/4, 18/5, 18/6, 18/7, 18/8 and 18/9. Or in other words, the overall SS for
8/11/2019 The antecedents and consequences of the variability in job satisfaction
the time series was partitioned into sums of squares that are accounted for by N/2
(i.e., 9) different cyclic components with periods of 18, 9, 6, 4.5, 3.6, 3, 2.57, 2.25
and 2. Frequency is the inverse of period and is the proportion of a cycle that occurs
during one observation. Frequencies are equally spaced, and thus are orthogonal
(Warner, 1998).
In the current study periodogram analyses were conducted on the time series
residuals. These analyses were conducted separately for each person for each time
series variable (i.e. the five facets of job satisfaction and job stressors). Periodogram
intensities (sums of squares) were used to identify the period that accounted for thelargest proportion of time-series variance. So for example, consider participant #1. If
a relatively large proportion of the variance of participant 1’s coworker satisfaction
time series corresponds to a period of 18/4, than a cycle length of 4.5 is concluded to
best explain variability in person 1’s coworker satisfaction time series. Thus
participants’ frequency scores were identified by finding the period/frequency
components that explained the largest amount of time-series variance for each of the
time series variables. Harmonic analyses were then conducted to further model the
identified cyclic components.
Harmonic analysis is designed to more specifically model a cyclic component.
Using the previously identified period, harmonic analysis estimates the mean, phase,
and amplitude that maximize fit to the observed time series (Warner, 1998). Cosine
and sine functions that represented the previously identified period/cycle were
computed for each person for each time series variable. For example, if the cycle of
participant 1’s coworker satisfaction time series was identified to be 4.5 then, the
8/11/2019 The antecedents and consequences of the variability in job satisfaction
cosine function was calculated as cos(2time/4.5) and the sine function was
calculated as sin(2 time/4.5). Next, for each person OLS regression procedures
were used with the calculated cos and sin variables as predictors of the residual time
series variables. The mean of the time series is the resulting intercept. The estimated
amplitude was calculated using cosine and sine coefficients. The amplitude estimate
indicates that the cycle has peaks and troughs that are roughly X points above and
below the overall mean. In sum, conducting both periodogram and harmonic
analyses for each person separately and for each time series variable resulted in
frequency and amplitude scores for each individual for each facet of job satisfactionand for job stressors. Tables 4 and 5 display the means and standard deviations of
the frequency and amplitude scores for each facet of job satisfaction. As can be
seen in these tables, the average frequency and amplitude variation scores do not
differ greatly by facet.
Bivariate correlations between study variables can be seen in Table 6. This
initial examination of the data reveals that relationships between study variables are
generally small. However, there are some significant correlations that are worthy of
further mention. Looking at the relationships between personality and job satisfaction
variability, the frequency with which pay satisfaction varies significantly relates to
both NA and neuroticism, and the frequency with which nature of work satisfaction
varies significantly relates to both NA and PA. To elaborate, greater frequency
variation in pay satisfaction relates to lower NA (r = -.21, p < .05) and to higher
emotional stability (i.e. lower neuroticism; r = .21, p < .05). Greater frequency
variation in nature of work satisfaction relates to lower NA (r = -.21, p < .05), to higher
8/11/2019 The antecedents and consequences of the variability in job satisfaction
PA (r = .29, p < .05) and though not significant, to higher emotional stability (i.e.,
lower neuroticism; r = .20). It should also be pointed out that non-momentary (i.e.
cross-sectional assessments) coworker satisfaction significantly negatively relates to
NA and to neuroticism. Additionally, considering non-momentary nature of work
satisfaction, higher NA and neuroticism (i.e. lower emotional stability) relate to lower
satisfaction and higher PA relates to higher satisfaction. Examining the relationships
between job satisfaction variability and job stressor variability, the only significant
correlation is between the frequency with which coworker satisfaction varies and the
frequency with which job stressors vary such that higher frequency variation in jobstressors relates to higher frequency variation in coworker satisfaction (r = .22, p <
.05). Finally, an initial assessment of the relationships between job satisfaction
variability and organizational outcomes reveals a significant correlation between the
frequency with which nature of work satisfaction varies and performance. More
specifically, higher frequency variation in nature of work satisfaction relates to higher
performance (r = .25, p < .05). It should also be noted that all non-momentary job
satisfaction facets negatively and significantly correlate with turnover intentions and
non-momentary work satisfaction positively relates to performance.
In sum, this initial assessment revealed few significant relationships. However,
while the amplitude with which job satisfaction varies did not significantly correlate
with personality, stressors, nor organizational outcomes, significant relationships
were found between the frequency with which the different facets of job satisfaction
vary and the variables of interest in this study. While these correlations provide a first
8/11/2019 The antecedents and consequences of the variability in job satisfaction
look at the data, as described below, a series of regressions were conducted to test
the specific hypotheses set forth.
Before presenting the results of these regressions, it should be pointed out
that because large numbers of regressions were needed to fully address the
hypotheses of the current study, the chances of making a Type 1 error were inflated.
Because of this, instead of evaluating the results in terms of their statistical
significance, results were evaluated in terms of their magnitude or strength. In other
words, when conducting multiple tests, as in the current study, there is an increased
chance of erroneously concluding that results are significant. In addition, proceduressuch as the Bonferroni correction, the Šidák-Bonferroni procedure and Holm’s
method that control Type 1 error by adjusting the alpha that should be used for each
individual test lack power (e.g., Keppel & Wickens, 2004). As such, in the current
study statistics were evaluated in terms of effect size. More specifically, Cohen’s f 2
was used to evaluate the R2 of each regression model. By convention, 0.02, 0.15,
and 0.35 are considered small medium and large effects, respectively (Cohen,
1988). Aside from evaluating the regression model as a whole, it is also important to
evaluate the individual predictors. The square of the t value for a predictor i is related
to the increment in R2 due to predictor i (Bring, 1994). So, in the current study, it is
argued that if the t for a predictor was at least 2.23, which roughly means a 5%
percent increase in R2 can be attributed to that predictor, then the effect is worth
talking about. In sum, the current study uses Cohen’s f 2 and the t-statistic to explore
the magnitude of relationships vs. focusing on the statistical significance of the
results.
8/11/2019 The antecedents and consequences of the variability in job satisfaction
To test hypotheses 1-4 as well as the exploratory aspects of the present
study, ten regressions were conducted to test how the set of predictors (i.e.,
neuroticism, extraversion, openness to experience, positive and negative affectivity)
relate to the each of the ten criteria (i.e., amplitude and frequency scores for each of
five facets of job satisfaction). As previously discussed, because this family of
analyses includes 10 separate tests, the chances of making Type 1 error when
evaluating statistical significance is 40%! Results obtained from these regressions
can be seen in Table 7.
Looking at Table 7, it can be seen that the amount of variability that the set ofpersonality predictors explained in the criteria ranged from 0.3% to 11.4% depending
on the criterion. An examination of the calculated effect sizes reveals that all but one
of the effects are considered small. In the only effect worth reporting, the set of
personality predictors accounting for 11.4% of the variability in the frequency with
which nature of work satisfaction varies (Cohen’s f 2 = .1287), which is approaching a
medium effect size. Looking more closely at the individual predictors in this model,
the t-value for positive affectivity was 2.329. This suggests that roughly a 5.5%
percent increase in the 11.4% of the variability that the set of personality variables
accounts for can be attributed to positive affectivity. Thus, positive affectivity can be
considered an important predictor of the frequency with which nature of work
satisfaction varies. This relationship is positive, suggesting that high levels of PA are
predictive of high levels of frequency variation in nature of work satisfaction. In
addition, while the effect of the overall model was small (f 2 = .0695), openness to
experience was found to be an important predictor to consider in understanding
8/11/2019 The antecedents and consequences of the variability in job satisfaction
coworker satisfaction frequency variation (t = -2.215). Around a 4.9% increase in the
6.9% of the variance that personality accounts for in coworker satisfaction frequency
variation can be attributed to openness to experience. However, this negative
relationship is contrary to what was expected as it suggests lower levels of openness
to experience to predict higher levels of frequency variation in coworker satisfaction.
In sum, regression analyses did not support hypotheses 1 and 3 which
suggested neuroticism and NA to significantly relate to the intrainidividual variability
in job satisfaction. Yet, as previously discussed, bivariate correlations did reveal
significant relationships between NA and neuroticism and the frequency with whichpay and nature of work satisfaction vary. However, these bivariate relationships were
opposite what was hypothesized and so do not lend further support to hypotheses 1
and 3. Hypothesis 2 was also not supported as results were opposite what was
anticipated. Finally, hypothesis 4 received some support in that positive affectivity
was found to be an important predictor of the frequency with which nature of work
satisfaction varies within individuals.
Hypothesis 5 suggested the intraindividual variability in job stressors to relate
to the intraindividual variability in job satisfaction. Ten regressions were again
conducted to test this hypothesis. More specifically each analysis regressed one of
the ten criteria (i.e., amplitude or frequency scores for five different facets of
satisfaction) on amplitude job stressor variability or frequency job stressor variability
respectively. This family of analyses requires 10 regressions, thus the probability of
committing a Type 1 error was again inflated. Unexpectedly, results provide little
support for hypothesis 5. As can be seen in Table 8, the within-person variability in
8/11/2019 The antecedents and consequences of the variability in job satisfaction
job stressors never accounts for more than 5% of the variance in the within-person
variability in job satisfaction, regardless of the type of variation (i.e. frequency or
amplitude) and of the facet of job satisfaction being investigated. All effects are
small. Although the overall effect is small, (f 2 = .0515), job stressor frequency
variation seems to be an important predictor of coworker satisfaction frequency
variation (t = 2.250). Thus, there is some support for the notion that higher job
stressor frequency variation predicts higher job satisfaction frequency variation, or
more specifically in this case higher coworker satisfaction frequency variation.
However, job stressor variation was not found to be a predictor of interest for any ofthe other 9 facets of job satisfaction variability. Therefore, hypothesis 5 is not
generally supported.
In order to assess hypotheses 6-9 and the associated exploratory analyses,
which suggest personality to moderate the relationship between job stressor
variability and job satisfaction variability, a series of hierarchical regressions were
performed. Job stressor variability (either amplitude or frequency) and personality
(either neuroticism, extraversion, openness to experience, NA, or PA) were entered
into the first step, and the respective multiplicative term was entered into the second
step of the regression. In addressing these hypotheses and exploratory analyses, a
total of 50 hierarchical regressions were performed (i.e. there are five potential
moderating relationships to test for each of the ten criteria). Because of the 50
regressions required, familywise error was calculated to be an exorbitant .92. As
such, as with the other analyses, results will be interpreted in terms of magnitude
and not in terms of statistical significance. All independent variables were centered
8/11/2019 The antecedents and consequences of the variability in job satisfaction
before the calculation of interaction terms, as is generally recommended (Cohen,
Cohen, West, & Aiken, 2003). In addition, to obtain the correct standardized
regression coefficients for the interaction terms from SPSS, the regressions were
also run with standardized variables and their respective interaction terms (Cohen et
al., 2003). Results of these hierarchical regressions can be found in Table 9 – Table
18.
Examining the magnitude of the change in R2 from the first step to the second
step of each hierarchical regression, all effect sizes were small (see Tables 9 – 18).
This indicates that none of the 50 interactions investigated had a meaningful impacton the explained variance of the various criteria. In other words, no interactions were
worthy of further exploration. Even so, glancing the through the t-statistics associated
with each interaction term, there was one interaction term for which a 4.33%
increase in R2 could be attributed to the interaction (t = 2.081). This interaction term
of interest can be seen in Table 10. Though, not reaching the 5% that was deemed
an important effect, the interaction seemed worthy of further exploration.
To examine this interaction, an unpublished Microsoft excel spreadsheet
program developed by Bing and LeBreton (2001) that is designed to graph the
regression interactions for two continuous variables was used. This program graphs
continuous interactions using the formulas presented in Cohen and Cohen (1983).
As one can see in Figure 1, the relationship between frequency variation in jobstressors and frequency variation in promotion satisfaction becomes stronger as
extraversion increases. More specifically, the expected positive b-weight (i.e., higher
job stressor variability should lead to more job satisfaction variability) increases from
8/11/2019 The antecedents and consequences of the variability in job satisfaction
In addition, hierarchical regression was used to investigate whether job
satisfaction variability is a significant predictor of these organizational outcomes (i.e.,
performance and intent to turnover) above and beyond ones mean level of job
satisfaction measured cross sectionally. To test these hypotheses (H12 and H13)
non-momentary job satisfaction scores were entered into the first step of the
regression and job satisfaction variability scores (either amplitude or phase) were
entered into the second step. This procedure, allowed for the determination of
whether job satisfaction variability predicts performance and turnover intentions
above and beyond non-momentary assessments. In order to fully address thesehypotheses and exploratory analyses, a total of 4 hierarchical regressions were
performed.
First considered was the impact that frequency variation has on the prediction
of job performance above and beyond cross-sectional assessments of job
satisfaction. Looking at Table 21, it can be seen that in step 1 the non-momentary
facets of job satisfaction explained 10.4% of the variance in job performance. The
frequency variation terms were added in step 2 resulting in the explanation of 21.6%
percent of the variance in job performance. The effect size associated with this
change in R2 (11.2%) can be considered close to medium in strength (f 2 = .1261).
This suggests that the inclusion of the frequency variation terms added meaningfully
to the prediction of job performance. In investigating the individual predictors, it can
be seen that non-momentary nature of work satisfaction and frequency variation in
work satisfaction contribute most to the prediction of job performance (t = 2.750 and t
= 2.655) respectively. Again, because the addition of the frequency terms adds
8/11/2019 The antecedents and consequences of the variability in job satisfaction
meaningfully to the prediction of job performance, it can be concluded that frequency
variation in nature of work satisfaction adds explanatory power to the equation
predicting job performance above and beyond the contribution of non-momentary
nature of work satisfaction. Thus hypothesis 12 received some support. However, it
is also important to point out that this relationship was not in the expected direction.
Next considered was the potential impact that amplitude variation in job
satisfaction has above and beyond cross-sectional assessments of job satisfaction to
the prediction of job performance. Again looking at Table 21, the non-momentary job
satisfaction measures explained 12.2% of the variance in job performance. Theaddition of the amplitude variation job satisfaction measures only added 3.3% to the
variance explained, which is a small effect (f 2 = .0341). In addition, non-of the
amplitude variation measures added meaningfully to the prediction of job
performance. Therefore there is no evidence to suggest that amplitude variation in
job satisfaction predicts job performance above and beyond non-momentary
assessments of the job satisfaction. When considering amplitude variation,
hypothesis 12 is not supported.
As explained above, to test hypothesis 13 which suggests the variability in job
satisfaction to predict intent to turnover above and beyond non-momentary
assessments of job satisfaction the non-momentary job satisfaction scores were
entered into the first step of the regression and the job satisfaction variability scores
(either amplitude or phase) were entered into the second step. These results can be
seen in Table 22. First considering frequency variation, the five facets of job
satisfaction measured cross-sectionally explained 30.2% of the variance in intent to
8/11/2019 The antecedents and consequences of the variability in job satisfaction
turnover. Adding in the frequency variation of each facet of job satisfaction only
added 1% to the amount of variance explained, a small effect (f 2 = .0101). Similarly
when considering amplitude variation, the non-momentary job satisfaction measures
together accounted for 32.4% of the variation in intent to turnover, but adding the
amplitude variation predictors only added an additional 1.1% to the variance
explained, again a small effect (f 2 = .0111). Thus, neither frequency nor amplitude
variation in job satisfaction added to the prediction of ones intent to turnover above
and beyond cross-sectional assessments of job satisfaction. In other words, there
was no support for hypothesis 13.DISCUSSION
The present study was conducted to gain a more thorough understanding of
the intraindividual variability in job satisfaction. The present investigation attempted
to contribute to the growing body of literature on within-person job satisfaction
variability in several ways. First, past research has conceptualized variability in terms
of standard deviation (e.g., Ilies & Judge, 2002) and no study to date has explored
both amplitude and frequency conceptualizations of job satisfaction variability.
Someone who varies frequently, but with little magnitude may have the same within-
person standard deviation as an individual who varies with great magnitude but
infrequently. Therefore, the current investigation argued that it may be important to
distinguish between frequency and amplitude variation. As will be discussed in moredetail below, this study reveals that, in fact, this distinction may not be as useful for
modeling short-term variation as originally thought.
8/11/2019 The antecedents and consequences of the variability in job satisfaction
job satisfaction. More specifically, it was expected that individuals high in these
aforementioned personality traits would vary with greater frequency and/or amplitude
given that these personality traits have been shown to intensify individuals’ reactions
to their environment. In addition, it was expected that these relationships may be
different for different facets of job satisfaction as the different facets represent
different aspects of the working environment.
It was surprising to find that positive affectivity was the only personality trait to
relate to job satisfaction variability in the expected way. It was found that having high
levels of positive affectivity is predictive of greater frequency variation in nature ofwork satisfaction. In other words, those characterized as having high levels of
energy, optimism, enthusiasm, and social interest (Watson, Clark, & Tellegen, 1988)
were shown to fluctuate more frequently in their attitudes toward the nature of their
work than those low in PA. Positive affectivity was not found to predict the amplitude
with which individuals fluctuate, thus, PA seems to impact how often individuals
change in their attitudes toward their work but not the magnitude or degree to which
they change their attitudes toward their work. PA was also not predictive of the
variation in any other facet of job satisfaction. This lends some support to the notion
that personality relates differentially to the variation of different facets.
Openness to experience was suggested to be an important predictor of the
frequency with which ones satisfaction with their coworkers varies. In the current
study, this relationship was opposite what was expected in that high levels of
openness to experience were predictive of low levels of frequency variation in
coworker satisfaction. This is not in-line with previous research suggesting that
8/11/2019 The antecedents and consequences of the variability in job satisfaction
openness to experience relates to increased within-person variability (Velting &
Liebert, 1997). However, because individuals high in openness to experience have
an orientation that is creative, curious, and flexible and further have an affinity for
situations involving novelty, diversity, and change (McCrae & Costa, 1989), it is
possible that they are more accustomed to variable situations, and thus do not react
affectively (i.e., do not change in their levels of job satisfaction). In other words, in
context of the current significant finding, while interactions with coworkers may
fluctuate frequently, individuals high in openness to experience may not react
affectively (i.e. change their attitudes) because they are comfortable with, and in factenjoy, the variable interactions. Openness to experience was not found to relate to
the amplitude of variation and did not relate to any other facet of job satisfaction.
Contrary to expectation neither NA, neuroticism nor extraversion were found
to be important predictors of frequency or amplitude variation for any facet of job
satisfaction. While at the bivariate level, both NA and neuroticism significantly
correlated with the frequency with which pay and nature of work satisfaction varied,
these relationships were opposite what was expected. In other words, it was
hypothesized that individuals high in neuoriticism and NA would be more variable in
their job satisfaction, but correlational analyses suggest the opposite. One potential
explanation for these unexpected bivariate correlations is that individuals who are
high in NA and/or neuroticism have generally very low levels of job satisfaction.
Perhaps, there is not as much room for these individuals to vary in their levels of
satisfaction because they are at the bottom of the job satisfaction spectrum. In the
current study, as well as in previous research, cross-sectional assessments of job
8/11/2019 The antecedents and consequences of the variability in job satisfaction
satisfaction relate negatively to cross-sectional assessments of both NA and
neuroticism, providing some support for this notion.
It should be pointed out that the only meaningful relationships observed
utilizing regression were observed for the facets of coworker and nature of work
satisfaction. It is possible that one’s coworkers and the actual nature of one’s work
are more salient/observable in the job environment than pay, promotion, or one’s
supervisor. As such, perhaps personality is relating to the variation of these facets
because, in the actual environment individuals are more aware of changes in their
interactions with their coworkers and of changes in their work and thus personality ismore likely to impact how individuals affectively react to these more cognizant
changes. Said differently, maybe these facets are somewhat easier to affectively
react to, thus personality has a stronger influence on them.
environmental job stressors are more predictive of changing attitudes regarding ones
coworkers than of changing attitudes regarding ones pay, promotion, work or
supervisor. Again, perhaps this is because coworkers are more obvious in the work
environment and thus are more relatable to perceived changes in stressors in the
environment. Job stressor frequency variation did not relate to any other facets of job
satisfaction frequency variation, and it is also important to point out that no significant
relationships were found between the amplitude with which job stressors vary and
amplitude variation in job satisfaction.
It was also expected that personality would moderate the relationship betweenstressor variability and satisfaction variability. This was expected because job
satisfaction was suggested to vary as the situation varies and personality was
suggested to impact how individuals react to situational change. Of the 50 potential
moderating effects, only one was identified as being worthy of further discussion.
More specifically, extraversion was found to moderate the relationship between job
stressor frequency and promotion satisfaction frequency such that this relationship
was stronger for those high in extraversion than for those low in extraversion.
Individuals high in extraversion were expected to react to situational variability more
than those low in extraversion. In the current study, the frequency of situational
change was found to predict the frequency of change in individuals’ attitudes toward
promotion for those high in extraversion. This relationship was not as strong for
those low in extraversion. While this moderating effect was as expected, extraversion
was not found to moderate the relationship between stressor variation and
satisfaction variation for any other facet of satisfaction. In addition, no other
8/11/2019 The antecedents and consequences of the variability in job satisfaction
personality trait was found to moderate the stressor variation – satisfaction variation
relationship.
The Predictive Power of Job Satisfaction Variability
The final goal of the current study was to identify whether frequency and/or
amplitude variation in job satisfaction is important to the prediction of the important
organizational outcomes of job performance and intent to turnover. To provide more
detail, it was expected that increased variation in the facets of satisfaction would be
predictive of lower job performance and higher intent to turnover.
Results did not generally support the hypotheses set forth. While thefrequency of with which nature of work satisfaction varied was positively related to
job performance, and in fact, was positively predictive of job performance above and
beyond cross-sectional assements of job satisfaction, this was opposite of what was
expected. Given the arguments presented in the current study, it may be that if
nature of work satisfaction is varying with great frequency then the actual nature of
ones work may be changing with great frequency. In other words, there is potentially
greater variety in the nature of ones work. Similar to the concept of skill variety, this
nature of work variability can potentially be conceptualized as a motivator (Hackman
& Oldham, 1976), and thus could potentially result in improved performance.
Therefore, this unexpected result can be, somewhat, theoretically supported. It is
also important to point out that this significant result is again with nature of work
satisfaction, a facet that was also shown to relate to personality. Amplitude variation
in nature of work was not predictive of performance, nor was variation in any other
facet of satisfaction. Job satisfaction variation was also not predictive of intent to
8/11/2019 The antecedents and consequences of the variability in job satisfaction
While results of the current study were not generally supportive of the
hypotheses, there are some things that are important to highlight. First, it is likely
important to continue to investigate the intraindividual variability of job satisfaction at
the facet level as the variation of different facets related differentially to the predictors
and criteria in this study. The facets of coworker satisfaction and nature of work
satisfaction seem to be particularly important. As already eluded to, this may
because these facets are the most obvious in the work environment and are thusparticularly important to consider when investigating within-person variation which is
theorized to result from reactions to the environment as impacted by personality.
Second, it should be pointed out that all of the meaningful relationships
discovered involved frequency variation in job satisfaction, none involved amplitude
variation. One potential conclusion is that the frequency with which individuals vary is
likely more important than the magnitude of their variation. Or, perhaps given the
time frame of the current study, meaningful changes in frequency were easier to
capture than meaningful changes in amplitude. In other words, large amplitude
changes may not occur very often and thus may be potentially difficult to accurately
model over short time frames. This provides some indication that the
amplitude/frequency distinction may not be useful for modeling short-term, within-
person attitude variation. And again, because coworkers and the nature of ones work
are more situationally obvious, they are more susceptible to frequent changes vs.
pay, promotion or supervisors which may not be as susceptible to frequent changes.
8/11/2019 The antecedents and consequences of the variability in job satisfaction
varying frequently and with great amplitude at the same time. Also, as previously
mentioned, it is possible that amplitude variation is more difficult to truly capture over
short-time frames. If the goal of the research is to understand what impacts and what
is impacted by short-term, within person variation in job satisfaction, the frequency
and amplitude distinction may not be important and actually, standard deviation may
be a more inclusive and less restrictive measure of variability.
It is also important to consider how the testing itself may have impacted the
current findings. Practice, familiarity or other forms of participant reactivity impact the
inferences about the observed covariation between the variables of interest(Shadish, et al., 2002). This is of particular concern in the present study given that
participants completed the same measure at multiple time points within the same
day. Individuals may have intentionally altered their responses on subsequent job
satisfaction assessments to either seem consistent or intentionally different from
their previous assessment. However, in order to try and reduce potential participant
reactivity, participants were instructed to think of the job satisfaction at the current
moment in time. It is also important to consider the effects of attrition (Shadish et al.,
2002). There were several participants that did not remain in the study for the
duration and thus were not included as participants. It is possible that different
results would have been obtained had those participants remained.
Additional Limitations and Future Research
As with all empirical research, there are several limitations of the current study
that warrant further discussion. First and foremost, the procedures used to model
frequency and amplitude variation were not as useful as originally expected for
8/11/2019 The antecedents and consequences of the variability in job satisfaction
understanding short-term within-person variation. The primary goals of the present
study were to gain insight into how personality and situational variables interact to
predict job satisfaction variation and further to better understand the predictive power
of this variation. These were important questions to address given the growing
popularity in assessing job satisfaction intraindividually. Based on the results of the
current study, one may be inclined to suggest that neither personality nor situational
variability are important to the prediction of job satisfaction variability and that job
satisfaction variability is of little importance to the prediction of organizational
outcomes. However, this conclusion may be premature given that there were somemeaningful relationships unveiled and that a different methodology for modeling
variability may have captured variability more accurately. Regardless, the present
study did not provide conclusive evidence regarding the correlates of job satisfaction
variability and additionally raised several interesting questions. It is suggested that
the hypotheses put forth in the current study be investigated utilizing standard
deviation as a measure of variation. Looking at such results in concert with the
results of the current study would provide a more comprehensive picture of the
antecedents and consequences of the within-person variation of job satisfaction.
A second limitation of the present study results from the time frame chosen for
the measurement of job satisfaction. More specifically, the present study modeled
variability in job satisfaction over the course of a three week period. Measurement
using this time frame does not allow for modeling changes that may occur over
longer time frames than three weeks. For example, the chosen measurement period
did not allow for the modeling of changes in amplitude that occur outside of the three
8/11/2019 The antecedents and consequences of the variability in job satisfaction
week period assessed. However, given that within-person variability seems to
primarily exist within the same day (Young & Baltes, 2006) and that within-person
variability was previously measured over similar time frames (e.g., Ilies & Judge,
2002), it seems that the time frame chosen was appropriate for investigating short-
term within-person variation in job satisfaction.
In addition, participants of the present study consisted of university staff with
continuous access to a computer. Participants primarily worked in clerical and
administrative positions, thus it is not clear whether job satisfaction variability would
be modeled in the same way for other positions (e.g., sales or manufacturingpositions). Thus, future research should also investigate these questions using more
diverse samples.
This study is further limited in that there are likely additional individual
difference variables that are important to the prediction of job satisfaction variability.
For example, locus of control, or the extent to which individuals feel that they are in
control of their own destiny (i.e. have an internal locus of control) vs. holding the view
that things happen randomly or by chance (i.e., have an external locus of control)
may impact the way in which they react to situational change. Thus, locus of control
may be important to consider. There are also many additional organizational
variables that job satisfaction variability may affect (e.g., organizational citizenship
behaviors, organizational commitment). Thus future research should continue to
investigate the individual difference predictors of and outcomes of job satisfaction
variability.
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Similarly, utilizing a measure of job stressors as a means of assessing the
work environment (i.e., the situation) is, of course, not the only way that the work
environment can be measured. To provide a few examples, future research could
assess situational change by utilizing momentary measures of the five core job
characteristics of Hackman and Oldham’s (1976) job characteristics model. To
provide another possibility, future research could code qualitative self-reports of the
work environment. While these are also potentially interesting methods of assessing
the situation, they are conceptually different from assessing the momentary
environmental stressors present at work. Thus future research should continue toinvestigate how situational change and job satisfaction variability are related for
different operationalizations of situational change.
Summary
Though many of the hypothesized relationships were not supported and this
study is not without limitations, there are a few potentially important take home
points. Frequency variation in job stressors, as well as three of the five personality
traits, were shown to be important in explaining the frequency variation of three
different facets of job satisfaction. Therefore it seems worthwhile to continue to
investigate the roles that personality and situational change play in describing the
within-person variation in job satisfaction. In addition, variation in the facet of work
satisfaction was important to the prediction of job performance above and beyond
mean levels of work satisfaction measured cross-sectionally, lending some support
to the notion that variation in job satisfaction contributes to the prediction of job
performance above and beyond the effects of cross-sectional assessments. Again,
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Extra -.077 .149 -.059 -.517ttNeur .070 .187 .053 .376
Open -.134 .216 -.067 -.619Note. ttNeuroticism is scored such that high scores indicate high emotional stability (i.e. low neuroticism) *p < .05. Pay Freq =
Frequency variation in pay satisfaction, Pay Amp = Amplitude variation in pay satisfaction, Pro Freq = Frequency variation in
promotion satisfaction, Pro Amp = Amplitude variation in promotion satisfaction, Sup Freq = Frequency variation in supervisor
satisfaction, Sup Amp = Amplitude variation in supervisor satisfaction, Co Freq = Frequency variation in coworker satisfaction,
Co Amp = Amplitude variation in coworker satisfaction, Wrk Freq = Frequency variation in work satisfaction, Wrk Amp =
Amplitude variation in work satisfaction, NA = Negative affectivity, PA = Positive affectivity, Extra = Extraversion, Neur =
Neuroticism, Open = Openness to Experience.
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Summary of Hierarchical Regression Analyses Examining the Moderating Effect of Personality on the Relationship Between Frequency Variation in Job Stressors and Frequency Variation in Pay Satisfaction
Note. ttNeuroticism is scored such that high scores indicate high emotional stability (i.e. low neuroticism) *p < .05. f 2 = effectsize for R2. Strs Freq = Frequency variation in job stressors
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Summary of Hierarchical Regression Analyses Examining the Moderating Effect of Personality on the Relationship Between Frequency Variation in Job Stressors and Frequency Variation in Promotion Satisfaction
Note. ttNeuroticism is scored such that high scores indicate high emotional stability (i.e. low neuroticism) *p < .05. f 2 = effectsize for R2. Strs Freq = Frequency variation in job stressors
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Summary of Hierarchical Regression Analyses Examining the Moderating Effect of Personality on the Relationship Between Frequency Variation in Job Stressors and Frequency Variation in Supervisor Satisfaction
Note. ttNeuroticism is scored such that high scores indicate high emotional stability (i.e. low neuroticism) *p < .05. f 2 = effectsize for R2. Strs Freq = Frequency variation in job stressors
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Summary of Hierarchical Regression Analyses Examining the Moderating Effect of Personality on the Relationship Between Frequency Variation in Job Stressors and Frequency Variation in Coworker Satisfaction
Note. ttNeuroticism is scored such that high scores indicate high emotional stability (i.e. low neuroticism) *p < .05. f 2 = effectsize for R2. Strs Freq = Frequency variation in job stressors
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Summary of Hierarchical Regression Analyses Examining the Moderating Effect of Personality on the Relationship Between Frequency Variation in Job Stressors and Frequency Variation in Nature of Work Satisfaction
Note. ttNeuroticism is scored such that high scores indicate high emotional stability (i.e. low neuroticism) *p < .05. f 2 = effectsize for R2. Strs Freq = Frequency variation in job stressors
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Summary of Hierarchical Regression Analyses Examining the Moderating Effect of Personality on the Relationship Between Amplitude Variation in Job Stressors and Amplitude Variation in Pay Satisfaction
Note. ttNeuroticism is scored such that high scores indicate high emotional stability (i.e. low neuroticism) *p < .05. f 2 = effectsize for R2. Strs Freq = Frequency variation in job stressors
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Summary of Hierarchical Regression Analyses Examining the Moderating Effect of Personality on the Relationship Between Amplitude Variation in Job Stressors and Amplitude Variation in Promotion Satisfaction
Note. ttNeuroticism is scored such that high scores indicate high emotional stability (i.e. low neuroticism) *p < .05. f 2 = effectsize for R2. Strs Freq = Frequency variation in job stressors
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Summary of Hierarchical Regression Analyses Examining the Moderating Effect of Personality on the Relationship Between Amplitude Variation in Job Stressors and Amplitude Variation in Supervisor Satisfaction
Note. ttNeuroticism is scored such that high scores indicate high emotional stability (i.e. low neuroticism) *p < .05. f 2 = effectsize for R2. Strs Freq = Frequency variation in job stressors
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Summary of Hierarchical Regression Analyses Examining the Moderating Effect of Personality on the Relationship Between Amplitude Variation in Job Stressors and Amplitude Variation in Coworker Satisfaction
Note. ttNeuroticism is scored such that high scores indicate high emotional stability (i.e. low neuroticism) *p < .05. f 2 = effectsize for R2. Strs Freq = Frequency variation in job stressors
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Summary of Hierarchical Regression Analyses Examining the Moderating Effect of Personality on the Relationship Between Amplitude Variation in Job Stressors and Amplitude Variation in Nature of Work Satisfaction
Note. ttNeuroticism is scored such that high scores indicate high emotional stability (i.e. low neuroticism) *p < .05. f 2 = effectsize for R2. Strs Freq = Frequency variation in job stressors
8/11/2019 The antecedents and consequences of the variability in job satisfaction
1. At this moment, I feel I am being paid a fair amount for the work I do.2. At this moment, there is really too little chance for promotion on my job.
3. At this moment, my supervisor is quite competent in doing his/her job.
4. At this moment, I am not satisfied with the benefits I receive.
5. At this moment, I feel that I receive the recognition that I should for doing a
good job.6. At this moment, many of the rules and procedures make doing a good job
difficult.
7. At this moment, I like the people I work with.
8. At this moment, I feel my job is meaningless.
9. At this moment, communications seem good within this organization.
10. At this moment, I feel that raises are too few and far between.
11. At this moment, those who do well on the job stand a fair chance of being
promoted.
12. At this moment, my supervisor is unfair to me.
13. At this moment, I feel that the benefits we receive are as good as most other
organizations offer.
14. At this moment, I do not feel that the work I do is appreciated.
15. At this moment, my efforts to do a good job are seldom blocked by red tape.
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Cohen, J. (1988). Statistical Power analysis for the behavioral sciences ( 2nd ed.).
Mahwah, NJ: Erlbaum.
Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for
the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Earlbaum
Associates.
Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied Multiple
Regression/Correlation Analysis for the Behavioral Sciences, 3 rd Edition.
Lawerence Erlbaum Associates: Mahwah, NJ.
Connolly, J. J., & Viswesvaran, C. (2000). The role of affectivity in job satisfaction: Ameta-analysis. Personality and Individual Differences, 29, 265-281.
Cranny, C. J., Smith, P. C., & Stone, E. F. (Eds.) (1992). Job Satisfaction: Advances
in Research and Applications. The Free Press: New York.
Dormann, C. Fay, D., Zapf, D., & Frese, M. (2006). A state-trait analysis of job
satisfaction: On the effect of core self evaluations. Applied Psychology: An
International Review, 55, 27-51.
Fisher, C. A. (2003). Why do lay people believe that satisfaction and performance
are correlated? Possible sources of commonsense theory. Journal of
Organizational Behavior, 24, 753-777.
Fisher, R. A. (1929). Tests of significance in harmonic analysis. Proceedings of the
Royal Society, Series A, 125, 54-59.Fisher, C. D., & Gitelson, R. (1983). A meta-analysis of the correlates of role conflict
and role ambiguity. Journal of Applied Psychology, 68, 320-333.
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Heller, D., Judge, T.A. & Watson, D. the confounding role of personality and trait in
the relationship between job and life satisfaction. Journal of Organizational
Behavior, 23, 815-835.
Hepburn, L., & Eysenck, M. W. (1989). Personality, average mood and mood
variability. Personality and Individual Differences, 10, 975-983.
Ilies, R., & Judge, T. A. (2002). Understanding the dynamic relationships among
personality, mood and job satisfaction: a field experience sampling study.
Organizational Behavior and Human Decision Processes, 89, 1119-1139.
Ilies, R. & Judge, T. A. (2004). An experience sampling measure of job satisfactionand its relationships with affectivity, mood at work, job beliefs, and general job
satisfaction. European Journal of Work and Organizational Psychology, 13(3),
367-389.
Jamal, M. (1990). Relationship of job stress and Type A behavior to employees’ job
satisfaction, organizational commitment, psychosomatic health problems, and
turnover motivation. Human Relations, 43(8), 727-738.
Judge, T. A., & Bono, J. E. (2001). Relationship of core self-evaluations traits- self-
esteem, generalized self-efficacy, locus of control, and emotional stability-
with job satisfaction and performance: A meta-analysis. Journal of Applied
Psychology, 86, 80-92.
Judge, T. A., Bono, J. E., & Locke, E. A. (2000). Personality and job satisfaction: Themediating role of job characteristics. Journal of Applied Psychology, 85, 237-
249.
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Kammeyer-Mueller, J. D., Wanberg, C. R., Glomb, T. M., & Ahlburg, D. (2005). The
role of temporal shifts in turnover processes: It’s about time. Journal of
Applied Psychology, 90, 644-658.
Katzell, R. A., Thompson, D. E. & Guzzo, R. A. (1992). How job satisfaction and
performance are and are not linked in C. Cranny, P. Smith, & E. Stone (Eds.),
Job Satisfaction: How People Feel About Their Jobs and How it Effects Their
Performance. New York: Lexington Books.
Keppel, G. & Wickens, T. D. (2004). Design and Analysis: A Researcher’s Handbook
4 th
ed. Englewood Cliffs, NJ: Prentice Hall.Kemey, E. R., Mossholder, K. W. & Bedian, A. (1985). Role stress, physical
symptomatology and turnover intentions: A causal analysis of three alternative
specifications. Journal of Occupational Behavior, 8, 11-23.
Larsen, R. J., & Ketelaar, T. (1991). Personality and susceptibility to positive and
negative emotional states. Journal of Personality and Social Psychology, 61,
132-140.
Lazarus, R. S. (1990). Theory-based stress measurement. Psychological Inquiry, 1,
3-13.
Lee, T. W., & Mitchell, T. R. (1994). An alternative approach: The unfolding model of
voluntary employee turnover. Academy of Management Review, 19, 51-89.
Lee, T. W., Mitchell, T. R., Holtom, B. C., McDaniel, L. S., & Hill, J. W. (1999) Theunfolding model of voluntary turnover: A replication and extension. Academy
of Management Journal, 42, 450-462.
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Organ, D. W., & Ryan, K. (1995). A meta-analytic review of attitudinal and
dispositional predictors of organizational citizenship behavior. Personnel
Psychology, 48, 775-802.
Parkes, K. R. (1990). Coping, negative affectivity, and the work environment:
Additive and interactive predictors of mental health. Journal of Applied
Psychology, 75, 399-408.
Podsakoff, P. M., & Williams, L. J. (1986). The relationship between job performance
and job satisfaction. In Generalizing From Laboratory to Field Settings, Edwin
A. Locke (ed.), Lexington, MA: Lexington Books.Quarstein, V. A., McAfee, R. B., & Glassman, M. (1992). The situational occurrences
theory of job satisfaction. Human Relations, 45(8), 859-873.
Raaijmakers, Q. A. (1999). Effectiveness of different missing data treatments in
surveys with likert-type data: Introducing the relative mean substitution
approach. Educational and Psychological Measurement, 59, 725-748.
Rain, J. S., Lane, I. M., & Steiner, D. D. (1991). A current look at the job
satisfaction/life satisfaction relationship: Review and future considerations.
Human Relations, 44, 287-307.
Roth, R. L, & Switzer, F. S. (1995). A monte carlo analysis of missing data
techniques in a HRM setting. Journal of Management, 21, 1003-1023.
Roznowski, M., & Hulin, C. (1992). The scientific merit of valid measures of generalconstructs with special reference to job satisfaction and job withdrawal. In C.J.
Cranny, P.C. Smith & E. F. Stone, (Eds.), Job Satisfaction, Lexington Books,
New York.
8/11/2019 The antecedents and consequences of the variability in job satisfaction
Given the recent increase in the investigation of the intraindividual variability of
job satisfaction (e.g., Ilies & Judge, 2002) the present study sought to further our
understanding of this topic through more comprehensively investigating the potential
predictors and criteria of the frequency and amplitude with which individuals vary in
their levels of job satisfaction over short time frames. It was hoped that by including
both personality (i.e., NA, PA, Neuroticism, Extraversion, and Openness toExperience) and situational predictors (i.e. job stressors), results from this study
would help clarify how individual differences interact with situational characteristics to
explain patterns of within-person variability. In addition, this study also sought to
determine the extent to which job satisfaction variability is important to the prediction
of employee job performance and turnover intentions. Results from the current study
suggest that it may not be useful to distinguish between the frequency and amplitude
with which individuals vary in their levels of job satisfaction over short time frames.
Aside from this general conclusion, the situational predictor and three of the five
personality traits investigated were shown to be important in explaining the frequency
8/11/2019 The antecedents and consequences of the variability in job satisfaction
with which individuals’ coworker, nature of work and promotion satisfaction varied.
Variation in the facet of work satisfaction was also found to be an important predictor
of job performance, and this remained true after accounting for mean levels of work
satisfaction measured cross-sectionally. Finally, results lend some support to
suggest the facets of coworker and nature of work satisfaction to be particularly
important when considering short-term within-person variation. While many of the
hypothesized relationships were not supported, given the current results, it seems
worthwhile to continue to investigate the roles that personality and situational change
play in describing the within-person variation of different facets of job satisfaction aswell as the role of this within-person variation in predicting organizational outcomes.
8/11/2019 The antecedents and consequences of the variability in job satisfaction