'Flow* at work: An experience sampling approach - E. Kevin Kelloway
'Flow* at work: An experience sampling approach - E. Kevin Kelloway
'Flow* at work: An experience sampling approach - E. Kevin Kelloway
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
595<br />
British<br />
Psychological<br />
Journal of Occup<strong>at</strong>ional and Organiz<strong>at</strong>ional Psychology (2009). 82, 595-615 n d U H c • h,<br />
((•' 2009 The British Psychokigic<strong>at</strong> Society *-'^^ bociety<br />
The<br />
www. bpsjournals.co.uk<br />
<strong>'Flow*</strong> <strong>at</strong> <strong>work</strong>: <strong>An</strong> <strong>experience</strong> <strong>sampling</strong> <strong>approach</strong><br />
Clive J. Fullagar'* and E. <strong>Kevin</strong> <strong>Kelloway</strong>^<br />
'Kansas St<strong>at</strong>e University. Manh<strong>at</strong>tan, Kansas, USA<br />
^Saint Mary's University, Halifax, Canada<br />
One of the core constructs of the positive psychology movement is th<strong>at</strong> of 'f<br />
oroptimal<strong>experience</strong>. The current study investig<strong>at</strong>ed the rel<strong>at</strong>ionship between 'flow', the<br />
core job dimensions, and subjective well-being (SWB). as well as distinguishing between<br />
the st<strong>at</strong>e and trait components of flow. Experience <strong>sampling</strong> methodology (ESM) was<br />
used to track 40 architectural students over a 15 week semester while they engaged in<br />
studio <strong>work</strong>. Hierarchical linear modelling (HLM) indic<strong>at</strong>ed th<strong>at</strong> 74% of the variance<br />
in flow was <strong>at</strong>tributable to situ<strong>at</strong>ional characteristics compared to dispositional factors.<br />
Results also indic<strong>at</strong>ed th<strong>at</strong> academic <strong>work</strong> th<strong>at</strong> was high in skill variety and autonomy was<br />
associ<strong>at</strong>ed with flow. Flow was found to be correl<strong>at</strong>ed with positive mood. Cross-lagged<br />
regression analysis showed th<strong>at</strong> momentary flow was predictive of momentary mood and<br />
not vice versa. The strengths and limit<strong>at</strong>ions of using ESM to study subjective <strong>work</strong><br />
<strong>experience</strong>s and well-being are discussed, as well as the implic<strong>at</strong>ions of the study of flow<br />
or optimal <strong>experience</strong> for industrial/organiz<strong>at</strong>ional psychology.<br />
,.<br />
A repercussion of Martin Seligman's (1999) call for a 'positive psychology' can be<br />
witnessed in the upswing in organiz<strong>at</strong>ional research th<strong>at</strong> emphasizes optima!<br />
<strong>experience</strong>, transcendent pertbrmance, excellence, and positive deviance ((Cameron,<br />
Dutton, & Quinn, 2003; Luthans, 2001). One concept th<strong>at</strong> would seem to be particularly<br />
appropri<strong>at</strong>e to positive or^ianiz<strong>at</strong>iona! scholarship isth<strong>at</strong> of "flow'. Flow' was a term first<br />
coined by Csikszentmihal) i (1975) wbo defined it as the holistic sens<strong>at</strong>ion tb<strong>at</strong> people<br />
feel when they act witb total involvement (p. 4)'. Flow bas been defined as the<br />
<strong>experience</strong> of <strong>work</strong>ing <strong>at</strong> full capacity, with intense engagement and effortless action,<br />
where personal skills m<strong>at</strong>ch required challenges (Nakamura & Csikszentniihalyi, 2002).<br />
It is regarded as an 'optimal <strong>experience</strong>' to such an extent tb<strong>at</strong> the two terms are often<br />
used interchangeably.<br />
Originally Csikszentmihalyi studied flow among painters, rock climbers, dancers,<br />
and musicians. Most research has focused on the <strong>experience</strong> of flow in voluntary<br />
and pleasurable leisure and sporting activities. However, the <strong>experience</strong> of flow<br />
has also been frequently reported wbile engaging in <strong>work</strong> rel<strong>at</strong>ed tasks as opposed to<br />
leisure activities (Csikszentmibalyi & LeFevre, 1989; Fave & Massimiiii, 1988). This is not<br />
* Correspondence should be addressed to Dr Clive j. Fullagor, Department of Psychology, Kansas St<strong>at</strong>e University, Manh<strong>at</strong>tan,<br />
KS 66506, USA (e-mail: fullagar@ksu.edu).<br />
DOI: 10.1348/096317908X357903<br />
I
596 Give }. Fullagar and L <strong>Kevin</strong> Keihway<br />
siir|:)rising given tii<strong>at</strong> many of the precursors of flow (such as immedi<strong>at</strong>e feedback,<br />
comiiiensur<strong>at</strong>e challenges and skills, and clear goals) are more likely to be found in <strong>work</strong><br />
activities. Despite the fact th<strong>at</strong> flow <strong>experience</strong>s have been reported in <strong>work</strong>, rel<strong>at</strong>ively<br />
little research has studied flow in the <strong>work</strong> context. The current research is a<br />
longitudinal investig<strong>at</strong>ion of some of the antecedents and outcomes of flow while<br />
engaging in <strong>work</strong> activities.<br />
Oper<strong>at</strong>ionaliz<strong>at</strong>ions of flow<br />
A remarkable consistency has been found in the description of flow across diverse<br />
settings (C^sikszentmihalyi, 1975). The dominant characteristic of early models of flow<br />
was the emphasis on the balance between challenges and skills (Csikszentmihalyi, 1975).<br />
However, qualit<strong>at</strong>ive and quantit<strong>at</strong>ive research has identified several key elements<br />
inherent in the <strong>experience</strong> of flow (Csikszentmihalyi, 1990,1993;Jackson, 1996:Jackson<br />
& Marsh, 1996). Specifically flow has been characterized as consisting of nine<br />
component st<strong>at</strong>es (Csikszentmihalyi, 1990, 1993). The first is a challenge-skill halunce<br />
and it refers to the m<strong>at</strong>ch th<strong>at</strong> must be present between the skills required to peribrni<br />
the task and challenges of the situ<strong>at</strong>ion. If the challenges in the activity exceed the<br />
individual's skill level then faistr<strong>at</strong>ion r<strong>at</strong>her than flow occurs. If the activity does not<br />
challenge the individual's skills then the result is boredom. If both challenges and skills<br />
are balanced but do not exceed a specific level of difficulty and complexity then ap<strong>at</strong>hy is<br />
likely to occur. Flow only occurs when the balance between challenges and skills<br />
exceeds a level th<strong>at</strong> is topical for daily <strong>experience</strong>s. Second, there is a merging of action<br />
and <strong>at</strong>vareness. One s involvement in the task is such th<strong>at</strong> behaviour becomes autom<strong>at</strong>ic<br />
and spontaneous and there is little awareness of the self other than wh<strong>at</strong> one is doing.<br />
Third, there is a strong sense of wh<strong>at</strong> has to be done and clarity of goals. Fourth, the<br />
activity itself provides clear, immedi<strong>at</strong>e and unambiguous feedback concerning<br />
progress towards one's goals. Fifth, there is an extremely high degree of involvement<br />
with, focus on, and concentr<strong>at</strong>ion on the task <strong>at</strong> hand. Attention and energies are<br />
exquisitely focused on the task and there is an absence of distraction. Sixth, there is a<br />
sense of exercising control without having to try and be in control thai is termed the<br />
paradox of control. People in flow report they feel in control but as soon as their<br />
<strong>at</strong>tention shifts to tr> ing to maintain control they lose the sense of flow.Seventh, there is<br />
a loss of self-consciousness as all concern for self disappears and the person becomes<br />
one with the activity. Eighth, there is a transform<strong>at</strong>ion of time th<strong>at</strong> seems to occur in<br />
which there is loss of time awareness. Finally, the ninth dimension of flowis the autotelic<br />
<strong>experience</strong>. Csikszentmihalyi (1990) coined the term autotelic <strong>experience</strong> from the two<br />
Greek words auto, meaning self, and telos meaning goal. Such optimal <strong>experience</strong>s are an<br />
end in themselves and are so enjoyable th<strong>at</strong> they become intrinsically motiv<strong>at</strong>ing. Flow is<br />
described as such an enjoyable <strong>experience</strong> th<strong>at</strong> one is motiv<strong>at</strong>ed to return to this st<strong>at</strong>e.<br />
Considered together these nine dimensions of flow provide a comprehensive<br />
measure of optimal <strong>experience</strong> (Jackson & Ecklund, 2004). In the most recent research<br />
on flow in the <strong>work</strong>place flow has been measured as either a global construct<br />
(e.g. Demerouti, 2006) or the challenge/skill dimension of flow has been focused on<br />
(e.g. Eisenberger. Jones, Stinglhaniber, Slianock, & Randall, 2005). Such unidimensional<br />
<strong>approach</strong>es have been found to be psychometrically acceptable üackson & Ecklund,<br />
2002; Jackson & Marsh, 1996; Marsh & Jackson, 1999), Indeed, the nine subscaies of<br />
flow have been found to be highly correl<strong>at</strong>ed and there appears to be a substantial<br />
proportion of shared variance (Jackson & Ecklund, 2004).
Flow <strong>at</strong> <strong>work</strong> 597<br />
Ffow: St<strong>at</strong>e or trait<br />
Theodore Mischel (1969) has st<strong>at</strong>ed th<strong>at</strong> 'clarific<strong>at</strong>ion of key concepts must be part of<br />
the busiticss of theory construction in psychology' (p. 264). <strong>An</strong> important piece<br />
of this clarific<strong>at</strong>ion process is distinguishing whether a construct is a trait or a st<strong>at</strong>e<br />
(Alien & Potkay. I98I). A trait construct is a behaviour th<strong>at</strong> individuals perform regularly<br />
- it is a reliably permanent internal disposition, A st<strong>at</strong>e construct, on the other hand,<br />
is more ephemeral and is a behaviour th<strong>at</strong> is performed occasionally in response to<br />
situ<strong>at</strong>ional contingencies (Allport & Odbtrt, 193(>)-<br />
There is some deb<strong>at</strong>e as to whether flow is a st<strong>at</strong>e or a trait. Flow as a st<strong>at</strong>e is defined<br />
as a psychological st<strong>at</strong>e of mind th<strong>at</strong> is transitory, existing <strong>at</strong> a given moment in time and<br />
<strong>at</strong> a partictilar level of intensity, and <strong>experience</strong>d while performing a specific activity.<br />
However, Csikszcntmihaiyi (1990) has also referred to the 'autotelic personality" in th<strong>at</strong><br />
certain individuals may have psychological characteristics th<strong>at</strong> make them more prone<br />
to the <strong>experience</strong> of flow regardless of the situ<strong>at</strong>ion. Allen and Potkay (1981) have<br />
pointed out th<strong>at</strong> the distinction between tniit and st<strong>at</strong>e constructs is frequently and<br />
arbitrarily made by varying the instructions on scales th<strong>at</strong> contain the same items. Wben<br />
a st<strong>at</strong>e construct is being assessed participants are asked to r<strong>at</strong>e their st<strong>at</strong>e of minti<br />
'today' or 'now'. When traits are measured, subjects are required to r<strong>at</strong>e their st<strong>at</strong>e of<br />
mind 'in general". This arbitrary distinction can be found in oper<strong>at</strong>ionaliz<strong>at</strong>ions of flow<br />
(Jackson & Ecklund, 2004). Dispositional assessments of tbe fiow <strong>experience</strong> (e.g. tbe<br />
dispositional fiow scale, Jackson & Ecklund, 2004) measure the general tendency to<br />
<strong>experience</strong> flow during an activity usually selected by the participant. Respondents are<br />
required to recall and r<strong>at</strong>e the frequency' with which he/she <strong>experience</strong> tbe various<br />
characteristics of flow G^ckson & Ecklund, 2004), The premise bebind this <strong>approach</strong> is<br />
th<strong>at</strong> individuals who report higher frequencies of flow characteristics have a gre<strong>at</strong>er<br />
predisposition to <strong>experience</strong> How, St<strong>at</strong>e measures of flow (e,g. the flow st<strong>at</strong>e scale (FSS),<br />
Jackson & Ecklund, 2004) are usually administered immedi<strong>at</strong>ely following an activity',<br />
and assess the level of flow characteristics during the activity. By surveying participants<br />
close to the conclusion of an activity it is assumed thai a more accur<strong>at</strong>e assessment of<br />
the flow <strong>experience</strong> is assessed, one th<strong>at</strong> is less susceptible to recall bias. As Allen and<br />
Potkay (19K1) note, changing the instructions presumes th<strong>at</strong> there is a change in the<br />
respondent's orient<strong>at</strong>ion from st<strong>at</strong>e to trait. Using such minor changes in the wording of<br />
survey instructions to differenti<strong>at</strong>e between st<strong>at</strong>e and trait constructs is an arbitrary and<br />
tenuous distinction.<br />
Tbere are other, more valid ways of distinguishing between st<strong>at</strong>e and trait constrticts,<br />
Zuckerman (1983) has suggested th<strong>at</strong> st<strong>at</strong>e and trait variables behave in ver)' different<br />
ways. First, trait variables should be expected to exhibit low vari<strong>at</strong>ion from situ<strong>at</strong>ion to<br />
situ<strong>at</strong>ion. Th<strong>at</strong> is they should have high test-retest reliabilit)' and shotild n
598 Clive j. Fullagar and L <strong>Kevin</strong> Kellaway<br />
propensity to regularly <strong>experience</strong> flow Oackson, Kitniecik, Ford, & Marsh, 1998).<br />
Most research th<strong>at</strong> has investig<strong>at</strong>ed both the trait and st<strong>at</strong>e aspects of flow has<br />
proposed th<strong>at</strong> the rel<strong>at</strong>ionship between situ<strong>at</strong>ional characteristics and st<strong>at</strong>e-based flow<br />
is tnoder<strong>at</strong>ed by dispositional eletnents. For example, Kimiecik and Stein (1992) have<br />
suggested th<strong>at</strong> ibr <strong>at</strong>hletes tbe rel<strong>at</strong>ionship between <strong>experience</strong> of flow and engaging<br />
in sporting activities is contingent upon whether tbe <strong>at</strong>hletes have a task- or egt><br />
involved goal orient<strong>at</strong>ion. More recently, two studies th<strong>at</strong> have investig<strong>at</strong>ed flow <strong>at</strong><br />
<strong>work</strong> found tb<strong>at</strong> tbe rel<strong>at</strong>ionship between flow (challenge-skill balance) and mood, task<br />
interest and performance was moder<strong>at</strong>ed by need ibr achievement ( Eisenberger et al.,<br />
2005), and th<strong>at</strong> tbe rel<strong>at</strong>ionship between flow and in-role and extra-role performance<br />
was dependent on conscientiousness (Demerouti, 2006).<br />
Although personality characteristics stich as achievement motiv<strong>at</strong>ion, aestheticism,<br />
and inquisitiveness have been associ<strong>at</strong>ed with flow (Csikszentmibalyi, R<strong>at</strong>bunde, &<br />
Wlialen, 1993: Jackson, 1984) no research bas been conducted tb<strong>at</strong> sifts out tbe trait<br />
versus st<strong>at</strong>e components of optimal <strong>experience</strong>. The first aim of the current study was<br />
to track flow over <strong>at</strong>i extended period of time in order to assess the extent to which<br />
it varied from task to task and across individuals. If flow is a dispositional construct,<br />
tben one would expect vari<strong>at</strong>ion between tbe individual's predisposition to <strong>experience</strong><br />
flow, and little moment-to-moment variance. On the otber hand, if tbere is a large st<strong>at</strong>erel<strong>at</strong>ed<br />
component in flow, tben one should expect to find considerable vari<strong>at</strong>ion in<br />
momentary flow.<br />
Hypothesis /: R<strong>at</strong>ings of flow will exhibit both within-individual variance (st<strong>at</strong>e) and betweenindividual<br />
(trait) variance with the fornner accounting for most of the variance in r<strong>at</strong>ings.<br />
Task characteristics and fiow<br />
lhere has been a considerable amount of research th<strong>at</strong> has investig<strong>at</strong>ed the links<br />
between specific aspects of a person s job and his/her well-being (see Warr, 1999, for a<br />
stimmary). One frame<strong>work</strong> th<strong>at</strong> has been particularly influential in tinderstanding this<br />
rel<strong>at</strong>ionship is the job characteristics model (Hackman & Oldham, 1980; Oldham, 1996).<br />
Turner, Barling, and Zachar<strong>at</strong>os (2002) have argued the utility of tbis model in understanding<br />
the concept of "healthy' <strong>work</strong> and in promoting psychological and physical<br />
well-being.<br />
The JCM outlines five core job cbaracteristics th<strong>at</strong> influence <strong>work</strong>ers' <strong>at</strong>titudes and<br />
behaviour (Hackman & Oldbam. 1975, 1980). Skill variety refers to tbe degree to wbich<br />
tbe job requires different activities and skills to carry out the task. Task identity is the<br />
extent to wbich the job requires completion of a whole and identifiable piece of <strong>work</strong>.<br />
Task significance indic<strong>at</strong>es the degree to which the job bas a meaningful impact on<br />
otber people. Autonomy refers to the extent to which the <strong>work</strong>er has independent<br />
discretion in determining the schedule and process of <strong>work</strong>. Finally, feedback is tbe<br />
extent to which the <strong>work</strong>er is provided with inform<strong>at</strong>ion concerning bow well be/she Is<br />
performing. Tbese live core job characteristics are proposed to impact on critical<br />
psychological st<strong>at</strong>es tb<strong>at</strong> then influence affective and motiv<strong>at</strong>ional outcomes.<br />
Although the majority of research has supported tbe validity of tbe JCM (Fried &<br />
Ferris. 1987; Taber SL Taylor, 1990), most studies have not tested tbe full, tbreestage<br />
model (Behson, Eddy, & Lorenzet, 2000). Specifically, tbe medi<strong>at</strong>ing effects of tbe<br />
critical psycbological st<strong>at</strong>es have been neglected, witb the preponderance of research<br />
assessing the direct impact of the core job characteristics on a varietj' of outcomes
Row <strong>at</strong> <strong>work</strong> 599<br />
(Behson et al., 2000; Johns, Xie, & Fang, 1992). Recently there has been some metaanaljtical<br />
evidence to suggest th<strong>at</strong> the critical psychological st<strong>at</strong>es in the J("M do explain<br />
some additional variance in the outcome measures (Behson et al.. 2000). One consistent<br />
conclusion regarding research on the JCM is th<strong>at</strong> more research <strong>at</strong>tention needs to take<br />
into account p,sychological .st<strong>at</strong>es in the development of an understanding ofthe impact<br />
of <strong>work</strong> on employees Oohns et al., 1992; Renn & Vandenberg, 1995).<br />
The aim of the present research was not to valid<strong>at</strong>e the JCM but to use it as a<br />
nomological fr<strong>at</strong>ne<strong>work</strong> to establish the validit)' of flow as a critical psychological st<strong>at</strong>e<br />
th<strong>at</strong> medi<strong>at</strong>es the rel<strong>at</strong>ionsliip between certaiti core job characteristics <strong>at</strong>id well-being<br />
among individuals who are engaged in <strong>work</strong> activities. Recently, Demerouti (2006)<br />
found th<strong>at</strong> motiv<strong>at</strong>ing job characteristics are indeed positively rel<strong>at</strong>ed to How <strong>at</strong> <strong>work</strong>.<br />
However, Demerouti did not ascertain the differential effects of each ofthe five core<br />
job dimensions on flow. Following Fried and Ferris' (1987) guidelines she calcul<strong>at</strong>ed a<br />
single, unweighted additive index th<strong>at</strong> assessed the motiv<strong>at</strong>ing potential score of the<br />
job. We believe th<strong>at</strong> curtain core job characteristics have stronger conceptual links to<br />
flow and th<strong>at</strong> others will be weakly rel<strong>at</strong>ed or unrel<strong>at</strong>ed to optimal <strong>experience</strong>.<br />
For example, the utiliz<strong>at</strong>ion of a variety of skills to meet the challenges of <strong>work</strong> has<br />
consistently been found to be associ<strong>at</strong>ed with job s<strong>at</strong>isfaction and motiv<strong>at</strong>ion (Ciavin &<br />
Axelrod, 1977; O Brien. 1983). Csikszentmihalyi's flow theory also holds th<strong>at</strong> an optimal<br />
balance between challenges and skills inherent in a task is crucial for the intrinsically<br />
motiv<strong>at</strong>ing and s<strong>at</strong>isfying subjective st<strong>at</strong>e th<strong>at</strong> he termed flow (Csikszentmihalyi, 1990).<br />
Those tasks th<strong>at</strong> require the utiliz<strong>at</strong>ion of many skills antl abilities will be the most<br />
intrinsically rewarding and conducive of flow (('sikszentmihaiyi & R<strong>at</strong>himde, 1992).<br />
C;onseqiiently, one would expect a strong associ<strong>at</strong>ion between the core job characteristic<br />
of skill variety and the <strong>experience</strong> of flow.<br />
Not only is it important th<strong>at</strong> the task be challenging to facilit<strong>at</strong>e flow, but there must<br />
also be clear goals so th<strong>at</strong> the performer can remain fully engaged in the task. Feedback<br />
becomes the mechanism th<strong>at</strong> enables individuals to clarifj- their progress towards<br />
these goals (Jackson & Bcklund, 2004). Csikszentmihalyi (1997) st<strong>at</strong>es th<strong>at</strong> immedi<strong>at</strong>e<br />
feedback is a necessary component to <strong>experience</strong> flow. Feedback is also an important<br />
core dimension outlined by the job characteristics model. However, unlike the flow<br />
<strong>experience</strong> where feedback is intrinsic to the task itself. Hackman and Oldham (1975)<br />
emphasize feedback from extrinsic sources such as supervisors, co-<strong>work</strong>ers and clients.<br />
Nonetheless, one would expect the two forms of feedback to be rel<strong>at</strong>ed and th<strong>at</strong><br />
feedback from outside sources should be rel<strong>at</strong>ed to the overall <strong>experience</strong> of flow.<br />
Autonomy and a sense of control also have been argued to be central elements of<br />
both flow (Csikszentmihalyi, 1997) and the core job dimensions (Hackman & Oldham,<br />
1975). There is also some empirical evidence indic<strong>at</strong>ing th<strong>at</strong> autonomy is an antecedent<br />
of flow <strong>experience</strong>s (Bakker. 2005, 2008). More recently a tneta-analysis of locus of<br />
control <strong>at</strong> <strong>work</strong> (Ng, Sorensen. & Eby. 2006) found internal locus of control to be<br />
significantly associ<strong>at</strong>ed with positive task characteristics and motiv<strong>at</strong>ion. Consequently<br />
on would expect high levels of autonomy to be positively correl<strong>at</strong>ed witb flow.<br />
However, task significance and task identity refer to broader perspectives of <strong>work</strong><br />
concerning whether a job has a substantial impact on the lives of other people and<br />
the extent to which the job entails an entire product. Flow has been associ<strong>at</strong>ed with<br />
individual engagement in specific, isol<strong>at</strong>ed, activities th<strong>at</strong> do not necessarily have an<br />
impact on any one else beyond the individual performing the task or consist of a<br />
meaningful whole. Consequently one would not expect these job characteristics to be<br />
associ<strong>at</strong>ed with optimal <strong>experience</strong>.
600 Clive j. Fullagar and E. <strong>Kevin</strong> <strong>Kelloway</strong><br />
Hypothesis 2: The <strong>experience</strong> of flow will be positively rel<strong>at</strong>ed to (a) skill variety, (b) feedback,<br />
and (c) autonomy.<br />
F/ow and well-being<br />
One of the repercussions of positive psychology' has been an upswing in interest in<br />
well-being. A distinction has been made between hedonic and eudaimonic well-heing<br />
(Kyan & Deci, 2001; W<strong>at</strong>erman, 1993). Hedonic <strong>approach</strong>es are pleasure-based and<br />
define well-being in terms of tbe presence of pleasure and tbe absence of pain. Well-being<br />
frequently is assessed by measuring momentary positive aJfecl or mood (e.g. Diener, Sub,<br />
Lucas, & Smitb, 1999; Kabneman, Diener. & Schwarz, 1999). Eudaimonic <strong>approach</strong>es,<br />
on tbe other hand, emphasize optimal functioning and personal expressiveness<br />
(W<strong>at</strong>erman, 1993). For example, Ryff has defined psychological well-being from a<br />
eudaimonic perspective and as consisting of six dimensions: autonomy; personal<br />
growth; seli-acceptance; life purpose; mastery; and positive rel<strong>at</strong>edness (Ryff. 1989,<br />
1995; Ryff & Keyes, 1995). Flow could also be regarded as a momentary form of<br />
eudaimonic well-being. It consists of many of the same aspects of psychological wellbeing;<br />
a sense of control and mastery over the task, a clarity of purpose, a belief<br />
th<strong>at</strong> one possesses the skills to perform well, engagement in and rel<strong>at</strong>edness to the task,<br />
and a sense of enjoyment (Csikszentmihalyi & Csikszentmihalyi, 1988). There is evidence<br />
to suggest th<strong>at</strong> well-being is best conceptualized as a multidimensional construct th<strong>at</strong><br />
includes botb hedonic and eudaimonic aspects (King & Napa, 1998; Mcíiregor & Little,<br />
1998; Ryan & Deci, 2001). However, emotional positivity is regarded as being<br />
independent of eudaimonic well-being. Indeed, positive affect has been suggested to be<br />
an outcome of eudaimonic processes. Tbere is some longitudinal research tb<strong>at</strong> suggests<br />
th<strong>at</strong> positive affectivity is an outcome of both eudaimonic functioning (Reis, Sbeldon,<br />
Gable, Roscoe, & Ryan, 2000; Ryff & Singer, 1998; Sheldon, Ryan, & Reis, 1996) and flow<br />
(Han, 1988; Hull, 1991; Massimini & Carli, 1988) further strengthening the convergent<br />
validity argument tb<strong>at</strong> eudaimonic well-being and flow are conceptually similar.<br />
Hedonic well-being has been primarily oper<strong>at</strong>ionalized as subjective well-being<br />
(SWB) or bappiness (Diener & Lucas, 1999; Kabneman. 1999; Ryan & Deci. 2001).<br />
<strong>An</strong> important component of SWB is the presence of positive mood. The otber<br />
component is life s<strong>at</strong>isfaction. In the current study we focused on assessing momentary<br />
positive affectivity r<strong>at</strong>her tban life s<strong>at</strong>isfaction as the l<strong>at</strong>ter bas been found to be<br />
rel<strong>at</strong>ively stable across time and to only slightly fluctu<strong>at</strong>e around an individual set-point<br />
(Fujita & Diener, 2005). Flow <strong>experience</strong>s have been ibund to be associ<strong>at</strong>ed with such<br />
positive moods and emotions as excitement, joy, ecstasy, happiness, and pride (Bloch,<br />
2002) and SWB (Br>ce & Hawortb, 2002; Clarke & Haworth, 1994; Fave & Massimini,<br />
1988; Seongyeul, 1988). LeFevre (1988) found tb<strong>at</strong> the more time people spent in (low<br />
the more positive affect tbey <strong>experience</strong>d. Tbe little empirical researcb tb<strong>at</strong> has been<br />
undertaken on flow in a <strong>work</strong> context suggests tb<strong>at</strong> people who enjoy tbeir jobs (jften<br />
report <strong>experience</strong>s of flow while <strong>at</strong> <strong>work</strong> (Hryce & Haworth, 2002; c;sikszentmihalyi,<br />
1990; Csikszentmihalyi & LeFevre, 1989; Haworth & MiU. 1992). More recently flow<br />
(speciflcally bigh challenge skill balance) bas been ibund to be associ<strong>at</strong>ed witb gre<strong>at</strong>er<br />
positive mood and bigher performance, particularly among bigb need for achievers<br />
(Eisenberger et al., 2005).<br />
However, much (îf the research tb<strong>at</strong> has investig<strong>at</strong>ed tbe rel<strong>at</strong>ionsbip between<br />
flow and mood is confounded by a methodology th<strong>at</strong> makes for a tautt)logous r<strong>at</strong>ionale<br />
for the rel<strong>at</strong>ionsbip between flow and well-being. Csikszentmibalyi ( 1975) initially asked
Flow <strong>at</strong> <strong>work</strong> 601<br />
subjects to describe highly enjoyable <strong>experience</strong>s or activities in order to define<br />
the characteristics of flow. Also there has been some research to suggest th<strong>at</strong> the<br />
rel<strong>at</strong>ionship between flow and well-being may not be as strong as originally proposed.<br />
Clarke and Haworth (1994) found th<strong>at</strong> college students perceived <strong>work</strong> where skills<br />
exceeded challenges to be more enjoyable than flow activities where skills m<strong>at</strong>ched<br />
challenges. Csikszentmihalyi (1997) has commented on the paradoxical n<strong>at</strong>ure of flow<br />
in th<strong>at</strong> although individuals report positive <strong>experience</strong>s with flow <strong>at</strong> worit, they may not<br />
necessarily feel happy <strong>at</strong> <strong>work</strong>. The current research surveyed a wide variety of <strong>work</strong><br />
activities (not just enjoyable ones) in order to determine the rel<strong>at</strong>ionship between flow<br />
and hedonic well-being.<br />
Finally, although flow has been associ<strong>at</strong>ed with positive mood, the causal direction<br />
of this rel<strong>at</strong>ionship has yet to be established, AH of the research on flow <strong>at</strong> <strong>work</strong><br />
and míMid has been cross-sectional in n<strong>at</strong>ure. In contrast, in the current study we adopt<br />
a longitudinal design in order to sift out the lagged effects of momentary flow on<br />
momentarj- mood and vice versa. Consistent with our view of flow as a critical<br />
psychological st<strong>at</strong>e, we, we hypothesize th<strong>at</strong>;<br />
Hypothesis 3: The <strong>experience</strong> of flowwill be positively rel<strong>at</strong>ed to positive mood.<br />
[ • .<br />
The current study<br />
Thus, the current study is an examin<strong>at</strong>ion of the rel<strong>at</strong>ionships between flow, task<br />
characteristics and hedonic well-being. We propose th<strong>at</strong> flow is primarily a st<strong>at</strong>e-like<br />
variable (111 ), th<strong>at</strong> flow is rel<strong>at</strong>ed to task characteristics (H2), and th<strong>at</strong> the <strong>experience</strong> of<br />
flow is associ<strong>at</strong>ed with hedonic well-being or mood (H3). To examine these hypotheses<br />
we use an <strong>experience</strong> <strong>sampling</strong> methodology (ESM) in a sample of architecture students<br />
engaged in studio <strong>work</strong>.<br />
Architectural students were chosen as the focus for our study for two reasons. First,<br />
architectural students spend long hours <strong>work</strong>ing on cre<strong>at</strong>ive projects in studio th<strong>at</strong><br />
would tend to indic<strong>at</strong>e th<strong>at</strong> they are engaged in tasks th<strong>at</strong> have a high motiv<strong>at</strong>ing<br />
potential and th<strong>at</strong> are conducive to flow. Second. 5th year students engage in studio<br />
projects th<strong>at</strong> are represent<strong>at</strong>ive of the <strong>work</strong> th<strong>at</strong> they would do as professional architects,<br />
so the ecological validity of the <strong>work</strong> tasks is high. In our study, all participants<br />
were enrolled in a Sth-year studio in architectural design.<br />
Studio sessions focus on the synthesis of basic social, functional, teclinical, and<br />
aesthetic factors in design and students are required to <strong>work</strong> independently on projects<br />
th<strong>at</strong> visually represent design ideas. Being the final studio in the architectural<br />
programme, students engage in <strong>work</strong> higlily similar to the activities of their profession.<br />
Architectural studio consists of two components; 3 weekly, 4 hour formal sessions,<br />
where students collectively meet with their professor to discuss their projects; and more<br />
frequent and inf(ïrmal independent sessions where students <strong>work</strong> on their own on their<br />
projects. These l<strong>at</strong>ter sessions usually take place <strong>at</strong> night and continue into the early<br />
hours of the morning. Our d<strong>at</strong>a collection focused on these sessions in which<br />
individuals <strong>work</strong>ed freely on their own architecture projects.<br />
Methodology<br />
Experience samp/ing methodology<br />
The current study utilized an HSM. ESM Is a method of d<strong>at</strong>a collection where participants<br />
are assessed <strong>at</strong> repe<strong>at</strong>ed moments over the course of time while functioning in their
602 Clive }. Fuilagar and £ Ksvin Kelioway<br />
n<strong>at</strong>ural settings. D<strong>at</strong>a were collected by handheld personal digital assistants (PDAs).<br />
Participants were required to carry the PDAs for a specific <strong>sampling</strong> period and to<br />
record their behaviours, <strong>at</strong>titudes, and <strong>experience</strong>s. The PDAs were configured using a<br />
eustomized software called <strong>experience</strong> <strong>sampling</strong> program (ESP; Barrett & Feldman<br />
Barrett, 2000).<br />
Participants and procedure<br />
Participants in the study consisted of 40, 5th-year architectural students from the school<br />
of architecture <strong>at</strong> a large American university in the Midwest of the United St<strong>at</strong>es. At the<br />
beginning of the semester tour studio classes were solicited ibr 40 volunteers.<br />
Particip<strong>at</strong>ion in the study was completely voluntary and individuals who fully particip<strong>at</strong>ed<br />
received an honorarium ($50) in return for their particip<strong>at</strong>ion. The research<br />
protocol was reviewed and approved by the university Institutional Review Board.<br />
The d<strong>at</strong>a were collected across 15 weeks of the Fall semester of 2004. During an initial<br />
session participants were asked to provide the researcher with three days and time<br />
periods during the coming week when they were most likely to be <strong>work</strong>ing<br />
independently on their projects in studio.' Each student was then given a PDA th<strong>at</strong> was<br />
pre-programmed to signal participants <strong>at</strong> two random moments during the three<br />
stipul<strong>at</strong>ed weekly studio periods. Students were instructed to complete the <strong>work</strong> task<br />
th<strong>at</strong> they were involved in before completing a short survey on the PDA. Once tbe initial<br />
signal had beeped, the alarm repe<strong>at</strong>ed itself every 5 minutes for 15 minutes until the<br />
survey bad been completed. The average time between the initial signai and the<br />
participant s response was 4.7 minutes. Bu-ticipants were informed th<strong>at</strong> they should<br />
only respond when beeped during a <strong>work</strong>-rel<strong>at</strong>ed activity and th<strong>at</strong> they should respond<br />
only once during any single studio session. This process was repe<strong>at</strong>ed t)n a weekly basis<br />
until 25 trials had been completed (giving 1,000 d<strong>at</strong>a points in total). Tbe average time<br />
between measurement periods was 2.3 days.<br />
Experienced-sampied measures<br />
Eacb PDA was programmed with an <strong>experience</strong> <strong>sampling</strong> form (ESF) th<strong>at</strong> was designed<br />
to elicit a wide range of inform<strong>at</strong>ion on tbe participant s activities, mood, perceptions of<br />
optimal performanee, and psychological st<strong>at</strong>e of flow.For the researcb process to be the<br />
least disruptive, it was important th<strong>at</strong> each trial should not take longer than two<br />
minutes. Each PDA was programmed to beep participants <strong>at</strong> two random times during<br />
every block of diarized studio time, with the restriction th<strong>at</strong> no two signals would be less<br />
tban 60 minutes apart.<br />
Row was measured using nine items from tbe FSS-2 (Jackson & Ekiund, 2004;<br />
Jackson & Marsh, 1996). Although this scale was designed to assess flow <strong>experience</strong>s<br />
within a particular physical activity, it was deemed appropri<strong>at</strong>e for tbe current research.<br />
Like sports activities (the tiriginal focus for measurement), studio <strong>work</strong> occurred during<br />
a clearly defined time period witb clearly defined activities comprising the 'event'. The<br />
scale was theoretically grounded in Csikszentmihalyi's (1990) nine-dimensional<br />
oper<strong>at</strong>ionaliz<strong>at</strong>ion of flow. It would have been too intnisive and disruptive to include<br />
all 36 items of FSS-2 scale for each of the 25 trials. Moreover, althougli designed to<br />
' These times did not include the regular formal meeting times of studio which occurred on Mondays, Wednesdays, and Fridays<br />
from 1:00 p.m. to 5:00 p.m.
flow <strong>at</strong> <strong>work</strong> 603<br />
measure nine dimensions of flow,it is common to take an 'overall' flowscore as the sum<br />
of the dimensional scores for use in fiirther analysis (see e.g. Jackson & Marsh, 1996).<br />
Consequently, nine items were chosen on tbe basis of their loadings on each of the nine<br />
dimensions derived in the original construct valid<strong>at</strong>ion study (Jackson & Eklund, 2004).<br />
The flow questions were preceded by tbe momentary st<strong>at</strong>ement, 'Please answer tbe<br />
following questions in rel<strong>at</strong>ion to your <strong>experience</strong> in the <strong>work</strong> activity' th<strong>at</strong> you were<br />
engaged in when you were beeped'. R<strong>at</strong>ings were obtained by tapping with the<br />
stylus one of the following options; strongly disagree, disagree, neither agree nor<br />
disagree, agree, strongly agree which were scored 1-5 accordingly. Each item had to be<br />
responded to lor tbe trial to progress. The order in wbich the flowitems appeared in ihe<br />
survey was randomized across situ<strong>at</strong>ions in order to reduce response set. A total flow<br />
score was computed tb<strong>at</strong> consisted of the sum of tbe scores on the nine flow items. The<br />
internal consistency coefficient of the flow scale (computed on witliin-individual mean<br />
item r<strong>at</strong>ings) was s<strong>at</strong>isfactory (a = .83).<br />
job characteristics model<br />
Fifteen items from tbe job diagtiostic survey were used to assess participants' perceptions<br />
of tbe task characteristics associ<strong>at</strong>ed witb studio <strong>work</strong> (Hackman & Oldbam, 1975).<br />
Tlie job diagnostic survey is thet)reticaliy grounded in tbe job characteristics model<br />
and assesses each of the five core task dimensions (skill variety, task identity, task<br />
significance, autonomy, and feedback from the job) witb six items. Items were reworded<br />
to be appropri<strong>at</strong>e for studio <strong>work</strong> (e.g. How much were you left on your own to do your<br />
own <strong>work</strong>?'; "How repetitious was the <strong>work</strong> in .studio?) The measure was administered <strong>at</strong><br />
the end of the semester. Internal consistency was s<strong>at</strong>isfactory for all fivecore dimensions<br />
(skill variety. .71; task identity, .85; task significance, .60; autonomy, .83; and feedback<br />
from tlie job, .81).<br />
Hedonic well-being<br />
Mood was measured witb 10 positive emotions (alert, bappy, cheerfiil, strong, active,<br />
sociable, involved, excited, clear, and relaxed) and 10 neg<strong>at</strong>ive emotions (drowsy, sad,<br />
irritable, weak, passive, lonely, detaehed, bored, confused, and tense) tb<strong>at</strong> were<br />
arranged as bipolar items (e.g. alert-drowsy, bappy-sad). These emotions were selected<br />
as they have been used to conceptually represent the major forms of pleasant and<br />
unpleasant emotion (Csikszentmihalyi & Csikszentmihalyi, 1988; Diener, Smith, &<br />
Fujita. 1995). A shorter list was selected as more appropri<strong>at</strong>e fbr <strong>experience</strong> samplinf;<br />
(Scollon, Diener. Oishi, &. Biswas-Diener. 2005). Tbe mood items were form<strong>at</strong>ted as<br />
'slider' questions, where respondents bad to move a slider towards the bipolar adjective<br />
th<strong>at</strong> best described their mood <strong>at</strong> the time th<strong>at</strong> tbey were beeped. Tbe slider questions<br />
were scored on a scale of 1-10, with higher scores being indic<strong>at</strong>ive of more positive<br />
mood. Again the order of tbe mood items was randomly varied for eacb trial. The<br />
internal reliability of the mood scale was found to be s<strong>at</strong>isfactory (Cronbach'sa — .96).<br />
I<br />
<strong>An</strong>alyses<br />
To investig<strong>at</strong>e tbe effect of job characteristics on flow and fiow on hedonic well-being,<br />
HLM was used (Bryk & Raudenbush, 1992). The HLM modelling <strong>approach</strong> is an<br />
extension of multiple regression th<strong>at</strong> allows investig<strong>at</strong>ion of tbe rel<strong>at</strong>ionships between<br />
variables manifested <strong>at</strong> two levels of analysis. In longitudinal analyses there are a series
604 Clive ]. Fulhgar and £ <strong>Kevin</strong> Kelhway<br />
Table I. Means (M), standard devi<strong>at</strong>ions (SD), and intercorrel<strong>at</strong>ions across individuals for ali study<br />
variables (N = 40)<br />
M<br />
SD<br />
2<br />
3<br />
4<br />
5<br />
6<br />
7<br />
1. Flow<br />
2. Mood<br />
3. Si
Flow <strong>at</strong> <strong>work</strong> 605<br />
Table 2. Parameter estim<strong>at</strong>es and variance components of null models tested {N = 999)<br />
Model equ<strong>at</strong>ions -yoo "¥io ""^ Too<br />
Null model 1^<br />
Row^ = ßoi + U 32.30 - 22.81 8.19**<br />
Mood,j = ßoj + r^¡ 5.90 - 2.22 1.45**<br />
ßoj ~ "Yoo "t" "Oi<br />
"ßoj is the average level of flow/mood for individual j; -yoo is the grand mean of flow/mood scores;<br />
a^ — var{r,j) the within-individual variance in flov^i/mood (computed as the average squared distance<br />
from individual momentary scores and the individual's mean score); and TOO — v^riUoj) the betv/eenindividual<br />
variance in flow/mood (the variance of the 40 average flov//mood scores).<br />
*p < .05; **p < .01.<br />
between-individual variance accounts for 26% of the total variance in flow. This would<br />
suggest th<strong>at</strong> 74% of the variance is <strong>at</strong>tributable to witbin-individual vari<strong>at</strong>ion. For mt>od,<br />
40% of the total variance was accounted for by between-individual variance, indic<strong>at</strong>ing<br />
th<strong>at</strong> 60% of the variance was due to witbin-individual vari<strong>at</strong>ion in momentary' mood.<br />
Tbese results suggest th<strong>at</strong> How varies considerably from situ<strong>at</strong>ion to situ<strong>at</strong>ion :md tb<strong>at</strong> it<br />
behaves predominantly like a st<strong>at</strong>e construct.<br />
We also tested whether individuals' flow and mood nitings were randomly<br />
distributed or serially dependent (Hies & Judge, 2002). Comparing the level I variance<br />
components of the lagged regression model with th<strong>at</strong> of tbe null models indic<strong>at</strong>ed th<strong>at</strong><br />
tbe lagged measures explained fbur and less tban one percent of the within-individual<br />
variance for How and mood respectively.^ Lagged flow was found to be a significant<br />
predictor of momentarj- flow (Flow: -YK, =^ 0.20, p < .05), but mood was serially<br />
independent of its lagged effect (Mood: ^n, = 0.08, p > .05). All subsequent models<br />
where flow was the outcome variable were controlled for tbe lagged effect.**<br />
Hypothesis 2: Job characteristics and flow<br />
Hypotbesis 2 suggested th<strong>at</strong> the individual variance in flow would be associ<strong>at</strong>ed witb<br />
tbe job characteristics outlined by Hackman and Oldbam's model (1975). Based on<br />
tbeoretical arguments, the ñve job characteristics were bierarchically entered into the<br />
model as ibilows: skill variety; feedback; autonomy; task identity; and task significance.<br />
In order to test tbis hypothesis, two models were calcul<strong>at</strong>ed: a random-coefficient<br />
regression model and an intercepts-as-outcomes model (see Table 3). Tests were<br />
undertaken to ascertain the proportion of variimce in tbe intercept terms tb<strong>at</strong> was<br />
explained by the five job characteristics <strong>at</strong> level 2 (see Table 3). Tbe t tests for the<br />
fixed effects in tbe intercepts-as-outcomes model assessed the significance of<br />
the rel<strong>at</strong>ionships between tbe five job characteristics and momentary' flow, wbile<br />
contniJling for the lagged effects of flow. Tbese tests indic<strong>at</strong>ed th<strong>at</strong> botb skill variety<br />
(7i, = 0.05, t(M) = 2.64, p < .05) and autonomy (7,, ^ 0.03, i(34) = 2.12, p < .05)<br />
were significant predictors of between-group variimce in flow after controlling for<br />
'Percent = in^(nullmodel) - (T^OaggedeffectsmodelWtr'inulimodel).<br />
•* Following Hies and judge (2002) these results arc not presented in the main models as they ore not of substantive<br />
inleresi fiere.
606 Clive }. Fullagar and £ <strong>Kevin</strong> <strong>Kelloway</strong><br />
.0 rt<br />
Si<br />
o<br />
C V)<br />
a<br />
01<br />
I 1<br />
I<br />
I<br />
L.<br />
00<br />
gi<br />
><br />
—<br />
m<br />
a.<br />
0<br />
Ti<br />
II<br />
ci<br />
M<br />
C<br />
o<br />
^<br />
1.<br />
a.<br />
01<br />
• - ~ ^<br />
CQ.<br />
ç<br />
y<br />
V<br />
a.<br />
"S T) 0Ë<br />
L.<br />
"S<br />
•n<br />
resit<br />
(SI<br />
"33 ><br />
c<br />
'^0)<br />
u<br />
o c<br />
01<br />
T3<br />
o<br />
E<br />
I<br />
I<br />
"O 0) _ ::?<br />
S £ - ^<br />
"= — 0) t<br />
t<br />
I<br />
cf<br />
^ a.<br />
w<br />
o<br />
S O<br />
c E<br />
5 >-<br />
0) 0)<br />
•y<br />
••<br />
V<br />
> re "= y<br />
•s 0 ^ ^<br />
g<br />
V<br />
0. .- - ^<br />
C Q]<br />
i l A
Row <strong>at</strong> <strong>work</strong> 607<br />
serial dependence effects. Neither feedback, task significance nor task identity were<br />
significant predictors.<br />
To determine tbe magnitude of the rel<strong>at</strong>ionship between the two significant job<br />
characteristics and flow, we calcul<strong>at</strong>ed separ<strong>at</strong>e intercepts-as-outcomes models. Using<br />
the residual variance in tbe intercepts, we found th<strong>at</strong> skill variety' accounted for 1 S% of<br />
tbe between-individual variance in flow (TQÜ ^ 23.91, X"C39) ^ 58.67, p < .05), and<br />
autonomy explained 8% of the variance (TOO = 25.72, x^(39) = 59.45, p < .05).^<br />
Hypothesis 3: Flow and mood<br />
It was predicted th<strong>at</strong> momentary flow would be rel<strong>at</strong>ed to momentary positive mood<br />
both across and witbin individuals. The results of the hierarchical linear model th<strong>at</strong><br />
was tested are presented in Table 4. Tlie pooled slope parameter was significantly<br />
different from zero (7io ^ 0.19, i(39) ^ 16.04, p < .01). The direction of the pooled<br />
slope of the regression of flow on to mood (710) indic<strong>at</strong>es th<strong>at</strong> higher levels of flow<br />
are associ<strong>at</strong>ed with more positive mood. The magnitude of this rel<strong>at</strong>ionship was<br />
calcul<strong>at</strong>ed by comparing the within-individual variance in mood from tbe first null<br />
model with the within-individual variance after controlling for flow.^' Flow accounted<br />
for i6% of tbe within-individual variance in mood. Similarly, comparing the level 2<br />
variance components of random regression model and the null model indic<strong>at</strong>ed th<strong>at</strong><br />
flow explained 30% of the between-individual variance in mood. "'^ In sum, results<br />
tended to support tbe bypothesis th<strong>at</strong> momentary flow was assoei<strong>at</strong>ed with<br />
momentary positive mood.<br />
We also tested the causal direction of the rel<strong>at</strong>ionship between flow and mood using<br />
lagged measures of mood and flow,respectively. Results suggested tb<strong>at</strong> lagged fitiw was<br />
predictive of mood (7,0 = 0.11, i(39) = 2.28, p < .05), and th<strong>at</strong> lagged mood was not<br />
associ<strong>at</strong>ed with l<strong>at</strong>er now (7,0 = 0.10, i(39) = 0.74, p > .05).<br />
Discussion<br />
The current study tried to understand tbe n<strong>at</strong>ure of flow <strong>at</strong> <strong>work</strong> by longitudinally<br />
tracking the flow st<strong>at</strong>e in a sample of architecture students <strong>work</strong>ing on studio projects.<br />
Specifically, we want to sift out the st<strong>at</strong>e and trait components of flow, to ascertain<br />
wbich <strong>work</strong>-rel<strong>at</strong>ed characteristics were associ<strong>at</strong>ed with flow, and to determine the<br />
rel<strong>at</strong>ionship between momentary flow and positive mood. Our results suggested th<strong>at</strong><br />
botb within-individual and between-individual variance comprised flow with tbe<br />
former accounting for 74% of the total variance. Thus, flow bas both st<strong>at</strong>e and trait<br />
components with the former predomin<strong>at</strong>ing. Second, our d<strong>at</strong>a suggest th<strong>at</strong> skill variety<br />
and autonomy are the two job characteristics th<strong>at</strong> make unique independent<br />
contributions to the predictions of flow. The remaining three job characterizitics<br />
*R^ = (Toolrandomregression) - Too(Jntercepts-os-outcomes))/T(jo(ran
608 Clive J. Fullagar and E. <strong>Kevin</strong> Keltoway<br />
Table 4. Parameter estim<strong>at</strong>es and variance components testing rel<strong>at</strong>ionship between flow and mood<br />
{N = 999)<br />
Model equ<strong>at</strong>ions -yoo Tio f^^ Too TK<br />
Ran do m-coefficient regression model t'<br />
LI:Mood„ = ßo, + ß,^(Flow,^) + r,j<br />
L2: ßoj = 7oo + ^o,<br />
L2:ß|^ = ^,o + U„<br />
-0.27 0.19* 1.41 2.95 0.00<br />
"•yoo is the mean of the intercepts across groups: 7io is the mean of the slopes across groups; (T^{ri,) is<br />
the level I residual variance; TOO(UO,) is the variance in intercepts; and 7| i (U|,) is the variance in slopes.<br />
*p< .01.<br />
(feedback, task significance, and task identity) did not contribute to the prediction.<br />
Finally, our results showed th<strong>at</strong> the <strong>experience</strong> of flow preceded changes in mood.<br />
Howard Ciardner (2002) has pointed out th<strong>at</strong> one of the reasons th<strong>at</strong> psychologists<br />
have neglected to study the actual <strong>experience</strong> of <strong>work</strong> has been the lack of an<br />
appropri<strong>at</strong>e methodology to scientifically study internal feelings. In this context, it is<br />
appropri<strong>at</strong>e to note th<strong>at</strong> our results emerge from the use of a rigorous methodolog)' th<strong>at</strong><br />
allows the testing of longitudinal rel<strong>at</strong>ionships and addresses the known problems of<br />
cross-sectional survey d<strong>at</strong>a (e.g. Scolion, Kim-Prieto, & Diener. 2003; Smyth & Stone,<br />
2003). We believe th<strong>at</strong> the use of the ESM methodology substantially advances the study<br />
of How and offers considerable advantage in other areas of organiz<strong>at</strong>ional research. ESM<br />
has been used to assess flow <strong>experience</strong>s in daily life among elementar)' and secondary<br />
school sttidents (Csikszentmihalyi & Hunter, 2003), students (Carli. Delle Fave, &<br />
Massimini, 1988; Csikszentmihalyi & Urson, 1984), family members (Larson & Richards,<br />
1994), psychi<strong>at</strong>ric p<strong>at</strong>ients (Massimitii, Csikszentmihalyi, & Carli, 1987), and nursing<br />
home residents (Voelkl, 1990). The product of such a methodology- is a den.se and<br />
system<strong>at</strong>ic description of the external conditions and subjective <strong>experience</strong> of flow in<br />
different situ<strong>at</strong>ional contexts and <strong>at</strong> different moments in time. Despite the iaet th<strong>at</strong> ESM<br />
allows for the study of the interactions between person and situ<strong>at</strong>ional variables without<br />
tbe limit<strong>at</strong>ions of traditional research designs (Hormuth, 1986; Hies & Judge, 2002),<br />
rel<strong>at</strong>ively little use has been made of ESM in organiz<strong>at</strong>ional psycbological research.<br />
Tlie current study had three main findings.First, we bave argued th<strong>at</strong> it is important<br />
to distinguish between the st<strong>at</strong>e and trait components of flow. Tbe results of the present<br />
research suggest th<strong>at</strong> flow, or optimal <strong>experience</strong>, is predominantly a situ<strong>at</strong>ional st<strong>at</strong>e<br />
of mind r<strong>at</strong>her tban a trait or disposition. We found tb<strong>at</strong> approxim<strong>at</strong>ely tbree quarters of<br />
tbe variance in flow was <strong>at</strong>tributable to situ<strong>at</strong>ion-io-situ<strong>at</strong>ion fluctu<strong>at</strong>ions. Dispositional<br />
variables or traits tend to be stable across time and place (Zuckerman, 1983).<br />
Consequently such variables are less amenable to manipul<strong>at</strong>ion. St<strong>at</strong>e variables on the<br />
other hand are dynamic and cbange across time and situ<strong>at</strong>ion. St<strong>at</strong>e variables are capable<br />
of being manipul<strong>at</strong>ed and are likely the result of an interaction between personal<br />
dispositions and tbe environment. Previous research on flow (e.g. Demerouti, 2006;<br />
Eisenberger et ai, 2005) has tended to blur the st<strong>at</strong>e-trait distinction. Flow has been<br />
tre<strong>at</strong>ed as a st<strong>at</strong>e variable, in tb<strong>at</strong> situ<strong>at</strong>ional variables have been correl<strong>at</strong>ed with it,<br />
but measured as a tniit construct, in tb<strong>at</strong> participants' flow <strong>experience</strong>s have been<br />
assessed across a broad range of activities. Furthermore, both of these studies were
Fiow <strong>at</strong> <strong>work</strong> 609<br />
cross-sectional in n<strong>at</strong>ure making it impossible to sift out between-individual vari<strong>at</strong>ion<br />
from within-individual variance.<br />
Our study allowed this separ<strong>at</strong>ion :md resulted in the finding th<strong>at</strong> most of the<br />
variance in fiow was within-individual giving rise to the possibility th<strong>at</strong> the <strong>experience</strong><br />
of flowis governed by characteristics ofthe task or situ<strong>at</strong>ion. In the particular context of<br />
the <strong>work</strong>place, our results suggests th<strong>at</strong> tasks can be designed to enhance the likelihood<br />
of experiencing flow.Just wh<strong>at</strong> aspects of task design are important is informed by our<br />
second finding.<br />
Tiie second finding of the current research was th<strong>at</strong> certain <strong>work</strong> or task<br />
characteristics are associ<strong>at</strong>ed with optimal <strong>experience</strong>. Earlier research by Demerouti<br />
(2006) has found th<strong>at</strong> motiv<strong>at</strong>ing job characteristics are correl<strong>at</strong>ed with the <strong>experience</strong><br />
of flow <strong>at</strong> <strong>work</strong>. However, Demerouti did not specifically identify which job characteristics<br />
were associ<strong>at</strong>ed with flow. Our finditigs indic<strong>at</strong>e th<strong>at</strong> flow was associ<strong>at</strong>ed<br />
with studio <strong>work</strong> th<strong>at</strong> was perceived as requiring different skills and th<strong>at</strong> allowed<br />
the student cre<strong>at</strong>ive choice and discretion in their <strong>work</strong>. Th<strong>at</strong> these two job<br />
characteristics should be correl<strong>at</strong>ed with flow may be due to the n<strong>at</strong>ure of the <strong>work</strong><br />
th<strong>at</strong> we studied. Architectural <strong>work</strong> is cognitive in n<strong>at</strong>ure, with students having to use<br />
logic and reasoning to identifv' the strengths and weaknesses of altern<strong>at</strong>ive solutions or<br />
<strong>approach</strong>es to problems. Architecture studio requires students to engage in complex<br />
problem solving where they have to independently identify problems and review<br />
rel<strong>at</strong>ed inform<strong>at</strong>ion to develop and evalu<strong>at</strong>e options and implement unique solutions.<br />
The architectural student is more isol<strong>at</strong>ed, less monitored, and engaged in activities th<strong>at</strong><br />
are more cognitively complex and cre<strong>at</strong>ive in n<strong>at</strong>ure than many other kinds of <strong>work</strong>.<br />
Even though a different st<strong>at</strong>e of mind was investig<strong>at</strong>ed in the current research, the<br />
results lend support to Hackman and Oldham's (1975) contetition th<strong>at</strong> core job<br />
characteristics are important in gener<strong>at</strong>ing critical psychological st<strong>at</strong>es (such as flow).<br />
Thus, organiz<strong>at</strong>ions wishing to enhance employees' <strong>experience</strong> of. or likelihood<br />
of experiencing flow, are well-advised to design ta.sks th<strong>at</strong> incorpor<strong>at</strong>e these elements<br />
of skill variety and autonomy. Our findings suggest th<strong>at</strong> engaging in tasks th<strong>at</strong> require<br />
complex skills, resolving challenging problems and expressing cre<strong>at</strong>ivity leads to the<br />
<strong>experience</strong> of flow.<br />
Finally, the current findings also indic<strong>at</strong>ed a significant and reliable rel<strong>at</strong>ionship<br />
between flow and pt)sitive mood. Students who <strong>experience</strong>d higher levels of flow also<br />
reported being more alert, happy, involved, excited, and so forth. This is consistent with<br />
previous research th<strong>at</strong> has found fiow to be associ<strong>at</strong>ed with hedonic well-being, or<br />
positive mood (Eisenberger fi rt/., 2005; Nakamura & Csikszentmihalyi. 2002). However<br />
the r<strong>at</strong>ionale for this rel<strong>at</strong>ionsliip has been criticized for its circularity. Research on flow<br />
has tended to focus on self-reports of single moments of well-being. The current study<br />
<strong>at</strong>tempted to overcome this methodological shortcoming by randomly sur\'eying a<br />
broad range of <strong>work</strong> activities, not just optimal ones, over several weeks. I.eFevre ( 1988)<br />
found th<strong>at</strong> motiv<strong>at</strong>ion, activ<strong>at</strong>ion, and s<strong>at</strong>isfaction were higher in flow <strong>experience</strong>d <strong>at</strong><br />
<strong>work</strong>, hut th<strong>at</strong> affect was not. In fact mildly neg<strong>at</strong>ive mood was reported by engineers<br />
and managers who had <strong>experience</strong>d flow.Positive affect was more likely to be rel<strong>at</strong>ed to<br />
fiow in leisure activities. The difference between these and the current findings may<br />
again be in the n<strong>at</strong>ure of <strong>work</strong> th<strong>at</strong> was studied, ln architectural studio <strong>work</strong> many ofthe<br />
tasks th<strong>at</strong> are associ<strong>at</strong>ed with high levels of flow are the same as those <strong>experience</strong>d in<br />
leisure activities (e.g. drawing <strong>at</strong>id model making).<br />
The findingth<strong>at</strong> flow is associ<strong>at</strong>ed with positive mood has significant theoretical and<br />
practical implic<strong>at</strong>iotis. Prederickson (1998, 2001) has argued tli<strong>at</strong> positive emotions
610 Clive J. Fullagar and £ <strong>Kevin</strong> Kelhway<br />
broaden individuai resources and tbougbt-action repertories. There is growing evidence<br />
th<strong>at</strong> tbe psychological well-being of <strong>work</strong>ers is predictive of employee performance, job<br />
s<strong>at</strong>isfaction, and turnover (Cropanzano &Wrigbt, 2001; Staw & Barsade, 1993; Wright &<br />
Bonett, 2007; Wright & Staw, 1999). We have argued th<strong>at</strong> flow is a form of eudaimonic<br />
well-being. Our findings suggest tb<strong>at</strong> flow may be a critical psycbological st<strong>at</strong>e th<strong>at</strong> is<br />
associ<strong>at</strong>ed with positive mood, a core component of psycbological well-being.<br />
Furtbermore, we established th<strong>at</strong> the majtirity of variance in flow is determined by<br />
situ<strong>at</strong>ional characteristics r<strong>at</strong>berthan individual dispositional factors. This has important<br />
practical implic<strong>at</strong>ions for the design of <strong>work</strong>. Providing <strong>work</strong> tb<strong>at</strong> is bigh in autonomy<br />
and skill variety is more likeiy to induce flow whicb is likely to have beneficial<br />
consequences for the organiz<strong>at</strong>ion. Flow has been found to be positively associ<strong>at</strong>ed with<br />
optimal ftmctioning. specifically with botb in-role and extra-role performance<br />
(Demerouti, 2006; Eisenherger et al., 2005). Eisenberger et ai. (2005) bave shown<br />
tb<strong>at</strong> tbe rel<strong>at</strong>ionship between flow and employee performance is partially medi<strong>at</strong>ed by<br />
positive mood. Earlier research has indic<strong>at</strong>ed th<strong>at</strong> the rel<strong>at</strong>ionship between positive<br />
mood and performance is particularly pertinent to tasks ih<strong>at</strong> require spontaneity and<br />
cre<strong>at</strong>ivity (Eisenberger & Rhoades, 2001; George & Brief, 1992). Specifically, positive<br />
mood has been .shown to promote cre<strong>at</strong>ive thinking. The longitudinal n<strong>at</strong>ure of our<br />
research contributes to the liter<strong>at</strong>ure tb<strong>at</strong> bas established a rel<strong>at</strong>ionship between<br />
optimal <strong>experience</strong>s <strong>at</strong> <strong>work</strong> and positive mood by suggesting tb<strong>at</strong> this associ<strong>at</strong>ion is<br />
consistently reliable across time and situ<strong>at</strong>ions. Clearly flow is an important positive<br />
psycbological variable tb<strong>at</strong> has been found to be associ<strong>at</strong>ed with flourishingand one tb<strong>at</strong><br />
can be influenced by tbe design of <strong>work</strong>.<br />
Potential limit<strong>at</strong>ions of the research<br />
Although the ESM used in tbe current study offers considerable advantage to tbe study of<br />
flow and other aspects of organiz<strong>at</strong>ional behaviour, our applic<strong>at</strong>ion of the method<br />
suggests several potential limit<strong>at</strong>ions to be addressed by future researcb. First,<br />
participants in tbe research were architecture students engaged in <strong>work</strong>-like but not<br />
<strong>work</strong> tasks suggesting the need to replic<strong>at</strong>e our findings in the context of paid<br />
employment. The <strong>experience</strong> of employment is more tban perfonning a set of tasks<br />
{<strong>Kelloway</strong>, Gallajíbcr, & Barling, 2005) and the extent to whicb our findingsgeneralize to<br />
the context of paid employment remains an empirical question.<br />
Second, it couid be argued th<strong>at</strong> tbere was not much vari<strong>at</strong>ion in the n<strong>at</strong>ure of the<br />
<strong>work</strong> th<strong>at</strong> we investig<strong>at</strong>ed, th<strong>at</strong> is architectural studio <strong>work</strong>. Cxmsequently any withinsubject<br />
variance could be <strong>at</strong>tributable to dispositional vari<strong>at</strong>ions r<strong>at</strong>her than a change in<br />
the <strong>work</strong> situ<strong>at</strong>iijn. However our sample was drawn from three different studio sessions.<br />
Students were responsible for designing unique solutions to architectural problems, Tbe<br />
n<strong>at</strong>ure of the <strong>work</strong> varied considerably and included a variety of tasks such as reading,<br />
drawing, computer programming, model building, writing, computer rendering and<br />
d<strong>at</strong>a analysis. Nonetheless, the variance in task characteristics was rel<strong>at</strong>ively small.<br />
However, this would indic<strong>at</strong>e tb<strong>at</strong> the strength of rel<strong>at</strong>ionships between tbe job<br />
cbaracteristics and flow were underestim<strong>at</strong>ed in our sample.<br />
Third, in this study we focus on a single 'narrow-band:' indic<strong>at</strong>or of well-being<br />
(i.e. mood). In doing so, we parallel earlier studies which have also focused on tbe<br />
rel<strong>at</strong>ion between flow and mood (e.g. Eisenberger et ai, 2005) Broader conceptualiz<strong>at</strong>ions<br />
of well-being would suggest going beyond measures of affect to include<br />
measures sucb as aspir<strong>at</strong>ion or self-efficacy (e.g. Warr, 1987) and future research would
Flow <strong>at</strong> <strong>work</strong> 611<br />
be well advised to consider broadening the domain of inquiry to more fully understand<br />
the rel<strong>at</strong>ionship between flow and well-being.<br />
Finally, we used a short composite measure of flow in our research. This was<br />
primarily dict<strong>at</strong>ed by the ESM th<strong>at</strong> we used, th<strong>at</strong> necessit<strong>at</strong>es the use of short scales th<strong>at</strong><br />
are minimally disruptive of the <strong>experience</strong> being assessed. There seems to be confusion<br />
as to an appropri<strong>at</strong>e psychometric form with which to assess flow. There is a growing<br />
body of theory to suggest th<strong>at</strong> flow is a multidimensional construct (Jackson & Marsh,<br />
1996; Marsh & Jackson. 1999), yet these dimensions are highly correl<strong>at</strong>ed. Furthermore,<br />
composite oper<strong>at</strong>ionaliz<strong>at</strong>ions of flow fail to distinguish between the preconditions<br />
or antecedents of flow (e.g. challenge-skill balance, clear goals) and the components<br />
of optimal <strong>experience</strong> (e.g. enjoyment of. and concentr<strong>at</strong>ion on the activity).<br />
Future research needs to move away from uni-dimensional conceptualiz<strong>at</strong>ions of flow,<br />
and to distinguish between the antecedents and qualities of the construct.<br />
Conclusion<br />
Recently there has been an interest in the applic<strong>at</strong>ion of positive psychology to<br />
understanding organiz<strong>at</strong>ional behaviour (Cameron et al., 2003; Luthans, 2001, 2002;<br />
Turner et al.. 2002). Flow is a construct th<strong>at</strong> is <strong>at</strong> the centre oí the ptjsitivc psychology<br />
movement (Seligman & Csikszentmihalyi, 2000) and provides researchers with a<br />
valuable conceptual frame<strong>work</strong> with which to study the n<strong>at</strong>ure and process of w
612 Clive J. Fultagar anà E. Ksvin Ketloway<br />
Bakker, A. B. (2005). Flow among music teachers and their students: The crossover of peak<br />
<strong>experience</strong>s. Joitmal of Voc<strong>at</strong>ional Behcwior, 66. 26-44.<br />
Bakkcr, A. II. (2008). The <strong>work</strong> rel<strong>at</strong>ed ilow inventor)': Construction and initial valid<strong>at</strong>ion of the<br />
\i/OLf. Joumtii of Voc<strong>at</strong>ional Behavior, 72, 400-414.<br />
Barrett, D. J., & Feldman Barrett, L. (2000). The <strong>experience</strong>^ampUng program (ESP). Retrieved<br />
from http://www2.bc.edu/ ~ barrelli/esp/<br />
Behson, S. J., Eddy, E. R., & Lorenzet, S. J. (2000). The importance of the critical psychological<br />
st<strong>at</strong>es in the job characteristics model: A meta-analytic and structural equ<strong>at</strong>ions modeling<br />
examin<strong>at</strong>ion. Ciirrenl Research in Social Psychology, 5(12), 170-189.<br />
Bloch, C. (2002). Moods and the quality of liie. Journal of Happiness Studies, 3, 101-128,<br />
Bryce, j., it Haworth, J. (2002). Well-being and flow in sample of male and female office <strong>work</strong>ers.<br />
Leisutv Studies, 21, 249-263.<br />
Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models: AppUc<strong>at</strong>iom and d<strong>at</strong>a<br />
analysis methods. Ihousand Oaks. CA: Sage.<br />
Bryk, A. S., Raudenbush, S. W., & Congdon. R. X, Jr. (2000). HLM S for windows. Chicago, IL:<br />
Scientific Software Intern<strong>at</strong>ional.<br />
Cameron, K, S., Dutton, J. E., & Quinn. R. E. (Bds.), (2003)- Pf>sitire orficniiz<strong>at</strong>ional scholarship:<br />
Found<strong>at</strong>ions of a new discipline. San Fransisco, CA: Berrett-Koeliler Publishers.<br />
Carli, M., Delle Fave, A., & Massimini, F (1988). Tlie quality of <strong>experience</strong> in tbe flow channels:<br />
Comparison of Italian and U.S. students. In M. Csikszcntniibaiyi & 1. S. ([isikszentmihaiyi (Eds.),<br />
Optimalexpciienci': I'sychoiogicalstudiesofßtnc in conscioiis/uvs {pp. 26(i-50(^). New York:<br />
Cambridge University Press.<br />
Clarke. S. G., & Haworth, J. T. (1994). Flow' <strong>experience</strong> in ibe daily lives of sixth-form college<br />
students. British Journal of Psychology. 85, 511-523.<br />
Cropanzano, R..iܣ Wrigbi. T. A. (2(M>1 ), When a happy' <strong>work</strong>er is a productive' woriver: A review<br />
and further refinement of the happy-productive <strong>work</strong>er thesis. Consulting Psychology<br />
Journal: Practice and Research, 53, 182-199.<br />
Csikszentmihalyi, M. (1975). Bey-oml boredom and anxiety: Experiencing flow in <strong>work</strong> and play.<br />
San Fransisco, CA: Josscy-Bass.<br />
Csikszentmihalyi, M. (1990). Flour The psychology of optimal <strong>experience</strong>. New York: Harper<br />
and Row.<br />
Csikszentmihalyi, M. (1993). The evolving self. New York: Harper and Row,<br />
Csikszentmihalyi, M. (1997). Finding ßow: The pr^chology of engagement with everyday life.<br />
New York: Basic BcMiks.<br />
(Csikszentmihalyi, M., & Csikszentmihalyi, I. S. (1988). Introduction to part IV. In<br />
M. Csikszentmihalyi & I. S. Csikszenimihalyi (Eds.), Optimal <strong>experience</strong>: Psychological<br />
studies ofjlow in consciousness (pp. 251-265). New 'Vork: Cambridge Univcrsit)' Press.<br />
Csikszentmihalyi. M., & Hunter, J. (2003). Happiness in everyday life: The uses of <strong>experience</strong><br />
<strong>sampling</strong>. /o//í77í-í/ of Happiness Studies, 4, 185-199.<br />
Csikszentmihaiyi, M., & Lar.son, R. (1984). Being adolescent: Conflict and growth in the teenage<br />
years. New York: Basic B(H)ks.<br />
Csikszentmihalyi, M., & LeFevre, J. (1989). Optimal <strong>experience</strong> In <strong>work</strong> and leisure. yoMrn«/o/<br />
Personality and Social Psychology, 56, 815-822.<br />
Csikszentmihalyi, M., & R<strong>at</strong>hunde, K. (1992). The measurement of flow in everyday life: Toward a<br />
tlieor>' of emergent motiv<strong>at</strong>ion. In J. E. Jacobs (Ed.), Developmental perspectives on<br />
niotir<strong>at</strong>ion: Voiume if) of the Nebraska symposium on motiv<strong>at</strong>ion (pp. 5^-97). Lincoln,<br />
Nebraska: University of Nebra.ska Press.<br />
Csikszentmihalyi, M., R<strong>at</strong>hunde. K., & Whalen, S. (1993). Talented teenagers: A longitudinal<br />
study of their development. New York: c:ambridge University Press.<br />
Demerouti, E. (2006). Job characteristics, flow, and performance: Tbe moder<strong>at</strong>ing role of<br />
conscientiousness. yo/írf¡«/ of Occup<strong>at</strong>ional Health Psychology, 11, 266-280.
Flow <strong>at</strong> <strong>work</strong> 613<br />
Diener, E., & Lucas, R. E. (1999). Personality and subjective well-being. In D. Kahneman, E. Diener,<br />
& N. Schwarz (Eds.), Welt-being: The found<strong>at</strong>ions of hedonic psychoiog)' (pp. 213-229).<br />
New York; Russell Sage.<br />
Diener, E., Smith, H., & Fujita, F. (1995). The personalit>' structure oíañect. Journal of Person<strong>at</strong>it)'<br />
and Sociai Psycholog)', 69, 130-141.<br />
Diener, E., Suh, E., Lucas, R, E., & Smith, H. L. (1999). Subjective well-being; Three deeadcs of<br />
progress. Ps)>chological Bulletin. 125. 276-302.<br />
Eisenberger, R., Jones, J. R..Stinglhamber, F.Shanock, L.,&RandaU. A. T (2()0S). Flow <strong>experience</strong>s<br />
<strong>at</strong> <strong>work</strong>: For high need achievers alone? Journal of Organiz<strong>at</strong>iotial Behavior, 26, 755-775.<br />
Eisenberger. R., & Rhoades, L. (2001). Incremental effects of reward on cre<strong>at</strong>ivity-./owrwrt/ of<br />
Personality and Sociai Psychology. St. 728-741.<br />
Fave, A. D., & Massimini, F (1988). Moderniz<strong>at</strong>ion antl the changing contexts of flow in <strong>work</strong> and<br />
leisure. In M. Csikszentmihahi & I. S. Csikszentmihalyi (Eds.), Optimal <strong>experience</strong>:<br />
Psychologicai studies of flow in consciousness (pp. 193-213). New York: Cambridge<br />
University Press.<br />
Frederickson, B. L. (1998). Wh<strong>at</strong> good are positive cmotionsi* Review of Gênerai Psycholog)\ 2,<br />
300-319. !<br />
Frederickson. B. L. (2001). The role of positive emotions ¡n positive psychology': The Broaden-and-<br />
Build rheor>' i)f positive emotions. American Psychologist. 56, 218-226.<br />
Fried, Y, & Ferris, G. R. (1987). The validity of the job characteristics model: A review and metaanal)'sis.<br />
Personnel Psychotogy, 40, 287-322.<br />
Fujita, E, & Diener, E. (2(M)'>). Ufc s<strong>at</strong>isfaction set point: Stability and change. Jourtial of<br />
Personality and Social P.sychotogy, 88. I 58-164.<br />
Gardner, H. (2002). Good <strong>work</strong>, well done: A psych()l(ïgic;il study. Chronicle ofHigljer Educ<strong>at</strong>ion.<br />
February 22, 7.<br />
Gavin. J. F, & Axelrod, W. L. (1977). Managerial stress and strain in a mining oi^aniz<strong>at</strong>ion.<br />
Journal of Voc<strong>at</strong>ion<strong>at</strong> Behavior. II. 66-74.<br />
Geoi^, J. M., 8t Brief, A. P (1992). Feeling good-doing good: A conceptual analysis of the mood<br />
<strong>at</strong> <strong>work</strong>-organiz<strong>at</strong>ional spontaneity rel<strong>at</strong>ionship. Psychological Bulletin, 112, 310-329.<br />
Hackman. J. R.. & Oklham, (i. R. (197S). Development of the joh diagmïslic ?.\.iv\ey. founud of<br />
Appiied Psychotogy, 60, 159-170.<br />
Hackman. J. R., & Oldham, G. R. (1980). Work redesign. Reading. MA: Addison-We.sley.<br />
Han, S. (1988). The rel<strong>at</strong>ionship between life s<strong>at</strong>isfaction and fiow in eldtrl)' Korean immigrants.<br />
ln M. Csikszentmihalyi ¿i 1. S. Csikszentmihalyi (Eds.), Optttn<strong>at</strong> e.xperience: Psychoiogicai<br />
studies ofjlow in counscisousness (pp. 138-149). New York: (Cambridge University Press.<br />
Haworth, J. T., & Hill, S. (1992). Work, leisure, and psychological well-being in a sample of young<br />
adults. Jourtial of Communit)- and Applied Social Psychoiog): 2(2). 147-160.<br />
Hofmann, D. A., Griifin, M. A., & Gavin, M. B. (2000). The applic<strong>at</strong>ion of hierarchical linear<br />
modeling to organiz<strong>at</strong>ional research. In K. Klein & S. Kozlowski (Eds.), Multiievei theory.<br />
researcij. and methods in organiz<strong>at</strong>ions: Found<strong>at</strong>ions, extensions, and new directions<br />
(pp. 467-511). San Francisco. CA: Jossey-Bass.<br />
flormiith. S, H. (1986). Tlie SAmpMngoftxpt^úcnccs in situ. Journ<strong>at</strong> of Perstin<strong>at</strong>ity. 5Í. 262-293.<br />
Hull, H. B. (1991). Mood as a product of leisure: Causes and consequences. In B. L. Driver,<br />
P J. Brown, & G. L. Peterson (Eds.), Benefits of teisure (pp. 250-262). St<strong>at</strong>e College, PA:<br />
Venture Publishing.<br />
Ilies. R., & Judge. T A. (2002). Understanding the dynamic rel<strong>at</strong>ionships among personality, mood,<br />
and joh s<strong>at</strong>isfaction: A field <strong>experience</strong> <strong>sampling</strong> study. Organiz<strong>at</strong>ional Behavior and Human<br />
Decision Processes, 89. 1119-1139-<br />
Jackson, D. (1984). Personality research form manual. Port Huron, MI: Research Psycho<br />
logists Press.<br />
Jackson, S. A. (1996). Toward a conceptual understanding of the flow <strong>experience</strong> in elite <strong>at</strong>hletes.<br />
Research Quarterly for Exercise and Sport. 67(1), 76-90.
614 Clive ]. Fullagar and E. <strong>Kevin</strong> <strong>Kelloway</strong><br />
Jackson, S. A., & Eklund. R. C. (2002). Assessing flow in physical activity: The flow st<strong>at</strong>e scale-2 and<br />
dispositional flow scaic-2. Jounud of Sport and E.xercise Psychotogy, 24. 133-l''O.<br />
Jackson, S, A., ik Kklund, R. C. (2004). The Jlow scides manual. Morgantown, WV: Fitness<br />
Inform<strong>at</strong>ion Technology.<br />
Jackson. S. A., Kimiecik, J, C;.. Ford. S., & Marsh, H. W. (1998). Psychologicai correl<strong>at</strong>es of How in<br />
span. Journal of Sport and Exercise Psychotogy, 20, 358-378.<br />
Jackson, S. A., & Marsh, H. W. (1996). Development and valid<strong>at</strong>ion of a seale to measure optimal<br />
<strong>experience</strong>: The flow st<strong>at</strong>e scale. Joumal of Sport and Exercise Psycho¡o¡>y. 18, 17-35.<br />
Johns, G., Xie,J. L., &Fang, Y. (1992). Medi<strong>at</strong>ing and moder<strong>at</strong>ing effects in job desigj}. Joumal of<br />
Management, 18, 657-676.<br />
Kahneman, D. (1999). Objective happiness. In D. Kahneman, E. Diener, & N. Schwarz (Eds.),<br />
Well-being: We fotmd<strong>at</strong>ions of hedonic psy^chology (pp. 3-25). New York: Russell Sage.<br />
Kahneman, D., Diener, E., & Schwarz, N. (Eds.), (1999). Well-being: The found<strong>at</strong>itms of hedonic<br />
psychology. New York: Russell Sa};e.<br />
<strong>Kelloway</strong>, F. K.. (iallagher, D., & Barling, J. (2005). Work employment and the individual.<br />
In B. Kaufman (Ed.), Theoretical perspectives on <strong>work</strong> and the employment rel<strong>at</strong>ionship<br />
(pp. 105-132). Champaign, IL: Industrial Rel<strong>at</strong>ions Research Associ<strong>at</strong>ion.<br />
Kimieeik, J. C, & Stein, (i. L. (1992). Examining flow <strong>experience</strong>s in sports contexts: Coneeptual<br />
issues and methodological concerns. Journ<strong>at</strong> of Applied Sports Psychoiog^\ 4, 144-160.<br />
King, L. A., & Napa, C. K. (1998). Wh<strong>at</strong> makes life good? Joumal of Personality and Social<br />
Psychology. ^5. 156-165.<br />
Larson, R., & Richards, M. H.(1994). Divergent realities: The emotional lives of mothers, f<strong>at</strong>hers,<br />
and ¿idolescents. New York: Basic Books.<br />
LeFevre. J. (1988). Flow and tht qualitj' of <strong>experience</strong> during <strong>work</strong> and leisure. In<br />
M. Csikszentmihalyi & I, Csikszentmihalyi (Eds.), Optimal <strong>experience</strong>: Psychological studies<br />
ofjtow in corLsciousness (pp. 307-518). New York: Cambridge University Press.<br />
Lutbans, F (2001). i'be case tor positive organiz<strong>at</strong>ionai bebavior (.POB). Current Issues in<br />
Management, /, 10-21.<br />
Lutbans, F (2002). Positive organiz<strong>at</strong>ional behavior: Developing and managing psycbological<br />
strengths. Academy of Management Executive, /6(I), 57-72.<br />
Marsh, H. W, & Jackson, S. A. (1999). Flow <strong>experience</strong>s in spon: Constnict valid<strong>at</strong>ion of<br />
multidimensional, hierarchical st<strong>at</strong>e and trait responses. .Structurai Equ<strong>at</strong>ion Modelling, 6,<br />
343-371.<br />
Massimini, F, & Carii, M. (19H8). The system<strong>at</strong>ic assessment of flow in daily lift-. In<br />
M. Csikszentmihalyi & I. Csiks7x'ntniibalyi (Eds,), Optimai <strong>experience</strong>: Psychotogic<strong>at</strong> studies<br />
of Jlow in counsdsousness (pp. 266-287). New York: Cambridge University Press.<br />
Massimini, F., Csikszentmihalyi, M., & CarIi, M. (1987). Tbe monitoring of optimal <strong>experience</strong>: A tool<br />
for psychi<strong>at</strong>ric rehabilit<strong>at</strong>ion.yo//ri;rt/ oj'Nervous and Mental Disease. I ~5(9), 545-549.<br />
McC>regor, I.. & Litlle, B. R. (1998). Personal projects, happiness, and meaning: On doing well and<br />
being yourself. Journai of Persomitity and Snci<strong>at</strong> Psychology, 74, 494-512.<br />
Miscbel, T. (1969). Scientific and philosophical psychology: A Iiistorical introduction. In T. Mischel<br />
(Ed.), Human action (pp. 1-40). New York: Academic Press.<br />
Nakamuni, J., & Csikszentmihalyi, M. (2002). TIic concept of flow. In C. R. Snyder &J. S. Lopez<br />
(Eds.), Handbook of posittvepsychology (pp. 89-105). New York: Oxford University Press.<br />
Ng, T. W. H., Sorenseon, K, L., & Eby, L. T. (2006), Locus of control <strong>at</strong> <strong>work</strong>: A meta analysis.<br />
Joumal of Organiz<strong>at</strong>ional Behavior. 21, 1057-1087.<br />
O Brien. (}. E. (1983). Skill-iitiliz;uion. skill-variety <strong>at</strong>id the job characteristics model. Australian<br />
Joumal of Psychology, J5. 455-460.<br />
Oklbam, G. R. (1996). Job design. In C. L. Coopérât I. T. Robertson (Eds.), Intem<strong>at</strong>ional review of<br />
industrial and organiz<strong>at</strong>ional psychology' CWoX. 11, pp. 33-60). Cbichester, UK: Wiley.<br />
Reis, H. T., Sheldon, K. M., Gable, S. L., Roscoc, J.. & Ryan. R. M. (2(KH)). Daily well being: Tbe<br />
role of autonomy, competence, and rel<strong>at</strong>edness. Personality and Social Psychology Bulletin,<br />
26,419-435.
Flow <strong>at</strong> vifork 615<br />
Renn. R. W., & Vandenberg, R. J. (1995). Tlie critical psychological st<strong>at</strong>es: <strong>An</strong> urulerrcpresented<br />
component in job characteristics model research. Jotirtiiil of Manu}'eweut. 21, 279-303.<br />
Ryan. R. M.. & Deci, E. L. (2001). On happiness und human potentials: A review of research on<br />
hedonic and eudaimonic well-being. <strong>An</strong>nual Review of Psycb(tl(>}>}\ ^2. 1-Í I-U>6.<br />
Ryff, C. D. (1989). Happiness is everything, or is it? Explor<strong>at</strong>Ums on the meaning of psychologicul<br />
well-being. Joumal of Peisoncttity and Soci<strong>at</strong> Psychology, 57, 1069-1081.<br />
Ryff. C^ D. (1995). Psychological well-being in adult life, entrent Directions in Psycbological<br />
Science, 4,99-104.<br />
Ryff, C. D., & Keyes, C. L. M. (1995). The structure of p.sychological well-being revisited.youma/<br />
of Persnnaiit}' anil Social i\s;i'cholog}', 69, 719-727.<br />
Ryff. C. D.. & Singer, B. ( 1998). The contours of positive human health. Psycbological Inquiry,<br />
9, 1-28.<br />
ScoUon, C. N., Diener, E., Oishi, S., & Biswas-Diencr. R. (2(H)5). <strong>An</strong> <strong>experience</strong> saniplinfi and crosscultural<br />
investig<strong>at</strong>ion of the rel<strong>at</strong>ion between pleasant and unpleasant affect. Cognition ami<br />
Emotion. /9, 27-52.<br />
ScoUon. C N.. Kim-Prieto. C, & Diener, E. (2003). Experience <strong>sampling</strong>: Promises and pitfalls,<br />
strengths and weaknesses, yo/zr««/ of Happiness Studies, 4, 5-34.<br />
Seligman. M. E. P (1999). The presidents address. American Psycbologist. 54, 559-562.<br />
Seligman. M. E. P. & Csikszentmihalyi, M. (2000). Positive psychology: <strong>An</strong> introduction. American<br />
Psycboiogist. 55(1), 5-14.<br />
Seongyeul. H. (1988). The rel<strong>at</strong>ionship between life s<strong>at</strong>isfaction and flow in elderly Korean<br />
immigrants. In M. Csikszentmihalyi & I. S. Csikszcntmiiialyi (Eds.). Optimal <strong>experience</strong>:<br />
Psycbotogical studies of flow in conscious7iess (pp. 158-149). New York: Cambridge<br />
University Press.<br />
Sheldon, K. M.. Ryan. R. M.. & Reis, H. T. ( 1996). Wli<strong>at</strong> makts fbr a good day? Competence and<br />
autonomy in the day and in the person. Personality and Social Psychology Bulletin, 22,<br />
1270-1279.<br />
Smyth, J. M.. & Stone, A. A. (2003). Ecological momentary assessment research in behavioral<br />
mciWcine. Journal of Happiness Studies. 4. 35-^2.<br />
Staw, B. M.. & Barsade, S. G. (1993). Affect and majierial performance: A test of the sadder-butwiser<br />
vs. happier-and-smarter hypotheses. Administr<strong>at</strong>ife Scietue Quarterly, ,^8. 304-331.<br />
Taber, T D., & Taylor, E. (1990). A review and evalu<strong>at</strong>ion of the psychometric properties of the job<br />
diagnostic survey. Personnel Psycbology. 4,i. 467-')OO.<br />
Turner, N.. Barling, J., & Zachar<strong>at</strong>os, A. (2002). Positive psycholog>' <strong>at</strong> <strong>work</strong>. In C. R. Snyder &<br />
S. J. Lopcx (Eds.), Handbook of positive psychology (pp.71S-728). New York: Oxford<br />
University Press.<br />
Voelkl, J. E. (1990). The challenge skill r<strong>at</strong>io of daily <strong>experience</strong>s among older adults residing<br />
in institutional environments. Therapeutic Recre<strong>at</strong>ion Journal, 24i2), 7-17.<br />
Warr. P B. (1987). Work, unemployment und mental healtb. Oxford: Oxford University Press.<br />
Warr, P B. (1999). Well-being in the <strong>work</strong>place. In D. Kahneman. E. Diener, & N. Schwarz (Eds.),<br />
Well-being: The found<strong>at</strong>ions of bedonic psycbology (pp. 392-412). New York: RusscI Sage.<br />
W<strong>at</strong>erman, A. S. (1993). Two conceptions of happiness: (x)ntra.sts of personal expressiveness<br />
(eudaimonia) and hedonic enjoyment. Journal of Personality and Social Psychology, 64,<br />
678-91.<br />
Wright, T. A., & Bonett. D. G. (2007). Job s<strong>at</strong>isfaction and psychological well-being as non-additive<br />
predictors of <strong>work</strong>place turnover./o/zr;/«/ of Management, .Î.Î, 141-160.<br />
Wriglit, T A.. & Staw. B. M. (1999). Affect and favorable <strong>work</strong> outcomes: Two longitudinal tests of<br />
the happy-productive <strong>work</strong>er thesis./o//r«i// of Organiz<strong>at</strong>ional fiehavior. 20, 1 -23<br />
Zuckerman, M. (1983). The distinction between trait and st<strong>at</strong>e .scales is not arbitrary: Comment<br />
on Allen and Potkay's 'On the arbitrary distinction between traits and St<strong>at</strong>es'. Journal of<br />
Personality and Social Psychology, 44, 1083-1086.<br />
4 April 2008; revised version received 13 August 2008