Professional Documents
Culture Documents
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Sage Publications, Ltd. is collaborating with JSTOR to digitize, preserve and extend access to Sociology.
http://www.jstor.org
RESEARCH NOTE
Introduction
Identifying the factorsthat affectthe qualityand quantityof employees'job
performanceis an importantissue in the studyof work and the workplace.
Social scientistsand managersroutinelymake assumptionsabout the charac-
teristicsthatenhance or detractfromperformance.It is generallyassumed,for
example, that greaterorganisationalcommitmentcontributesto betterper-
formanceas more committedemployeesshould be more motivatedto work
hard on theirorganisation'sbehalf(Kalleberg and Marsden 1994). Moreover,
people withbetterskillsand greaterabilitiesare usually thoughtto be more
capable of performingtheir work tasks well, irrespectiveof their levels of
organisationalcommitment.Performanceis also assumed to be affectedby
structuraland task characteristicssuch as whetheremployees are able to
exerciseautonomyand discretionin theirwork,the extentto whichtheirtasks
are clearlydefined,and whethertheyare rewardedforhard work.
Despite the importance of understandingthe determinantsof job per-
formance,empirical research on this topic is conspicuously scarce (see
Kalleberg and Marsden 1994, for a review). Particularlysparse are studies
based on data fromdiverse and broadly representative samples of workers.
Cross-nationalstudiesof performanceare practicallynon-existent;most of the
extantstudiesof performanceare based on data fromthe United States.
The Americandata come fromthe 1991 General Social Survey (GSS). The
GSS is a nearlyannual,multi-topicsurveyadministeredto an area probability
sample of roughly1500 English-speaking Americans(Davis and Smith 1992).
It includes a wealth of socio-demographicdata on the background and
currentstatusof respondents,as well as manyattitudinaldata. The 1991 GSS
included a topical module focused on 'work organisations'which asked
questions about job performanceas well as on commitment,abilityand task
This data set is the onlynationallyrepresentative
characteristics. sample in the
U.S. of which we are aware that contains informationon job performance.
Other existingdata, usually collected in a single occupation or organisation,
are based on restrictedsamples whichhâve littlegeneralisability.
The Britishdata are drawn fromthe 1992 Employmentin Britainsurvey
(EIB). The survey was sponsored by a consortium of industries,U.K.
Departmentof Employment,EmploymentService and LeverhulmeTrust. It
was conducted under the directionof Duncan Gallie of NuffieldCollege
(Oxford),and Michael Whiteof the PolicyStudies Institutein London (Gallie
MeasuringJobPerformance
Our measures of the qualityand quantityof an employee'sjob performance
are based on the respondent'sself-reports of his/herperformancecomparedto
otherswho do similarwork.Table 1 givesthe two questionsthatwere used.
Self-reportmeasures such as these have been used previouslyby others
(Pruden and Rees 1972; Busch and Busch 1978; and Darden et al. 1989).
Subjective measures may be criticisedon the grounds that people cannot
evaluatetheirperformanceaccuratelydue to biases such as poor introspection
(Locke et al. 1988). Nevertheless,self-reportmeasures have been found to
correlatehighlywith more objective measures. In a study comparing self-
reportedjob performanceratingsand ratingsgiven by their superiors,for
example, Heneman (1974) found that self-reportmeasures had less halo
error,restrictionof range and leniencythan the more objective measures.
He concluded that self-reportmeasures are more appropriatefor research
purposes than for organisationalevaluations. In addition, more objective
measures,such as supervisor'sratingsand outputmeasures,are usefulonlyin
specific settingsand cannot be applied to the whole labour force (Steers
1977b; Judgeand Ferris 1993).
In the U.S., the mean for Question 1 (WORKWELL) is 4.07, while for
Britainit is 3.58. For Question 2 (WORKMUCH) the U.S. mean is 3.83; the
Britishmean is 3.46. That the mean scores are above 3 for both samples
indicatesthatrespondentsin both the U.S. and Britaingenerallyevaluatetheir
performancelevels to be above average. This result parallels that usually
obtained for self evaluations generally,and recalls Garrison Keillor's des-
criptionof the children in mythicalLake Wobegon, who were all 'above
average'. Why this is the case is less clear. One explanationmightbe that
respondentsin these surveys do indeed performbetter than people with
whom theycompared themselves.In that case, respondentstend to compare
themselveswithpeople whose performancelevel is below theirs.It mightalso
be thatpeople tend to inflatetheirown performancerelativeto others,and/or
to use different standardsof evaluation(see Kallebergand Marsden 1994).
U.S. respondents report higher average scores on both dimensions of
performancethan respondentsin Britain.The countrydifferencesin mean
scores (0.49 forWORKWELL, 0.37 forWORKMUCH) are both statistically
significantand may indicatethatthereis a higherlevel of job performancein
the U.S. than in Britain: American employees might performbetter and
produce more than theirBritishcounterpartson average.On the otherhand.
Table 1
Measures ofJobPerformance,U.S. and Britain
U.S. Britain
N (%) N (%)
Question1
'Comparedto otherpeople
whodo thesameor similar
kindofworkthatyoudo,
howwellwouldyousay
youdo yourjob?'
(WORKWELL)
1. Muchworse 0 (0) 5 (0.1)
2. Somewhat worse 2 (0.3) 20 (0.6)
3. Aboutthesame 188 (25.7) 1688 (48.8)
4. Somewhat better 287 (39.3) 1179 (34.1)
5. Muchbetter 241 (33.0) 377 (10.9)
Missing 13 (1.8) 192 (5.5)
Mean (S.D.) 4.07 (0.78) 3.58 (0.70)a
Question2
'Comparedto otherpeople
whodo thesameor similar
kindofworkthatyoudo,
howmuchworkwouldyou
sayyoudo?'
(WORKMUCH)
1. Muchless 1 (0.1) 10 (0.3)
2. Somewhat less 18 (2.5) 83 (2.4)
3. Aboutthesame 251 (34.3) 1839 (53.1)
4. Somewhat more 276 (37.8) 1083 (31-3)
5. Muchmore 170 (23.3) 261 (7.5)
Missing 15 (2.1) 185 (5.3)
Mean (S.D.) (0.82) (0.69)a
3^3 3^6
aU.S.- Britaindifference at
significantp < 0.001 (2-tailedtest)
Correlates
ofJobPerformance
Motivation
Mowday,Porterand Steers definedorganisationalcommitmentas:
Ability
We do not have directmeasuresof 'ability',so we use a varietyof ascribedand
Table 2
DescriptiveStatisticson ExplanatoryVariables,Mean StandardDeviation
U.S. Britain
Ability
Education 0.15 (0.46) 0.12 (0.35)
Class 3.27 (1.68) 3.16 (1.69)
Income 5.47 (2.93) 5.50 (2.87)
Full-time 0.85 (0.36) 0.79 (0.41)a
Unemployed 0.16 (0.37) 0.13 (0.34)
Employertenure 6.63 (7.47) 6.61 (7.12)
Supervisor 0.35 (0.48) 0.44 (0.50)a
Motivation
Effort 3.22 (0.65) 3.01 (0.67)a
Promotion Importance 2.65 (1-30) 2.28 (0.96)a
WorkOrganisation
WorkIndependently 3.06 (0.95) 3.26 (0.85)a
in Decisions
Participation 2.78 (0.92) 2.55 (0.98)a
TaskSignificance 3.17 (0.69) 3.00 (0.75)a
PerformanceVisibility 3.04 (0.66) 2.92 (0.67)a
PayforPerformance 0.68 (0.47) 0.39 (0.49)a
Size ofWorkplace 418 (517) 195 (193)a
UnionMembership 0.15 (0.36) 0.40 (0.49)a
is significantly
Coefficient betweenU.S. and U.K. at p < 0.001 (2-tailed
different
t-test)
TaskCharacteristics
Previousresearchhas found that autonomy(freedomto workindependently)
and discretion (participationin decision-making)enhance both motivationand
job performance(Darden et al. 1989). Workerswithgreaterautonomytend to
be more involvedin theirjobs (Britishworkersreportedgreaterautonomy
than U.S. workers). In addition, discretionor decision-makingbinds the
interestsof individual employees to the outcome of work (U.S. workers
reported greater discretionthan British workers). There are likely to be
substantialcross-nationaldifferences in the linksbetweenthese taskcharacter-
istics and performance.Lincoln and Kalleberg (1990) found autonomyand
discretionto be generallymore importantfor American workersthan for
Japanese in accounting for organisationalcommitment,for example, while
Dore (1973) attributedthe relativelypoorer performanceof Britishindustry
partly to a lower level of organisational integrationthrough employee
participation.
Employeesare also assumed to workbetteriftheirjobs are well-definedand
if theyare recognisedwhen theydo theirworkwell. We use two measuresto
indicate the degree of visibilityof individualemployeesin the organisation.
One representsthe respondent'sevaluationof tasksignificance , i.e., the import-
ance of his/herwork for the success of the organisation.The second is the
respondent'sperceptionof performance , i.e., how easy it is forhis/her
visibility
efforts to be noticedby others.On both measures,U.S. workersscored higher
than British.
An additionaltask-relatedmeasureis the perceivedlinkbetweenperformance
and pay rises(Americanswere more likelythan the Britishto feel thatthese
were linked). We hypothesisethat people whose effortsare duly rewarded
financiallyare likelyto performbetterin theirwork.
We also include indicatorsof establishment size (Americansworkedin larger
organisationsthan the Britishin these samples) and unionmembership(the
Britishwere more likelyto be unionised) as proxiesforworkstructureslikely
to affectperformance.
BivariateCorrelations
Table 3
CorrelationbetweenPerformanceMeasures and ExplanatoryVariables
U.S. Britain
WORKWELLWORKMUCHWORKWELLWORKMUCH
Ability
Education -0.047 -0.073* -0.023 -0.018
Class 0.039 0.092* 0.071*** 0.096***
Income 0.124*** 0.142*** 0.111*** 0.087***
Full-time 0.064t 0.124*** 0.114*** 0.110***
Unemployed -0.043 -0.066f -0.015 -0.041*
Employertenure 0.087* 0.042 0.074*** 0.039*
Supervisor 0.126*** 0.173*** 0.127*** 0.135***
Motivation
Effort 0.095** 0.161*** 0.089*** 0.120***
Promotion Importance 0.106** 0.197*** 0.075*** 0.096***
WorkOrganisation
WorkIndependently 0.132*** 0.131*** 0.096*** 0.091***
in Decisions 0.131***
Participation 0.100** 0.126*** 0.138***
TaskSignificance 0.123*** 0.175*** 0.087*** 0.084***
PerformanceVisibility 0.138*** 0.104** 0.076*** 0.051**
PayforPerformance -0.007 0.050 0.036* 0.03If
Size ofWorkplace 0.018 0.017 0.078*** 0.038*
UnionMembership -0.028 -0.035 0.003 0.002
***:p < 0.001; **: p < 0.01; *: p < 0.05; f: P < 0.10
Determinants : MultivariateResults
ofJobPerformance
QualityofPerformance
How do we account for differencesamong employeesin the qualityof their
job performance?And how do we explain the differences in qualityperform-
in
ance the U.S. and Britain?Table 4 presentsregressioncoefficients fromour
multivariateanalysisof qualityperformanceforemployeesin both countries.
It givesour assessmentof the effectsof motivation,abilityand task character-
istics on quality controls for all other explanatoryvariables and individual
characteristics.We test whether there is a statisticallysignificantcountry
difference betweena pair of coefficients by means of a (two-tailed)t-test.2
Task Characteristics
. Employees with more autonomyin work also report
better quality performance.However, the effectof discretion on quality
performanceis not significant.A possible explanation of this pattern is
suggestedby results obtained by Gallie, Kalleberg and White (1994), who
found that discretionis more stronglyrelated to organisationalcommitment
than autonomy.It may be that the effectof discretionon performanceis
primarilyindirect,via the effortdimension of commitment,leaving only a
weak, direct effectof discretion on performanceonce commitment(i.e.,
effort)is controlled.
Respondentswho feltthattheyare noticed when theyworkwell were more
likelyto report better performance.However, respondentswho considered
theirworkto be significant forthe success of the organisationwere not more
likelyto reportbetterqualityperformance.
Contraryto expectations,workerswho reportthat hard work leads to pay
risestend to reportpoorerqualityperformance.This resultis surprising, since
pay rises are presumablyused by employersto stimulatebetterperformance.
We speculate that this negative effectmight reflecta resentmentagainst
pecuniaryrewardsbased on performanceevaluation,especiallyamong those
employeeswho receiverelativelylowerpay rises than theircolleagues. Union
membershipis unrelated to performancequality.Britishworkersin larger
establishments - but not in the U.S. - are more likelyto reportthattheirwork
is of higherquality.
Table 4
Regressionsof QualityofJobPerformanceon Determinants,
U.S. and Britain
U.S. Britain
B (S.E.) B (S.E.)
Male -0.045 (0.076) 0.048 (0.037)
White -0.205 (0.095)* -0.185 (0.059)**
Ability
Education -0.082 (0.063) 0.017 (0.047)
Class -0.014 (0.024) 0.014 (0.011)
Income 0.031 (0.016)* -0.001 (0.008)b
Full-time -0.066 (0.110) 0.048 (0.049)
Unemployed 0.056 (0.106 0.045 (0.047)
Employertenure -0.0003 (0.005) 0.006 (0.002)**
Supervisor 0.102 (0.075) 0.087 (0.034)**
Motivation
Effort -0.022 (0.063) 0.041 (0.025)t
Promotion Importance 0.053 (0.03l)f 0.041 (0.016)**
WorkOrganisation
WorkIndependently 0.105 (0.042)** 0.058 (0.019)**
in Decisions
Participation -0.0007 (0.047) 0.011 (0.019)
TaskSignificance 0.042 (0.058) 0.024 (0.022)
PerformanceVisibility 0.104 (0.056)f 0.064 (0.024)**
PayforPerformance -0.182 (0.075)* -0.046 (0.032)b
Size ofWorkplace(X1000) -0.005 (0.06) 0.20 (0.08)**a
UnionMembership -0.041 (0.095) -0.054 (0.034)
N 515 2210
Table 5
of of
Regressions Quantity JobPerformanceon Determinants,
U.S. and Britain
U.S. Britain
B (S.E.) B (S.E.)
Male -0.205 (0.079)** -0.012 (0.036)a
White -0.095 (0.099) -0.110 (0.058)f
Ability
Education -0.099 (0.066) -0.016 (0.046)
Class 0.004 (0.025) 0.004 (0.011)
Income 0.024 (0.017) -0.005 (0.008)
Full-time 0.015 (0.117) 0.114 (0.048)*
Unemployed -0.045 (0.110) -0.048 (0.046)
Employertenure -0.004 (0.005) 0.001 (0.002)
Supervisor 0.170 (0.078)* 0.104 (0.034)**
Motivation
Effort 0.036 (0.065) 0.076 (0.025)**
Promotion Importance 0.113 (0.032)*** 0.034 (0.016)*a
WorkOrganisation
WorkIndependendy 0.113 (0.044)** 0.048 (0.019)**
in Decisions - 0.057 0.037 b
Participation (0.049) (0.019)
TaskSignificance 0.140 (0.060)* 0.026 (0.022)b
PerformanceVisibility 0.027 (0.058) 0.013 (0.023)
PayforPerformance -0.028 (0.078) -0.069 (0.031)*
Size ofWorkplace(X1000) 0.03 (0.07) 0.07 (0.08)
UnionMembership 0.037 (0.099) -0.047 (0.033)
Intercept 2.709 (0.272) 2.819 (0.125)
R2 (adj) 0.077 0.036
N 512 2217
***:p < 0.001; **: p < 0.01; *: p < 0.05; f: p < 0.10
a: U.S. - Britaindifference
significant at p < 0.05;
: U.S. - Britaindifferencesignificant at p < 0.10
QuantityofPerformance
Motivation . British employees- but not Americans- who say they are
willingto exert extraeffortto help theirorganisationsucceed are more likely
also to report that they do more work than comparable workers(but this
countrydifferenceis not significant).On the other hand, employees who
attachgreaterimportanceto promotionalso say theyproduce more thantheir
colleagues. Americanemployeesare even more likelythan the Britishto do
this,and the countrydifference is statistically
significant.
Task Characteristics
. Autonomyis positivelyrelatedto quantityperformance
in both countries,but discretionis not. American- but not British- em-
ployees who perceivethemselvesto 'matter'forthe success of theirorganis-
ation also report better quantityperformance.This countrydifferenceis
statistically
significant.
Conclusions
Notes
1. It mightbe arguedthat motivation/commitment is an intervening
variable
betweenpersonalabilityand task characteristics, on the one hand,and job
performance, on the other.For instance, highlyqualifiedpeoplemaybe more
motivated and committed workers, who havea driveto succeedoccupationally
through hardworkand good performance. Moreover, tasksthatinvolvegreater
autonomy mayproducegreatercommitment (see Lincolnand Kalleberg1990)
as wellas betterperformance. This suggests thatabilityand taskvariableshave
directeffects
on performance, as wellas indirect effectsthroughmotivation.
2. The formula forthet-testis:
bļ - b2
t=
(Se12+se22)1/2,
where'V and '2' areregression coefficientsassociatedwiththeU.S. and Britain,
and se! and se2aretheircorresponding
respectively, standarderrors.
References
Blumberg,M. and Pringle, C. D. 1982. 'The MissingOpportunity in Organisa-
tionalResearch:Some Implications fora Theoryof WorkPerformance'.Academy of
Management Review7:560-569.
Busch, P, and Busch, R. F. 1978.'WomenContrastedto Men in theIndustrial
Sales
Force:JobSatisfaction,
Values,Role Clarity,
Performanceand Propensityto Leave'.
JournalofMarketingResearch 15:438-448.
Address:Yuan Cheng,PolicyStudiesInstitute,
100 ParkVillageEast,LondonNW1
3SR. Arne L. Kalleberg,Departmentof Sociology,University
of NorthCarolina,
ChapelHill,NC, USA 27599-3210.