group-based pay plan. Studies have shown, however, that these plans can impact
Wrm performance, especially when they include mechanisms eliciting employee
participation, rewards for individual contributions, cash payouts instead of deferred
monies, and payouts of a signiWcant increment based on relatively small proWt
centers (e.g. Kruse 1993 ;Long 2000 ; Mangan and St-Onge 2005 ).
In a series of well-designedWeld studies of gain-sharing plans, Arthur and
colleagues (Arthur and Jelf 1999 ; Arthur and Aiman-Smith 2001 ; Arthur
and Huntley forthcoming) have illuminated some of the intermediary mechanisms
underlying these plans’ eVectiveness. These studies found that gain-sharing plans
can improve labor relations (Arthur and Jelf 1999 ), organizational learning (Arthur
and Aiman-Smith 2001 ), and eYciency via employees’ suggestions (Arthur and
Huntley forthcoming).
Intriguing evidence also suggests that the aggregate productivity eVects observed
for group-based pay plans may be partly due to their eVects on lower performing
employees. Studies by Weiss ( 1987 ) and Hansen ( 1997 ) suggest the move from
individual incentive to group incentive plans results in employee performance
converging to a standard: the lowest performing employees show substantial
improvement following the introduction of the group plan, while the performance
of the most able workers tends to decrease. These results await further replication
and speciWcation.
Sorting eVects. The eVects of pay-for-performance plans may be enhanced by
what labor economists label as ‘sorting’—the inXuence that incentives have on
employee attraction and attrition. In perhaps the best examination of sorting
eVects to date, Lazear ( 1999 ) provides compelling results from a study of wind-
shield installers at Safelite Glass Corporation. DeWned in terms of units installed
per day per worker, he found that worker productivity increased by 44 percent
following implementation of a piece-rate incentive plan, with about one-half due
to incentive eVects, and the remainder due to sorting.
Research grounded in the person–organizationWt paradigm also provides evi-
dence related to sorting. Per Rynes ( 1987 ), compensation sends signals to prospect-
ive applicants regarding organizational culture and values. These ‘signals’ may
impact workforce composition in the form of ability and/or disposition. Using
experimental and/or policy-capturing studies, research has shown that the attract-
iveness of pay-at-risk and individual vs. group-based contingent pay depends on
individual diVerences such as cognitive ability, risk preference, self-eYcacy, and
other aspects of personality (cf. Bretz et al. 1989 ; Cable and Judge 1994 ; Stevens and
Ash 1998 ; Trank et al. 2002 ).
Research also suggests sorting eVects through turnover. A meta-analysis
performed by Williams and Livingstone ( 1994 ) found that pay-for-performance
reward contingencies magniWed the inverse relationship between performance and
turnover. A more precise illustration is provided by Harrison, Virick, and William
remuneration: pay effects at work 349