Reconciliation Between Monetary Incentives and Motivation Crowding-Out: The Influence of Perceptions of Incentives on Research Performance

Abstract

Motivation crowding theories suggest that the influence of performance-based monetary incentives on performance may depend on how employees perceive the incentives. To examine the crowding-out effect, this article analyzed a panel dataset of faculty’s research published over 9 years in a Korean university, focusing on the moderating role of perceptions of incentives. We found that, as the university increased financial incentives for research performance, academic researchers who perceived the incentives as supportive published more papers in higher impact factor journals. In contrast, the quantity and quality of research performance of those who perceived such incentives as controlling were not significantly associated with the increase in the incentives. To improve the performance of the performance-incentive system with potential crowding-out effects, administrators should communicate with employees to help them perceive incentives as supportive and positive.

Keywords:

• academics
• crowding-out motivation crowding theory
• performance-based incentives
• research performance

Notes

1 It should be noted that some studies proposed a more specific classification of intrinsic and extrinsic motivation. For instance, Ryan and Deci (2000) proposed that extrinsic motivation can vary greatly in its relative autonomy from the one involves more compliance with an external control to the case entails personal endorsement and a feeling of choice. Additionally, Lindenberg (2001) suggested using two types of intrinsic motivation. Enjoyment-based intrinsic motivation refers to the satisfying flow of activity, such as playing a game, while obligation-based intrinsic motivation is the activity undertaken to act appropriately, which can be relevant to public service or prosocial activity. However, because our study was based mainly on Frey’s motivation crowding theory, we used Frey’s broad definition of intrinsic motivation, i.e., “One is said to be intrinsically motivated to perform activity when one receives no apparent reward except the activity itself (Frey & Jegen, 2001, p. 591, as cited in Deci, 1971, p. 105). In our study, scientists’ motivation to conduct research because of the importance of the task or enjoyment can be regarded as intrinsic motivation.

2 On average, non-respondents published 0.87 papers in SCI-indexed journals per year after adjusting for the number of coauthors (S.D.=1.74), and the average impact factor was 1.98 (S.D.=1.54). Among 101 professors, about 44 percentage were from engineering fields, and 44% were from the science field.

3 We excluded clinicians from the sampling frame because the incentives for publications are much less important to them compared to the incentives for performing medical treatment at the university hospital.

4 The most notable example is the San Francisco Declaration on Research Assessment (DORA) initiated in 2012 (see https://sfdora.org/).

5 If the professor is the first author (or corresponding author) and there is n coauthor(s), the paper is counted as 1/(n + 2); if the professor is neither the first author nor the corresponding author, and there is n coauthor(s), the paper is counted as 2/(n + 2).

6 The quality-based incentive index was calculated based on the product of the proportion of publications in incentive category at t − 1 and the amount of cash bonus for the category at t.

Table A. The quality-based incentives for incentive category ($100)

Table

*Nature, Science, Cell, Lancet, and New England Journal of Medicine.

7 We used lagged financial incentives variable due to the delayed adjustment in professors’ behaviors. In case of promotion requirement, however, the university applied a grace period for professors to adjust their behaviors before the assessment; hence, we used the promotion requirement without a lag.

8 For the analysis, we used subgroups divided by the median rather than zero because, in the latter, the number of professors in the controlling group was only 30 compared with 145 in the supportive group. We concerned that, if we used subgroups divided by zero, there could be arguments that the non-significant results of the controlling group may come from the small sample size rather than the influence of perception. In addition to the results presented in Tables 3 and 4, we also examined the same fixed-effect models using subgroups divided by zero, and the signs and levels of significance for the variables of our interest were the same as those for the analyses with subgroups divided by the median.

Additional information

Funding

This work was supported by a grant from Kyung Hee University in 2017 (KHU-20171281).

Notes on contributors

Do Han Kim

Do Han Kim is an associate professor of public administration at Kyung Hee University, Korea. His research focuses on the performance management system in public and nonprofit organizations. He is currently conducting research on academics’ strategic responses to performance-based incentives such as gaming, goal-replacement, and crowding effects, as well as the empirical test of the superstar theory.

Hee-Je Bak

Hee-Je Bak is a professor of sociology and director of the Center for Science, Technology and Society at Kyung Hee University, Korea. His main research interests are research governance and the structural transformation of academic science. He is currently conducting research on the effects of quantitative performance-measures on Korean universities and social responsibilities of science.