24 March 2016
A new study led by the University of Oxford and Cornell University finds that giving workers feedback on their level of performance by ranking them according to how they compared with the rest of the workforce affects the effort they put into future tasks. Researchers found that those who received the best and worst ranking put in the most effort afterwards. Those in the middle still put in more effort than those who had received no feedback at all, but less than those ranked first and last. The working paper, ‘First-place Loving and Last-place Loathing’, shows a ‘U-shaped response’ in effort from workers who are ranked, and suggests the findings could help firms in how they communicate their appraisals or incentivise staff.
A total of 18 experimental sessions took place in a laboratory study involving 300 students, who were divided into groups made up of 17 people for the various tasks. They were given verbal and numerical tasks and ranked afterwards on how well they had done, with their score in the tasks corresponding to ‘effort’. The researchers found that those who learnt they were in first place significantly boosted their effort in tasks afterwards by over one-fifth (21%) compared with the average level of effort for the tasks overall. Those placed last boosted effort in later tasks by 12% compared with the average. However, those ranked in the middle (ranked 9th or 10th) put in the least effort of those who had received feedback on their ranking, putting in 11% less effort than the average. As their rank increased from 17th to 10th, their effort fell as they improved and safely moved away from bottom place. The reverse happened once they started to climb up the rankings from 9th to first, says the study.
The paper shows that the effects of ranking are universal, regardless of gender, race or age. The experiments were designed to show the ‘pure’ effect of ranking on effort, aside from any financial or reputational reward. They found no significant difference in the subsequent effort of study participants who had been told where they ranked privately as compared with those told in a group. No monetary rewards were provided to the best-ranked performers, with everyone receiving the same flat wage for taking part in the experiment.
The paper suggests that ranking is particularly effective in incentivising individuals who put in very good or very poor performances at work. The results highlight the value of awarding symbolic prizes to the high performers or scheduling regular appraisals with the worst performers. However, it says ranking might demoralise those in the middle if done too often, cautioning that middle-ranking employees are often the most loyal (possibly because they are less likely to be poached or fired), and that this type of feedback might not be the best approach in work environments where teamwork and cooperation between workers of different abilities are particularly important.
David Gill, Associate Professor in the Department of Economics at the University of Oxford and the Roger Van Noorden Fellow at Hertford College, said: ‘The findings show that people love being in first place and will do the best they can to stay there. People also loathe being in last place and will do all they can to climb up. Ranking performance is a popular approach for the business world and even the public sector, in hospitals and schools, today. It is important to know what effect this feedback has on those receiving the news and how this feeds into how best to communicate levels of pay, promotion decisions, appraisals and other rewards at work. Despite its popularity, there has been little consensus on how workers respond to this type of feedback. This research finds the specific rung that someone is on in the ranking ladder determines how much effort they put in afterwards. We also show that the effects of ranking are huge, highlighting just how important it is to get it right as there are also economic implications for the firms involved.’
Associate Professor Victoria Prowse, from the School of Industrial and Labor Relations at Cornell University, said: ‘Bonuses, promotions, performance appraisals and symbolic awards often depend on how well employees rank relative to other similar workers in the firm. In this paper, we show that workers respond strongly to being ranked. The pattern of response that we find has implications for organisational design. For instance, firms might want to divide workers into small comparison groups, adopt a decentralised organisational structure or design jobs that are more highly specialised in order to reduce the number of middle ranks. We hope that our results will also help improve the design of transparency policies, such as public disclosure of official hospital, school and university league tables in the United Kingdom, or public disclosure of income tax records in Scandinavian countries.’
For more information, contact the University of Oxford News Office on email@example.com ; or Tel: +44 (0) 1865 280534.
Alternatively, contact firstname.lastname@example.org; or Tel: 001 607 255 4723 (US number)/ mobile: 07576 04 4954.
Notes for Editors:
- The paper, First-place loving and last-place loathing: How rank in the distribution of performance affects effort provision, is a working paper published in March 2016.
- The authors are: David Gill, from the Department of Economics and Hertford College, University of Oxford; Zdenka Kissová, from PricewaterhouseCoopers; Jaesun Lee and Victoria Prowse from the Department of Economics, Cornell University.
- The research was funded by the Economic and Social Research Council and the British Academy.
- As is standard in the experimental literature on effort provision, the researchers measured ‘effort’ by performance or output in a work task rather than the associated cost of producing the output. In each of six rounds, the subjects worked on a verbal task (based on word spotting) for three minutes and on a numerical task (based on adding pairs of two-digit numbers) for a further three minutes. A subject’s effort in a given round is measured by the sum of the number of words spotted and pairs of two-digit numbers correctly added up in that round.
- The subjects ranked first increased effort by 21% compared to the average, while the middle-ranked subjects lowered effort by 11%. Relative to the middle-ranked subjects, the subjects ranked first increased effort by 35%.