I’ve always mistrusted rewards for students.  Students should be doing things for their own sake – turning up, learning, being nice – and I understood extrinsic rewards to undermine their intrinsic motivation to do so.  Consequently, I feel uncomfortable about the reward systems found in most schools: merits, house points, or attendance awards (84% of secondary schools reward attendance).  However, there is evidence that rewards can encourage sustained behaviour change: one review describes them as a ‘central component’ of habit formation (Wood and Neal, 2016, p.73).  In this post I review the evidence for and against rewards, when and why they can be effective, and my misgivings.  I’d welcome your thoughts on how you’ve seen reward systems succeed or fail.

Rewards can help people change.  One study offered students $175 either to:

  • Visit the gym eight times in a month
  • Visit the gym once
  • Complete an unrelated activity

Students offered money for visiting repeatedly did – unsurprisingly – more importantly, they continued going afterwards, losing weight and reducing their Body Mass Index.  Each group was equally enthusiastic about exercise, but multiple visits initiated a new habit: “monetary compensation for a sufficient number of occurrences… appears to move some people past the ‘threshold’ needed to engage in an activity (Charness and Gneezy, 2009, p.911).”  This study encouraged me to examine rewards in education; these have had mixed effects (reviewed by Fryer, 2016):

  • Paying seven-year-olds to read books increased reading achievement for native English speakers, but decreased it for students speaking English as a second language
  • Paying ten-year-olds to complete maths homework increased maths achievement but decreased reading achievement
  • Paying secondary school students for interim test results and for effort had limited or no effect on outcomes
  • Educational Maintenance Allowance – paying poorer students up to £40 per week for attendance – increased participation in education by over 6 percentage points by Year 13 (particularly for boys; Middleton et al., 2003) and led to higher A level results (Chowdry, Dearden and Emmerson, 2007)

These mixed results suggest that rewards can change behaviour and increase attainment, but their design is crucial.  Three features affect the impact rewards have:

Specificity

We encourage habits by rewarding specific behaviours, not general achievement.  Rewards which are “too broad to promote specific habits” are “less successful at habit formation”:

Overly general rewards include symbolic trophies, prizes that recognize strong performance, or temporal landmarks such as birthdays or the kickoff of a new calendar year. Only rewards that promote the repetition of specific actions contribute to habit formation (Wood and Neal, 2016, p.75).”

In education, rewards for inputs (like doing homework) lead to better results than rewards for outputs (like exam results); one study found rewards improved performance in maths but not in other subjects, possibly because it was clearer to students how to improve (Gneezy, Meier and Rey-Biel, 2011).  The most pertinent test of specific rewards in England offered GCSE students money, or a trip, for meeting weekly thresholds for attendance, behaviour, homework and classwork: on average, this had no effect, but it helped in maths, particularly for lower-attaining pupils (Sibieta, Greaves and Sianesi, 2014).  The evidence is suggestive rather than conclusive, but it aligns with advice on using praise: “the more actionable the thing you reinforce, the more students can replicate their success (Lemov, 2015, p.435).”  Specific rewards rely on specific goals, and on identifying the activities which matter most: rewarding every desirable action would be expensive, time-consuming and would no longer encourage the most important.  So, while we can acknowledge and welcome any desirable behaviour, rewards should focus on behaviours which are crucial and can be identified reliably.  If what matters most in Year 7 is building study habits and learning critical ideas, we could reward regular practice (through online tests, for example), acknowledging other desirable behaviour without rewarding it.

Predictability

Rewards which are less predictable may be more effective: “habits form best when rewards are powerful enough to motivate behavior but are uncertain in the sense that they do not always occur (Wood and Neal, 2016, p.75).”  Students become desensitised to predictable rewards: they come to see a merit as an entitlement, expecting it irrespective of their effort, or demanding something more for additional effort.  Intermittent rewards remain effective; they encourage people to focus on what they hope will happen and “powerfully motivate repetition and habit formation”:

Slot machines are a good example of uncertain rewards. People keep paying money into the machines because sometimes they win, sometimes they don’t… E-mail and social networking sites have similar effects: people keep checking on them because sometimes they are rewarded with interesting communications, but other times they get only junk. The key is that rewards are received probabilistically, meaning not for every behaviour (Wood and Neal, 2016, p.75).”

Rewards may be even more effective if they are eye-catching: offering a chance to win one big prize may be more motivating than offering small prizes to everyone, as people tend to overestimate the probability they will win; a lottery offering a £5000 prize increased electoral registration by 4.2% (Service et al., 2014).  We may offer students a predictable reward to start a habit (as in the gym attendance experiment), then offer surprise rewards to maintain the habit after a few weeks (more on how long it takes to form a habit here).*

Signalling

The signal students receive must match the signal we wish to send by offering a reward:

  • Intrinsic vs extrinsic motivation: Rewards may signal that an activity is unattractive, difficult or beyond students’ capacity; this may diminish their efforts after the reward is withdrawn, or encourage them to pursue the reward itself, not the learning it is designed to catalyse (Gneezy, Meyer and Rey-Biel, 2011).  This may explain why people often stop pursuing a behaviour once a reward is withdrawn (Mantzari et al., 2015).  Equally, some important things may never be intrinsically motivating and both extrinsic and intrinsic motivation have a role (Ryan and Deci, 2000).  We may promote intrinsic reward by helping students find value and satisfaction in an activity (Gneezy, Meier and Rey-Biel, 2011; Wood and Neal, 2016), or offer indirect rewards (as attempted by a study which offered students access to tempting audiobooks while they were at the gym; Milkman, Minson and Volpp, 2014), while balancing these approaches with more explicit rewards where these are insufficient.
  • Giving up: As a tutor, I found Educational Maintenance Allowance motivated students to attend.  It was paid for a full week’s attendance however: some students who were late on Monday made a point of being late for the rest of the week: “I’ve lost my EMA, there’s no point.”  We may adapt rewards if we notice such effects: paying EMA per day, rather than per week, for example, would have ensured each day mattered.
  • Crowding out: Rewarding one activity may reduce students’ motivation to pursue another: this may explain why students paid to do maths homework performed worse in reading (Fryer, 2012).  Recognising this encourages us to reward the actions which matter most and accept the consequences this may have for other desirable activities.
  • Diverging effects: Rewards affect individuals differently.  In education, girls and higher-attaining students tend to benefit more (Gneezy, Meier and Rey-Biel, 2011), but the English experiment described above found the opposite effect (Sibieta, Greaves and Sianesi, 2014).  Again, this comes with a risk: the successful gym attendance study actually diminished attendance slightly among those who were previously attending regularly – even though the overall effect was positive (Charness and Gneezy, 2009).  If we see rewards working for some students, we may wish to alter them to work for all students, changing the thresholds to achieve them, or focusing on personal bests.

Conclusion

“We have a reward system because for a few students, it really works,” one school principal told me.  In an ideal world, I’d do without; where we can help students value learning, attendance and kindness for its own sake, we should avoid diminishing this through extrinsic rewards.  The evidence beyond this seems inconsistent, but I think this is because the design of reward systems (and studies) is inconsistent too.  Under the right circumstances, targeting the right things, designed in the right way, rewards can have a positive influence.  However, I’m not convinced the average school reward system meets these criteria: most are too general, too vague and too predictable; often, the result is straight bribes: “Ten Vivos if you work all lesson.”  Specific, unpredictable and carefully-modified rewards may offer better prospects: if this helps students do better, maybe we should pursue it further.  I would welcome comments which would persuade me either way.

This is a draft excerpt from Habits of success: getting every student learning.  It’s available from Amazon & Routledge.

If you found this interesting, you may like…

Learn more about the guide I’m writing, here.

Read the overall framework for behavioural change I’m working with, here.

Learn more about habit formation, here.

* This suggestion is about rewards; it should not be confused with the value of predictable sanctions)..

References

Charness, G. and Gneezy, U. (2009). Incentives to Exercise. Econometrica, 77(3), pp.909-931.

Chowdry, H., Dearden, L., Emmerson, C. (2007). Education Maintenance Allowance: Evaluation with Administrative Data: The impact of the EMA pilots on participation and attainment in post-compulsory education.

Fryer, R. (2016) The Production of Human Capital in Developed Countries: Evidence from 196 Randomized Field Experiments. NBER Working Paper No. 22130.

Gneezy, U., Meier, S., & Rey-Biel, P. (2011). When and why incentives (don’t) work to modify behavior. The Journal of Economic Perspectives, 25(4), pp.191–209.

Lemov, D. (2015). Teach like a champion 2.0. San Francisco: Jossey-Bass.

Mantzari, E., Vogt, F., Shemilt, I., Wei, Y., Higgins, J. P., & Marteau, T. M. (2015). Personal financial incentives for changing habitual health-related behaviors: A systematic review and meta-analysis. Preventive Medicine, 75, 75-85. http://dx.doi.org/10.1016/j.ypmed.2015.03.001

Middleton, S. Maguire, A., Ashworth, K., Legge, K., Allen, T., Perrin, K., Battistin, E., Dearden, L., Emmerson, C., Fitzsimons, E. and Meghir, C. (2003). The evaluation of Education Maintenance Allowance Pilots: three years evidence: a quantitative evaluation. Department for Education and Skills research report ; 499. Annesley, Nottingham : DfES Publications.

Milkman, K., Minson, J. and Volpp, K. (2014). Holding the Hunger Games Hostage at the Gym: An Evaluation of Temptation Bundling. Management Science, 60(2), pp.283-299.

Ryan, R. and Deci, E. (2000). Self-Determination Theory and the Facilitation of Intrinsic Motivation, Social Development, and Well-Being. American Psychologist, 55(1), pp.68-78.

Service, O., Hallsworth, M., Halpern, D., Algate, F., Gallagher, R., Nguyen, S., Ruda, S., Sanders, M. (2014). EAST: Four simple ways to apply behavioural insights. Behavioural Insights Team.

Sibieta, L., Greaves, E. and Sianesi, B. (2014). Increasing Pupil Motivation: Evaluation Report and Executive Summary. Education Endowment Foundation.

Wood, W., and Neal, D. (2016). Healthy through habit: Interventions for initiating & maintaining health behavior change. Behavioral Science & Policy, 2(1), pp. 71–83.