Loss Aversion as Incentive to Study Guglielmo Volpe School of Economics and Finance Queen Mary University of London Developments in Economics Education Conference, Birmingham 10th-11th September 2015 The Broad Idea Can we use insights from Behavioural Economics to inform teaching, learning and assessment strategy in a way that academic and (possibly) other aspects of students’ performance are enhanced? The Talk “Heterogeneous Gender Effects under Loss Aversion in the Economics Classroom: A Field Experiment” Apostolova-Mihaylova M., Cooper W., Hoyt G., Marshall E.C. Southern Economic Journal, 2015, 81(4), 980-994 The Experiment Framing Loss Aversion in assessment strategy • Loss aversion • Individuals respond asymmetrically when they are faced with the prospect of gains or losses • Individuals are much more unhappy about losing an asset as they are happy about acquiring the same asset • The hypothesis • Induce higher student performance by framing grades as a reduction from the maximum points allowed, as opposed to a gain in score from the minimum points possible Field Experiment Design • Treatment: Penalty contract-aversion • Students are given ‘points’ when the course commences and progressively lose points throughout the semester for incorrect answers • Control: Reward contract-reciprocity • Students earn points as various assessments are completed throughout the year Results • Framing does not have an overall impact on students’ outcomes • Differentiated response to loss aversion by gender • In the treatment group, males score between 2.88 and 4.19 percentage points higher than the control group • In the treatment group, females score between 3.30 and 4.33 percentage points lower than females in the control group Time Inconsistency and Loss Aversion “Student Effort in Preparing for Exams: Intertemporal Preferences and Loss Aversion” Wüst K., Beck H. Decision Sciences Journal of Innovative Education, 2012, 10(2), 245-262 Hypotheses Hyperbolic discounting: time inconsistency • Students behave time inconsistently and prefer the largerlater reward i.e. a good grade, in the long run, but in the short term sacrifice the prospects of a good grade for immediate leisure time Loss Aversion • Students will be more eager to avoid a deterioration of their grade which would be perceived as a loss than they are to sacrifice time or money to improve their grade which would be perceived as a gain Method • Questionnaire distributed at start and end of term (before final exam) • Students asked about the total time they planned to spend (or had spent) on preparation of final exam • Two versions of questionnaire: • A: students asked how much extra time they would invest to improve their grade by 0.3 • B: students asked how much time they should be ‘given back’ in return for a grade which was 0.3 worse than their acquired one Results • Students behave time-inconsistently • At the beginning of the term, students understand that it is necessary to learn, but the planned number of learning hours decreases significantly as the end of the term approaches • Students show loss-aversion • Students are willing to study more to improve their grade than they would spend more leisure time in exchange for a deterioration of their grade What did I do • Year 1 Statistical Methods in Economics module • Assessment • 7 very short (5 minutes) in class tests worth 2 marks each (last test is worth 3 marks) • Statistical project (15 marks) • Final exam (70 marks) • Frame ‘Loss Aversion’ in assessment in order to induce more effort/better performance • Change introduced first in 2013/14 Framing Loss Aversion In class test results presented in the following way: Test Test Test Test Test Test Test Tot. Maximum 1 2 3 4 5 6 7 Available mark you Marks can achieve by end of semester Max % mark you can achieve by end of semester after nth test • In 2013/14 Students could see each other’s results • In 2014/15 Students could see only their own results Summary statistics Average assessment marks over the years n In Class tests Project Exam 2011/12 136 68.7 66.2 53.3 2012/13 186 62.9 71.2 61.3 2013/14 191 74.0 69.7 53.6 2014/15 197 70.2 67.8 56.6 Testing for differences in means • The average in-class test results are significantly higher under the ‘loss aversion’ framework • The 2014/15 average test scores are significantly (at 5% level) lower than in 2013/14 • Does comparing relative performance spur incentives to work harder? • There are no significant differences in the project average scores • There is a significant difference in average exam scores between 2012/13 and 2013/14 • It is likely that this difference has to do with my exam paper design…. • But, generally, no evidence of impact on exam marks…. Random thoughts….. • Could loss-aversion play a role? More statistical analysis…. • Need to check for gender differences? • Treatment and Control Groups • Borrow insights from behavioural economics? • Potentially costless changes that have positive impact? • E.g.: group work, learning diaries, learning contract etc. to ameliorate ‘time inconsistency’ problem?