Students' concern about indebtedness: a rank based social norms account

Abstract

This paper describes a new model of students' concern about indebtedness within a rank-based social norms framework. Study 1 found that students hold highly variable beliefs about how much other students will owe at the end of their degree. Students' concern about their own anticipated debt – and their intention of taking on a part-time job during term time – was best predicted not by the size of the anticipated debt, but by how they, often incorrectly, believed their debt ranked amongst that of others. Study 2 manipulated hypothetical debt amounts experimentally and found that the same anticipated debt was rated as 2.5 times more concerning when it ranked as the second highest being considered than when it was the fifth highest. Study 3 demonstrated that the model applies to evaluation of different types of debt (income contingent loans versus general debt).

Keywords:

• students' indebtedness
• context effects
• attitudes towards debt
• range frequency theory
• decision by sampling
• social norms

Acknowledgements

This work was supported by the Economic and Social Research Council [grant numbers RES-062-23-2462, ES/K002201/1] and the Leverhulme Trust [grant number RP2012-V-022].

Notes

1 By including in the present study those students who indicated that they anticipated no debt at the end of the degree we would have artificially inflated the relationship between debt and concern about it: if a student expects to owe £0 at the end of his/her degree, s/he won't worry about it at all. This would reduce overall variability in the data without adding any useful information.

2 The same results were observed when (a) the non-logarithmically transformed data were analysed, as only subjective rank and gender were significant predictors, and (b) when more flexible regression equations were used. For the latter, for instance, if the interaction term between anticipated debt and gender was entered into the analyses – thus allowing for debt levels effects to vary by gender – rank was still a significant predictor for both worry, B = 1.33, Wald = 6.16, p = .013, and difficulty to repay, B = 1.78, Wald = 11.01, p < .001, while anticipated debt was not, B = 1.01, Wald = 1.82, p = .177 and B = 0.62, Wald = 0.58, p = .447, respectively. The same pattern was observed when we allowed instead the term fees to interact, as rank was still a significant predictor for both dependent variables, B = 1.32, Wald = 5.98, p = .017 and B = 1.72, Wald = 10.16, p < .001, while anticipated debt was not, B = 0.33, Wald = 0.16, p = .690 and B = 1.12, Wald = 1.84, p = .175. Also, subjective mean did not enter as a significant predictor in any of these additional analyses (all p > .748).

3 We repeated this analysis by including the 44 students who had already a part-time job at the time of testing; intention towards taking a part-time job was coded as 0 = not planning to work during term-time and 1 = either working already or planning to take a job. The results from the logistic regression confirmed that only subjective rank was a significant predictor, while the effect of gender again approached significance.

4 The same results were observed when the non-logarithmically transformed data were analysed, as subjective rank was the only significant predictor and the effect of gender again approached significance.

5 We again collected data only from students who anticipate some debt at the end of their degree, as it is mostly those students who engage in financial considerations about debt and its impact on their lives, and thus represent the most informative sample for the current research purposes.

6 The same results were observed regardless of gender and fees.

7 As in Study 2, data were collapsed also across gender and fees as neither variable impacted on the results.