Economics students lying more . . . again?

In a study by Shalvi, Dana, Handgraaf, and De Dreu (2011) it was convincingly demonstrated that psychologically, the distinction between right and wrong is not discrete, rather it is a continuous distribution of relative ‘rightness’ and ‘wrongness’. Using the ‘die-under-the-cup’ paradigm participants over-reported high numbers on the roll of a die when there were financial incentives to do so and no chance of detection for lying. Participants generally did not maximise income, instead making moral compromises. In an adaptation of this procedure in a single die experiment 9% of participants lied that they had rolled a ‘6’ when they had not compared to 2.5% in the Shalvi et al. study suggesting that when the incentive is donation to charity this encourages more dishonesty than direct personal gain. In a follow-up questionnaire study where sequences of three rolls were presented, lying increased where counterfactuals became available as predicted by Shalvi et al. A novel finding is reported where ‘justified’ lying is more common when comparative gains are higher. 

An investigation of individual differences revealed that economics students were much more likely to lie than psychology students. Relevance to research on tax evasion, corporate social responsibility and the ‘credit crunch’ is discussed. 

That is the abstract of the paper "Drawing the line somewhere: An experimental study of moral compromise" by Lewis et al (Journal of Economic Psychology, August 2012)  (alas, no full paper is available online). If we assume that this is a stylized fact about economics students, how can we fix it or minimize it (beyond the obvious answer of teaching ethics, or reading ethical codes)? Is there a self-selection problem? How generalizable is this behavior?