Abstract
While field experiments are valuable in their ability to objectively analyse patterns of discrimination, research that focuses on reducing discrimination is systematically lacking. The present study addresses this gap. Building on both cognitive and behavioural processes that steer discriminatory decision-making and insights from diversity literature, we develop and test the effectiveness of a training intervention which is tailored to the specific context of the housing market, with 113 students in real estate. We develop our training around three theoretical mechanisms of discrimination that are measured through a specified vignette experiment in pre- and post-tests. Our results show a reduction in statistical and customer taste-based discrimination after the training, but no change in agent taste-based discrimination. The driver for this reduction in discrimination is being more selective on majority candidates rather than increasing the chances of ethnic minority candidates. Additionally, we found that training has no effect on taste and that most of these effects remain present on the long term.
Disclosure statement
No potential competing interests were reported by the authors.
Notes
1 However, some discussion is at stake in the housing discrimination literature as to what extent customer taste based discrimination could be limited to taste based discrimination. Some scholars consider this a form of taste based discrimination (e.g. (Ahmed et al., Citation2010; Auspurg et al., Citation2017; Baldini & Federici, Citation2011), others interpret it as a form of statistical discrimination because it is more economic rational (e.g. Hanson & Hawley, Citation2014). However, the argument for customer statistical discrimination, made by Hanson and Hawley (Citation2014), relates to the agent’s past experiences to formulate the expected payoff for each potential home-seeker and selects a lessee by profit maximization. However, no lack of information is at stake here and a direct, discriminatory, personal request is being made by the client.
2 For more detailed information on the vignette experiment, we refer to the study by Ghekiere et al. (Citation2022).
3 Discrimination exists for both training and control group, which implies the important notion that these two groups are rather similar in their assessment of the rental candidates at Time 1, before the training.
4 We ran additional analysis where we compare the results at T1, T2 and T3 for the same 28 respondents but found no significant results. This can most likely be addressed to the small sample size of T3 respondents.
5 Additional analysis show that the tested group, after the training, favor high income candidates when a discriminatory question is asked. Moreover, a discriminatory request not only eases the way for discrimination based on ethnic origin but also based on income. This shows the strong intersectionality of ethnicity and financial strength