This is how we do research ethics at the SFM (update)

A while ago, I shared how we do research ethics at the Swiss Forum for Migration and Population Studies (SFM). Given that we often do commissioned research at the SFM, it’s important that the administrative burden is kept low, but we do and want to take ethical questions serious. Here I share the updated guidelines that I have put together for the institute, recently further streamlined. The aim remains to encourage all researchers to think about and take research ethics serious, and the guidelines are a synthesis of other ethics guidelines (duly acknowledged).

The guidelines begin with a short flowchart to deal with the most common cases. The list of exempt cases is now more explicit, with the understanding that if researchers identify ethical questions in seemingly benign approaches such as a literature review (e.g. because of the way the research question is posed, or because of the funder) can require a more thorough reflection (and thus the checklist to be filled in).

The core of the guidelines remains a checklist with 11 question. Each question — like “Does the research involve sensitive topics?” — comes with a few examples, and there are three possible responses: yes, uncertain, no. Researchers can tick the appropriate boxes, but it proved useful to use numbers for “yes” and “uncertain” answers to facilitate cross-referencing with part 2 of the guidelines.

Where some of the answers are “yes” or “uncertain”, researchers fill in part 2. Now more detail is required, including a brief description of the work, but normally the longest part is “Steps taken to address ethical issues”. Here the cross-references come in handy. If I identified 3 issues in part 1, I can now refer to them by number.

An uncontroversial list of ethical principles like “no harm to subjects and researcher” or “informed consent should normally be obtained” is included at the end of the checklist.

With the streamlined design, for some projects the ethical checklist takes only a short moment (e.g. literature review, analysis of secondary data where individuals are not identifiable). For other projects, we can typically handle the situation at the institutional level (e.g. interviews), while occasionally we want to have a thorough examination by the ethics commission of the university (e.g. field experiments).

Correspondence Test Shows Discrimination by School Administrators

A new correspondence study from Denmark shows discrimination by school administrators against parents with ‘Muslim’ names. They sent letters to schools across the country to ask whether they could move their son to that particular school (implying that they were not happy with the current school). 25% of fathers with a ‘Danish’ name (i.e. Peter Nielsen) received a positive answer, compared with 15% of fathers with a ‘Muslim’ name (i.e. Mohammad Osman).

In addition to holding everything constant by using men only (fathers enquiring about their sons), they had a variation in whether the son was a ‘diligent’ student. An interesting qualitative detail is that ‘Muslims’ are more often subjected to additional questions by e-mail (simple questions like verifying they actually live in the catchment area), while the ‘Danes’ were more often asked to call.

I find it interesting that their point of reference were studies on discrimination by public officials (typically politicians), but did not reflect methodological innovations from other correspondence tests, like stimulus sampling (!), or considerations of unmatched designs. I find it disappointing to find that the pre-registration at EGAP leads to a “page not found” error, especially since footnote 1 contains this interesting teaser: “We diverge from the preregistration to limit our focus only to the two variables that were subject to experimental manipulation and causal inference rather than those conditional on posttreatment responses.”

Continue reading “Correspondence Test Shows Discrimination by School Administrators”