Spring semester ended ten days into Ramadan, and a quarter of my evening advanced academic reading & writing class was fasting. This means that they took their finals after not eating or drinking anything for about 12 hours. Did this affect their scores? Should we care, since this is a nominally optional practice?
The short answer: it depends/possibly/needs further study, but we should definitely care.
This class makes a decent model to study this question for 3 reasons: a relatively high percentage of the class was observant Muslims (at least as far as fasting and refusing candy that contains gelatin goes), there are 6 big tests (3 reading and 3 writing) during the semester which are similiar in format to one another, and 2 of those tests (1 each of reading and writing) are in finals week – that is, during Ramadan. If fasting students were hurt academically, it should show up on those last 2 tests.
Without posting my students’ actual test scores here, I will just say that on the reading test, fasting students faced a 12% drop compared to the previous reading test, while the rest of the class had only a 2% drop. That is, the final reading test was harder for everyone, but much harder for fasting students. On the other hand, fasting students did 1% better on the final writing test than on the previous writing test while the rest of the class did 5% worse. All of this should be taken with a heaping tablespoon of salt, since the sample size for the while class was 22, producing a hilariously high t (significant at p<1).
So before any conclusions can be drawn, someone with access to administrative-level amounts of data should take a look at whether students who fast during Ramadan suffer disadvantages on tests or assignments offered during that time (the mid-May to mid-June, prime exam time at many schools).
And yes, we should care about this, first since a demographically specific disadvantage during a certain time of the year reduces the validity of the test. We don’t want to assess how well our students know a subject at 7:00 pm during Ramadan for the same reason we don’t want to assess the average temperature of an area at 3:00 am during January. Also, as I pointed out in my last end-of-the-semester statistics post, community colleges exist specifically to allow access to higher education to nontraditional student populations, including working people, older students, and immigrants (who are often older, working, and with family obligations as well). Obviously, this applies even moreso to ESL. If a quarter of the students our departments specifically exists to serve have a reliable disadvantage on certain tests because of factors that we can change, we should probably change those factors.