For my experience, the biggest factor that students will care about is what days do they have exams in their other classes. If the students have two exams the same day, they won't be able to properly study for either of them. Whether you can arrange this or not depends on how similar their schedules are.
E.g. You said you are teaching an upper level social science elective. It may happen then that 80 - 90% of your students are also taking the exact same required social science class. If that class has an exam on a Tuesday, then you should not have your the same day.
Now, it may also happen that your students have completely different schedules. Some have exams on a Tuesday, others have on a Thursday. In that case it doesn't matter that much. Just pick one. Flip a coin if necessary.
To figure this out, you may just want to ask the students what day is best for them and see if there is a consensus or not.
Edit:
@Buffy's point about incentivizing cramming is interesting. I decided to search if there was any research on the effect of time between exams. I found this paper: Devin G. Pope, Ian Fillmore. "The impact of time between cognitive tasks on performance: Evidence from advanced placement exams"
Economics of Education Review, Volume 48, 2015, Pages 30-40, https://doi.org/10.1016/j.econedurev.2015.04.002
It is AP exams, not college midterm exams, but the conclusions should transfer I think. Here is the abstract:
Students are often required to perform several mental tasks in a short period of time, and their performance is likely to depend on how closely the tasks are scheduled. We examine this phenomenon in a particular context: Advanced Placement (AP) exams in the United States. We exploit variation in the AP exam schedule from year to year which results in a student who takes two exams in one year having more or less time between the exams than a student who takes the same two exams in a different year. We find evidence that more time between exams results in higher scores, particularly on the second exam, and that this effect varies across different types of students. Our estimates suggest that a student taking two exams ten days apart is 6–8% more likely to pass them both than a student taking the same exams only one day apart.
Here is a relevant quote from the conclusions section:
One could imagine various mechanisms for why more time between exams leads to better outcomes. For example, one possible explanation for our results is simple fatigue. Taking an AP exam is mentally and physically exhausting and it may be difficult to perform at peak ability when taking two exams in close succession. Another possible explanation is that last-minute preparation for exams (“cramming”) is important and more difficult when exams are close together. A third and related mechanism is that when exams are close together, students foresee their possible fatigue or lack of cramming time, and preemptively allocate their energies to just one exam. Our data do not allow us to identify a specific mechanism behind our findings, but may provide some clues. For example, we find that the detrimental effect of temporally proximal exams is primarily associated with the second exam taken. Our fatigue mechanism predicts this effect. The cramming mechanism may also predict this effect, but not so directly. For example, if two exams are close together and a student has to do last-minute cramming for both exams at the same time, this could arguably affect both the first and the second test score. Thus, this evidence is suggestive of fatigue, but cannot rule out a cramming effect
(emphasize mine).
So I would still stand by my suggestion of not having two exams on one day, to avoid mental fatigue of the students, even though it could potentially incentivize cramming to a certain degree.