I've been teaching a course on experimental philosophy to third-year undergraduates at Oxford Brookes University. The course is a mix of teaching the basics about statistical thinking and using statistical methods, and an introduction to experimental philosophy as an approach to answer philosophical queestions. Of all the courses I am teaching, this one has the most positive teaching evaluations, and also is the course that students put most effort in (as several students told me, they worked very hard on this course).
How the course is set up
I teach this in weekly sessions for 2.5 hours in a seminar room with computers.
The course alternates between discussions about recent papers in experimental philosophy and practical sessions. The discussion sessions are in a seminar format. We look at recent works in experimental philosophy, such as Markus Kneer and Edouard Machery's recent paper on moral luck. We also discuss papers on what experimental philosophy is, such as Regina Rini's and Antti Kauppinen's work. I start each session by briefly walking students through the experimental setup of these papers, and the broader philosophical implications (because, not everyone does the reading although my sense is students have at least skimmed these papers). We then discuss whether the papers under consideration successfully establish what they set out to do. I also let students individually, or in groups, think about alternative ways to test the same claim (e.g., folk views on moral luck, whether being a utilitarian makes you morally better).
The aim of these theoretical sessions is to give students a good sense of how experimental philosophy works and what its limitations are. Usually, the discussions are engaged and occasionally students come up with wonderful ideas.
The practical sessions I start with a one-hour introduction (I know this is long, but it really does take some time, at least for me) to explain a particular statistical technique. I look at the history of the technique, who invented it, what it has been used for. We go into quite some details about the mathematics - this really engages some students but it is fair to say some students find this boring. The majority seem at least mildly engaged.
I also take the opportunity here for students to obtain more statistical literacy, so we look at recent concrete cases, for example: why did polls predict that in the Referendum on the UK's membership of the EU in 2016 Remain would win by a few percentage points, whereas Leave won by about 2 percentage points. Such concrete situations help students think through such things as margin of error, social desirability effects, sampling size, and other factors. We also look at spurious correlations such as the number of people dying by being tangled in their bedsheets and the US pro capita consumption of cheese, and I ask students to think about how this is different from e.g., correlations between number of storks on your roof and number of children (there is indeed a correlation, and there is a common causal factor, namely rural living), or correlations between per-capita income and fertility across countries. This explanation of stats involves a lot of writing formulas and graphs on the white board.
The second part of the practical session for the remaining 1.5 hours the students do exercises on SPSS. Their main dataset is a replication of Knobe's side effect effect. Students have collected these data by surveying people early on in the course. They do t-tests, chi-square tests, etc. on the dataset that I collated from their individual responses (10 for each student, so for about 20 students this makes 200). I first let them practice on other datasets (from open access sources) and they make the exercise on their own dataset as a graded exercise that they submit (big plus for this course, the grading goes so quickly!). I let them do significance testing, having explained the possibilities and limitations it offers, and the epistemological problems and controversies surrounding it. We also do confidence intervals, testing for normality, and we make graphs (bar charts, box plots, scatter plots). We do not do tests for effect sizes, except when this is inherent in the test already (e.g., correlations). My main reason for this is lack of time. The course only runs 12 weeks, and the final week is not a teaching week.
Teaching materials
It is important to have a clear manual on statistics for this course. I use Sarah Boslaugh's Statistics in a Nutshell which is very clear, with lots of examples and thorough on theory, but also affordable. I provide it as background reading. Some students really seem to get into it, others get by on my notes and handouts. I provide handouts on each statistical test we do, with examples of when the test is appropriate and exercises for them to go through.
For the specifics on experimental philosophy, I use Justin Sytsma's and Jonathan Livengood's The Theory and Practice of Experimental Philosophy.
Next to this, we look at papers in experimental philosophy (this changes every year. So as to make prep not too onerous I keep the exercises the same but we change the discussions on papers in experimental philosophy).
The coursework
The coursework for this course consists of 4 graded exercises, which include data collection (not only inputting the data they collect in SPSS, but also a narrative on how they collected the data) an independent-sample t-test, chi-square test/Fisher's exact, and correlations. Students consistently get high grades for this, but unfortunately, each year I have one or two students who just don't seem to get it.
The majority of marks go to a large essay, a 3000-word piece where students replicate Shaun Nichols' study on disgust and Erasmus' norms. The original paper presented coders with a series of norms as outlined by the 16th century author Erasmus in his Good Manners for Boys, norms such as not wiping your nose on your sleeve after sneezing (still a norm today) and crossing yourself after sneezing (no longer a norm today). Shaun Nichols' coders judged that norms that prohibited actions that are elicit core disgust were more likely to be part of the norms today. They also judged violations of norms that prohibit disgusting actions as more serious norm violations. This was taken by Nichols to be evidence that affect, particularly disgust, plays a role in how norms are transmitted over time.
The quality of coders that my students recruit seems quite variable. In spite of offering my students guidelines on how to train their coders, it seems some coders hardly pay attention or struggle with Erasmus' difficult prose. To mitigate these variations, and also to facilitate my grading, I combine all my students' coder files into one single file by taking the mode (most common response) to each of the items. Students get information on how to structure an experimental philosophy essay (roughly, this follows conventions in psychology papers with sections on literature background, method, results, discussion, conclusion but with more attention for philosophy). They also see several examples of how the structure of an experimental philosophy paper works through the seminar discussions.
Overall, students work hard and do their best to write excellent replications. With permission of the student, I am here putting an example graph that one of my students (Sophie Rees) designed to show the difference between Nichols' original findings and our results last year.
In conclusion
Students enjoy this course, it does require quite a bit of prep work (given the seminar-style nature of the course), but it is easy to mark. Students particularly enjoyed doing coursework that was a bit different from what they are used to (note in the UK you start the major right away, in your first year), and they also hoped that learning the statistical techniques would improve their employability prospects.
Comments