The Coding Personal Ads Project

by Ben Le

I still do this in my lab course (in fact, I'm about to run off to meet with my students regarding analyzing and writing up this project in a few minutes). In the version I had posted a couple of years ago, I laid out the coding strategy, provided ads photocopied from a local paper, and basically dictated what the students did for their project. This wasn't ideal because a) the coding strategy didn't yield any interesting sex differences in the ads I selected, and b) the students didn't have any ownership of the project and missed out on the important process of planning the project. So, I've revamped this project be simply leaving it open: we sit down during lab time and lay out several hypotheses based on student interests. They we (as a group) find a source for ads (they like this!). This year we used ads from http://www.datingfaces.com, which seemed to be a really nice source for ads because there were predefined categories to code, plus open-ended questions that could be coded using a strategy that the students come up with. Also, using the web allows for coding of pictures that are provided in the ad (we're finding some interesting effects regarding who provides pictures).

(all along in this process they are expected to be doing background reading that supports they hypotheses that they are interested in; I provided them with 2-3 papers of studies employing personal ads, and they are expected to do additional literature searches)

After generating a few hypotheses, coming up with a source for the ads, and developing a coding strategy, the students divide up into pairs, and each member of the team codes a number of ads (this year it was 50, which took them about 5-7 hours to do), such that all the ads are double coded. This is important (for me) because one of the goals of the project is to learn about interrater reliability. We then combine the data from each of the teams of coders, thus assembling a data set that includes (in our case this semester) ~300 ads that have each been double coded.

I then do a SPSS tutorial (for some students this is their first experience with SPSS) where students learn how to navigate the data set, get basic info (e.g., descriptives and frequencies), and learn how to run some basic tests (Chi-square and t-tests, and correlations for interrater reliability; more advanced students are allowed to do other more complicated stuff if they want).

Finally, they write up a lab report in the form of a MS in APA style. Again, this is the first experience with these tasks for some students, so I dedicate a lab period to talking about the structure of an empirical psychology paper and learning APA style (including an activity where they proofread and mark-up a MS that is full of formatting mistakes). They then write up a final paper that is typically about 10 pages of text (and this year I'm going to be introducing a "peer review" process where each person's paper is read by 2 other students).

The whole process including discussion of hypotheses, development of a coding strategy, coding of ads, analysis, the APA style activity, and writing takes about 6 weeks (my lab course meets 1.5 hours a week).