Syndicate content

Add new comment

Submitted by Slawa Rokicki on

I conducted a list experiment in Ghana with adolescent girls about sexual behavior as part of a larger RCT and it was, frankly, a complete failure. The results were a mess. I think part of this was due to the fact that it was self-administered and some girls were either confused or just not interested (some evidence of that as I made sure to put in innocuous items I knew for sure should be true for them and some answered 0). Another problem was power (I had 3 arms with 800 girls, but clustered into 34 schools). I think a bigger problem with research on sexual behavior of adolescents is that adolescent responses to these questions are inherently full of inconsistencies, recall error, and misunderstanding. When I conducted qualitative work after the RCT, that became very clear to me. It wasn't that girls were shy about providing answers about their sexual health; it was more that they misremember, they say completely inconsistent things minutes apart in the interview, and they interpret questions in different ways. This, to me, is the big problem in research on adolescent sexual health. Perhaps the list experiment can work in other contexts, but I wouldn't do it again with adolescents.