I am running a crowd-sourced experiment where two participants have to engage in a text-based conversation (about movies or TV shows). My goal is to understand linguistic dynamics of the conversation, e.g. the sentiment, the similarity to other utterances, or the pragmatic function of the utterances. I really want to understand motivations and reasoning, not just a list of movies.
There are two prompts, and they have 8 minutes for each. The problem is that a handful of the participants in my pilot study barely said anything. In the entire 16 minutes, they typed 10-15 utterances.
I’ve reproduced the first prompt below, but I’m looking for suggestions as to what I can modify or add to encourage more actual conversation. I don’t want to use a disincentive such as “If you produce less than 30 lines of dialogue you won’t receive full compensation…” I’ve been thinking about making the prompts more specific, or adding in the instructions that they should have a lively conversation. I just don’t want to artificially constrain my experiment too much.
Any thoughts at all?
- [Participant 1], first get to know [Participant 2]'s tastes. What kinds of movies or TV shows do they like and dislike?
- Do not hesitate to express strong opinions, about genres, actors, etc. you especially like or don’t like. Try your best to not provide any “spoilers.”
- You will have 8 minutes to discuss the prompt below, and a warning will appear when there is 1 minute left.
[Participant 2] has had a long week at work, and would like to relax and watch a movie or TV show to unwind. [Participant 1], what movies or TV shows would you recommend and why?