Review for Full-Context Videos for First-Time, Non-Literate PC Users

The authors developed a text-free job search system for non-literate domestic workers in Bangalore. Due to reticence by participants, they altered the system to include a video tutorial on system usage. Finally, they developed a "full-context video" (a five-minute sort-of propaganda commercial) that acted out how information was inputted into the system and how it could benefit users in addition to providing a tutorial. Participants who viewed the "full-context video" completed the tasks more quickly and with fewer mistakes than those who did not watch the video. Additionally, the effect was emphemeral in that after two hours, participants would have trouble using the same system without the video.

I loved this paper. It is an excellent story of how design evolves with user feedback. Their solution is simple, but surprisingly effective. Furthermore, I find it very interesting that they find the video effect is ephemeral for 61% of subjects.

I was also interested in reading the qualitative observations and descriptions of levels of engagement, for example, criticizing the movie couple for being too cheap with wages.

I think the authors should consider comparing their approach to informercials, commercials, or other propaganda. While these approaches are less research and more marketing research, I think it is a pertinent comparison--informercials "sell" a product, convincing users to buy something by showing its use in everday situations and motivating the audience to find value in the product.

One problem I have with the paper is that it seems to imply that the technological solution is superior to the informal social networks these individuals form (_The Social Life of Information_ argument). From the screenshots, I believe that the system does not convey alternative measures of endorsement, like ratings of employers or their history of paying on time. Some of this kind of information may not be pertinent to people who are desperate for money, but it seems like something people would naturally convey when discussing jobs.

A few small suggestions: the avg completion time for both tasks (4.59 min) in condition C is probably wrong, it seems like it should be 7.77 min in Table 1. Also, I think "Task completed" is actually referring to people not tasks, so maybe "Num Participants that Completed Task" would be a better label.