Text Messaging Shows Promise as a Survey Tool

October 05, 2009

Eszter Hargittai explains how cellphone text messages can be used for social-science research in her new book.

Text messages can be a powerful survey tool even if they are typed on cellphones and are short and grammatically iffy, says Ms. Hargittai, an associate professor of communications studies at Northwestern University, in Research Confidential: Solutions to Problems Most Social Scientists Pretend They Never Have (University of Michigan Press), which she edited.

She and a graduate student asked 60 students at the University of Illinois at Chicago to text them every hour with information on what they were doing, who they were with, and what sorts of digital media they were using.

Q. Do you think the respondents would have been more likely to answer the questions if they had been in a room with a pencil and a piece of paper to fill out?

A. If someone's sitting in a room, they've already come in to do it. I think a better comparison might be if you e-mailed people a survey, will they ever take it? We were quite impressed by how many responses we got throughout the day. The other thing that was a really pleasant surprise was just how much information they managed to fit into 140 or 150 characters.

Q. Did you ever come across any confusing abbreviations in the responses?

A. It was totally fine for them to give us short, abbreviated texts. In fact, I don't think they could have told us all that they did if they hadn't done that in some cases.

Q. Do you think people who text and people who don't are also divided along racial, economic, or other kinds of social lines? Should researchers consider that when deciding whether to use texting to collect data?

A. It's unlikely you'd find that kind of difference [among college students] because everyone's doing it.

Q. What were some problems you ran into while using texts to collect data?

A. In the testing phase, we had all sorts of issues. We didn't anticipate that sometimes people have their phones off. If they've been off for five hours and we were texting hourly, then what? We wanted to anticipate these issues so that we could tell them ahead of time, if you're off the grid, and then you get three [text messages], this is how you deal with it. As much as you can do a lot with text messaging for collecting this type of data, we definitely needed other means of communication to set up this study. Initially, we had sent the respondents a longer e-mail that explained things. This was really necessary.

Q. Do you think text messaging could be used in a broad range of research?

A. I think it could be used in all sorts of research, absolutely. I think it has a lot of potential certainly for projects that have this time-diary data. It's certainly the kind of method where you could have very concrete messages that you're sending and you just want yes-no answers back.