A Q&A with a UCIPT Postdoctoral Researcher
What are the most common topics of discussion for the Freshman 200 study participants? How closely do Twitter Trends and/or top headlines on Yahoo News align with Freshman 200 study participant discussion topics?
Not surprisingly, one of the most common topics of discussion among the Freshman 200 study participants was their academic studies. Overall, 12% to 15% of tweets were related to academics. The percentage of academic tweets increased to about 20% during exam periods. Other common topics included tweets related to students’ personal health (~10% of all tweets) and social life (~6% of all tweets). There were also tweets that about current events and news headlines. I think it will be interesting to study students’ attitudes toward Twitter Trends and/or top headlines in the near future.
How were the study participants recruited? How many of the Freshman 200 are followers of each other on Twitter?
The study participants were recruited on campus via flyers and on social media pages. The recruitment took place during the months of September and October 2015. We are not exactly sure the number of participants who follow each other on Twitter, but we do have plans to conduct a social network analysis. In our preliminary analysis, we already found that the sentiment of tweets differs depending on the audience (e.g., friends, celebrities, strangers). We found that the sentiment was more positive when individuals tweet at celebrities or strangers compared with close friends. We will post the results once we have completed our analysis.
Has the fact that UCIPT is monitoring students’ tweeting affected their posting behaviors? Have you a noticed self-consciousness on the part of participants? Are there studies on this?
With this type of study, there is going to be a Hawthorne effect. This is a type of reactivity in which individuals modify an aspect of their behavior in response to their awareness of being observed. We acknowledge this as a limitation of our study. Nevertheless, we have tried to minimize the Hawthorne effect by 1) avoiding student contact via Twitter during the study period, and 2) ensuring that students’ data are kept confidential. We plan to validate our results in a larger sample and increase the observation time down the road. This will also help minimize the Hawthorne effect.
How do you communicate with participants during the study? Do you communicate via Twitter or email?
We communicate with the participants via email or telephone during the study. This was because we wanted to minimize interference with the participants’ Twitter behaviors. Most of our communication only involved reminding the participants to complete their weekly surveys.
When study participants communicate with third parties, do you disregard or use third-party communications as part of your algorithms? For example, a son arguing with their son or a boyfriend fighting with a girlfriend might happen publicly over Twitter. Do you only use your participant’s tweets or the third party’s tweets as well?
We did not include third-party communication in our analysis. I think those types of exchanges could be useful in our prediction models, but it is also important for us to work closely with our institutional review board to ensure that we are not breaching user privacy. Currently, we are only using participants’ tweets in our analysis. We are working with our institutional review board to come up with a plan so that we can analyze all publicly available tweets. It is important to keep in mind that Twitter is different from other social media platforms (e.g., Facebook, Snapchat, Instagram) in that its contents are publicly available. We hope to integrate third-party communications as part of our prediction algorithms in the near future.