When Subjects Interpret the Data: Social Media as a Case for Adapting the Delphi Method to CSCW

Citation: Eric P.S. Baumer, Xiaotong Xu, Christine Chu, Shion Guha, Geri K. Gay (2017/02/25) When Subjects Interpret the Data: Social Media as a Case for Adapting the Delphi Method to CSCW. Proceedings of CSCW 2017 (RSS)
DOI (original publisher): 10.1145/2998181.2998182
Semantic Scholar (metadata): 10.1145/2998181.2998182
Sci-Hub (fulltext): 10.1145/2998181.2998182
Internet Archive Scholar (search for fulltext): When Subjects Interpret the Data: Social Media as a Case for Adapting the Delphi Method to CSCW
Tagged:

Summary

This paper describes the Delphi Method and demonstrates how it might be applied to research questions of interest to the Computer Supported & Co-operative Work research community. The Delphi method is a generalized framework for involving research participants in the interpretation of research data which summarizes their responses: that is, researchers ask participants both to answer questions and to offer interpretations of a summary report about how participants responded to the question more generally. The research team then compares their own analysis with the participant response.

To illustrate this method, the authors conducted a study investigating how people feel about their level of Facebook use, whether they have tried to quit, and whether they have adjusted privacy settings. Participants were also asked to predict the response of others to these questions. Informed by the Delphi Method, the authors then shared mean response values with participants, asked them if they were surprised by these results, and why. The authors followed up their survey with an interview protocol to test the validity of their findings with a new group of participants.

This study found that participants, on average, were accurate in their perceptions of people's ability to control their use of Facebook. However, participants were inaccurate when predicting whether or not people considered quitting, and in predicting other people's use of privacy settings. In those cases where participants' perceptions of others were not accurate, they were not surprised by the inaccuracy of their assessments. The authors concluded that when social media behaviors are visible, individuals tend to build their predictions of other people's behavior based on those observations. However, when social media behaviors are not visible, individuals tend to build their predictions of other people's behavior based on their own personal experience.

Theoretical and Practical Relevance

This article is a demonstration of the use of the Delphi Method for engaging participants with the process of interpreting data collected via surveys. This work also contributes to the understanding of how individuals build their perceptions of how social media is used.