The second presentation at the ISA panel that I’m chairing will feature research by Fabien Miard on mobile phones as facilitators of political activism (see previous post for first presentation). Fabien will be sharing the findings from his recent MA thesis (PDF), which I have read with great interest.
Fabien’s research examines whether the number mobile phones affect political activity by drawing on a large-N quantitative study. This is an area in much need of empirical analysis since “little systematic research beyond loose collections of case studies has been done so far.” Furthermore, as I have noted in my own dissertation research, the vast majority of social science research on information and communication technologies (ICTs) is focused on the impact of the Internet exclusively.
The large-N study draws on the proprietary Cross-National Time-Series Data Archive (CNTS) for data on three forms of political activism: anti-government demonstrations, riots and major government crises. This dataset is derived from articles published in the New York Times (NYT). The data on the number of mobile phone subscribers is provided by the International Telecommunication Union (ITU). Control variables include: GDP per capita and population. The data he used spanned 191 countries between 1991-2006 but “a third of these dropped out due to missing values.”
Fabien uses negative binomial regression (with one year time lag) to test whether the number of mobile phone subscribers is a statistically significant predictor of political activism.
The results indicate that mobile density has no significant effect on anti-government demonstrations when the control variables are included. The same is true when using riots or major government crises as dependent variables. GDP per capita is small and insignificant except for riots, where it has a significant negative effect. Population has an effect on all three variants of political activism variables.
Fabien therefore concludes that mobile connectivity is neither negatively nor positively associated with political activism. This implies that existing case studies “are overrated and that generalization by means of a global comparative case study is not possible.” He suggests that future quantitative research take into account the following two recommendations:
- Compare the impact of mobile phones on democratic versus oppressive regimes;
- Analyze the combined impact of mobile phones and the Internet in addition to traditional technology variables;
These suggestions are spot on. One large-N quantitative study that I recently co-authored at the Berkman Center takes the first recommendation into account by comparing the impact of Internet and mobile phone users on measures of governance and democracy in both democratic and autocratic regimes (stay tuned for a blog post on this).
In my own dissertation research, I plan to compare the impact of Internet and mobile phone users on protests frequency in highly repressive versus midly repressive regimes. I also take into account Fabien’s second recommendation by adding Internet users and landlines. Furthermore, I include unemployment rate as a control variable which Fabien omits in his analysis.