The Potential of Internet Mobilization: An Experimental Study on the Effect of Internet and Face-to-Face Mobilization Efforts

Abstract

In recent years, voluntary associations and political organizations have increasingly relied on Internet-based mobilization campaigns, replacing traditional forms of face-to-face recruitment and mobilization. Within the literature, one can observe an intensive debate about the possible consequences of this transition. Most importantly, the question is whether political mobilization through the Internet is just as effective as mobilization in a face-to-face setting. In this article, we report on a mobilization experiment using both traditional (face-to-face) and modern (Web site) incentives for mobilization. The experiment was conducted among undergraduate students in Belgium and Canada and included a test of medium-term mobilization effects. Results suggest that the Internet is successful in transferring knowledge and raising issue salience among respondents, but neither experimental manipulation led to significant behavioral changes. We do not find any indication that among this experimental sample Internet-based mobilization would be less effective than traditional face-to-face forms of mobilization.

Keywords:

• mobilization
• political participation
• Internet
• global warming

Notes

1. Tests on the main variables were performed in order to investigate the successful randomization of respondents. These tests did not lead to any anomalies, so we can assume that randomization was successful.

2. Because of several logistical difficulties, a limited number of respondents took part in the experiments in smaller groups. Eliminating these participants from the analysis, however, did not lead to different results.

3. The Web sites were set online only during the experimental activities, with the researchers having already posted various forum entries under fictitious names.

4. The logistic regression was able to explain 6% of the observed attrition in Canada and 3% in Belgium. Dependent variables in the analyses were participation in the Web condition, participation in the face-to-face condition, knowledge in Wave 1, salience in Wave 1, and behavior in Wave 1. Not a single one of these variables was significant at the .05 level, showing that initial measurements and the experimental condition did not have an effect on dropout between Wave 2 and Wave 3. There also were no significant gender differences between Wave 2 and Wave 3.

5. All of the analyses were also conducted with this method (comparing means), and these confirmed the results of the regression analysis (results available from the authors).

6. An alternative method is to examine the knowledge gains of the respondents (the score in Wave 2 minus the score in Wave 1). A hierarchical regression with these knowledge gain scores as a dependent variable reveals the same results as the reported analysis (r ² = .25). Compared to the control group, the Internet (β = 0.64, p < .001) and the face-to-face condition (β = 0.44, p < .001) contributed significantly to the prediction of the knowledge gain scores. Participants learned more from the Web-based mobilization than from face-to-face mobilization (β = −0.17, p < .01).

7. A 3 (experimental conditions) × 2 (country) between-subjects analysis of variance indicated a significant interaction effect between experimental condition and country. For Canadian students, face-to-face stimuli are as effective in stimulating knowledge as Web stimuli. For the Belgian students, Web stimuli are more effective in enhancing environmental knowledge compared to face-to-face stimuli. We also tested for an interaction effect between gender and experimental conditions, but this was found to be statistically insignificant.

8. A hierarchical regression on the donation gain scores for environmental NGOs (R ² = .053) revealed similar results, except for the face-to-face condition. Regression on the gain scores controlling for country indicated a significant main effect of both experimental conditions: Web, β = 0.29, p < .001, and face-to-face, β = 0.18, p < .05.