Abstract
The recent proliferation of online videos captured at the scene of news events begs the question: Do news videos have a meaningful impact on citizens’ political opinions and behaviors that is different than that of news articles? This was examined in an experiment using carefully matched videos and articles about the European refugee crisis. Findings show that articles generated stronger intentions to help refugees than videos, and this was mediated by the depth with which the news story was processed. Despite their increasing prominence and intuitively impactful qualities, news videos do not deliver more powerful effects than news articles.
Acknowledgments
This research was funded by The Amsterdam School of Communication Research and the Department of Communication, University of Vienna.
Notes
1 On average, each of the 6 article conditions contained 87 participants and each of the 6 video conditions contained 66 participants. The higher number of participants in the article conditions is due to these conditions being used in a separate study that is not related to the present experiment. To provide an indication of the power of the study we conducted a post-hoc power analysis using G*Power (Erdfelder, Faul, & Buchner, Citation1996). Taking the finding for the ANOVA testing RQ1 (N = 306): there was a 93% chance of detecting a small to medium effect size (equivalent to the observed ƞp2 = .03 translated to those defined by Cohen, 1992, as d = .2) as significant at the 5% level (two-tailed). Separately, experimental groups did not differ on age, gender, education, religious background, place of birth, nor on prior attitudes toward refugees, knowledge of the refugee crisis, general political interest, preferred political party, and political ideology (all p > .1). Note also that we chose not to include a control condition for the visual frame condition. This is because it is not clear what would comprise a neutral and balanced visual to serve as an adequate equivalent to the text control condition. Due to this, and to achieve a more parsimonious experimental design, we opted not to include a control visual condition.
2 Success of the frame manipulations was also confirmed in the final experiment using two questions: “to what extent did the video/article portray refugees as 1) victims, 2) intruders?” An independent samples t-test showed a significant difference between the victim and intruder visual frame conditions (Mvictim = 2.87, SDvictim = 2.63, Mintruder = 2.02, SDintruder = 2.82, p < .001). For the verbal frame conditions, one-way ANOVA (using post-hoc comparisons) with one-tailed significance testing showed a significant difference between the victim (M = 2.66, SD = 2.74) and intruder (M = 2.25, SD = 2.92) verbal frames (p = .036). There was no significant difference between the control (M = 2.44, SD = 2.60) and victim (p = .165) and the control and intruder (p = .199) verbal frames. However, as intended, the mean score for the control condition fell between that of the victim and intruder frames.
3 Although it would be possible to enter the emotion variables and the processing depth variable in parallel as mediators in these models, we chose to model them separately as this more accurately reflects our theoretical derivation of H2a and H2b.
Additional information
Notes on contributors
Thomas E. Powell
Thomas E. Powell (Ph.D., University of Amsterdam) is an assistant professor of Political Communication and Journalism at the University of Amsterdam.
Hajo G. Boomgaarden
Hajo G. Boomgaarden (Ph.D., University of Amsterdam) is professor of Empirical Social Science Methods with a Focus on Text Analysis at the Department of Communication at the University of Vienna.
Knut De Swert
Knut De Swert (Ph.D., University of Antwerp) is a member of the Political Communication and Journalism program group at the University of Amsterdam.
Claes H. de Vreese
Claes H. de Vreese (Ph.D., University of Amsterdam) is professor and chair of Political Communication and Journalism at the University of Amsterdam.