Replication Projects

Kim Witte's Extended Parallel Process Model (EPPM) is one of the most influential theories in health communication. Cited over 5,000 times and taught in textbooks worldwide, it underlies major public health campaigns — yet its foundational 1994 empirical test was conducted with only 146 students, never included a 6-week follow-up in its replication record, and has not been tested for generalisability across cultures and populations.

This project conducts the first large-scale, preregistered, multilab replication of the EPPM, recruiting labs across five countries to test whether the core threat × efficacy interaction holds at scale, what the unbiased effect size is, and what moderates cross-lab variation. Stage 1 preregistration will be submitted to a journal offering Registered Replication Reports — guaranteeing publication regardless of outcome.

This methodological paper develops a taxonomy of replication types organized by replicator motivation — verification, study generalizability, theory generalizability, and extension. The core argument: direct replications in communication science require deviations from originals (updated stimuli, platforms, timeframes) to remain scientifically meaningful. The paper walks through seven study elements (sample, setting, procedure, design, measures, stimuli, analyses) and articulates rules for when deviations are updates (necessary for verification) versus deliberate deviations (associated with generalizability tests).

The negativity bias hypothesis in political communication contends that people are more aroused by negative than by positive news. Soroka et al. (2019) provided evidence for this bias in a study spanning 17 countries across six continents. In a conceptually close, well-powered, and preregistered replication, we find suggestive evidence for Soroka et al.'s central finding that negativity causes an increase in skin conductance levels.

We extend the original work in three ways. First, we theorise, test, and confirm that negative (vs. positive) news causes an increase in activity of the corrugator supercilii muscle above the eyebrow (measured via facial electromyography) and is associated with negative affect. Second, we find that people self-report negative news as causing negative affect, but that positive (rather than negative) news increases self-reported arousal. Third, we test Soroka et al.'s argument in a new context, the Netherlands. Overall, our findings suggest that negative news especially drives negative affect — contributing to the negativity bias debate in political communication while underscoring the importance of replication in empirical communication research.

Inoculation theory was introduced 60 years ago, after McGuire and Papageorgis (1961) published their first study on how resistance to persuasion can be induced. They demonstrated that people who are pre-exposed to weakened arguments against an attitude or position they currently hold (i.e., inoculated) are less affected by a subsequent strong counter-attitudinal message than people who are pre-exposed to arguments consistent with their attitude (i.e., supportive defense treatment) or to no arguments. Although these results significantly impacted both science and practice on a general level, rigid tests of the key theoretical propositions were lacking.

The authors conducted a highly powered replication study (N = 679) as a registered report and found that an inoculation treatment is more effective in increasing resistance toward persuasion compared to a supportive defense treatment and a no-treatment control condition. Their results were mostly consistent with McGuire and Papageorgis's original work.

There is much need to verify the robustness of published findings in the field of communication — particularly regarding the effects of persuasive emotional appeals about social issues. To this end, this project presents the results from a preregistered, direct replication of Skurka, Niederdeppe, Romero-Canyas, and Acup (2018). The original study found that a threat appeal about climate change can increase risk perception and activism intentions and that a humor appeal can also increase activism intentions with a large sample of young adults.

Using the same stimuli, measures, and experimental design with a similar sample, the authors failed to replicate these main effects. They did, however, replicate age as a moderator of humor's effect on perceived risk, such that the humor appeal only persuaded emerging adults (ages 18–21.9). The paper considers several explanations for the discrepant findings, including the challenges (and opportunities) that persuasion researchers must navigate when communicating about rapidly evolving social issues.