The first-person effect. A reconsideration of two meta-analyses

The third-person effect describes a tendency to estimate the influence of mass communication on others ("third persons") as being stronger than on oneself and this has been well documented in previous research. Though a first-person effect has also been postulated for desirable mass communication messages (for ex. non-profit advertisements or public service announcements (PSAs)), for which reporting more influenceability of the self as compared to others should be a means to self-enhance, it has not been found in the two named meta-analyses. One cause might have been ambiguities in the meaning of "impact" of desirable messages. For ex., whereas the content of the message might intend a desirable impact (for ex. a plead against violence), it can collide with low preferences for the respective message context (for ex. rap music) and thus a reported low impact of the message can result. We assume that this ambiguity with respect to the function of messages is considerably lower if only advertisements are considered because they have one main function: persuasion. We thus present reanalyses of data from two meta-analyses though restricted to studies on the impact of advertising. With data from the first meta-analysis, we not only replicate the well-known third-person-effect for undesirable messages (d = .83; 95%CI [0.72, 0.94]; k = 27), but we also find a first-person-effect for desirable messages (d = -.47; 95%CI [-0.70, -0.24]; k = 7). From the studies included in the second meta-analysis, we reanalyzed the results of studies with PSAs (all socially desirable messages). After the exclusion of some studies due to methodological problems, we find a first-person-effect for PSAs (d = -.16; 95%CI [-0.27, -0.04]; k = 33). Thus, contrary to the conclusions of both meta-analyses, we confirm the existence of a reliable first-person-effect. Replicability of meta-analytic results, the necessity to exclude studies due to methodological problems, and the meaning of "impact of socially desirable messages" are briefly discussed.