Do Salient Social Norms Moderate Mortality Salience Effects? A (Challenging) Meta-Analysis of Terror Management Studies

Abstract

Terror management theory postulates that mortality salience (MS) increases the motivation to defend one’s cultural worldviews. How that motivation is expressed may depend on the social norm that is momentarily salient. Meta-analyses were conducted on studies that manipulated MS and social norm salience. Results based on 64 effect sizes for the hypothesized interaction between MS and norm salience revealed a small-to-medium effect of g = 0.34, 95% confidence interval [0.26, 0.41]. Bias-adjustment techniques suggested the presence of publication bias and/or the exploitation of researcher degrees of freedom and arrived at smaller effect size estimates for the hypothesized interaction, in several cases reducing the effect to nonsignificance (range gcorrected = −0.36 to 0.15). To increase confidence in the idea that MS and norm salience interact to influence behavior, preregistered, high-powered experiments using validated norm salience manipulations are necessary. Concomitantly, more specific theorizing is needed to identify reliable boundary conditions of the effect.

Not so terrifying after all? A set of failed replications of the mortality salience effects of Terror Management Theory

“…These attention- and engagement-check systems are, of course, not perfect, but have been found to detect most poor response quality in previous work [ 90 , 92 ]. Given established effects of researcher [ 30 , 35 ] and experimental style [e.g., 54 ], it may be that this specific method does not work in this context. Indeed, many, but not all, of the recent failures to replicate using this method have used online or crowd sourced samples [e.g., 47 , 49 , 50 ].…”

“…When using age as a moderator, the impacts of MS attenuated and even reversed (e.g., greater levels of egalitarian attitudes emerged) in samples of older adults [e.g., 28 , 29 ]. In a more recent meta-analysis, Schindler et al [ 30 ] identified a smaller, but significant overall pooled Hedge’s g of 0.34, or approximately r = .168. Schindler and colleagues focused on methods more than individual differences, the latter of which are discussed further below.…”

“…There have also long been failures to replicate some of the basic effects seen in TMT literature [e.g., Ochsman, personal communication, cited in 27 ], even with meticulous methods, faithful experimental conditions, and enviable sample sizes [ 31 , 32 , 47 – 50 ]. Although meta-analyses indicate a fairly robust effect of TMT on a host of dependent variables [ 19 ], Schindler and colleagues [ 30 ] found that the prototypical dependent variable—worldview defense—is not robustly found across studies when accounting for publication bias, control conditions, researcher effects (namely, researcher degrees of freedom), and other key components [ 30 ]. In other words, these “conceptual replications,” among others, were inconsistently successful, especially in more recent studies [ 30 , 50 ].…”

“…Although meta-analyses indicate a fairly robust effect of TMT on a host of dependent variables [ 19 ], Schindler and colleagues [ 30 ] found that the prototypical dependent variable—worldview defense—is not robustly found across studies when accounting for publication bias, control conditions, researcher effects (namely, researcher degrees of freedom), and other key components [ 30 ]. In other words, these “conceptual replications,” among others, were inconsistently successful, especially in more recent studies [ 30 , 50 ]. Others have accused previous researchers of cherry-picking data, presenting false-positives, and otherwise engaging in misleading or less-than-honest behavior (intentionally or otherwise) in studies of TMT [ 30 , 49 – 51 ].…”

“…In other words, these “conceptual replications,” among others, were inconsistently successful, especially in more recent studies [ 30 , 50 ]. Others have accused previous researchers of cherry-picking data, presenting false-positives, and otherwise engaging in misleading or less-than-honest behavior (intentionally or otherwise) in studies of TMT [ 30 , 49 – 51 ]. However, lapses in rigor and difficulty with replication are far from unique to TMT-related work [ 52 , 53 ].…”