Estimating the Effect of Talking About Death on the Rejection of Holocaust Deniers’ Civil Liberties: A New Analysis of “Old” Data from the Edmonton Subsample of the 1992 All Alberta Survey

This paper reports the results of a reanalysis of data from the Edmonton subsample of the 1992 All Alberta Survey. The data were generated by a terror management theory (TMT) experiment embedded in the survey. The independent variable, a mortality salient (MS) induction, had respondents in the experimental condition reflect on their mortality prior to administering the questions used to measure the dependent variable. The dependent variable consisted of nine questions about support for the civil liberties of three targets — a revolutionary, a white supremacist, and a holocaust denier — who could express their views in three contexts — a public meeting, a high school classroom, and a book in a public library. The TMT hypothesis was that respondents in the MS condition are less like to support the civil liberties of all three targets in all three contexts. The initial results of the TMT test were disappointing. The only significant MS effect was decreased support for holocaust deniers. An interesting unexpected finding emerged, however. Respondents in the control condition were less likely to provide “valid” responses to the civil liberties items, and some predictors of this missingness appeared to be specific to respondents in the control condition. The addition of a dual processing model of consciousness to TMT and developments in the software for structural equation models (SEM) motivated my decision to reanalyze the data. Using the dual processing model, I reasoned that answering the first six civil liberties provided the distraction for the death thoughts of respondents in the MS condition to move from focal consciousness to non-conscious accessibility — the state required for worldview defense. To test this revised TMT hypothesis, I specified two latent change SEMs: one that assumed data missing at random (MAR) and one that allowed for non-ignorable missingness (MNAR) by conducting a two-class latent class analysis (LCA). The results for both models provided support for the revised TMT hypothesis. Respondents in the MS condition of the MAR model and respondents in the non-missing class of the MS condition of the MNAR model both exhibited an increase in the rejection of civil liberties for holocaust deniers. The results for the missing classes of the MNAR model were unexpected, however. I interpreted increased support for holocaust deniers in the MS condition as representing a worldview sympathetic to holocaust denial. I suggested that increased rejection of holocaust deniers’ civil liberties in the control condition as due to a combination of a predisposition to high death threat accessibility (DTA) of these respondents coupled with worldview threat posed by the administration of the civil liberties questions.