I believe that one of the most interesting findings of Terror Management Theory research is that death anxiety increases out-group aggression (and potentially support for extremism and terrorism). One of the most cited articles at this regard was written in 2006 by Pyszczynski, Abdollahi, Solomon, Greenberg, Cohen and Weise (basically the academics who created TMT). The study was published in the Personality and Social Psychology Bulletin: you can find it here. The article has been cited 278 times (Google Scholar, 6/05/2015).
In Study 1 40 undergraduates from 2 Iranian Universities were randomly assigned to Mortality Salience (MS) or control (dental pain), and then they were asked to indicate their opinions about questionnaires supposedly completed by fellow students one supporting and one opposing martyrdom attacks. The authors found that while students exposed to dental pain rated about 1.5 out of 9 their willingness to join pro-martyrdom causes, students exposed to death reminders rated their willingness to join pro-martyrdom causes on average 6 out of 9. Also, while students in the control condition evaluated on average about 2 out of 9 a person supporting martyrdom attacks, students in the MS condition evaluated the same person almost 7 out of 9 (see figure below).
The effect of the manipulation was impressive and led the authors to write “thoughts of death led young people in the Middle East who ordinarily preferred a person who took a pacifist stance to switch their allegiance to a person who advocated suicide bombings” (p.530).
This statement may be too bold as it is based on a single study. But what makes it even more problematic is the design and analyses.
The authors described the experiments procedures as it follows:
- “the design was a 2 × 2 mixed factorial.” (p.528)
- “After reading each questionnaire (presented in counterbalanced order), participants indicated their impressions of the student” (p.529)
The analyses were presented as it follows:
“A 2 (MS vs. control) × 2 (pro- vs. antimartyrdom) ANOVA yielded a significant main effect for MS, F(1, 38) = 19.86, p < 0001, and more important, a significant MS × Martyrdom Attitude interaction, F(1, 38) = 66.04, p < .0001. Pairwise comparisons revealed that although participants
preferred the student who opposed martyrdom attacks over the one who supported martyrdom attacks in the dental pain control condition, t(38) = 5.47, p < .0001, MS led to a dramatic reversal of this pattern such that after being reminded of their mortality participants preferred the student who supported martyrdom attacks over the one who opposed them, t(38) = 6.02, p < .0001.” (p.529)
However, I find the following point unclear:
- If the pro and anti-martyrdom conditions were presented to ALL participants, why do the authors talk about a 2 x 2 design and analyses? How many people per cell did they consider, 10 or 20? The 2 x 2 design would suggest 10, but the analysis 20. Does the counterbalanced order mean anything in the analyses and, if yes, why is it unclear? Shouldn’t be incumbent on authors and editors to be clear on details such as how many manipulations they used and how many people per cell they got?
- Can we really ground a statement like “thoughts of death led young people in the Middle East to switch their allegiance to a person who advocate suicide bombings” on an experiment with 10/20 people per cell?
- Who are the participants, what are their demographic, psychological, social and political characteristics? We only know that 14 were women and 26 men, mean age 22.46. Is this enough to describe the participants?
What do you think about it?