Resenting moral rebels

ANOVA
We usually applaud people who rebel against the status quo for moral reasons – but those involved in the status quo do not. An experiment attempts to understand why this is.
Author

Alex Reinhart

Published

February 24, 2017

Data files
Data year

2015

Motivation

In 2008, three psychologists published a paper in the Journal of Personality and Social Psychology about “moral rebels”: people who, despite being under pressure to conform, instead rebel to do what they think is morally right. Consider Frank Serpico, for example, a New York Police Department officer in the 1960s and early 70s who, aware many fellow officers were corrupt, finally spoke out and brought evidence to the New York Times, leading to reform; or Oskar Schindler, a Nazi Party member who nonetheless worked to save the lives of 1,200 Jewish workers during the Holocaust.

The psychologists observed that moral rebels, though widely admired by outsiders, are often not well-appreciated by their peers who were involved in the activities they rebelled against. (Frank Serpico, for example, was shot, possibly in a set-up by other police officers.) They hypothesized that peers of moral rebels feel threatened: accepting the rebel would mean admitting their own behavior was wrong, and it is easier to condemn the rebel than to admit you have done wrong.

The psychologists performed several experiments to test their hypothesis. We’re interested in experiment 4, which tested the hypothesis that self-affirmation—performing some activity to make you feel good about yourself—would make it easier to accept the rebel. Their experiment found evidence for this hypothesis.

Here’s the experiment. Participants (undergraduate students) were assigned to one of three experimental conditions: obedient, rebel control, or rebel self-affirmation. They were then asked to imagine that a burglary had just occurred, and they were then given three photographs of potential suspects, along with descriptions of each suspect. The participants, on the basis of the photographs and descriptions, were supposed to choose which suspect was most likely the burglar, and fill out a form indicating why they made this choice.

Two of the photographs depicted white men. The third was “Steven Jones”, an African American man, whose description indicated he had a previous criminal record and no alibi—the descriptions were designed to make him the most likely suspect.

After making the choice and filling out the form, participants were shown a form purportedly from someone else who participated in the study (though it was actually prepared by the experimenters). Those in the “obedient” condition were shown a form picking Steven Jones as the burglar. Those in both “rebel” conditions were shown a form by someone stating “I refuse to make a choice here” on the grounds that it’s “offensive to make a black man the obvious suspect”. Participants were then asked a series of questions about the other participant whose form they viewed.

The crucial difference was between the “rebel control” and the “rebel self-affirmation” groups. Participants in the “rebel self-affirmation”, between filling in their own form and viewing the fake form, were asked to write a short essay about “a recent experience in which you demonstrated a quality of value that is very important to you and which made you feel good about yourself.” Participants in the other two groups, as a placebo, were just asked to describe what they had eaten in the past two days.

This dataset presents the results of a replication of this experiment, conducted as part of the Reproducibility Project: Psychology, a massive project to redo experiments from 100 different papers from psychological journals, to determine if their results hold up when tested again. Our goal is to see if the self-affirmation task made participants feel better about the rebel than those in the control group.

Data

The data file contains results from 75 participants: 20 in the obedient condition, 28 in rebel control, and 27 in rebel self-affirmation. (Assignment to conditions was done randomly.) The replication, instead of using undergraduate students, used participants from Amazon’s Mechanical Turk service, who completed the task entirely online.

Data preview

moral-rebels.csv

Variable descriptions

Variable Description
Condition Which group the participant was in (Obedient, Rebel Affirmed, or Rebel Control)
Gender Participant’s self-reported sex (1 = male, 2 = female)
Ethnicity Participant’s self-reported ethnicity (1 = White, 2 = Asian, 3 = Hispanic, 4 = Native American, 5 = Hawaiian/Pacific Islander, 6 = African American)
YOB Participant’s year of birth
Other.Work “How much would you like to work on a project with the other participant?”
Other.Friend “How much would you like the other participant as a friend?”
Other.Respect “How much do you respect the other participant?”
AboutGuess “What do you think this study was about?”
FoodList For those in the “obedient” and “rebel control” groups, the foods they ate over the past two days
ExplainChoice Explanation of their choice of suspect in the task
ValueDescription For those in the “rebel self-affirmation group”, the story they wrote about the value they demonstrated
ValueName For those in the “rebel self-affirmation” group, the value they demonstrated
ValueImportance How important that value is to them
Choice.happy How happy participants were with their choice (after seeing the other participant’s choice)
Self.Work Along with Self.Friend and Self.Respect, their answer to how they felt the other participant would feel about working with them, being their friend, or respecting them

There are a range of columns like Intelligent, Strong, Moral and so on, which reflect the participant’s ratings of the other (fake) participant. Those labeled Self.Moral, Self.happy and so on are their ratings of themselves.

The original psychologists combined Other.Work, Other.Friend, and Other.Respect into a combined score by averaging them.

Questions

  1. Is there a difference in how participants viewed the other participant (based on the Other.Work, Other.Friend, and Other.Respect scores), depending on which Condition they were in? Does the data support the psychologists’ original hypothesis?

  2. What about participants impressions of how moral the other participant was? What does this mean for the research hypothesis?

  3. Repeat your analysis from question 1, but using Self.Work, Self.Friend, and Self.Respect. Interpret and explain your results.

  4. Think about the experimental design. Does it adequately test the psychological hypotheses? Are there any threats to its validity or generalizability?

References

B Monin, PJ Sawyer, MJ Marquez (2008). The rejection of moral rebels: resenting those who do the right thing. Journal of Personality and Social Psychology, 95.1: 76-93. https://doi.org/10.1037/0022-3514.95.1.76

Replicated as part of: Open Science Collaboration (2015). Estimating the reproducibility of psychological science. Science, 349 (6251), aac4716. https://doi.org/10.1126/science.aac4716

Replication data available from https://osf.io/pz0my/.