Download data and study materials from OSF
Principal investigator:
Daniel Bergan
Michigan State University
Email: bergan@msu.edu
Homepage: http://cas.msu.edu/people/faculty-staff/staff-listing/name/daniel-bergan/
Sample size: 1206
Field period: 09/19/2011-03/06/2012
The Flexible Correction Model holds that when made aware of potential sources of bias, people use naive theories to correct for that bias. We tested whether people instructed to correct for the influence of party labels attempt to correct for those biases, and if these attempts at correction are moderated by theories about the influence of party labels. Subjects were exposed to a short reading about a proposed health reform in which party labels attached to the reform were randomly assigned. Subjects were also randomly assigned to bias correction instructions or no instructions.
We test whether a) people instructed to correct for the influence of party labels are less likely to be influenced by party labels and b) whether instructions to correct for bias interact with attitudes about the influence of party labels.
The experiment consists of 6 conditions: 3 (party label: Democrats, Republicans, or none) X 2 ( bias correction instructions, no instructions). Subjects will be first exposed to a news article about a novel policy proposal claimed to be cut and pasted from The New York Times’ online edition. Three different party labels will be attached to this policy proposal (Democratic, Republican, or “policymaker” in Congress (i.e. the last condition has no party label)). Then half of the participants will be exposed to the instruction “Please try not to let irrelevant factors influence your response or bias your judgments” before answering questions, but the other half people don't receive this instruction. Respondents will then answer questions about their attitude toward the novel policy, their perceived bias of party label, etc.
An attitude scale created from 4 Likert-type questions about support for a health reform measure.
The hypotheses were not supported by the data.