We run an experiment that studies belief updating biases in inference and extrapolation problems.
Intervention
We study biases in two kinds of belief updating problems: inference and extrapolation.
Design details
Randomization method
Randomization done by a computer
Randomization unit
Individual
Clusters
No cluster
Observations
660 subjects
Arms
200 subjects in the baseline treatment, 140 subjects in a treatment with binary signals, 100 subjects in a treatment where subjects answer the inference question and extrapolation question sequentially, 100 subjects in treatments with two performance indices
To explore the role of the conceptual distance between the updating question and the information provided, we will run two more treatments. One treatment (60 subjects) make explicit the equivalence between high quality and good performance. The other treatment (60 subjects) uses similar words to frame the firm quality and the firm performance.
Primary outcomes
Subjects' prior and posterior beliefs
Cited by
No citation data available for this registration yet.