Willett, Ciara Louise
(2023)
Why Correlation Doesn’t Imply Causation: Improving Undergraduates Understanding of Research Design.
Doctoral Dissertation, University of Pittsburgh.
(Unpublished)
This is the latest version of this item.
Abstract
Understanding when it is appropriate to make causal inferences from a statistical result is a fundamental skill for science literacy. Prior research has concentrated on erroneous causal judgments about observational studies, but there is little research on whether people understand that experiments provide stronger justification for causal claims. Our study tested the efficacy of an intervention at improving students’ ability to discriminate between correlation and causation. Students were taught how to use causal diagrams to illustrate possible explanations for a statistical relation in an experiment versus an observational study. To evaluate the intervention’s efficacy, intro psych (Experiments 1-3) and research methods (Experiment 1) students decided whether to make causal inferences about hypothetical observational studies and experiments. In Experiment 1, we tested multiple methods of instruction to see which worked best. Intro psych students learned more when they completed practice problems that involved generating self-explanations, whereas research methods students learned more from making analogical comparisons or reading worked examples. Critically, we found that students struggled with identifying the study design, which is the first step in correlation-causation discrimination. In Experiment 2, we added instructions to the Self Explanation intervention about how to identify observational studies versus experiments. Our modifications did not improve this skill nor students’ ability to discriminate between correlation and causation. The most successful intervention was in Experiment 3, which explicitly pointed out that people often make errors when evaluating evidence from observational studies and repeated the importance of considering study design when making causal judgments. A second goal of Experiment 3 was to test the influence of students’ expectations about the direction of the statistical relationship on their evaluation of the evidence. Students made more causal inferences about study outcomes that were in the same direction as their prior beliefs than outcomes they thought were implausible. After the intervention, students still used their prior beliefs to decide whether to make a causal judgment, but they also more strongly considered the study design in their evaluation of evidence. In general, our intervention improved students’ understanding of causality, but its efficacy may also depend on their prior knowledge.
Share
| Citation/Export: |
|
| Social Networking: |
|
Details
| Item Type: |
University of Pittsburgh ETD
|
| Status: |
Unpublished |
| Creators/Authors: |
|
| ETD Committee: |
|
| Date: |
26 January 2023 |
| Date Type: |
Publication |
| Defense Date: |
24 August 2022 |
| Approval Date: |
26 January 2023 |
| Submission Date: |
9 December 2022 |
| Access Restriction: |
No restriction; Release the ETD for access worldwide immediately. |
| Number of Pages: |
129 |
| Institution: |
University of Pittsburgh |
| Schools and Programs: |
Dietrich School of Arts and Sciences > Psychology |
| Degree: |
PhD - Doctor of Philosophy |
| Thesis Type: |
Doctoral Dissertation |
| Refereed: |
Yes |
| Uncontrolled Keywords: |
causal learning |
| Date Deposited: |
26 Jan 2023 15:21 |
| Last Modified: |
26 Jan 2023 15:21 |
| URI: |
http://d-scholarship.pitt.edu/id/eprint/44077 |
Available Versions of this Item
-
Why Correlation Doesn’t Imply Causation: Improving Undergraduates Understanding of Research Design. (deposited 26 Jan 2023 15:21)
[Currently Displayed]
Metrics
Monthly Views for the past 3 years
Plum Analytics
Actions (login required)
 |
View Item |
![[feed]](/images/feed-icon-32x32.png){kind=icon}