Reliability of Affective Bias Tasks
Author Information
Author(s): Pike Alexandra C., Tan Katrina H. T., Tromblee Hoda, Wing Michelle, Robinson Oliver J.
Primary Institution: University College London, UK
Hypothesis
The study aims to examine the test-retest reliability of two affective bias tasks.
Conclusion
The study concludes that measures from these affective bias tasks are moderately reliable, but further work is needed to improve their reliability.
Supporting Evidence
- Summary statistics showed generally moderate to good test-retest reliability.
- Parameters from computational modeling were often less reliable than summary statistics.
- Embedding the covariance matrix within the estimation procedure increased reliability estimates.
Takeaway
This study looked at how consistently people perform on tasks that measure their emotional biases over time, finding that the tasks are somewhat reliable but could be better.
Methodology
Participants completed two affective bias tasks twice, with at least 14 days between sessions, and their performance was analyzed using Pearson’s correlations and intra-class correlations.
Potential Biases
Participants who were inattentive were excluded, which may introduce bias as inattention is a symptom of anxiety and mood disorders.
Limitations
The study used only two tasks, which may not generalize to all affective bias measures, and the sample size may limit the precision of reliability estimates.
Participant Demographics
Participants had a mean age of 28, with 34 females, 23 males, and 1 nonbinary individual.
Statistical Information
P-Value
p<0.001
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website