Related Experiment Videos
A new approach to exploring biases in educational assessment
I Dennis1, S E Newstead, D E Wright
1Department of Psychology, University of Plymouth, UK.
British Journal of Psychology (London, England : 1953)
|November 1, 1996
Summary
This study reveals supervisor marking biases in psychology dissertations. Supervisor-specific factors, potentially personal knowledge, influenced marks, particularly for male students.
Area of Science:
- Psychology
- Educational Assessment
Background:
- Marking biases can affect academic assessment fairness.
- Understanding these biases is crucial for equitable evaluation.
Purpose of the Study:
- To investigate marking biases in undergraduate psychology dissertations.
- To quantify the influence of supervisor-specific factors on marks.
Main Methods:
- Employed multi-sample confirmatory factor analysis.
- Analyzed marks from two independent markers for 255 dissertations.
- One marker was the student's supervisor.
Main Results:
- Approximately 30% of supervisor mark variance stemmed from supervisor-specific influences.
- These influences were orthogonal to joint assessment of dissertation merit.
- Supervisor bias disproportionately affected male students' marks, elevating them slightly.
Conclusions:
- Supervisor's personal knowledge may contaminate dissertation marks.
- Demonstrates the utility of multi-sample confirmatory factor analysis for bias detection.
- Findings highlight potential gender-based disparities in academic marking.