More or Less: Behind the Stats Same data, opposite results. Can we trust research?
Nov 7, 2021
Professor Martin Schweinsberg, an Assistant Professor of Organizational Behaviour at ESMT Berlin, discusses the subjective nature of data analysis. He shares his experience of drawing different conclusions from the same data as his peers, raising questions about research trustworthiness. Schweinsberg unpacks the challenges of replicating studies, the complexities of crowd-sourced analysis, and the interpretations that lead to conflicting results. He emphasizes the need for humility in data interpretation, shedding light on the murky waters of research integrity.
AI Snips
Chapters
Transcript
Episode notes
Power Posing Study
- Amy Cuddy's power posing study and its subsequent failed replications highlight the issue of differing research outcomes.
- This sparked self-criticism in psychology, often focusing on data gathering biases rather than analytical approaches.
Data Discrepancy
- Martin Schweinsberg questioned his competence when his research conclusions differed from peers despite using the same data.
- This led him to design an experiment to investigate the impact of subjective assumptions on data analysis.
Subjectivity in Analysis
- Subjectivity plays a significant role in data analysis, even with identical datasets.
- Analysts' inherent biases and assumptions influence their choices of methods and interpretations, leading to varied conclusions.