Project info
Work package
- Synthesis
Sustainability threat
- Feedback Cycles
Challenge
- Shared responsibility and sustainable cooperation
Study info
Description of Study
We find that a highly influential meta-analysis by Schuch et al. (2016) uses trim and fill to correct for publication bias in the incorrect direction (i.e., towards larger effect sizes). This is in disagreement with known theory on publication bias and accepted practices in using this statistical technique. We use Schuch et al.'s raw data to create a re-analysis using correct settings, which greatly decreases the effect sizes observed by Schuch et al. We express our concern about statistical practices in meta-analyses.
Study research question
Collection provenance
- External data
Collection methods
- Experiment
- Other
Personal data
No
External Source
Source description
File formats
Data types
- Structured
Languages
Coverage start
Coverage end
Spatial coverage
Collection period start
—
Collection period end
—
Variables
Unit
Unit description
Sample size
Sampling method
Experimental Group
Effect sizes reported in studies on the effects of exercise interventions on clinical depression ('MDD') and depressive symptoms
25
—
Hypothesis
Theory
Variable type
Variable name
Variable description
Dependent variable
Standardized mean difference (SMD)
Standardized effect size outcomes (SMD) created from the difference between control group and intervention group, with higher values indicating better results in the intervention groups
Discipline-specific operationalizations
Conflict of interest
Data packages
Publications
Letter to the Editor: Comment on Schuch et al., “Exercise as a treatment for depression: A meta-analysis adjusting for publication bias”
Journal of Psychiatric Research
Documents
Filename
Description
Date
Ethics
Ethical assessment
No
Ethical committee