Simulation Results for Two-Stage Mendelian Randomization on Count Data

Supplementary materials for a peer-reviewed journal article compare the performance of two-stage predictor substitution and two-stage residual inclusion methods in Mendelian randomization simulations. The tables contain performance metrics like bias, RMSE, and coverage for null and positive causal effects across varying instrument strengths and confounding levels. This dataset directly addresses a methodological gap in applying instrumental variable techniques to count-based exposures and outcomes common in biomedical research.

Use Cases

• Benchmarking two-stage MR method performance on simulated count data based on reported bias and RMSE metrics
• Assessing method validity under weak instrument conditions based on tables varying instrument strength (IS=0.001–1)
• Evaluating the impact of confounding on causal inference based on tables varying confounding effects (0.05–1.2)
• Comparing coverage and Type I error rates for null and positive causal effects

Strengths

• Peer-reviewed supplementary materials for a published journal article, indicating academic scrutiny
• Contains specific performance metrics (Bias, RMSE, 95% CI Width, Coverage) for methodological comparison
• Simulation parameters are explicitly detailed (instrument strength 0.001–1, confounding 0.05–1.2)

Limitations

• Column names and exact row counts are not provided, limiting immediate structural understanding
• Data is simulation output, not empirical observations, limiting direct application to real-world studies
• The dataset is a single 40,819-byte DOCX file, which suggests the tables are embedded in a document rather than a structured data file

Provenance

Source

Supplementary materials for the article 'Comparison of two-stage methods for count data in Mendelian randomization: a simulation study' in the Journal of Comparative Effectiveness Research

Collection Method

Simulation study output, methodology described in the associated peer-reviewed article

Time Range

Simulation study timeframe, not specified

Freshness

Last updated 2026-03-19

Geography

Not applicable for simulation data