Simulation methods to estimate design power: an overview for applied research
2011

Using Simulation to Estimate Study Power

Commentary Evidence: moderate

Author Information

Author(s): Arnold Benjamin F, Hogan Daniel R, Colford John M Jr, Hubbard Alan E

Primary Institution: University of California, Berkeley

Hypothesis

Can computer simulation be used to estimate statistical power for complex study designs?

Conclusion

Simulation methods provide a flexible way to estimate statistical power for both standard and complex study designs.

Supporting Evidence

  • Simulation methods can reproduce conventional power estimates for simple randomized designs.
  • The approach is applicable to a broad range of study designs and outcomes.
  • Computer code for simulations is provided for both R and Stata.

Takeaway

This study shows that using computer simulations can help researchers figure out how many people they need for their studies, especially when the study design is complicated.

Methodology

The authors review simulation techniques for estimating study power in randomized trials and provide examples.

Potential Biases

The simulation approach may not account for non-random losses to follow-up or systematic measurement error.

Limitations

The results depend on the assumptions made about outcome variability and the data generating model.

Digital Object Identifier (DOI)

10.1186/1471-2288-11-94

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication