New in StataNow 19.5: Power Analysis for Logistic Regression
With the May 21, 2025 update, we’re excited to introduce the first major highlight in StataNow 19.5: power analysis for logistic regression. This powerful feature helps researchers make informed decisions about study design and sample size—without wasting time or resources.
What’s New?
Whether you’re analyzing marketing conversion rates (purchased vs. not purchased), medical recovery outcomes (recovered vs. not recovered), or digital user behavior (subscribed vs. not subscribed), the new power logistic command helps determine the optimal sample size for logistic regression models.
This command allows you to:
-
Calculate sample size, power, or effect size for testing a coefficient in a logistic regression model.
-
Specify multiple parameter values and generate both tabular and graphical outputs.
For example, suppose your binary outcome variable Y is influenced by two binary covariates—X as the variable of interest and Z as a nuisance variable. To calculate the sample size needed for 80% power with a 5% significance level, you can simply type:
But that’s not all: the command also supports power analysis for more complex models, allowing up to 20 nuisance covariates drawn from 11 different distributions.
What Makes This Feature Special?
The method implemented in power logistic is recognized as one of the most accurate available, and Stata is one of the few statistical packages to offer it. It combines precision with Stata’s well-known ease of use and flexibility—making even complex model specifications straightforward.
Who Is This For?
This new feature is especially useful for users in medicine, epidemiology, public health, and ecology who frequently perform logistic regression analysis. They benefit from an accurate and user-friendly tool to conduct power analysis efficiently—whether for study planning or for analyzing complex relationships. And for anyone working with binary outcomes in other fields, this feature adds real value.
Learn more about the new feature at: stata.com/statanow/power-analysis-logistic-regression/