PowerLMM.js (alpha)

Sample size planning for longitudinal models with missing data

Power analysis for longitudinal data with dropout

This app calculates statistical power for a linear mixed-effects model, comparing the rate of change between two groups over time. It supports standard null hypothesis testing, as well as non-inferiority and equivalence tests. The model accommodates random slopes and accounts for participant dropout. Adjust the parameters using the inputs on the left.

Power Curve

Power for varying sample sizes. The dashed line represents power without any dropout. You can adjust the maximum value and step size using the settings button.

05101520253035404550556065707580859095100110120130140150160170180195Sample size (per group)0%25%50%75%100%Power80 % Power

Power (complete data)

Power (with dropout)

Explore the Model

Use the buttons bellow to explore model implied parameters.

Proportion Remaining

This plot shows the proportion of participants remaining at each time point. The dropout process is monotonic, with no intermittent missing data.

012345678910Time0.000.250.500.751.00Proportion Remaining

Power Summary (WIP)

About

Created by Kristoffer Magnusson. This app is currently under active development and is not yet ready for use in planning real studies. Please contact me with any ideas or suggestions.