Power analysis for balanced, between-subjects designs using the rules from Rainey (2026) “Power Rules: Practical Advice for Computing Power (and Automating with Pilot Data).” The preprint is on OSF.
Quick Start
The main function is power_rules(), which interactively walks you through the logic presented in the paper, step by step. At the end, it gives you a one-line from_*() |> find_*() pipeline that mirrors the conceptual logic from the paper. You can save this pipeline as a record of your inputs and outputs.
# install from github
pak::pak("carlislerainey/powerrules")
# run the interactive guide
powerrules::power_rules()Overview
Rainey (2026) argues for the value of simple rules of thumb that let you predict statistical power with paper, a pencil, and a pocket calculator. Predicting power involves guesswork, and paper-and-pencil calculations better match the spirit of the task than specialized software that brings an air of false exactness. Pencil-and-paper work also constantly reminds you how inputs relate to outputs, which matters when you are designing an experiment and have freedom to make different choices.
powerrules automates these rules while preserving their transparency. Every computation shows its work, so you can see exactly how inputs relate to outputs. The package tries to maintain the feel of paper-and-pencil.
Using the Package
power_rules() is the primary entry point. It asks about your design, collects the numbers it needs, shows the computation with rule citations, and gives you a copyable from_*() |> find_*() pipeline. The pipeline is the permanent artifact you save to your script.
shiny_rules() is a browser-based alternative. It performs the same computation, presented as a fill-in-the-blank board, and also outputs the pipeline.
Both guides produce a one-line pipeline in the form from_*() |> find_*(). Here is an example:
library(powerrules)
from_sd(sd_y = 20.8) |>
find_mde(n = 500)## -- Power Analysis ------------------------------------------------------
## Design: balanced, between-subjects
## Source: reference population SD
## CI level: 90% (size-0.05 test of directional hypothesis)
##
## Inputs:
## SD(Y) = 20.8
## n = 500 per condition (1,000 total)
##
## Predicted SE = 2 * 20.8 / sqrt(2 * 500) = 1.32 [Rule 3]
## MDE (80% power) = 2.49 * 1.32 = 3.27 [Rule 5]
## MDE (95% power) = 3.29 * 1.32 = 4.33 [Rule 5]
##
## -- Manuscript sentence (edit as needed) --------------------------------
## For a balanced, between-subjects design with 500 respondents per
## condition (1,000 total), assuming a standard deviation of 20.8, the
## predicted standard error is 1.32. Using a one-sided test at the 0.05
## level, the experiment has 80% power to detect a treatment effect of
## 3.27 units and 95% power to detect a treatment effect of 4.33 units.
##
## Note: The paper rounds the MDE factor to 2.5 for 80% power and 3.3 for
## 95% power. This software uses exact values (2.49 and 3.29), so results
## differ slightly from hand calculations using the rounded factors.The output shows the inputs, step-by-step formulas with rule references, the MDE at 80% and 95% power, and a manuscript-ready sentence you can copy into a pre-registration or methods section.
Vignettes
Each find-from combination has a dedicated vignette with a fully worked example.
| Find the MDE | Find the power | Find the sample size | |
|---|---|---|---|
| From a reference SD | vignette | vignette | vignette |
| From an existing study | vignette | vignette | vignette |
| From pilot data | vignette | vignette | vignette |
Additional vignettes:
- Interactions — 2x2 factorial designs
Verify Your Results
powerrules can generate standalone code to verify any calculation three ways:
-
write_manual()generates base R code using the paper’s formulas. [vignette] -
write_DeclareDesign()generates a DeclareDesign simulation. [vignette] -
write_loop()generates a for-loop simulation. [vignette]
Each function takes a pipeline result and prints copy-pasteable code that reproduces the from_*() |> find_*() result with an alternative approach.
Citation
Rainey, Carlisle (2026). “Power Rules: Practical Advice for Computing Power (and Automating with Pilot Data).” DOI: 10.31219/osf.io/5am9q_v3.
BibTeX: