calculate_power
Read-only
Power and sample size for a planned experiment: how many replicates a test needs to detect an effect, or what power, detectable effect or significance level a design that is already fixed gives. Leave ONE quantity out with solve_for and it is solved; everything else is given. Thirteen tests in two families: Cohen's standardised effect sizes (pwr 1.3-0: t one/two-sample/paired and unequal n, one and two proportions by arcsine h, normal mean, correlation, chi-square w, one-way ANOVA f, linear-model f²) and raw units (R's power.t.test from a difference and SD, power.prop.test from two proportions, power.anova.test from variances). Every number is R's, reproduced: the solved value is uniroot's root at R's tolerance, and an unreachable target is refused with R's own message. Returns the power curve around the answer and the R call that reproduces it.
Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| N | No | total observations across every cell. Only applies when test is chisq. | |
| k | No | number of groups. Only applies when test is anova. | |
| n | No | observations per group (pairs, for paired; the sample, for one-sample and correlation). Only applies when test is t or two_proportions or one_proportion or normal_mean or correlation or anova or t_raw or proportions_raw or anova_raw. | |
| u | No | numerator degrees of freedom: the number of predictors tested. Only applies when test is linear_model. | |
| v | No | denominator degrees of freedom: n − u − 1 for n observations. Only applies when test is linear_model. | |
| df | No | degrees of freedom of the chi-square test: (cells − 1) for goodness of fit, (rows − 1)(columns − 1) for independence. Only applies when test is chisq. | |
| n1 | No | observations in the first group. Only applies when test is t_unequal or two_proportions_unequal. | |
| n2 | No | observations in the second group. Only applies when test is t_unequal or two_proportions_unequal. | |
| p1 | No | the proportion in group 1. Only applies when test is proportions_raw. | |
| p2 | No | the proportion in group 2. Only applies when test is proportions_raw. | |
| sd | No | the standard deviation, in the same units (of the differences, for paired); R's default is 1. Only applies when test is t_raw. | |
| test | Yes | which test the experiment will be analysed with; the power answer belongs to that test and no other. Values: t (t test: one-sample, two-sample or paired, Cohen's d), t_unequal (two-sample t test, unequal group sizes, Cohen's d), two_proportions (two proportions, equal n, Cohen's h), two_proportions_unequal (two proportions, unequal n, Cohen's h), one_proportion (one proportion against a null value, Cohen's h), normal_mean (normal mean, known variance, effect d), correlation (Pearson correlation, r), chisq (chi-square goodness of fit or independence, Cohen's w), anova (balanced one-way ANOVA, Cohen's f), linear_model (linear model F test (multiple regression), Cohen's f²), t_raw (t test in raw units: difference and standard deviation), proportions_raw (two proportions in raw units: p1 and p2), anova_raw (balanced one-way ANOVA in raw units: between- and within-group variance). | |
| type | No | the t test's design. Only applies when test is t or t_raw. | two.sample |
| power | No | 1 − β, the chance of detecting the effect if it is real. No default: the target is yours to choose. | |
| _files | No | The project: record references shaped like _subject. Their records become library documents a tool names in sequences/parts/document; a trace or alignment is found by file name. | |
| effect | No | the standardised effect: Cohen's d for t and normal-mean tests, h for proportions, r for correlation, w for chi-square, f for ANOVA, f² for the linear model. Give a value or `effect_size` instead. Only applies when test is t or t_unequal or two_proportions or two_proportions_unequal or one_proportion or normal_mean or correlation or chisq or anova or linear_model. | |
| groups | No | number of groups. Only applies when test is anova_raw. | |
| strict | No | two-sided raw-unit tests: also count rejections in the wrong tail, as R's strict = TRUE. Only applies when test is t_raw or proportions_raw. Advanced. | |
| _subject | No | The record to act on: exactly one of sequence, accession (+source), sample (an id from list_samples) or file ({name, text} or {name, base64}); record_index picks one record of several. | |
| sig_level | No | α, the type I error rate; R's default. | |
| solve_for | No | the one quantity to solve. Left unset it is the test's sample size (n; n2 for unequal groups, N for chi-square, v for the linear model). `effect` solves the smallest detectable effect. | |
| difference | No | δ, the true difference in means, in the measurement's units. Only applies when test is t_raw. | |
| within_var | No | variance within groups. Only applies when test is anova_raw. | |
| alternative | No | two-sided, or the direction a one-sided test looks in (the raw-unit t and proportion tests take only two.sided or one.sided). Only applies when test is t or t_unequal or two_proportions or two_proportions_unequal or one_proportion or normal_mean or correlation or t_raw or proportions_raw. | two.sided |
| between_var | No | variance of the group means. Only applies when test is anova_raw. | |
| effect_size | No | Cohen's conventional small, medium or large effect for this test (pwr's cohen.ES), when there is no estimate from data. A convention, not a measurement, and every answer built on it says so. Only applies when test is t or t_unequal or two_proportions or two_proportions_unequal or one_proportion or normal_mean or correlation or chisq or anova or linear_model. |
Output Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| power | Yes | the power of the design | |
| rCall | Yes | the R call that reproduces the answer | |
| value | Yes | the solved quantity's value | |
| solved | Yes | which quantity was solved | |
| values | Yes | every quantity of the test, the solved one included, by R's name | |
| warnings | Yes | R's own warnings for this solve, verbatim | |
| powerCurve | Yes | power against the sample size, every other quantity held | |
| wholeUnits | Yes | the sample size rounded up to whole units and the power that whole number gives |