v0.4.3
Wiki/Statistical Procedures

One-Way ANOVA & Post-Hoc Tests

Between-groups and within-groups variance partition, Levene's test, Welch/Brown-Forsythe adjustments, and Tukey HSD.

Verified vs SciPy (≤ 10⁻⁶)IBM SPSS Algorithms pp. 741–748

1. Sum of Squares Partition#

One-Way ANOVA decomposes total variance into between-groups and within-groups (error) sum of squares:

SStotal=SSbetween+SSwithinSS_{total} = SS_{between} + SS_{within}

The test statistic is evaluated as:

F=MSbetweenMSwithin=SSbetween/(k−1)SSwithin/(N−k)F = \frac{MS_{between}}{MS_{within}} = \frac{SS_{between} / (k - 1)}{SS_{within} / (N - k)}

2. Robust Tests of Equality of Means#

When homogeneity of variance is rejected by Levene's test, StatCraft provides Welch's F∗F^* and Brown-Forsythe's FBFF_{BF} statistics with adjusted fractional degrees of freedom.

3. Multiple Comparisons (Post-Hoc)#

Pairwise group differences are evaluated via Tukey's Honestly Significant Difference (HSD) based on the Studentized range distribution qq, as well as Bonferroni and Scheffé adjustments.

Canonical SPSS Syntax Example

SPSS SYNTAX (.SPS)
1ONEWAY salary BY educ
2 /STATISTICS=DESCRIPTIVES HOMOGENEITY
3 /POSTHOC=TUKEY BONFERRONI ALPHA(0.05).