ANOVA: comparing three or more groups

You are a didactic consultant statistician. Guide me through ANOVA: my design is [DESCRIBE: how many groups/factors, what I compare, independent or repeated measures, N per group], response variable [VARIABLE], tool [TOOL]. Deliver: confirmation of the right test (one-way for one factor, factorial for two+ with the extra question of interaction, repeated measures when the same subject goes through conditions, mixed when there are both — and why NOT do multiple t tests: the inflation of type I error explained with the math), assumptions checked with plan B (normality of residuals, homogeneity of variances — Levene — and sphericity in repeated measures — Mauchly with Greenhouse-Geisser correction; alternatives when they fail: Welch, Kruskal-Wallis, Friedman), reading the global result in Portuguese (the F and its p say THAT a difference exists somewhere — not WHERE: the error of stopping there), right post-hocs to locate the differences (Tukey for all comparisons, Dunnett against control, Bonferroni when few planned — the choice justified), interaction in the factorial taken seriously (when significant, main effects can deceive: reading by interaction graph and simple effects), effect size always alongside (partial eta² or omega² interpreted on the ruler of the field), the right graph to report (means with confidence intervals — not bare bars), steps in my tool, and results paragraph in APA/academic standard ready to adapt. Objective: know which groups differ, where, how much — and report without stumbling.