MarketingDistillery.com is about web analytics, data science and marketing strategy

3 responses to “Multiple A/B/n Tests in Marketing with ANOVA and R”

  1. Thomas Levi

    I’m unclear as to the values of f for “small” to “large”. The f-index or effect size is dependent on the distribution of means and the number of groups. One can find a lower bound for it based on the effect size, the (assumed) common variance and the number of variants or groups. Why not just use this value and either a middle or max value for the same to bound sample size estimates?

  2. Craig Marsden

    Particularly like the fact that you have included Barlett’s test for homoscedasticity in the explanation.

Leave a Reply

You must be logged in to post a comment.