This article supplements a review covering all (quasi)experimental studies of second language (L2) learning that appeared in Language Teaching Research (LTR) during the 19-year period 1997–2015. The dual focus of that review is practices in inferential statistics and the reporting of results. The present article sets out guidelines and recommendations reflecting findings reported in that review; and it includes additional discussion of topics of especial relevance to the analysis of data from L2 (quasi)experimental studies; for example, statistical power, effect sizes and confidence intervals for effect sizes, t-tests, the Wilcoxon–Mann–Whitney test, anova, contrast analysis in the context of anova, and control of the familywise Type I error rate. Discussed as well are two modern robust approaches to data analysis (i.e. randomization tests and bootstraps); ceiling and floor effects in pretest–posttest studies; and the ‘language-as-fixed-effect fallacy’.