Replicability of (Socio)Linguistics studies

Abstract

In this paper I address four methodological issues of relevance to the design of (Socio)Linguistics studies. Issues addressed in this paper are: the importance of a proper power analysis during the design stages, or the probability of correctly rejecting the null hypothesis when it is false in the population, the analysis of the structure of the data to select the proper statistical analysis to be used in these studies, polychotomization, or categorization, of continuous variables and its effect on the power of the study, and the determination of sample size to achieve an adequate level of power. After a careful analysis of the review of literature on copula + adjective and the lack of reports on power and sample size calculations to support the generalizations derived from each study, I discuss the design and replicability implications of studies where these design issues have not been addressed.