Abstract
Hypothesis testing and Spanish psychological research: Analyses and proposals. The purpose of this paper was to analyse the application of the most common statistical procedure for studying relationships among variables and empirical phenomena in psychology: The null hypothesis statistical test. In order to determine whether its use in psychological Spanish research is adequate, we carried out a power study of the papers published in Spanish journals. The analysis of the 169 experiments selected, with a total of 5,480 statistical tests, showed power values of 0.18, 0.58, 0.83, and 0.59 to low, medium, high, and estimated effect sizes, respectively. These values drastically decreased in about a 20% when the calculations were repeated controlling the Type I error inflation through Bonferroni adjustment. The results were very similar to those obtained in other international power studies and lead us to think about the need for a special attention for controlling the statistical power in designing a research. On the other hand, we discuss several complementary proposals to the use of significance tests that may improve the information obtained.
