2020 年 38 巻 2 号 p. 243-249
Psychologists have analyzed response time data by their rules of thumb. Modern advances of statistical methods promote to create a new practice. Traditionally, outliers were discarded prior to statistical test and skewed data were converted by logarithms. Fitting approach reminds that analyzers intend to estimate parameters for RT distributions. In statistical tests, psychologists often conducted ANOVA by aggregating data across different trials in the same condition. This practice loss precision information of measurement. Linear mixed models is changing the situation. While practical issues are remain for several aspects of applying linear mixed model, consensus among psychologists would be increasingly required.