2022 Volume 62 Issue 1 Pages 69-79
Regression analysis is a statistical method that attempts to quantitatively explain the variation in a dependent variable using the variations in independent variables. Regression analysis has three objectives: describing the associations between the dependent variable and independent variables, predicting the value of the dependent variable, and estimating the intervention effect of an independent variable on the dependent variable. However, Japanese mammologists are behind the curve in understanding the criteria for selecting independent variable to achieve the objectives of regression analysis. In addition, not only is the mean of a population of considerable interest but also the individuals that deviate from the mean. To date, regression analyses that focus on conditional expectations have been unconsciously adopted by Japanese mammologists. This study provides an overview of the dependent variable selection procedures for regression analyses with three analytical objectives. We further show the potential for introducing a quantile regression analysis method that would enable researchers estimate the conditional quantiles of a distribution of dependent variables in a linear model in mammalian studies.
