Premium
Nontraditional Regression Analyses
Author(s) -
Trexler Joel C.,
Travis Joseph
Publication year - 1993
Publication title -
ecology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.144
H-Index - 294
eISSN - 1939-9170
pISSN - 0012-9658
DOI - 10.2307/1939921
Subject(s) - multinomial logistic regression , regression diagnostic , binomial regression , statistics , logistic regression , categorical variable , regression analysis , cross sectional regression , mathematics , local regression , polynomial regression , variables , segmented regression , proper linear model , nonparametric regression , regression , econometrics , linear regression
Least—squares linear regression and multiple regression are among the most commonly used analytical techniques of ecologists. However, these techniques only address a portion of the possible applications of regression methods. We discuss two less commonly used regression analyses that could find wide application in ecology, logistic regression and LOWESS regression. Logistic regression is appropriate in cases where the dependent variable is categorical, dichotomous, or polychotomus. It can be used with continuous and/or discrete independent variables. Logistic regression is motivated by the underlying binomial or multinomial distribution of dichotomous and polychotomous dependent variables and transforms the data to explicitly model these distributions. Locally weighted regression scatterplot smoothing or LOWESS regression is used to model the relationship between a dependent variable and independent variable when no single functional form will do. LOWESS regression is motivated by the assumption that neighboring values of the independent variable are the best indicators of the dependent variable in that range of independent values.