An Example of Inaccuracies Inherent in Interpretation of Ecological Field Data

Publisher Website
Google Scholar
Abstract
A laboratory experiment was carried out to determine the relationship between the independent variables, temperature and population density, and the dependent variable, egg production, in populations of Tribolium confusum. A multiple regression equation describing this relationship was assumed to represent a complete description of a biological relationship, analogous to some component of production in a natural animal population. A 50-year series of observations was derived for temperature, assumed to be a normally distributed random variable, and for adult population density, assumed to be positively correlated with the previous year''s temperature and negatively correlated with the previous year''s population density. The temperature and population density values were used in the previously derived multiple regression equation to generate a 50-year series of egg production values. To each year''s egg production was added or subtracted a quantity, negatively correlated with the current year''s temperature, which represented the net effect of all other factors upon egg production. It was assumed that annual observations of temperature, adult population density, and egg production had been made during a 50-year field study. The observations, some subject to errors of measurement, were used in regression analyses of the underlying relationship. Comparison of the equations derived from "field data" with the "true" equation provided a concrete example of the distortions which may arise in descriptions of productivity relationships based upon data of field-level precision.

This publication has 1 reference indexed in Scilit: