ABSTRACT
The first condition means that correlation is inappropriate when planned experiments are analyzed since the X’s will be fixed and predetermined. The second condition takes us out of the realm of this text which considers only univariate statistics. Although correlation analysis is one of the most widely used statistical techniques, it is rarely used correctly. Texture measurements and the relationship between sensory and objective measurements are considerably more complex than the mathematical assumptions which underlie either correlation analysis or regression analysis. When more than two variables are involved in an analysis, it is usually of interest to calculate the partial correlation coefficients. A partial correlation coefficient indicates the relationship between two variables when all other variables are held constant. In a sense, these are all dependent variables and measurements of the crispness and texture of the raw apples. The data were ordered in terms of ascending panel scores.
