Uses and Philosophy of Correlation and Regression Analysis
Types of problems to which correlation and regression analysis have been applied
This chapter examines various types of research problems to which regression and correlation analysis have been applied, types of logical analysis made, and some of the pitfalls and difficulties encountered, and the kinds of conclusions reached.
When such a review was first presented in 1930, 77 individual studies were listed. In 1941, even a brief sampling of additional studies brought the number up to 95. In 1959, such studies have become so numerous that the titles alone would fill a book this large. In several fields (crop forecasting, supply and demand analysis, consumption studies, marketing studies, hydrology, forest mensuration, for example) whole books have been written about research methods in each field.
Brief History. The methods of correlation and regression analysis were first developed by students of heredity, notably Karl Pearson. The professional journal in this field, Biometrika, contains the original papers establishing the method, and many studies using it in the field of heredity. These include such studies as the relation of the stature of children to that of their parents. The very term "regression" itself comes from this initial use. When it was found that very tall or very short parents tended
-434-