Correlation Scatter Plotting as a Quick Substitute for Factor Analysis

The widespread use of analytical methods of factoring has tended to obscure the possibility of using more descriptive approaches to variable structuring. This paper demonstrates how simple correlation plotting can be used to provide a quick representation of variable inter-correlations. In the example offered, a study of transfer from primary to secondary, the correlation scatter plots show considerable similarity with the varimax factor analytic solution. Concluding comments suggest that correlation plotting could provide valuable assistance in teaching the principles of factor analysis, in arriving at quick evaluations of factor analytic studies, and in communicating results derived from factor analyses.

Factor analysis has achieved a considerable degree of support amongst behavioural researchers (see for example, Bennett and Bowers, 1976; Child, 1970; Kerlinger, 1979). Nevertheless there are situations where the computational complexity of the method may inhibit understanding of associated findings. In the early stages of educational training, for example, it is unrealistic to expect students to be able to devote the time and effort required for an adequate study of the method. It is also imprudent to expect students with such limited understanding to pursue projects which demand factoring procedures for successful analysis and interpretation. Moving to the area of practical applications, many consumers of research may become sceptical of findings and suggestions which rely heavily on a procedure whose rationale is obscure. And even with experienced researchers there are occasions when a quick, descriptive evaluation of data is required to assist subsequent, more probing analyses. Most statistical techniques have an associated graphical representation which can-be applied to summarise or describe data. It is suggested here that the method of plotting variable inter-correlations is more valuable than is generally assumed.