NCJ Number
52439
Journal
PSYCHOLOGY BULLETIN Volume: 82 Issue: 6 Dated: (NOVEMBER 1975) Pages: 887-903
Date Published
1975
Length
17 pages
Annotation
A STATISTICAL TEST FOR SPURIOUSNESS WHICH IS CALLED MORE RELIABLE THAN MULTIPLE REGRESSION ANALYSIS OR FACTOR ANALYSIS FOR CERTAIN TYPES OF LONGITUDINAL STUDIES IS DESCRIBED. TWO DETAILED EXAMPLES ARE PRESENTED.
Abstract
CROSS-LAGGED PANEL CORRELATION IS DEFINED AS 'THE COMPARISON OF CROSS-LAGGED CORRELATIONS WHICH CAN BE EXPRESSED AS A CROSS-LAGGED DIFFERENTIAL.' THE CONCEPT IS EXPLAINED IN DETAIL. LIKE ANY STATISTICAL METHOD CROSS-LAGGED ANALYSIS IS BASED ON A SET OF ASSUMPTIONS, IN THIS CASE SYNCHRONICITY AND STATIONARITY. SYNCHRONICITY ASSUMES THAT THE TWO CONSTRUCTS ARE MEASURED AT THE SAME POINT IN TIME. STATIONARITY MEANS THAT THE CAUSAL OR STRUCTURAL EQUATION FOR A VARIABLE IS NOT DIFFERENT AT THE TWO POINTS OF MEASUREMENT. THE PROBLEMS INHERENT WITH EACH ASSUMPTION ARE DISCUSSED INDEPTH, AND ARE ILLUSTRATED WITH TABLES AND FIGURES. THE TWO EXAMPLES ILLUSTRATE THE USE OF CROSS-LAGGED CORRELATION FOR PANEL (LONGITUDINAL) STUDIES AND POINT OUT ADVANTAGES OF THIS METHOD. IT IS CONCLUDED THAT THIS TECHNIQUE IS LARGELY AN EXPLORATORY STRATEGY OF DATA ANALYSIS. IT IS RECOMMENDED FOR UNCOVERING SIMPLE CAUSAL RELATIONSHIPS BETWEEN UNCONTROLLED VARIABLES AND IS SUGGESTED FOR USE IN BOTH LONGITUDINAL STUDIES AND EXPERIMENTS USING A PRETEST-POSTEST DESIGN. NOTES AND REFERENCES ARE APPENDED. (GLR)