Study of Behavior Slow Read Contents Search Post Reply ... Previous Up SB2 Overview of Chapter 2 (pp. 30_46) From: Paul M. Summitt (psummitt@summittnewmedia.com)
Date: 06 May 2001
Time:
Remote Name: Comments DEPENDENCY FACTOR ANALYSIS - "We must, to change the Greek epigram a little, ascend downward and descend upward, if we would reach truth, or any true persuasion of it."
Dependency analysis refers to the distinction between independent and dependent variables in conventional hypothesis_testing terminology. If sleep deprivation is thought to slow reaction times, then this can be tested by depriving a random sample of subjects (humans, white rats, or whatever) of sleep, and then seeing how their performances differ on some task (e.g., counting the number of vowels on a printed page, running through a maze, etc.). Sleep deprivation is the independent variable (X) which then determines performance on the dependent variable of reaction time (Y)... i.e., performance on Y is dependent_ upon levels of X, i.e., performance on Y can be predicted from X, i.e., performance on Y can be explained by X. It was against this tradition that Stephenson wished to advance his own thesis "that factor analysis can also be employed for dependency analysis" (p. 31), thereby restoring "simple factor methods ... for experimental purposes" (p. 33). His revision of Kendall's "multivariate tree" (p. 32) therefore pictures variance analysis and dependency forms of factor analysis as branches off the limb of dependency analysis, and with Q methodology incorporating both variance and factor analysis (dependency form) into a common methodology which he refers to as "postulatory dependency analysis." Rather than simply | |
|
