Methods for making predictions and for testing various hypotheses concerning processes of change are presented here. Applying these methods to the reanalysis of panel data on pre-election vote intention, we find that the tests given here permit more detailed analysis than did tests applied earlier analysis suggested that the process of attitude change in the July-August period (during which time the Democratic convention was held) differed from that in the August-September period, reanalysis indicates that the corresponding differences between the two periods were not statistically significant except for the data relating to those who "Didn't Know" in July and in August respectively; that is, apparent over-all differences in the process of change can be attributed to differences for those who "Didn't Know."
No takes yet. Share an insight, caveat, or question.
Leo A. Goodman (1962) studied this question.