skip to main content skip to footer

Item Characteristics for Measuring Primary and Secondary Processes

Alker, Henry A.
Publication Year:
Report Number:
ETS Research Memorandum
Document Type:
Page Count:
Subject/Key Words:
Coping, Defensiveness, Item Analysis, Personality Measures, Psychological Testing, Response Style (Tests), Social Desirability


The specific question raised in this study is: What kinds of items are most effective in identifying forms of primary process and how do these differ from items that are effective in identifying forms of secondary process functioning? The features of particular items that will be the subject of the analysis are: the judged desirability of the item; the controversality of the item (indexed by the standard deviation of the desirability judgments), the self-reference of the item, and the empirically keyed direction of response. The data sets for testing the hypotheses are Haan's (1965) empirically constructed "coping mechanism" and "defense mechanism" scales. Defense mechanisms successfully scaled (i.e., containing more items significantly correlated with criterion ratings than would be expected by chance) included: repression, displacement, regression, projection, denial, doubt, and intellectualizing. The coping mechanisms successfully included regression in the service of the ego, empathy, concentration, tolerance of ambiguity, logical analysis, intellectuality, and objectivity. The criteria against which these scales were constructed consisted solely of clinicians' ratings of the Oakland Growth Study adult sample based on extensive interviews. Both the MMPI and CPI were screened as item pools. Results include: 1) absolutely no mean difference between the desirability values of the coping and defense item sets; 2) low desirability items, keyed in different directions, items, keyed in a "yes" direction, do not; 3) self-referential statements are much more frequently found in the defensive behavior items; and 4) in the coping item set, high controversial items are more frequently predictive when keyed "yes" than are low controversial items. Overall, the working hypotheses, that two broad nonequivalent classes of behavior exist that require different assessment procedures for relevant effective measurement, has been supported by a variety of systematic and unexpected findings.

Read More