Construct Validity of the Sentence Completion Test of Ego Development

Abstract

Evidence for the construct validity of the Sen tence Completion Test of Ego Development (SCT), some of it previously unpublished, is reviewed. The substantive component is substantiated both by in tuitive plausibility and by the fact that rating skill can be communicated through written instructions. Internal consistency is shown by alpha coefficients of about .85. Evidence for sequentiality is provided by studies showing cross-sectional gains with age during adolescence, similar profiles of item scores for adjacent total protocol ratings, substantial cor relations over long time spans, gains proportional to retest intervals during adolescence, gains follow ing theory-relevant interventions of several months' length, and better comprehension of lower than of higher stages. The SCT has substantial correlation with tests of related conceptions, such as Kohl berg's test of moral maturity, and with measures of ego level applied to other behavior samples. Cor relations with isolated traits characteristic of spe cific stages range from negligible to positive. Studies of external validity in general yield positive results, though theory does not sharply define what results should be considered positive. At lower ego levels there is behavioral evidence confirming the syndromes; at higher levels differential evidence lies in attitudes and ideas. Overall, the test has ade quate validity for research purposes when ad ministered and scored with sufficient care, but not for individual diagnosis unsupported by other data. Evidence for construct validity is substantial, but falls short of clear proof of sequentiality.