Problems With Brief Rorschach Protocols

Abstract

Retest reliability coefficients were calculated for 72 pairs of Rorschach records. In the target group of 36 pairs, one of the tests contained fewer than 14 responses, and the second record in the pair contained more than 14 responses. The remaining 36 pairs of protocols were used as a control group, and both records in each pair contained at least 15 responses. The resulting correlations indicate that if a protocol contains at least 15 answers, there is a substantial probability that a retest will show essentially the same distribution of scores, but if a protocol contains less than 14 answers, only two variables, the χ + % and the Basic D Score, are likely to remain in the same interpretive range if a longer protocol is taken subsequently from the same subject. A comparison of pairs of protocols in which both records have Lambda values of .85 or less indicates that some brief records may be reliable for some variables, and some have interpretive value, but even that type of brief record should probably be approached with much caution.