Abstract
Brief and lengthy Rorschach records have been identified as common problems in protocol administration. Clinicians have debated how to prevent overly short and long records, but they have been reluctant to alter standardized administration for fear of introducing bias. The present study examines a nonintrusive method for constraining responses by prompting for an extra response when only one is offered per card and by removing the card after four responses are given. Among patients who typically produce brief records, consisting of a residential sample of civil and forensic patients with a range of disordered thinking, the alternative administration method demonstrated improved Comprehensive System validity in assessing thought disorder and eliminated the need to readminister the test due to fewer than 14 responses. The findings have clinical implications for protocol administration with thought-disordered populations that typically produce brief records.
ACKNOWLEDGMENTS
Kristin L. Dean is now with the University of Tennessee-Cherokee Health Systems Center of Excellence. This study was completed for Kristin L. Dean's doctoral dissertation at the California School of Professional Psychology at Alliant International University. Preliminary results were presented at the 2003 annual meeting of the Society of Personality Assessment, San Francisco, California. The authors thank the staff and residents at Central Prison, the North Carolina Correctional Institute for Women, and the Utah State Hospital for their support.
Notes
1This range was used for two reasons, both based on CitationExner and Erdberg's (2005) reference sample of 450 adults. First, in this sample the median value for R is 23.0, the SD is 5.68, and the interquartile range (IQR), a measure of dispersion or variability that is not influenced by outliers, is 5.0. Thus, the range from 18 to 28 roughly corresponds to the median ± 1 SD and it exactly corresponds to the median ± 1 IQR. Second, in this reference sample the cut-off values roughly separate the lowest 10% and highest 10% of the distribution from the central 80%. Specifically, the range from 14 to 17 includes 10.2% of the sample (n = 46) and the range from 29 and up includes 12.2% of the sample (n = 55), leaving the central range from 18 to 28 to include 77.6% of the sample.
* p < .05.
** p < .01.
*** p < .001.
* p < .05.
** p < .01.
*** p < .001.