| The study represented an attempt to create an instrument intended to be used to assess the validity of the Comprehensive System for the Rorschach Inkblot Test (CS; Exner, 2003). The Rorschach Statements Survey (RSS) was created to assess self-reported population base rates for a variety of traits, and to compare self-reports with CS data to explore the effects of confirmation bias and the Barnum effect on CS interpretation. It originally contained eight subscales including Affect, Depression, Ideation, Information Processing, Self Perception, Social Desirability, and Capacity for Control and Stress Tolerance. Two groups of individuals were recruited to complete the RSS in an attempt to establish the instruments' internal consistency and reliability. A nonpatient group was recruited from a housing complex in upstate New York and a patient group was recruited at a transitional living program in northern Virginia. During the scoring and data analysis, it was determined that two subscales, Capacity for Control and Stress Tolerance, and Information Processing, would be abandoned due to poor design and poor reliability, respectively. Six subscales emerged with acceptable reliability. It was hypothesized that significant gender differences would be revealed, and this hypothesis was not supported. It was also hypothesized that patients would report higher levels of psychological distress and lower levels of social desirability than nonpatient controls, although this was not supported either. Patients did report higher distractibility than nonpatient controls, however. An age and geography confound, as well as a likely socioeconomic status confound, lead to great caution in drawing any conclusions about the validity or explanation of these differences. It was finally hypothesized that the Social Desirability subscale would correlate negatively with the Depression subscale, and this hypothesis was supported. In fact, the Depression subscale was found to correlate negatively with all of the remaining five scales. Significant correlations also were revealed between the Self Perception and Ideation subscales and between the Affect and Interpersonal Perception subscales. Despite the study's limitations, including between group confounds, design problems, and a low N, it is believed that a revised RSS could be successfully used to explore the issues intended for the present RSS. |
