February 1984

Cautions in the Diagnosis of Borderline Schizophrenia

Author Affiliations

Department of Psychiatry University of Tennessee College of Medicine Suite 633, 66 N Pauline Memphis, TN 38105

Arch Gen Psychiatry. 1984;41(2):211-212. doi:10.1001/archpsyc.1984.01790130107017

To the Editor.—  In discussing schizotypy, Gunderson et al1 referred to a study2 of which one of us (P.J.K.) was an author and made incorrect statements about the design and statistics, stating that items in the study had "poor interrater reliability." A more careful examination of our raw data in Table II of that article would indicate that the Pearson's r coefficient is.834. This figure is found on page 144 and is a better measure of the reliability than the percentage agreement among raters that we also mentioned in Table II (range, 81% to 90%). Gunderson et al resorted to the same (percentage agreement) technique of reporting reliability and produced the figure of 65.7%, which did not deter them from making the statement about "poor interrater reliability" even though our data surpassed theirs in terms of precentage agreement.Second, they stated that "it is unclear how Welner made

First Page Preview View Large
First page PDF preview
First page PDF preview