Award Date


Degree Type


Degree Name

Doctor of Philosophy (PhD)



First Committee Member

Daniel Allen

Second Committee Member

Kimberly Barchard

Third Committee Member

Stephen Benning

Fourth Committee Member

Samantha John

Number of Pages



Considerable work has been devoted to characterizing the latent structure of cognition in schizophrenia (SZ) to understand important clinical outcomes associated with generalized or specific deficits but findings are limited in a number of ways. Previous work has not assessed bifactor and other complex models of cognition in SZ, which might provide a better understanding of the structure of cognitive abilities. It is also unclear whether the latent structure of cognitive abilities is similar between men and women with SZ or whether the latent structure of cognitive abilities is stable over time with repeated assessment. These limitations must be addressed before cognitive tests can be meaningfully interpreted in SZ, applied to investigate sex-based differences, or used longitudinally to judge changes in cognition in response to treatment and disease progression. To address these limitations, this dissertation conducted three studies aimed to: (I) assess a bifactor and other complex models of cognition in SZ, (II) compare the factor structure of men and women with SZ, and (III) examine the stability of the latent structure of cognition in SZ over repeated assessment. Each study used archival data from the National Institute of Mental Health Data Archive. The first study included 813 individuals with SZ who were divided into a baseline calibration sample (n = 413) and a cross-validation sample (n = 400). The second study examined whether the factor structure was the same (invariant) between men and women with SZ at a baseline assessment (men n = 612; women n = 201) and then again approximately two months later (men n = 549; women n = 198). The third study examined longitudinal invariance of the factor structure across four repeated assessments (n = 205) that included a baseline assessment and follow up assessments that occurred approximately two, six, and 18 months later. Results of these studies indicated that a bifactor seven-factor model that includes one general factor and seven specific factors (Processing Speed, Phonemic Fluency, Semantic Fluency, Reasoning, Working Memory, Verbal Memory, and Vigilance) best characterized the latent structure of cognition in SZ. The same bifactor seven-factor model was found to be invariant between men and women with SZ at two assessments. Subsequent comparison of the factor scores between men and women showed that women performed higher on Semantic Fluency, Verbal Memory, and General cognition, that men scored higher on Vigilance, and that there were no statistically significant differences between their performance scores on Processing Speed, Phonemic Fluency, Reasoning, and Working Memory. Further, the bifactor seven-factor model was longitudinally invariant across four assessments that spanned approximately 16.5 months. These results contribute to extant literature that has previously characterized cognition in SZ by demonstrating that cognition may be best understood by a morecomplex model that incorporates both general deficit and specific deficit conceptualizations. This model remained invariant between men and women allowing for direct comparisons of cognitive abilities and identification of a pattern of differences that was consistent with some prior literature. Finally, the longitudinal stability of the bifactor model suggests that tests are measuring equivalent latent constructs despite fluctuations in abilities that might be expected because of modifying factors of disorder state (e.g., symptoms, medication, course), and so can be appropriately used to investigate longitudinal changes in cognition.


bifactor; CATIE; cognition; factor analysis; schizophrenia


Clinical Psychology

File Format


File Size

2500 KB

Degree Grantor

University of Nevada, Las Vegas




IN COPYRIGHT. For more information about this rights statement, please visit