Using Profile Analysis via Multidimensional Scaling (PAMS) to identify core profiles from the WMS-III.
ABSTRACT Profile Analysis via Multidimensional Scaling (PAMS) is a procedure for extracting latent core profiles in a multitest data set. The PAMS procedure offers several advantages compared with other profile analysis procedures. Most notably, PAMS estimates individual profile weights that reflect the degree to which an individual's observed profile approximates the shape and scatter of latent core profiles. The PAMS procedure was applied to index scores of nonreplicated participants from the standardization sample (N = 1,033) for the Wechsler Memory Scale--Third Edition (D. Tulsky, J. Zhu, & M. F. Ledbetter, 2002). PAMS extracted discrepant visual memory and auditory memory versus working memory core profiles for the complete 16- to 89-year-old sample and discrepant working memory and auditory memory versus working memory core profiles for the 75- to 89-year-old cohort. Implications for use of PAMS in future research are discussed.
SourceAvailable from: David F Tolin[Show abstract] [Hide abstract]
ABSTRACT: Contemporary cognitive models of obsessive-compulsive disorder (OCD) emphasize the importance of various types of dysfunctional beliefs in contributing to OC symptoms, such as beliefs about excessive personal responsibility, perfectionism, and intolerance for uncertainty. The present study seeks to further our understanding of the role of these beliefs by identifying the common profiles of such beliefs, using profile analysis via multidimensional scaling (PAMS). In Study 1, a large student sample (N=4079) completed the 44-item obsessive beliefs questionnaire. One major profile, control of thoughts and perfectionism, was extracted. Study 2 examined profiles of the 87-item obsessive beliefs questionnaire in people with obsessive-compulsive disorder (OCD; n=398), other anxiety disorders (n=104), and a sample of undergraduate students (n=285). Inflated responsibility was a prominent subscale in the profiles of all three groups. Only control over thoughts was a unique subscale in the profile obtained for the OCD group, with this group having lower scores compared to the other groups. The results suggest that while inflated responsibility is a significant subscale in the profile of individuals with OCD, it is not a unique contributor; instead, control over thoughts is unique to OCD. The data, as well as recent research investigating obsessive beliefs, suggest the need to revise the contemporary cognitive model of OCD.Journal of anxiety disorders 03/2014; 28(4):352-357. DOI:10.1016/j.janxdis.2014.03.004 · 2.68 Impact Factor
[Show abstract] [Hide abstract]
ABSTRACT: The interpretation of subtest profiles from intelligence testing remains popular among many practitioners who use subtest performance to draw diagnostic conclusions, in spite of criticism by some researchers, who point to the low reliability and predictive validity of subtest scores in predicting achievement outcomes. Prior research outlines two approaches to the study of subtest variation: the examination of interindividual variation in specific cognitive domains or subtests as compared to a standard sample, and the examination of intraindividual strengths and weaknesses, regardless of overall level. The present study seeks to add to knowledge in this field with data from 567 children ages 5 to 10 years who exhibit meaningful subtest variation on a new test of intellectual abilities. Results from the present sample point to statistically significant utility, with small to medium effects, of intraindividual cognitive and motivational profile shape over and above profile level in predicting mathematical skills. We discuss implications for school psychological and educational assessment research.Psychology in the Schools 01/2013; 50(1). DOI:10.1002/pits.21659 · 0.72 Impact Factor
[Show abstract] [Hide abstract]
ABSTRACT: The current study aims to know what we can gain if we categorize continuous data, utilizing the dual scaling (or correspondence) paradigm, compared to analysis of the original continuous data. Different from our misconception that the continuous data analysis is more informative than the categorical data analysis, the correspondence analysis (of categorized continuous data) provides richer information, since it includes separate dimensional information for row and column categories, considering the categories as variables. This dual information for rows and columns is not available for the ordinary principal or factor analysis which usually provides only column-wise information about factors of variables. It is our hope that introduction of correspondence analysis opens a new arena for educational research.American Educational Research Association, New Orleans, Louisiana; 04/2011