Abstracts Category : Other

The multifaceted measurement of the individual through language

by Ryan Lee Boyd

Historically, research within the psychological sciences has adopted a classical approach to understanding the individual. This approach regularly involves the observation and measurement of specific, isolated psychological phenomena in an attempt to better understand psychological features, tendencies, and processes at varying levels of interest. While the scope of the traditional approach can vary depending on the construct under investigation, the core methodology and analytic strategy typically adheres to the isolate and/or manipulate doctrine for seeking knowledge. In recent years, however, technology has revolutionized researchers access to computational power, analytic techniques, and even the quality and quantity of data that can be used in scientific pursuits. This dissertation consists of 3 sets of studies that are either a) already published in peer-reviewed journals or b) are currently under review in peer-reviewed journals. The primary theme to be found in the included studies is a transition from classical methods of assessment to one where the individual is simultaneously quantified in high-dimensional space using language analysis techniques. This approach essentially constitutes the quantification of the individual as a cluster of traits/processes by means of psychological traces that are embedded in (and can be measured indirectly via) a persons language. This approach entails measuring psychological phenomena at both greater depth and breadth than commonly seen in the psychological sciences and, additionally, serves as a convenient and powerful replacement of traditional approaches to studying psychology in the real world. The studies included in this dissertation demonstrate the usefulness of a high-dimensional psychometric approach via language in realms of authorship attribution and value measurement. In 2 of the 3 studies, language analytic techniques are used to measure consistencies within the individual that can be capitalized upon in order to determine authorial identities. In the third study, the high-dimensional approach is applied to the realm of values, demonstrating greater utility in a classic research paradigm that vastly outperforms the traditional self-report method.Advisors/Committee Members: Pennebaker, James W. (advisor), Gosling, Samuel D (committee member), Swann, William B (committee member), Lease, Matthew (committee member).