Dimension Reduction on Measures of Impulsivity

<p>The main focus of this study was twofold: the first purpose being the assessment of convergent validity on six different measures of impulsivity, and the second being a comparison of dimension reduction techniques for quantitative data. Results from a Principal Components Analysis (PCA) were compared with results from a novel dimension technique known as Local Linear Embedding (LLE). LLE is an analysis of dimension reduction for nonlinear, high dimensional data. By computing neighborhood preserving embeddings, LLE aims to map newly constructed coordinates into a global coordinate structure of a lower dimension. Past research using LLE has solely been conducted on visual and auditory-oriented data. Thus, this is a novel approach, applying LLE to cognitive measures rather than visual or auditory data. This paper acts as a secondary analysis of data collected by a prior Masters student (Robles, 2016). The measures of impulsivity included in this study were: The Barratt Impulsiveness Scale (BIS), the Iowa Gambling Task (IGT), the UPPS-P Impulsive Behavior Scale, the Cued Go/No-Go Task, the Stroop Color-Word Interference Test, and the Delay and Probability Discounting Task (DPDT) – all tests performed in a Latin-square determined order by each of 151 students. Findings not only indicate that convergent validity was absent for certain measures of impulsivity, but that with future programming, LLE may be a suitable method of dimension reduction for quantitative data.</p>

Name thesis:degree_name

Master of Arts (M.A.)

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Psychology

Year dc:date.available

2018

Author dc:creator

• Schulz, Christina T

Contributors dc:contributor

• Adriana Espinosa
• Vivien Tartter
• Jay Jorgenson

• impulsivity
• impulsiveness
• dimension reduction
• local linear embedding
• principal components analysis
• Psychology
• Social and Behavioral Sciences

Repository record dc:identifier

https://academicworks.cuny.edu/cc_etds_theses/758

OAI identifier oai:identifier

oai:academicworks.cuny.edu:cc_etds_theses-1769