Analyzing ecological momentary data using growth mixture modeling
Item
-
Title
-
Analyzing ecological momentary data using growth mixture modeling
-
Identifier
-
d_2009_2013:1066140a4f27:10291
-
identifier
-
10192
-
Creator
-
Shiyko, Mariya P.,
-
Contributor
-
David Rindskopf
-
Date
-
2009
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Biostatistics | Public health | ecological momentary assessments | growth mixture modeling | intensive longitudinal data | latent classes | smoking cessation
-
Abstract
-
Real-time data capture, also known as ecological momentary assessment (EMA), is a unique data collection technique, which records moment-to-moment changes in human behavior as they occur in real time and in naturalistic settings. EMA is typically collected by electronic devices that prompt study participants to report behaviors (e.g., smoking) in real time, thereby minimizing problems associated with retrospective recall and reactivity. EMA has been heralded as a promising research tool in education, psychology, and behavioral medicine. It provides the needed data to examine patterns of behaviors as well as their temporal characteristics.;Growth mixture modeling (GMM) is a statistical solution to many challenges associated with analyzing intensive EMA data. GMM estimates individual developmental profiles and classifies them into latent homogenous groups based on similarities in trajectories. This dissertation is a secondary data analysis of daily smoking rate of 74 newly-diagnosed cancer patients, who were enrolled in a randomized smoking cessation clinical trial prior to their cancer-related surgery. Patients' daily smoking rate was recorded over an average period of two weeks, yielding 896 assessments in total. The exploratory data assessment demonstrated substantial differences in patterns of smoking reduction across individuals during the intervention period. The goal of the GMM analysis was threefold: (1) to identify distinct smoking cessation patterns in a sample of patients awaiting a cancer-related surgery, (2) to investigate whether differences in tapering profiles are associated with differential smoking abstinence at surgery, (3) to identify personal and situational characteristics that are associated with each of the smoking cessation approaches.;The final model identified three latent developmental classes, which included abrupt, gradual, and slow reducers, varying in their personal characteristics and smoking cessation rates. This model is contrasted with a single-class solution alternative. Challenges of model enumeration and model identification processes are discussed. While growth mixture modeling widens the spectrum of research questions that can be addressed, it also poses technical and conceptual challenges for future research.
-
Type
-
dissertation
-
Source
-
2009_2013.csv
-
degree
-
Ph.D.
-
Program
-
Educational Psychology
