My research focuses on evaluating and
applying advanced statistical methods to behavioral and
health data. My research falls into two broad categories: statistical
research and applied research.
GLiMs are a broad family of statistical models that extend linear
regression analysis to a wider range of outcome variable distributions
and error structures. I use simulations to explore the
performance of these models. I am particularly interested in the
statistical characteristics of counts outcome variables, especially in
statistical mediation models.
Coxe, S. (2019, March). Counts and Frequencies in
Psychological Research. At Blending Psychology and Methods: A Conference
in Honor of Professor Leona Aiken. Tempe, AZ. March 22, 2019.
Aiken, L. S., Mistler, S. A., Coxe, S., and West,
S. G. (2015). Analyzing Count Variables in Individuals and Groups:
Single and Multilevel Models. Group Processes and Intergroup
Relations, 18, 290-314. doi: 10.1177/1368430214556702
Coxe, S., Aiken, L. S., and West, S. G. (2013).
Generalized linear models. In T. Little (Ed.), Oxford Handbook of
Quantitative Methods, Volume 2. New York: Oxford University
Press.
Coxe, S. (2010). Mediation analysis of Poisson
distributed count outcomes: Standard estimates of the mediated effect
are not equivalent. Presented at the 8th Annual Society of Multivariate
Experimental Psychology Graduate Student Pre-conference. Atlanta, GA.
September 30, 2010.
Coxe, S., West, S. G., and Aiken, L. S. (2009). The
analysis of count data: A gentle introduction to Poisson regression and
its alternatives. Journal of Personality Assessment, 91,
121-136.
I am joint co-PI on TIDAL, an
integrative data analysis project funded by NIMH
combining data from 4 separate psychosocial interventions for teens who
have been diagnosed with ADHD. The combined dataset is publicly
available here.
Among the methodological challenges in this project were harmonization
of measures of ADHD symptoms (DSM-IV versus DSM-5) and parent depression
symptoms (various measures) using moderated nonlinear factor
analysis.
Coxe, S. & Sibley, M. H. (2023). Harmonizing
DSM-IV and DSM-5 versions of ADHD “A Criteria”: An Item Response Theory
Analysis. Assessment, 30, 606-617.https://doi.org/10.1177/10731911211061299
Zhao, X., Coxe, S., Sibley, M. J., Zulauf-McCurdy,
C.A., & Pettit, J. (2022). Harmonizing Depression Measures Across
Studies: A Tutorial for Data Harmonization. Prevention Science.
https://doi.org/10.1007
Sibley, M. H., Coxe, S., Stein, M.A., Meinzer, M.
C., Valente, M. (2022). Predictors of Treatment Engagement and Response
among Adolescents with ADHD: An Integrative Data Analysis. Journal
of the American Academy of Child and Adolescent and
Psychiatry.
Coxe, S., Sibley, M.H., & Becker, S.P. (2021).
Presenting Problem Profiles for Adolescents with ADHD: Differences by
Sex, Age, Race, and Family Adversity. Child and Adolescent Mental
Health, 26, 228-237.
Sibley, M.H. & Coxe, S. (2020). The ADHD Teen
Integrative Data Analysis Longitudinal (TIDAL) Dataset: Background,
Methodology, and Aims. BMC Psychiatry, 20, 1-12.
Statistical graphics and R shiny
I am interested in graphics as an aid to interpreting and
communicating statistical models. To this end, I have developed
packages and
shiny applications to calculate effect size for count
regression models and to estimate confidence intervals for mediation
models.
shiny applications are interactive,
web-based tools for statistical analysis and graphics.
countES: An
package to estimate effect sizes for regression models for count models
(Poisson regression and negative binomial regression). Standard
multiplicative effect size (rate ratio) and standardized mean difference
(Cohen’s d) effect sizes are provided. Monte Carlo confidence
intervals for each effect size are also given.
RcountD: An
shiny
app that implements the same functions
SimpleMediation:
Estimate indirect (mediated) effect with Monte Carlo simulated
confidence intervals. This app was developed for pedagogical
purposes.
I taught a course on Statistical Graphics and
Communication in the Psychology department at FIU. This course
covered a variety of topics, including R,
ggplot2, markdown, and
shiny. Most of the materials for the first
version of the course from 2019 are available here: https://stefanycoxe.github.io/graphics_2019/
Applied research
My applied research takes place largely within the Cancer
Institute at Cedars-Sinai Medical Center, focusing on behavioral
interventions within the Cancer Prevention and Control and the Health
Equity groups. I also have 20 years of experience with clinical and
psychosocial interventions in the Psychology departments at Arizona
State University and Florida International University. I have also
collaborated with faculty in the Center for Children and Families and
the departments of Epidemiology, Social Work, and Biomedical Engineering
at Florida International University. These studies often include complex
longitudinal designs, non-normally distributed outcome variables, and
extensive missing data.
Zhao, X., Coxe, S., Timmons, A., & Frazier, S.
L. (2022). Mental Health Information-Seeking Online: A Google Trends
Analysis of ADHD. Administration and Policy in Mental Health and
Mental Health Services Research, 49, 357 - 373. https://doi.org/10.1007/s10488-021-01168-w
Sibley, M.H., Coxe, S.J., Page, T.P., Pelham, W.E.,
Yeguez, C.E., LaCount, P.A., & Barney, S. (2020). Four-Year
Follow-up of High versus Low Intensity Summer Treatment for Adolescents
with ADHD. Journal of Clinical Child and Adolescent
Psychology.
Mukherjee, S., Coxe, S., Fennie, K., Madhivanan,
P., Trepka, M. J. (2017). Stressful life event experiences of pregnant
women in the United States: A latent class analysis. Women’s Health
Issues, 27, 83 – 92.
Bagner, D. M., Coxe, S., Hungerford, G. M.,
Linares, D., Barroso, N., Hernandez, J., Rosa-Olivares, J. (2016).
Behavioral parent training in infancy: A window of opportunity for
high-risk families. Journal of Abnormal Child Psychology, 44,
901-912. doi: 10.1007/s10802-015-0089-5