Name: Na Liu

School of Psychology – Ph.D. Dissertation Proposal Meeting

Date: Monday, September 29th

Time: 3:30 P.M. – 5:00 P.M.

Teams Meeting link: click here

Dissertation Chair/Advisor: 

James Roberts, Ph.D. (Georgia Institute of Technology)

Dissertation Committee Members: 

James Roberts, Ph.D. (Georgia Institute of Technology)

Audrey Leroux, Ph.D. (Georgia Institute of Technology)

Dingjing Shi, Ph.D. (Georgia Institute of Technology)

Mark Himmelstein, Ph.D. (Georgia Institute of Technology)

Hongli Li, Ph.D. (Georgia State University)

Title: Enhancing Precision in the Generalized Graded Unfolding Model (GGUM) Using Successive Interval Judgements Indicative of Item Location

Abstract:
This study introduces a Hybrid Generalized Graded Unfolding Model (HGGUM) to enhance precision in estimating item locations on a latent continuum for non-cognitive psychological constructs (e.g., attitudes, emotions, personality traits). Traditional GGUM-based attitude measures often require very large sample sizes (previously N > 750 recommended) to achieve stable parameter estimates. The proposed HGGUM integrates the Method of Successive Intervals (MSI) with the GGUM, leveraging additional favorability rating data to reduce sample size requirements without sacrificing estimation precision. To evaluate this hybrid model, simulation studies are conducted using a Markov Chain Monte Carlo (MCMC) estimation procedure to examine parameter recovery and estimate accuracy under varying sample sizes. In addition, the HGGUM is applied to real survey data on attitudes toward abortion collected from undergraduate students at Georgia Tech, demonstrating the model’s practical utility. Anticipated outcomes include more precise item location estimates and the ability to apply GGUM-based models in studies with smaller sample sizes, thereby broadening the model’s applicability in psychological research.