| 
  • If you are citizen of an European Union member nation, you may not use this service unless you are at least 16 years old.

  • You already know Dokkio is an AI-powered assistant to organize & manage your digital files & messages. Very soon, Dokkio will support Outlook as well as One Drive. Check it out today!

View
 

MCGILL FAMILY MEDICINE STUDIES (posters), 2018: 02 (redirected from MCGILL FAMILY MEDICINE STUDIES, 2018: 01)

Page history last edited by hao.zhang5@... 4 years, 11 months ago

Zhang, H., Gagnon, J., & Schuster, T. (2018 June). Collecting and aggregating prior information for Bayesian factor analysis. Poster session presented at the 2018 International Society of Bayesian Analysis (ISBA) World Meeting, Edinburgh, UK.

 

Download poster here 

 

Abstract 

Questionnaire validation is fundamental when developing instruments for assessing latent psychometric properties. To estimate item-domain correlations, Bayesian Confirmatory Factor Analysis (CFA) supports the utilization of expert prior knowledge with smaller sample sizes compared to classic frequentist CFA. This assumes that investigators are able to provide relevant prior information with respect to the latent model, which could be potentially challenging. We propose a qualitative approach to collecting and decoding such information and a quantitative method aggregating the belief of all experts.

 

We apply the proposed prior acquisition to data of a diabetes empowerment questionnaire (28 questions with priors input from 6 domain experts). Priors were selected by the experts from six graphs of beta distribution. We synthesize in total 28 priors each by integrating six beta densities. We compared the estimated factor loadings and credible intervals with the results of a classic CFA using the same data, yielding substantial gain in precision of estimated parameters. Feasibility and acceptability of the expert survey was high. 

 

We propose that qualitative approaches should be more routinely employed to inform Bayesian CFA for better knowledge translation and to improve the access of primary health care researchers to available and appropriate statistical methods. 

 

Comments (0)

You don't have permission to comment on this page.