What is it about?
Traditionally, researchers created scales using classical test theory (such as factor analysis) or item-response theory (such as Rasch models). However, these methods are not appropriate for certain kinds of data. The inputs for factor analysis and Rasch analysis also need to meet certain, restrictive assumptions. In this study, we used a new method to create scales that modeled input variables' joint probability function or density (JPD). The JPD method does not impose restrictive assumptions on data and can accommodate wildly different kinds of inputs. Compared to more traditional methods, we showed that the JPD method created more informative, granular scales.
Featured Image
Photo by Conny Schneider on Unsplash
Why is it important?
The JPD method gives researchers a completely new, extremely flexible method to create scales and indices while bypassing the challenges of more traditional methods.
Perspectives
I would consider myself more of a psychometric "dilettante," but I have worked on enough scales and indices to know how frustrating it is to run up against the limitations of classical test theory or item-response theory. Dr. Noorbaloochi and Master Statistician Clothier were obviously the main drivers of this study, but it was incredibly exciting to be part of something completely new. This method will really open up options for index creation. I see the biggest opportunities will be in applying this method to administrative health data and econometrics.
Dr. Maureen Murdoch
Minneapolis VA Health Care System
Read the Original
This page is a summary of: Using joint probability density to create most informative unidimensional indices: a new method using pain and psychiatric severity as examples, BMC Medical Research Methodology, August 2024, Springer Science + Business Media,
DOI: 10.1186/s12874-024-02299-y.
You can read the full text:
Contributors
The following have contributed to this page
