Application of Mokken Scale Analysis In Social Work Research

Mokken scale analysis (MSA) has been a useful psychometric tool for the construction of scales that measure people’s position on latent traits, such as depression, anxiety, and happiness. MSA is a probabilistic-nonparametric alternative to parametric item response theory (IRT) models with less strict assumptions. It is used similarly to other methods for data reduction that allow for the unidimensional measurement of latent variables.

MSA is a scaling technique mainly used for scaling test and survey data. It is based on nonparametric IRT models and consists of two parts: (1) an automated item selection procedure, which partitions a set of ordinal variables into Mokken scales, and (2) methods to investigate the goodness-of-fit and assumptions of models for the Mokken scales.

Models based on MSA have a number of advantages over other measurement models. For example, the relationship between the latent variable and respondent’s probability for a response category is not required to match a particular shape, as long as basic ordering requirements are met. Also, MSA includes an item parameter that shows how items differ in terms of distribution. It can also be applied when latent variables must be operationalized using a few items.

The MSA is especially relevant to social work research, since it serves as a flexible scaling procedure to derive measurement instruments from a (large) set of items based on observed item scores. Given the lack of both methodological and empirical studies based on the MSA in social work research, this workshop will discuss the steps in building a model based on MSA, strategies for model appraisal, and reporting guidelines based on real data.

Objectives

Upon completion of the workshop, participants will be able to:

1. recognize how the MSA is flexible and promising in scaling,
2. demonstrate understanding of the steps of model building based on MSA, and
3. use the mokken package in R to fit a model based on MSA.

Content

Based on the pedagogical techniques of learning-science-by-doing-science and problem-based learning, this workshop aims to address the following content:

1. summary of the characteristics and advantages of MSA,
2. a concise and systematic set of procedures to establish a model based on MSA,
3. application of MSA in social work research using a problem example based on real data,
4. a step-by-step demonstration of model construction, estimation, visualization, and diagnostics using the mokken package, and
5. how to draw inferences and conclusions.

Implications

Overall, this focus on Mokken scale analysis can be invaluable to social work researchers, faculty, students, and professionals who wish to leverage nonparametric item response theory models to develop measurement instruments empirically. Featuring real data analyzed with a cutting-edge statistical package, this workshop simplifies esoteric item response models making them more applicable to social work research. By becoming familiar with the best practices in applying models based on Mokken scale analysis, attendees can gain the expertise necessary to use a powerful statistical tool that can help enhance the examination of latent traits and advance research in social work.