A tutorial on how to do a Mokken scale analysis on your test and questionnaire data

Abstract

Over the past decade, Mokken scale analysis (MSA) has rapidly grown in popularity among researchers from many different research areas. This tutorial provides researchers with a set of techniques and a procedure for their application, such that the construction of scales that have superior measurement properties is further optimized, taking full advantage of the properties of MSA. First, we define the conceptual context of MSA, discuss the two item response theory (IRT) models that constitute the basis of MSA, and discuss how these models differ from other IRT models. Second, we discuss dos and don'ts for MSA; the don'ts include misunderstandings we have frequently encountered with researchers in our three decades of experience with real-data MSA. Third, we discuss a methodology for MSA on real data that consist of a sample of persons who have provided scores on a set of items that, depending on the composition of the item set, constitute the basis for one or more scales, and we use the methodology to analyse an example real-data set.

Patterns and Predictors of HIV-Status Disclosure Among Pregnant Women in South Africa: Dimensions of Disclosure and Influence of Social and Economic Circumstances

“…A special case of the model is the double monotonicity model which additionally allows for an ordering of items that have invariant item ordering [33]. Here, this would imply that the order in which individuals disclose to different categories does not differ across individuals [34].…”

“…Using the mokken package in R and the approach recommended by Sijtsma & van der Ark [33], we ran an automated item selection procedure ( aisp ) algorithm 12 times consecutively at increasing values of c ( c =0 to c =0.55), where c is a chosen positive coefficient representing the threshold of homogeneity above which an item is included in the scale. We calculated item ( Hj , representing how well each item discriminates between individuals’ overall disclosure) and total scale ( H , representing the accuracy with which items within the total scale are able to order individuals with respect to overall disclosure) scalability coefficients and standard errors; checked scale assumptions; and examined the possibility of invariant item ordering to assess whether or not women disclosed to different categories in the same order.…”

Development and validation of the Treatment Expectations in Chronic Pain Scale

“…The analysis was carried out in 10 steps as recommended by Sijtsma and van der Ark () using the R package Mokken (der van Ark, Straat, & Koopman, ) in the statistical software R version 3.4.2 (R Core Team, ).…”

Establishing the cut-off scores for the severity ranges of schizophrenia on the BPRS-6 scale: findings from the REAP-AP

“…Likewise, based on the findings of Bech et al [5], the 18-item Brief Psychiatric Rating Scale (BPRS-18) has not been scalable according to the Mokken scale analyses, whereas the BPRS-6 which consists of items on expressed delusions (unusual thought content), conceptual disorganization, hallucinations (hallucinatory behaviour), blunted affect, emotional withdrawal, and poverty of speech (retardation) has been scalable. The Mokken scale analysis, which is also based on the IRT, is used to confirm the optimal number of items for establishing a coherent scale [6]. The "unidimensionality" of the BPRS-6 has been regarded as an acceptable level, since a study has found that its Mokken homogeneity coefficient is 0.45.…”

“…Moreover, since the intra-class coefficient is 0.81, its inter-rater reliability is considered to be at acceptable levels. Therefore, BPRS-6 has been proposed as a promising psychometric method to evaluate the severity of schizophrenia [5][6][7].…”