Bias in Point Estimates and Standard Errors of Mokken’s Scalability Coefficients

Abstract

Mokken scale analysis uses three types of scalability coefficients to assess the quality of (a) pairs of items, (b) individual items, and (c) an entire scale. Both the point estimates and the standard errors of the scalability coefficients assume that the sample ordering of the item steps is identical to the population ordering, but due to sampling error, the sample ordering may be incorrect and, consequently, the estimates and the standard errors may be biased. Two simulation studies were used to investigate the bias of the estimates and the standard errors of the scalability coefficients, as well as the coverage of the 95% confidence intervals. Distance between item steps was the most important design factor. In addition, sample size, number of items, number of answer categories, and item discrimination were included in the design. Bias of the standard errors was negligible. Bias of the estimates was largest when all item steps were identical in the population, especially for small sample sizes. Furthermore, bias of the estimates decreased as number of answer categories increased and as item discrimination decreased. Coverage of the 95% confidence intervals was close to .950, but for small sample size coverage deteriorated. Coverage also became poorer as number of items increased, in particular for dichotomous items.

Establishing dimensionality of sexual behaviours in patients with regional brain dysfunction

“…From the pool of the remaining items, the item is selected that covaries positively with the items selected in Step 1 (Scale Condition 1); the procedure deems item pairs with negative inter-item covariances as unacceptable for the chosen domain, thus, for any set of items comprising a scale, the interitem covariances must be nonnegative (47). has an H i value with respect to the already selected items that is significantly larger than zero and is equal or larger than the predefined lowerbound c (Scale Condition 2), with the lowerbound default set to .30.maximizes the total H , considering the items already selected.…”

“…covaries positively with the items selected in Step 1 (Scale Condition 1); the procedure deems item pairs with negative inter-item covariances as unacceptable for the chosen domain, thus, for any set of items comprising a scale, the interitem covariances must be nonnegative (47).…”

Psychometric validation of Czech version of the Sport Motivation Scale

“…Mokken's MH model is considered appropriate to conduct a unidimensionality check in situations when a large number of different constructs are measured by a small number of indicators [81]. A basic premise in testing and constructing a Mokken scale is the role of homogeneity coefficients (also referred to as scalability coefficients, [81,83]), which, for a particular item i, are denoted as H i , and are defined as the ratio of the sum of all item's observed covariances over the sum of all item's maximal covariances. We calculated H i , for each item within all seven SMS subscales using the package "mokken" available in the freeware statistical computing environment R [84].…”

Standard errors of two‐level scalability coefficients

“…This method has a smaller burden of computation and is therefore applicable to larger data sets, for both dichotomous and polytomous items. Kuijpers, Van der Ark, Croon, and Sijtsma () showed that bias of the standard errors was negligible, and that the coverage of the 95% confidence intervals was satisfactory. The structure of two‐level data is more complex than the structure of single‐level data, so the method of Kuijpers et al .…”