Gideon J. Mellenbergh

The concept of validity

This article advances a simple conception of test validity: A test is valid for measuring an attribute if (a) the attribute exists and (b) variations in the attribute causally produce variation in the measurement outcomes. This conception is shown to diverge from current validity theory in several respects. In particular, the emphasis in the proposed conception is on ontology, reference, and causality, whereas current validity theory focuses on epistemology, meaning, and correlation. It is argued that the proposed conception is not only simpler but also theoretically superior to the position taken in the existing literature. Further, it has clear theoretical and practical implications for validation research. Most important, validation research must not be directed at the relation between the measured attribute and other attributes but at the processes that convey the effect of the measured attribute on the test scores.

The theoretical status of latent variables

This article examines the theoretical status of latent variables as used in modern test theory models. First, it is argued that a consistent interpretation of such models requires a realist ontology for latent variables. Second, the relation between latent variables and their indicators is discussed. It is maintained that this relation can be interpreted as a causal one but that in measurement models for interindividual differences the relation does not apply to the level of the individual person. To substantiate intraindividual causal conclusions, one must explicitly represent individual level processes in the measurement model. Several research strategies that may be useful in this respect are discussed, and a typology of constructs is proposed on the basis of this analysis. The need to link individual processes to latent variable models for interindividual differences is emphasized.

Item bias and item response theory

Abstract In this chapter the definition, detection, and explanation of item bias is discussed. Item bias is generally defined as conditional dependence; within the framework of item response theory the general definition implies that the item characteristic curves of two groups do not coincide. For detecting biased items, statistical tests and indices based on item response theory have been proposed. Several methods have been compared using empirical or simulated data. Methods combining information on the regression of item responses on latent trait or observed test score and information on the latent trait or observed test score distribution appear to do a good job in detecting biased items. Next to these general methods, specific methods are available that might perform better under specific assumptions and conditions. For explaining detected item bias, four strategies are described: (1) qualitative, (2) correlational, (3) quasi-experimental, and (4) experimental research.

Metric scaling : correspondence analysis

Introduction The Basic Structure of a Data Matrix Principal Components Analysis Multidimensional Preference Scaling Correspondence Analysis of Contingency Tables Correspondence Analysis of Non-Frequency Data Ordination, Seriation, and Guttman Scaling Multiple Correspondence Analysis

Contingency Table Models for Assessing Item Bias

Two strategies for assessing item bias are discussed: Methods based on Itemx Group interaction and methods that compare the probabilities of a correct response for different groups conditional on ability level. In latent trait models, correct response probabilties are compared conditional on latent ability; inScheuneman’s (1979) method these probabilities are compared conditional on the observed test score. Scheuneman’s method is modified to fit the general theory of loglinear and logit models for contingency tables. A distinction is made between uniform and nonuniform item bias, and a method to assess item bias and distinguish between uniform and nonuniform bias is described. Data reported by Scheuneman are used to demonstrate the method in detail; the differences obtained with Scheuneman’s method are discussed. In addition, the method is applied to two tests administered byvan der Flier (1980) in Kenya and Tanzania, and the results are compared to those obtained with Scheuneman’s approach.

A review of quality of life instruments used in dementia

Objective: To provide an overview of QOL measures applicable for research in dementia, the scale content, method of data collection, and their psychometric properties. Method: Literature research. Results: Six dementia-specific QOL measures were identified, eight generic measures were used in a demented population, and three dementia-specific measures related to QOL are described as well. Measures vary considerably in scale content, and method of data collection. Reliability indexes were always available, support of instrument validity was often reported, but reports of responsiveness to change were found only for two dementia-specific QOL measures. Conclusion: When the interest is primarily on people with dementia, a dementia-specific instrument is to be preferred. Further clarification of the concept of QOL and particularly its relation to disease severity is required.

Generalized linear item response theory

In this article generalized linear item response theory is discussed, which is based on the following assumptions: (a) A distribution of the responses occurs according to a given item format; (b) the item responses are explained by one continuous or nominal latent variable and p latent as well as observed variables that are continuous or nominal; (c) the responses to the different items of a test are independently distributed given the values ofthe explanatory variables; and (d) a monotone differentiable function g of the expected item response τ is needed such that a linear combination of the explanatory variables is a predictor of g(τ)

Social functioning in children with a chronic illness

Behavioural, cognitive, and affective aspects of social functioning of 107 children with a chronic illness were studied. The aim of the study was twofold: (1) to describe peer interaction of children with a chronic illness in comparison with normative data of healthy children; (2) to examine whether peer interaction was related to the illness characteristics physical restrictions and pain. Peer interaction was assessed with measures of social activities (CBCL), parent-reported social skills (CABS), child-reported social skills (MESSY), social self-esteem (SPPC), and social anxiety (SASK). Results showed no differences between diagnosis groups, suggesting that the social consequences of chronic illness are not diagnosis specific. Compared with healthy norms, chronically ill children reported less aggressive behaviour. The parent-report measures suggested a similar trend. Children with chronic illness also tended to display more submissive behaviour than healthy norms, as perceived by their parents. With regard to illness characteristics, both physical restrictions and pain were associated with restricted social activities, but not with other measures of social peer interaction. Children who display submissive behaviour and children who are restricted in their social activities should receive extra attention because they are especially vulnerable for problems in their social development.

Risk and preventive factors for cot death in The Netherlands, a low-incidence country

Abstract In the Netherlands an 18 months case control study into cot death was under-taken as part of the European Concerted Action (ECAS) on sudden infant death syndrome to determine the relative risk of prone sleeping and other sleep practices. Physicians in the Netherlands were asked to report to the study centre all sudden and unexpected deaths of children between 1 week and 2 years of age. Non cot death cases were deleted from further analysis after a consensus was reached by three pathologists, not primarily involved in the post mortem diagnosis. A positive response of families was achieved in 91% of cases registered in the Central Bureau of Statistics. The study comprised 73 cot deaths and 146 controls, two for each case and matched for date of birth. All families were visited at home for completion of a questionnaire. The cot death rate has dropped considerably over the past 10 years after the recommendations on supine sleeping to a low of 0.26 per 1000 live born infants. In addition to the ECAS objective, we wanted to establish whether previously found risk factors are still valid in the present situation or that new factors might have emerged, some of them possibly protective. Conclusion Placing an infant prone or on side on last occasion, secondary prone position (not placed prone but turned to prone), inexperienced prone sleeping and use of a duvet, leading to head and body being covered, were shown to be risk factors. Preventive factors were using a cotton sleeping-sack and a dummy. Even in a low incidence country, such as the Netherlands, there are indications that further prevention is possible.

DSM-IV Internal Construct Validity: When a Taxonomy Meets Data

The use of DSM-IV based questionnaires in child psychopathology is on the increase. The internal construct validity of a DSM-IV based model of ADHD, CD, ODD, Generalised Anxiety, and Depression was investigated in 11 samples by confirmatory factor analysis. The factorial structure of these syndrome dimensions was supported by the data. However, the model did not meet absolute standards of good model fit. Two sources of error are discussed in detail: multidimensionality of syndrome scales, and the presence of many symptoms that are diagnostically ambiguous with regard to the targeted syndrome dimension. It is argued that measurement precision may be increased by more careful operationalisation of the symptoms in the questionnaire. Additional approaches towards improved conceptualisation of DSM-IV are briefly discussed. A sharper DSM-IV model may improve the accuracy of inferences based on scale scores and provide more precise research findings with regard to relations with variables external to the taxonomy.