Clinical Psychometrics is an introduction to the long-term attempt to measure the psychiatric dimension of dementia, schizophrenia, mania, depression, anxiety, neuroticism, extraversion/introversion and health-related quality of life. The two psychometric procedures, classical factor analysis and modern item-response models, are presented for readers without any requirement for particular mathematical or statistical knowledge. The book is unique in this attempt and provides helpful background information for the dimensional approach that is being used in the forthcoming updates to the diagnostic classification systems, ICD-11 and DSM-5. The book is written for everyone who is interested in the origins and development of modern psychiatry, and who wants to be familiar with its practical possibilities; how it is possible to compare different individuals with each other, how one may determine the boundary between what is normal and what is disease, or how one may assess the clinical effect of the various forms of treatment, available to present day psychiatry.
Autorentext
Per Bech is the author of Clinical Psychometrics, 2nd Edition, published by Wiley.
Zusammenfassung
Clinical Psychometrics is an introduction to the long-term attempt to measure the psychiatric dimension of dementia, schizophrenia, mania, depression, anxiety, neuroticism, extraversion/introversion and health-related quality of life.
The two psychometric procedures, classical factor analysis and modern item-response models, are presented for readers without any requirement for particular mathematical or statistical knowledge. The book is unique in this attempt and provides helpful background information for the dimensional approach that is being used in the forthcoming updates to the diagnostic classification systems, ICD-11 and DSM-5.
The book is written for everyone who is interested in the origins and development of modern psychiatry, and who wants to be familiar with its practical possibilities; how it is possible to compare different individuals with each other, how one may determine the boundary between what is normal and what is disease, or how one may assess the clinical effect of the various forms of treatment, available to present day psychiatry.
Leseprobe
Chapter 1
Classical psychometrics
More than a century ago, psychology was defined as the science of human mental manifestations and phenomena. However, it was psychometrics (the science of measuring these mental manifestations and phenomena) that made psychology scientific. Thus, psychometrics is a purely psychological area of research.
From a historical point of view, psychology branched out from philosophy as an independent university discipline at the close of the nineteenth century. It all started in Leipzig in 1879. Here the philosopher Wilhelm Wundt (1832-1920) established his psychological laboratory at the university. Formally, however, his laboratory remained under the faculty of philosophy. Wundt succeeded in detaching psychology from philosophy, especially freeing it from the influence of Emanuel Kant, an extremely influential philosopher who stated that it is impossible to measure manifestations of the mind in the same way as physical objects (5). With his criticism of pure reason, Kant (1724-1804) established the very important distinction between 'the essential nature of things' (things in themselves) and 'things as they seem' (i.e., that which we sense or perceive as a phenomenon when faced with the object we are examining).
Figure 1.1 illustrates Kant's philosophical approaches with reference to present day psychiatry, according to which depression is understood to be a clinical phenomenological perception (shared phenomenology of depressive symptoms) as measured by the six depression symptoms contained in the Hamilton Depression Scale (HAM-D6, see Figure 3.1 ). Modern neuropsychiatry attempts to describe the depression behind the phenomenological perception, i.e., depression 'in itself', as we believe it to be present in the brain, for example, as a serotonin 1A receptor problem (impairment).
Figure 1.1 The philosophical background for the emergence of psychometrics
The area of research now known as brain research is just such an attempt to measure the processes presumed to be taking place in the brain, that is 'das Ding an sich'. As pointed out by Sontag, reality has increasingly grown to resemble what the camera shows us (6). It is reality itself when the neuropsychiatric camera demonstrates receptor binding in the brain, while clinical reality is increasingly becoming what the camera visualises for us by means of assessment scales or patient-related questionnaires.
The ability to describe reality as it is in itself, i.e., looking at the world unclouded by any preconception of it, has been debated by such neo-Kantentians as Wittgenstein and Quine (7). The quantification of endophenotypes or deep phenotypes is probably the most scientific image of the world. However, we do not have endophenotypes to tell us whether we indeed can describe reality, e.g., the brain, as it is itself. Wittgenstein tells us that he does not want to say whether we can or cannot describe reality as it is in itself. He wanted, as stated by Putman to bring our phenomenological items back to their home in clinical psychiatry. This is what clinical psychometrics is about (7).
Figure 1.2 shows a correlation between the so-called psychotic symptom items in an American rating scale (see Appendix) and serotonin 2A receptor binding, which it is now possible to measure by means of positron emission tomography (PET) scanning (8). The figure shows a correlation coefficient of _0.57; this is statistically significant but not clinically significant, as the variance on the ordinate axis (the 'psychosis' scale) can explain only about 32% of the variance on the axis of abscissas (serotonin 2A receptor binding). If the two patients at the far left are excluded as outliers, then the negative correlation value is halved, so that less than 10% of the variance is explained.
Figure
Inhalt
About the author, ix
Preface, x
Introduction, 1
1. Classical psychometrics, 3
Emil Kraepelin: Symptom check list and pharmacopsychology, 6
Charles Spearman: Factor analysis and intelligence tests, 10
Harold Hotelling: Principal Component Analysis, 13
Hans Eysenck: Factor analysis and personality questionnaires, 15
Max Hamilton: Factor analysis and rating scales, 20
Pierre Pichot: Symptom rating scales and clinical validity, 23
2. Modern psychiatry: DSM-IV/ICD-10, 27
Focusing on reliability, 27
Focusing on validity, 28
Quantitative, dimensional diagnosis, 29
3. Modern dimensional psychometrics, 32
Ronald A. Fisher: From Galton's pioneer work to the sufficient statistic, 32
Georg Rasch: From Guttman's pioneer work to item response theory analysis (IRT), 34
Sidney Siegel: Non-parametric statistics, 38
Robert J. Mokken: Non-parametric analysis for item response theory (IRT), 39
4. Modern psychometrics: Item categories and sufficient statistics, 43
Rensis Likert: Scale step measurements, 43
John Overall: Brief, sufficient rating scales, 45
Clinical versus psychometric validity, 48
Item-response theory versus factor analysis, 49
Jacob Cohen: Effect size, 50
5. The clinical consequence of IRT analyses: The pharmacopsychometric triangle, 53
Effect size and clinical significance, 53
The pharmacopsychometric triangle, 56
Antidementia medication, 59
Antipsychotic medication, 60
Antimanic medication, 65
<…