Psychometrics

from Wikipedia, the free encyclopedia

The Psychometrics is the field of psychology , the general with the theory and methodology of psychological measurement is concerned. The main tasks of research are primarily the development and improvement of theoretical approaches to psychological measurement, as well as the elaboration of basic methods for the development of measurement instruments and general procedures for psychological measurements, both for basic and application-oriented sub-areas of psychology. Currently, work on psychological test procedures , observation instruments and other methods of psychological assessments for more application-oriented sub-areas predominate, especially traffic psychology , personnel and occupational psychology, clinical neuropsychology , school psychology and legal psychology .

Definitions

Today's explanation defines psychometrics as "any branch of psychology that deals with psychological measurement". According to this view, every area of ​​psychology ultimately has psychometric components, or, to put it another way, psychometrics is one of the interdisciplinary, connecting, methodological approaches in psychology; Psychometry is therefore a “red thread” that runs through all areas of psychology. It thus unites all psychologists and is an essential element in the profile formation of the profession and psychology as a science.

Other psychologists prefer interpretative (qualitative) methods of psychology and qualitative social research , such as qualitative content analysis , differentiated methods of psychological interviews , and psychoanalytic interpretations in their professional practice . Mutual misunderstandings can often be traced back to the ambiguous concept of measurement ( scaling ) and to opposing convictions in basic questions of philosophy of science and psychological anthropology .

developments

Psychometry is ultimately home to a compilation of specific mathematical and statistical models and methods. These were developed to summarize and describe the empirical data obtained in the context of psychological research and to draw conclusions from them. Above all, they also serve to create psychological models , such as B. mathematical-statistical psychometric models over various cognitive functional areas and personality areas, which are derived and formalized from the corresponding basic theories.

This compilation can be roughly divided into three categories, some of which are interactively linked (that is, mutually influencing and enriching each other) and which in turn have interactions with other subject areas, in particular sociology and economics .

We distinguish three intertwined strands or categories:

It is important that psychometrics does not deal directly with the development of research methods, but with the general possibilities and problems of psychological measurement or making measurable.

The scaling has its origin in psychophysics , especially in the work of Fechner . The approaches of Thurstone , who was also involved in the development of factor analysis and his “Law of Comparative Judgment” and its further developments, should be emphasized here . This is in turn based on specific estimation models , as well as the development of multidimensional scaling , which in turn is linked to approaches of factor analysis. Further developments were methods of multiple scaling and those of non-dimensional scaling, such as B. the conjoint measurement . From the psychophysical approaches, other normative models were also developed, which then influenced the development of mathematical psychology . The summary of mathematical psychology and psychometrics is often referred to as quantitative psychology .

Approaches to factor analysis influenced the development of certain statistical measures of estimation, the analysis of covariance structures up to linear structural equation models . The test psychology was strongly influenced by both the scaling and the factor analysis. Here the item response theory developed over the classical test theory , influenced by normative models of scaling, which in turn influenced the development of latent class models and structural equation models, this under the mutual influence of quantitative social research and econometrics . Psychometrics is a fruitful direction for the further development of applied statistics . It therefore fertilizes many other quantitatively working sciences that deal with problems of measurement, just as these in turn influence psychometry and its developments.

Psychometric approaches, in particular the item response theory and the linear structural equation models, have been used in the field of school psychology in the context of international comparative school studies such as B. the PISA studies further developed. The same applies to their use z. B. in the context of the learning status surveys in North Rhine-Westphalia or comparable quality assurance measures such as school performance examinations in Berlin and other federal states.

criticism

Immanuel Kant already criticized the idea of ​​a precisely measuring psychology, although he only attacked a strictly introspectionist conception of psychology. Herbart , who re-established mathematical psychology in the 19th century, and initially Wilhelm Wundt wanted to contradict Kant, but overlooked the aim of his criticism. This debate has continued since then and often includes test psychology and the entire methodology of psychology. Instead of discussing the general question "Can psychological phenomena be measured?", A more precise distinction must be drawn between:

Basically, psychological test values ​​and other research data should accurately and clearly depict the empirical characteristic relationships, i.e. express them in adequate numerical ratios. A distinction is made between different scale levels, depending on whether only the frequency of features is counted, a comparative statement “A is more intelligent than B” or a measurement in the narrower sense takes place, i.e. H. with gradual differences on a scale that is formed from intervals of the same size as on a thermometer scale. The theoretical prerequisites and the admissible or impermissible arithmetic operations are treated in the general scaling theory or theory of measurement and are also fundamental for psychometrics.

An interval scale can usually be assumed for the objective intelligence and performance tests . For subjective statements about inner states, experiences and well-being, this assumption is highly questionable, because the equality of the intervals and of course the possibility of control by other observers or for direct comparison with the feelings of other people are missing. This objection also applies to the personality questionnaires and other questionnaires that are widely used in psychology , when individual statements are added to test values ​​during their evaluation as if they were metric levels on a scale.

The theoretical decisions are not arbitrary, but there are very different points of view. “The scale quality of a measurement is ultimately dependent on theoretical decisions, ie on interpretations”. Some methodologists understand the measurement of psychological variables as a test of structural hypotheses, i.e. establish a close relationship between measurement and psychological theory in the sense of a mathematical psychology . Others believe that the scale level of the output data does not have to be justified empirically or argumentatively in advance, it does not matter whether the scale level is "true", but whether the measurement model is useful.

Self-assessments do not provide measurement data in the narrower sense, but are subjective estimation methods with an unknown scale level, with pseudo-numerical reference systems that probably differ from individual to individual . The equality of the scale intervals is not given and consequently the item values ​​cannot simply be added to a test value. There are, however, great differences of opinion in the specialist literature about the consequences of this fact. Can the theoretical measurement assumptions of the intelligence tests and the scientific behavior analysis be easily transferred to introspection and self-assessment?

Quite a number of psychologists are epistemological criticism of what they consider to be an unreflected measurement and test theory and a pseudo-scientifically oriented psychology. Subjective-mental phenomena and also psychological properties would be reduced to numbers without clearly showing the deficits of this psychometry. The fundamental criticism of the "measurement of man" and the criticism of reductionism are often combined with arguments critical of society.

However, the opposite position is also discussed critically. The criticism that psychology cannot be researched experimentally and certainly cannot be described mathematically is a. rejected as a conclusion based on an understanding based on a preconceived concept of the soul.

literature

Measurement theory and mathematical psychology

criticism

  • Gerd Jüttemann: Psychology as a human science. A manual. Vandenhoeck & Ruprecht, Göttingen 2004, ISBN 3-525-46215-8 .
  • Jochen Fahrenberg: Theoretical Psychology - A System of Controversies. Pabst Science Publishers, Lengerich 2015, ISBN 978-3-95853-077-5 .
  • Paul Walter: The "measurement of the human being": Theoretical and methodological bases of psychological testing. In Siegfried Grubitzsch (ed.): Test theory - test practice: psychological tests and test procedures in a critical overview. 2nd edition. Klotz, Eschborn 1999, ISBN 3-88074-343-6 , pp. 98-127.

See also

Web links

Individual evidence

  1. ↑ wordnet.princeton.edu  ( page no longer available , search in web archives )  Info: The link was automatically marked as defective. Please check the link according to the instructions and then remove this notice.@1@ 2Template: Dead Link / wordnet.princeton.edu  
  2. ↑ Thomas Sturm: Kant and the human sciences. mentis, Paderborn 2009, ISBN 978-3-89785-608-0 .
  3. ↑ Wolfgang Schönpflug: History and systematics of psychology: a textbook for basic studies. 2nd edition Beltz, Weinheim 2004, ISBN 3-621-27559-2 .
  4. ↑ Jochen Fahrenberg: The scientific conception of psychology in Kant and Wundt In: e-Journal Philosophy of Psychology
  5. ↑ Jürgen Bortz, Gustav A. Lienert, Klaus Boehnke: Distribution-free methods in biostatistics. 2nd edition Springer, Berlin 2000, ISBN 978-3-540-74706-2 , p. 66.
  6. ↑ Mortensen, U. (2005): Understanding or explaining? The role of experimental and statistical methods in modern psychology