Thorndike’s Credo: Metaphysics in psychometrics

Endorsing a priori the conviction that any science worthy of the name must measure the attributes it investigates, psychometricians adopted a metaphysical paradigm (without acknowledging it as such) to secure its claim that mental tests measure psychological attributes, a claim that was threatened by the inadequacy of test data to secure it. The fundamental axiom of this paradigm was Thorndike’s Credo (“All that exists, exists in some amount and can be measured”; 1918, p. 16), which entails its central lemma, the psychometrician’s fallacy (“All ordered attributes are quantitative”; Michell, 2009, p. 41), and which, in turn, supplies psychometrics’ primary methodological principle (“interval scales can be derived from ordinal data”). Logically, this framework is flawed at every level: Thorndike’s Credo is metaphysical overreach; the psychometrician’s fallacy is just that—a logical fallacy; and their primary methodological principle, a prioristic thinking.

Quality as Subjectively Measured Value

The paper identifies the problem of ensuring the reliability of measurement results of quality characteristics as subjective values and their correct application in logical and mathematical models of making decisions. The purpose of this study is to increase the reliability of expert evaluation of individual characteristics of the quality of processes, products, systems.The article describes basic methodological approaches to subjective measurements represented by the classical, operational and representational theories of measurement. The most acceptable for the purposes of ensuring the reliability of the expert evaluation of the single quality characteristics a representative theory, suggesting that the subjective value can be measured only in nominal or ordinal scales was determined. The contradiction is established: the possibility of measuring of single quality characteristics in the ordinal scale does not meet the needs of specialists in the field of quality, whouse subjective measurements to solve problems of analysis and decision-making, requiring the use of logical and mathematical models; in that way estimates should be expressed at least in the interval scale. The article substantiates the best solution of this problem by use of the rating scale which has properties of both ordinal and interval scales.Within the framework of the expert methods of quality measurements development two fundamental elements of the methodology of subjective measurements of subjective values from the standpoint of representative theory are formulated: 1) the rating scale as a modified scale of ranks, 2) the method of organization of the measurement process as a method of alternative assessment of expert preferences.Much attention is given to axiomatic of the rating scale having properties of both ordinal and interval scales. The algorithm of implementation of alternative assessment's of expert preferences method which is based on a special two-stage plan of alternative expert survey and statistical criterion of preferences' stability was suggested. In conjunction, the methodologies of subjective measurements of subjective values ensure the correctness of the rating scale's formation and conversion of the values of quality characteristics in the form of ratings in the corresponding values expressed at least in the scale of intervals.

Colour preference, naturalness, vividness and colour quality metrics, Part 1: Experiments in a room

In Part I of this work, observers scaled colour preference, naturalness and vividness visually on interval scales (0–100) labelled by semantic categories (e.g. ‘moderate’, ‘good’ and ‘very good’) in the context of office lighting. Five customary light sources without object saturation effect illuminated a table with coloured objects in a real room. The observers’ assessments were predicted by recent colour quality indices and selected pairs of indices combined linearly. Criterion values of the indices for ‘good’ colour preference and vividness were determined to provide a usable acceptance limit for the spectral design and evaluation of light sources. To predict colour preference, correlated colour temperature turned out to be useful. In Part 2 of this work, another experiment with the same method but using multi-LED spectra with more object saturation will be analysed and the two datasets will be merged.

Colour preference, naturalness, vividness and colour quality metrics, Part 2: Experiments in a viewing booth and analysis of the combined dataset

In Part 2 of this work, observers scaled colour preference, naturalness and vividness visually on interval scales (0–100) labelled by semantic categories (e.g. ‘moderate’, ‘good’ and ‘very good’) in the context of food lighting using the same questionnaire as in Part 1. Seven multi-LED light sources with more or less object saturation effect illuminated a viewing booth with coloured food objects. The two datasets (Part 1: room + Part 2: viewing booth) were merged and the observers’ assessments were predicted by recent colour quality indices and CIELAB chroma differences. Linear combinations of selected pairs or triads of descriptors were used to predict the merged dataset. Criterion values to achieve ‘good’ preference, naturalness and vividness level were determined.

The use of numerical value of adverbs of quantity and frequency in the measurement of behavior patterns: transforming ordinal scales into interval scales

This paper presents a research on rating scales in response to different situations. It aims to improve the significance and accuracy of ordinal scales, transforming them into interval scales. To reach this objective, the presented scales combine quantitative and qualitative perspectives, joining the ease of the Likert scale and the Thurstone’s procedure. In this research, a sample of subjects was asked to indicate the numerical value of adverbs, in reference to a numerical scale. The results were subjected to statistical analysis, to assess their validity. Combining the qualitative dimension with a quantitative evaluation, this procedure can meet the biopsychosocial specificities of subjects, as required by the complexity paradigm. The results of this study seem to be an affirmative response to the questions about validity and reliability, and about the practicality of this procedure.

Fuzzy scales for the measurement of color

The color, with the particularity to be defined simultaneously as a physical quantity and as a psychophysical quantity, is one of the concepts that can link hard sciences and behavioural sciences. From the viewpoint of behavioural sciences, colors are basically measured with nominal scales. in hard sciences, colors are measured with interval scales. Our hypothesis is that the main relation that must be preserved during a color measurement is a metric. We suggest then that colors must be measured with metrical scales. The fuzzy metrical scale is preferred due to the possibility to define it like a nominal scale.