Scientific Center of the International Association of Personality Development Professionals
Introduction
The Max Luscher color test is a well-known tool for studying the psychological state, based on the interpretation of the order of preference by the respondent of the color cards offered to him. The test is very popular, there are many computer implementations, however, Max Luscher believes that these implementations cannot give a reliable result due to the inevitable distortions of the colors presented, associated not only with the difference in the color reproduction of monitors, but also with the fundamental mismatch of the RGB and CMYK color spaces, used by monitors and printers, respectively.
The purpose of the study was to test the assumption of the significance of the influence of deviations in the accuracy of color reproduction on the order of color selection by the respondents and to evaluate in principle the possibility of using computer implementations of the Luscher test to obtain reliable results.
Methods
For the study, a computer implementation of the full Luscher color test with some modifications was developed. Namely:
Additional scales instead of pairwise presentation were presented in a table to speed up the test. This was allowed due to the fact that the psychological interpretation of the test results was not included in the objectives of the study.
The test, as required by the method, contained two passes, however, the colors presented in the second pass were dosed changed according to the following algorithm:
The upper limit of the color difference was chosen, which was taken equal to 10%.
A random variable within this value was calculated.
The original color was converted to the equal contrast CIE Luv color space.
The calculated random change was randomly distributed among the three unit vectors of this space.
The resulting color was translated back into RGB space.
Despite the obvious losses during conversions between color spaces, the resulting colors at the maximum change differed quite noticeably by eye).
Table 1. Example of pairs of colors presented with max 10% change. The top line is the original colors, the bottom line is the modified colors.
The respondent was initially presented with unmodified sets of colors through a web page, which corresponded as closely as possible to the colors of the Luscher test. In the second pass, sets consisting entirely of altered colors were presented. Respondents were not informed about the fact of color change, nor were they informed about the study. The interpretation of the test was carried out programmatically on the first pass, solely for the moral encouragement of the respondents.
Data about changing colors and the order of choosing colors were recorded in a file.
The data consisted of 1726 passes by an unknown number of anonymous respondents who came to the test page through links posted on the Internet.
Interpretation of results
For each pair of colors, the real color distance was calculated in an equal-contrast space.
For each pass, the average color distance was calculated.
For each pass, the selection order discrepancy was calculated using the following algorithm:
For each color in each table, positions of the order of choice for the first and second passes were determined.
The difference between the position numbers was calculated.
Modulo position differences were summed up.
The result was divided by two.
In this way, two series of data were obtained: the average value of the deviation of colors along the passage and the corresponding differences in the order of selection of colors.
Diagram 1. Scatter plot of pass-averaged color distances in % of the color space axis and their corresponding changes in color selection order.
Spearman correlation coefficient 0
Coefficient of statistical significance p<.01>
Conclusions
The authors believe that a linear dependence of the change in the selection order on the magnitude of the color change can be assumed.
The authors believe that the standard polygraphic accuracy of color reproduction may be sufficient for the practical application of the Luscher color test.
The authors believe that the small value of the correlation coefficient suggests that computer implementations can be considered as a possible alternative to the traditional way of presenting stimulus material, although this does not mean that any computer implementation of the Luscher test will give the same reliable result as the original test. Of course, work is required to compare the colors proposed by Luscher and the corresponding colors that can be reproduced by the monitor, as well as to work out the presentation technique and test design.
The authors are planning an additional study, the purpose of which will be to determine the degree of influence of color distortions inherent in monitors on the order of color selection in the Luscher color test.