The construction of subjective brightness scales from fractionation data: a validation.

Abstract

An attempt has been made to validate the fractionation technic for the construction of sensory scales by statistical comparison of the results obtained from estimates of 1/2 and 1/3 as bright and 2 and 3 times as bright. The results for 24 Os, each of whom made each type of estimate, were subjected to an analysis of variance. Since the differences between the 1/2 and 1/3 estimates and between the double and triple estimates were found to be significant, the construction of scales based on these different ratios was possible. In addition, the order in which the different types of estimates were made, the brightness level, individual differences and the various interactions of these factors were found to be significant. The individual subjective scale values, obtained for each type of estimate by each O, were also treated by the analysis of variance. The variance due to individuals was again significant, but the variance for type of estimate was not significant. On the basis of these findings it is to be concluded that the estimation of fractional and multiple brightness is feasible and leads to consistent results for group means; the estimates are, nevertheless, very easily influenced by a large number of factors; individual differences are significant and in some cases may be quite large; even with this great variability, group estimates of 1/2 are significantly different from estimates of 1/3, and double estimates are significantly different from triple estimates; the various fractional and multiple estimates used in the present expt. lead to subjective scales which are not significantly different, hence the results are considered as a validation of the fractionation procedure for the construction of scales to represent sensory experience; and the subjective scale obtained from the present results is in essential agreement with the previous scale, and extends the range of physical brightness covered to an upper limit of about 500 millilamberts.