Theft reduction in a grocery store through product identification and graphing of losses for employees

Shoplifting and employee theft constitute a major problem for retailers. Previous research has described techniques for effectively reducing either type of theft but has not addressed the problem of thefts of unspecified origin. In a grocery store we evaluated the effect of identifying for employees frequently stolen products from three groups of items and graphing, twice weekly in the lunchroom, losses for the separate groups. After the products were identified and losses graphed, thefts from the three groups dropped from eight per day to two per day.     

THREE MODELS FOR THE ANALYSIS OF SUBJECTIVE RATIOS

SJÖBERG, L. Three models for the analysis of subjective ratios. Scand. J. Psychol. , 12, 217–240.–Circular area and heaviness were judged under four response modes: free ratio, per cent, similarity, and multiple estimation. Three log linear models were suggested for these data, each new model bringing in more free parameters to be fitted and assuming less strict types of invariance of subjective values. The most demanding, and traditional, model frequently had to be rejected. This "inconsistency" could be attributed both to the factor of relative size of the two stimuli in a comparative judgment and to the fixed vs. free factor. As a byproduct it was found that the Eisler-Ekman "similarity equation" was not supported by these data.     

THE DIMENSIONALITY PARADOX IN COMPARATIVE JUDGMENT: A RESOLUTION

S jöberg , L. The dimensionality paradox in comparative judgment: a resolution. Scand. J. Psychol ., 1968, 9, 97–108. — It has often been possible to erect unidimensional scales although stimuli have been clearly multidimensional. This fact is surprising since one expects the multidimensional structure to be reflected in data. It is suggested that this structure might be recovered from dispersions of comparative judgments. The suggestion is shown to be valid on three sets of paired comparison data. Implications for uni- and multidimensional scaling are discussed.     

Successive intervals scaling of paired comparisons

In this paper a method is developed for analyzing data resulting from the use of a modification of the usual paired comparisons procedure which allows for ratings of size of difference. The scaling model is essentially an extension of the Thurstonian successive intervals model. The method is applied to scaling of a set of nine immoral actions and the results agree rather well with those from conventional successive intervals scaling of the same stimuli.This study was supported by a grant from the Swedish Council for Social Science Research.