ABSTRACT
A Thurstone score based on the median position of each respondent is essentially a limen score indicating the subject's central response tendency towards the attitude. The Likert method of scoring by summing the responses in each category has commonly been found to yield higher reliability coefficients. C. I. Mosier investigated the relationship between the limen score and the summation score of a mental test, and found the reliability of the first to be 0.88 compared with 0.94 for the second. The coefficients yielded by our attitude scale are closely comparable. The lower reliability of true Scale—Product scoring can probably be explained, by the fact that the Thurstone scale values of items in parallel forms are close but not identical, and multiplying them by response weights exaggerates the differences. The method of weighted proportions also provides a good indication of validity.
