A further difficulty relating to scale construction is the imbalance in the number of
true and false items. In the L scale, all the items are scorable if answered "False"; on the K scale, 29 of 30 items are scored if answered "False"; and Scales 7>,8, and 9 have a ratio of approxirnare.ly 3 to 1 of true compared with false items. The danger of this imbalance is that persons having response styles of either acquiescing ("yea-saying") or disagreeing ("nay-saying") may answer according to their response style rather than to the content of the items. A more theoretically sound approach to item construction would have been to include an even balance between the number of true and false answers. Some authors (A. Edwards, 1957, 1964; D. Jackson et al., 1997) have even suggested that test results do not reflect psychological traits as much as generalized test-taking attitudes. Thus, a controversy has arisen over "content variance," in which an examinee is responding to the content of the items in a manner that will reflect psychological traits rather than «response style variance," in which responses reflect more the examinee ’s tendency to respond in a certain biased direction. In a review of the literature, Koss (1979) concluded that, although response sets can and do exist, the examinee’s tendency to respond accurately to the item content is far stronger. The MMPI-2 restandardization committee has also developed the Variable Response Inconsistency (VRIN) and True Response Inconsistency (TRIN) scales to help detect invalid profiles caused by inconsistent or contradictory responding. These scales have been specifically designed to detect either response acquiescence or response nonacquiescence and thus should help counter the potential complications resulting from imbalanced keying.