Abstract
Slightly different wordings are known to introduce important differences in the way people understand and answer survey questions and, moreover, in the quality of the items (Billiet 1991). This is what may happen also in the case of two wordings used to measure the attitude people express towards the effect of the women’s job on their children. The aim of this study is to assess which of the two almost similar items, assumed to tap this kind of attitudes, produces a better measurement in terms of validity, reliability and overall quality. For this purpose I employ original data and use OLS models as well as SQP analysis. The findings reveal some surprising differences between the two items. This study starts with a short introduction, followed by the description of the method and the presentation of the findings. A short discussion concludes the text, focusing on implications for future studies.