The article presents a reply to questions raised by Allport and Kerler (A&K) in their research note about theory and data in scale development. With the objective of creating a single scale consistent with an a priori construct definition, we choose principal components analysis as a means for initial data reduction. However, the study was indeed designed to have an initial set of items useful for data reduction or scale purification, as opposed to running tests to immediately suggest valid measures. An author suggested that the only way to evaluate the psychometric properties of the responses to rating scales with both positively and negatively worded items would be to use confirmatory factor analysis and structural equation model methods. Besides wording effects, A&K suggested that a response bias effect based on positive or negative framing might well be another possibility. To aid model improvement, the modification index for a parameter is an estimate of the amount by which the discrepancy function would decrease if the analysis were repeated with the constraints on that parameter removed.