What did IHDSC's Spencer-funded small conference on "non-cognitive predictors of children's academic achievement" accomplish? In preparing for the meeting, the invited participants addressed four key questions in brief format. These included (a) a list of the measures that the investigators currently use to assess children's self-regulation, relationships with teachers and attention/persistence, (b) discussion of developmental appropriateness of those measures, (c) discussion of past experience and concerns regarding the scaling of those measures, and (d) the issue of centering.
During the meeting, several partial solutions to the problems of scaling, centering, and standardizing metrics across different age periods were discussed. This discussion was led by Stephanie Jones. Jones introduced several options for measurement and modeling, using concrete examples of structuring data. Jones helped the group to consider ways that one measure collected at multiple time points across children's lives might serve as the "benchmark" in trying to establish a common metric or yardstick, across time. Reardon and others pointed out that Item Response Theory (or IRT) should ideally aid greatly in this effort to build a common metric across items that vary across time and developmental period. Additional discussion was carried by other members who have had extensive experience analyzing those data, including Duncan, O'Connor, and Bub. Additional methodological discussion was carried by Little and Reardon on the tradeoffs of different approaches to modeling change in children's non-cognitive skills over time.