Optimizing the 0-100 Number Line Estimation Task: Scale Reduction and Its Implications for Elementary Mathematical Cognition [Author Accepted Manuscript]

We investigate the optimal number of items for the 0-100 number line estimation task used in research on children’s mathematical cognition and learning. In this paper, we reanalyzed data involving N=234 students, applying an Item Response Theory- Graded Response Model to identify items with high discrimination parameters (>1.0), iteratively reducing the 23-item scale by including items with discrimination values close to 1.0 until the reduced scale produced comparable scores to the original. Our analysis identified a reduced scale of 15 items that maintained strong correlations with–and produced consistent patterns of developmental change and predictive capability compared to–the original scale. Our findings demonstrate that a reduced 0-100 number line estimation task can effectively measure numerical magnitude understanding (accuracy and linearity of estimates) from kindergarten through third grade while saving time and resources.