Supplemental Methods
To explore fit more directly, we compared cumulative distributions of observed and
predicted total number of correct responses using the Kolmogorov–Smirnoff distance (K-S D)
function at specific (marginal) combinations of item and task factor levels (cf. Swaminathan,
Hambleton, & Rogers, 2007). K-S D gives the maximum difference between two cumulative
distributions and ranges from 0 (good fit) to 1 (poor fit).
Supplemental Results
The K-S D values produced by the final model are very good (Supplemental Figure 1). In
no instance did the cumulative distributions of observed and predicted total number of correct
responses significantly differ from one another.
Supplemental References
Swaminathan, H., Hambleton, R., & Rogers, H. J. (2007). Assessing the fit of item response
theory models. In. C. R. Rao &S. Sinharay (Eds.), Handbook of statistics 26:
Psychometrics (pp. 683-718). Boston, MA: Elsevier North-Holland, 2007.
Supplemental Figure 1. Observed and predicted cumulative proportion of total correct responses over examinees for specific
combinations of N-back load by item type factor levels for the final generalized linear mixed model. K-S D = Kolmogorov–Smirnoff
distance (0 = good fit; 1= poor fit).