In the next step, we examined descriptive statistics and performed graphical analysis for these items to evaluate endorsement rates by age and study (see Figure 2). If all items represent a single factor, plots of endorsement rates should show similar developmental trends, each reflecting the developmental trend of the underlying factor. Similarly, study differences in endorsement rates should be consistent across item plots, reflecting study differences at the level of the factor. Dissimilar study trends in these item plots generally signal either that the item set is not unidimensional or that differential item functioning (DIF) is present. DIF (sometimes called factorial invariance) indicates that a given item does not reflect the underlying factor in the same way across individuals (e.g., at all ages or across all studies); in other words, DIF indicates that the endorsement of an item does not mean the same thing for all individuals, even if the item is worded the same way or has been harmonized to have an equivalent metric. Study DIF could arise due to differences in the harmonized items or simply due to
