Standardized tests often group items around a common stimulus. Such groupings of items are called testlets. The potential dependency among items within a testlet is generally ignored in practice, even though a basic assumption of item response theory (IRT) is that individual items are independent of one another. A technique called tree-based regression (TBR) was applied to identify key features of stimuli that could properly predict the dependence structure of testlet data. Knowledge about these features might help to develop item sets with small testlet effects. This study illustrates the merits of TBR in the analysis of test data.