Assessing the Effect of Language Demand in Bundles of Math Word Problems

Differential bundle functioning (DBF) analyses were conducted to determine whether seventh and eighth grade second language learners (SLLs) had lower probabilities of answering bundles of math word problems correctly that had heavy language demands, when compared to non-SLLs of equal math proficiency. Math word problems on each of four test forms (two at Grade 7 and two at Grade 8) were bundled together if they used the passive voice, conditional clause, relative clause, or a combination of any two. The results showed that the average total scores for SLLs was significantly lower than that for non-SLLs on each test form. However, only two bundles (passive voice and conditional clause at Grade 7 on Form 2) indicated statistically significant DBF against SLLs in favor of matched non-SLLs. An additional step was taken to determine whether the two bundles that showed statistically significant DBF against SLLs had biased the mean total scores for this group.The Walker, Zhang, Banks, and Cappaert (2012) procedure established for this purpose showed that this was not the case. Implications of the results are provided as well as suggestions for future research.