Anne Fleur Kortekaas-Rijlaarsdam

73 MPH EFFECTS ON MATH PERFORMANCE: INFLUENCE OF BEHAVIOR 3 outcomes, whereas previous work has suggested comparable efficacy of methylphenidate for children with and without comorbid learning disorders (Williamson et al., 2014). More research on this topic is necessary because in children with ADHD comorbidity with learning disorders is very high (DuPaul, Gormley, & Laracy, 2013) and academic problems are more stable in children with ADHD and comorbid learning disabilities (Willcutt et al., 2007). Although our results suggest that the effects of methylphenidate are at least partly dependent on ADHD diagnosis, future research could also shed light on the independent effects of methylphenidate on math performance for non-ADHD groups with learning challenges, although ethical concerns may be raised. Our results show no moderating role for ADHD symptom severity. Thus, medication efficacy is not larger for children with more ADHD symptoms and symptom severity is not a predictor of medication effectivity for academic improvement, which is in line with previous findings (Mattison & Blader, 2013). Limitations First, accuracy scores for mathematics on the SAT showed a ceiling effect, thereby limiting room for improvement in our sample. For the SAT accuracy we also analyzed only the 50% most difficult items per child, which also did not improve with methylphenidate. In addition we performed sensitivity analyses for all children (51%) that performed <95% accurate in the placebo condition. Also for this group the methylphenidate effect was non-significant (results available on request). Second, our trial had a limited duration (two weeks) and therefore has limited ecological validity and generalizability. Long-term placebo-controlled trials with treatment periods of at least two months allow for inclusion of school grades as measures of academic improvement and show higher correlations with both school dropout and college performance than standardized achievement tests (Allensworth & Easton, 2007; Zwick & Green, 2007) and may therefore be more predictive of future academic perspectives. However, long-term placebo-controlled trials are scarce because of ethical constraints and suffer from low medication adherence and high drop-out rates. Thirdly, medication dosage in our trial was identical to the clinical dosage prescribed by the treating physician. Because the majority of prescribing physicians uses classical titration procedures, this may possibly have resulted in lower methylphenidate efficacy in improving ADHD symptoms (MTA-group, 1999a). However, although methylphenidate-related improvements in academic accuracy and productivity may also be higher with controlled titration, we opted for this approach to maximize ecological validity and generalizability of our results. Further, we included primary school children only and although this limits the generality of the sample, we believe that the inclusion of primary school aged children is most relevant as the association