Anne Fleur Kortekaas-Rijlaarsdam

63 MPH EFFECTS ON MATH PERFORMANCE: INFLUENCE OF BEHAVIOR 3 for randomization using predefined randomization blocks to determine medication or placebo sequence. Randomization sampling occurred without restrictions. Due to the low drop-out risk randomization occurred without replacement. Researchers, children, parents and teachers were blinded to intervention order. Materials Academic performance measures: For all academic measures, parallel versions were used for the two measurement weeks to avoid learning effects. Mathematical accuracy and productivity were assessed using the Speeded Arithmetic Test (SAT; (De Vos, 1992)) and mathematical word problems (Janssen et al., 2010). The SAT is a commonly used task in research on mathematics an dyscalculia, measuring automation of mathematical calculations (Ghesquière & Ruijssenaars, 1994; Pieters, Desoete, Van Waelvelde, Vanderswalmen, & Roeyers, 2012). The SAT contains 200 mathematical sums of increasing difficulty. Five columns of 40 sums each are presented to the child, with each column containing addition, subtraction, multiplication, division and mixed sums requiring addition, subtraction, multiplication or division. Children were required to solve as many sums as possible within 5 minutes without taking notes. Productivity (percentage completed) and accuracy (percentage correct from items completed) were used as dependent variables. The SAT allows older children to perform better than younger children by completing more calculations within the given 5 minutes. Reliability of the SAT, as calculated in the current sample, was adequate (KR-20 = .68) and SAT accuracy and SAT productivity correlated significantly with math grades ( r = .33 and r = .28, respectively) and intelligence ( r = .25 and r = .34, respectively), supporting validity of the SAT. Mathematical word problems were mathematical exercises described in a story-format, often accompanied by a drawing. These mathematical word problems are typical for primary school math education. Different versions are available for different educational levels as developed by the Dutch pupil monitoring system (LVS, CITO). Children were required to complete as many exercises as possible in 5 minutes, with a maximum of 12 exercises. Reliability of this measure is high, with Cronbach’s alpha varying from .91 to .94 for different versions. Correlations with other mathematical tests are high ( r =.94-.98) in comparison to correlations with less related academic subjects such as reading ( r =.67), supporting validity of mathematical word problems (Janssen et al., 2010). For mathematical word problems, accuracy and productivity were used as dependent variables. Technical reading productivity was assessed with the One Minute Reading test (OMR; (Brus & Voeten, 1999)) The OMR consists of 116 real words with increasing length and difficulty. Participants were instructed to read the words out loud as accurately