This paper reports data from the evaluation of the numeracy component of a long-running educational intervention, covering the period from 2001 to 2008. "QuickSmart" is both an intervention and research project operating in Australian schools. It is a structured intervention program designed for middle-school students (ages 10 to 13 years) with significant learning difficulties in basic literacy and numeracy. The program aims to increase fluency (automaticity) in the most basic skills that underpin proficient performance in reading and in mathematical calculation and problem solving. The guiding principle is that building fluency and confidence in basic skills enables students to devote much more cognitive effort to the higher-order processes involved in reading for meaning and in solving mathematical problems. The "QuickSmart" project includes a strong research component that investigates the effectiveness of the intervention through collection of targeted students' pre-test and post-test scores on standardised achievement tests. These "QuickSmart" students' scores are compared to the pre-test and post-test results of average-achieving comparison students who come from their same class groups. The data indicate that the intervention has significantly improved the numeracy performance of more than 2,000 students with learning difficulties from over 90 schools in Australia. Students' average improvement has been of the order of two years' growth as measured by effect size calculations. The improvement of "QuickSmart" students on standardized scores has also helped "narrow the gap" between their performance and that of average-achieving students. The evaluation of this program across differing contexts indicates the importance of consolidating students' basic academic skills and providing ongoing support to instructors through a well-designed professional development program. (Contains 6 tables.)