Most past research investigating the effects of contextual interference (CI) have included only quantitative (product) measures of skilled performance. The present study examined the effects on skill form of CI on the learning of the push pass, flick and dribble by fifth graders. Participants (n=232) 5th graders, were classified as “high” or “low” skilled, practiced the 3 skills using a blocked, progressive-blocked or a random schedule (representing low, medium or high CI levels). The form of the movements (qualitative analyses) was the dependent measure assessed by three expert judges. A 3 x 2 x 4 (contextual interference level x skill level x time) ANOVA (with p set at <. 05) with repeated measures on time was performed for each skill. Analysis of the push pass demonstrated a significant main effect for time but no significant CI level x time interaction. There was a significant interaction for time x skill levels. The high skilled group’s performance was superior to the low skilled group at all testing times. Finally no significant CI x skill x time interaction was found. For the flick data a significant main effect for time and a CI level x time interaction was obtained. No significant differences between the groups was found at the post-test, however, at short-term retention, the progressive-blocked group was significantly better than the other groups. At long-term retention, there were no significant differences between the progressive-blocked and the random group, although both these groups were more accurate than the blocked group. There was no significant skill level x time interaction, nor any CI x skill x time interaction. For the dribble (for speed) a significant main effect for time, and CI level x time interaction was obtained. Although there were no significant differences between the groups both at post test and short-term retention, the progressive-blocked group had better form than the other groups at long- term retention. No significant skill x time interaction as well as no significant CI x skill x time interaction was obtained. The findings of this study extend previous findings about CI, by demonstrating changes in qualitative components (form) and suggesting that teachers use progressive-blocked and random groups as preferential to blocked practice for long-term retention.