The purpose of this study was to investigate the effects of an early numeracy intervention delivered by kindergarten teachers to students identified as having mathematics difficulties. A multigroup growth-modeling-with-random-assignment-to-intervention-condition design was employed. Thirty-two teachers were randomly assigned to the treatment or comparison condition. A total of 71 students participated in the study, 47 in the treatment group and 24 in the comparison group. Results indicated that the treatment condition students outperformed comparison students (g* = 0.99) and demonstrated statistically significantly higher scores on all proximal measures of early numeracy. Also, about 80% to 100% of the variance was accounted for at the student level. Performance on distal measures was less impressive, with no significant differences between groups; the effect size was 0.44. Teachers rated components of the intervention highly, reflecting strong teacher satisfaction.