This study examined the efficacy of a core kindergarten mathematics program, Early Learning in Mathematics (ELM), a 120-lesson program with four content strands: (a) number operations, (b) geometry, (c) measurement, and (d) vocabulary. The study utilized a randomized block design, with 129 classrooms randomly assigned within schools to treatment or control conditions. Measures of achievement were collected in the fall (pretest) and spring (posttest) in kindergarten and in the winter of first grade (follow-up). Although students in ELM classrooms did not differ significantly from students in control classrooms, we expected and found that the effects of ELM depended on students' initial achievement. Initial achievement moderated condition effects for Test of Early Mathematics Ability (TEMA) scores (p = 0.0039), but not the Early Numeracy-Curriculum-Based Measurement (p = 0.1887). We found no effects on follow-up first-grade scores from the Stanford Achievement Test Tenth Edition Problem Solving subtest (p = 0.9737) or Procedures subtest (p = 0.6336).