The purposes of this study were to replicate previous research on phonics-based tutoring in kindergarten and to compare treatment effects for students who received individual instruction compared to instruction in dyads. Thirty classroom teachers from 13 urban elementary schools referred at-risk students for participation. Students who met screening criteria were quasi-randomly assigned, within classroom, to one of three conditions: individual tutoring (n = 22), tutoring in dyads (n = 32), or no tutoring (n = 22, classroom instruction only). Twenty-one paraeducators provided 18 weeks of explicit instruction in phonemic skills and the alphabetic code to students during the latter half of kindergarten. Multilevel model results showed that tutored students outperformed non-tutored controls on posttest measures of phonological awareness, word reading accuracy, oral reading fluency, spelling, and comprehension. However, no significant differences were found between the two tutored groups on any measure, suggesting that code-oriented tutoring for pairs of students is a viable alternative to the gold standard of individual instruction.