The purpose of the present study was to examine the effects of a brief experimental intervention that integrated spelling practice into a systematic phonics approach to initial reading instruction for at-risk children. The effects of this intervention were studied by means of a randomized controlled trial design that compared the experimental condition to two trained control conditions and a further business-as-usual condition. The two trained control conditions were phonics-based interventions without spelling but with additional time spent on letter-sound practice. One emphasized letter-sound production, the other letter-sound recognition. Participants were 65 kindergartners with limited letter knowledge and no reading skills. Each participant was randomly assigned to one of the four conditions. Children were taught individually in four 20-min sessions in all three trained conditions. Analyses of the gains in abilities from pre- to posttest revealed that the integrated spelling condition was associated with significantly larger gains in phoneme awareness, spelling, and reading than were either the trained letter-sound recognition condition (d = 0.38–0.86) or the business-as-usual condition (d = 0.54–1.21). The results also favored the integrated spelling condition over the trained letter-sound production condition. Regarding the two trained control conditions, the letter-sound production condition was associated with slightly better reading and spelling outcomes than the letter-sound recognition condition. These findings indicate that integrated spelling may improve systematic phonics for children at risk of early reading difficulties, and that activities that encourage letter-sound production may be more beneficial than those which only require letter-sound recognition.