The present study compared the effects of different mediation routines provided to kindergartners from families of low socioeconomic status on the students' invented spelling attempts and on their gains obtained on spelling and other early literacy skills (letter naming, sounds of letters, word segmentation, and word decoding). The effects of the spelling mediation routines were assessed beyond the contribution of the students' self-regulation and baseline literacy levels. Participants ("N" = 197, mean age = 5.5 years) were randomly divided into four groups within each of 10 classrooms. Three groups underwent individual intervention sessions of inventing five words' spellings twice weekly for 16 weeks. Group 1 experienced process-product mediation, providing information both on the process of inventing spelling (sound-to-grapheme mapping) and on the product (naming the letters and spelling the word). Group 2 experienced product mediation only. Group 3 experienced spelling with no mediation. Group 4 (no intervention) experienced the regular kindergarten curriculum. Results indicate that the process-product mediation routine was most productive for all students, both in the short and long term, beyond their self-regulation and baseline early literacy levels. Students who started with higher self-regulation and poorer early literacy skills gained more from training. The results question the assumption, shared by previous studies, that promotion of invented spelling should be by providing students with one-step-up solutions to compare with those they produced, because the process-product mediation was most productive for students of a wide range of spelling levels. (Contains 4 tables.)