Research in sub-Saharan Africa investigating the effect of mother tongue (MT) literacy instruction at medium scale is limited. A randomized controlled trial of MT literacy instruction was implemented in 2013 and 2014 as part of the Primary Mathematics and Reading (PRIMR) Initiative in Kenya. We compare the effect of two treatment groups-the base PRIMR program teaching literacy in English and Kiswahili, and the PRIMR-MT program, which taught literacy in English, Kiswahili, and mother tongue-in two different language environments. Implementation of the MT program faced challenges because many educators were not speakers of the languages, some communities resisted mother tongue instruction, and some areas were more language heterogeneous. Effect sizes on MT literacy averaged between 0.3 and 0.6 standard deviations. The base PRIMR program also increased MT learning outcomes in some measures, but had smaller effects than the PRIMR-MT program in oral reading fluency and comprehension. In addition, we estimated the effect of the PRIMR-MT program on the instructional time using MT as a language of instruction in other subjects, as well as the impact on other subjects learning outcomes. Results showed very modest changes in language usage in other subjects and negligible or negative impacts on learning outcomes in mathematics. These results present policy dilemmas for countries considering implementing mother tongue programs to improve literacy as opposed to countries considering mother tongue language of instruction programs to improve outcomes in other subjects.