This study investigated whether different reading strategies used in Sepedi predict higher levels of reading comprehension among Grade 7 learners. The instrument used was a ‘within’ test (in a school) and a ‘between’ test (between three schools) that was designed by the researcher in order to compare the effect of dialects and comprehension variables (recall, inference and anaphoric resolution) in three rural schools. Descriptive statistics and Analysis of Variance (ANOVA) performed on data elicited from 150 (n = 150) participants revealed low reading levels in the participants’ home language, suggesting that the readers were still struggling at a decoding level. Using Linguistic Difference Coding Hypothesis (LDCH), it was argued that the results did not only predict major reading problems in their later school years, but also highlighted reading difficulties in second languages. This study presented useful recommendations for Sepedi reading in particular and generally African language reading trajectories in higher grades in order to address African language reading literacy deficiency in primary schools.