To examine the quality of rhythmic improvisations after learning Ugandan folksongs via notated or aural/oral means, we asked university music majors (N = 32) to practice two Ugandan folksongs via Western notation or while viewing a prerecorded video of an expert Ugandan performer singing the same song to mimic aural/oral tradition conditions. Subsequently participants heard an authentic performance of the song they had just learned and were asked to create a rhythmic accompaniment to that song. All conditions were counterbalanced by treatment and by song. Resulting improvisations were judged regarding whether the first was better than the second. Results indicated no significant difference between improvisations on the basis of how the song was learned (p = 0.2617), differences between the songs themselves (p = 0.1261) or the order of the songs (p = 0.7518). Participants improvised better when the song was learned under notation conditions (differences not significant), but 70.1% of participants preferred to learn the song via aural/oral means (p = 0.0041). Results are discussed in terms of challenges in assessment of improvisations and pedagogical implications for future research.