Training abstractive summarization models typically requires large amounts of\ndata, which can be a limitation for many domains. In this paper we explore\nusing domain transfer and data synthesis to improve the performance of recent\nabstractive summarization methods when applied to small corpora of student\nreflections. First, we explored whether tuning state of the art model trained\non newspaper data could boost performance on student reflection data.\nEvaluations demonstrated that summaries produced by the tuned model achieved\nhigher ROUGE scores compared to model trained on just student reflection data\nor just newspaper data. The tuned model also achieved higher scores compared to\nextractive summarization baselines, and additionally was judged to produce more\ncoherent and readable summaries in human evaluations. Second, we explored\nwhether synthesizing summaries of student data could additionally boost\nperformance. We proposed a template-based model to synthesize new data, which\nwhen incorporated into training further increased ROUGE scores. Finally, we\nshowed that combining data synthesis with domain transfer achieved higher ROUGE\nscores compared to only using one of the two approaches.\n