Algebra is considered a key gatekeeper for higher-level mathematics course-taking in high school and for college enrollment (Adelman, 2006; Gamoran & Hannigan, 2000). Yet, algebra pass rates are consistently low in many places (Higgins, 2008; Ham & Walker, 1999; Helfand, 2006), including Chicago Public Schools (CPS). This is of particular concern because academic performance in core courses during the first year of high school is the strongest predictor of eventual graduation (Allensworth & Easton, 2005). Offering online credit recovery courses is one strategy to deal with high failure rates. However, no rigorous evidence currently exists about the efficacy of online credit recovery courses. Understanding patterns of treatment effects may provide clues to the relative strengths and weaknesses of online and standard face-to-face (f2f) learning. A related policy question is whether district and school administrators should target online learning to certain students. This paper investigates these questions by exploring heterogeneity in the treatment effects of online algebra credit recovery using Chicago Public School students who failed second semester Algebra I in the spring of freshman year, and attempted credit recovery as part of the study in summer 2011 or 2012. [For a related abstract, see ED562834.]