We examined students' naturalistic decisions about spacing their study in an undergraduate course (N=185) and whether self-selected spacing predicted course performance. Usage of two study tools -- an online textbook and quiz tool -- was recorded daily. We operationalized spacing as how often the tools were used and the timing of their use relative to exams. We found that students increased their study near deadlines and exams, used the textbook more often than the quiz tool, and used the tools infrequently when they were optional (vs. required). Importantly, spaced retrieval practice (via quiz tool) predicted course performance and GPA, whereas spaced reading (via textbook) was a weaker predictor. That is, when students opted for more frequent and early quizzing, they earned higher grades, even controlling for time spent quizzing. Thus, self-selected spaced study -- especially spaced retrieval practice -- supports student achievement. [This is the in press version of a paper published in "Learning and Instruction."]