Two field observation studies conducted with American and Chinese students, aged 8 to 11, were used to examine developmental and cultural patterns in academic risk-taking (i.e., student selection of academic achievement tasks varying in difficulty) and to formulate hypotheses pertinent to this phenomenon. Data led to the following tentative conclusions: (a) Sex differences in academic risk-taking and failure tolerance are trivial. (b) Failure tolerance decreases with development. (c) Academic risk-taking is low relative to the theoretically optimum risk level of.50. (d) Developmental patterns in academic risk-taking vary with situational factors. (e) Academic risk-taking varies with content. (f) Academic risk-taking tends to be higher for American students than for Chinese students and higher for Chinese students from industrial settings in contrast to government-employment settings. Three hypotheses were formulated to explain the field observations: the variable payoff hypothesis, the accuracy-difficulty judgment hypothesis, and the external constraint hypothesis.