Following up on suggestions that attitudes toward science are multi-dimensional, we analyze nationally representative survey data collected in the United States in 2014 (N = 2006), and demonstrate the existence of a science confidence gap: some people place great trust in scientific methods and principles, but simultaneously distrust scientific institutions. This science confidence gap is strongly associated with level of education: it is larger among the less educated than among the more educated. We investigate explanations for these educational differences. Whereas hypotheses deduced from reflexive-modernization theory do not pass the test, those derived from theorizing on the role of anomie are corroborated. The less educated are more anomic (they have more modernity-induced cultural discontents), which not only underlies their distrust in scientific institutions, but also fuels their trust in scientific methods and principles. This explains why this science confidence gap is most pronounced among the less educated.