Study 1 analyzed 11 children’s science programs. Consistent with research on children’s storybooks, science information was often presented in a manner that might be confusing for children. Specifically, one-third of science lessons were presented alongside scientific misconceptions, only half of which were explicitly refuted. Additionally, one-quarter of lessons depicted or described science concepts in an anthropomorphized manner. These lessons mainly occurred during animated program segments and were often taught by non-human characters. Study 2 examined the effects of exposure to these inaccuracies on children’s science learning. A total of 78 3- to 5-year-old children watched two clips about two different science topics. Children were randomly assigned to one of three exposure conditions that varied the presence and type of inaccuracy embedded in the show, or to a no-exposure control condition where children answered all post-test questions prior to watching the show. Results indicated that children in the anthropomorphic condition provided the most accurate open-ended science explanations and close-ended endorsements. Children in the factual condition showed some improvements over the no-exposure condition, though these improvements were less consistent across all outcomes than those in the anthropomorphic condition. Children in the refutation condition did not differ from children in the no-exposure condition on any outcome. No condition modified children’s tendency to include anthropomorphic statements in their explanations. Study 3 examined the effects of exposure to these same clips on the accuracy of parents’ science explanations to their children. An online Mechanical Turk sample of 141 parents of preschoolers were randomly assigned to the same four conditions used in Study 2. Afterward, they were asked to imagine that their preschool child had posed a series of questions about the content and to respond as they normally would. Results indicated that exposure to science television (regardless of condition) improved parents’ science explanations to their children by increasing the number of facts they mentioned and reducing the number of misconceptions they mentioned (relative to no-exposure). Path analyses indicated that these effects were largely driven by parents’ higher endorsements of science facts as accurate and their lower ratings of misconceptions as informative to children.