Teaching Confidence Intervals: Problems and Potential Solutions

There are benefits of teaching inference via confidence intervals (CIs), rather than null hypothesis significance testing (NHST). However, CIs are not without misconceptions. First, we provide empirical evidence that CI presentations of data can help alleviate some typical misinterpretations of r esults, leading to more accurate conclusions and more justified decisions. Second, we demonstrate that CIs are also prone to parti cular types of misconceptions. Finally, we present interactive figures and simulations that, when used with guidelines for C I interpretation, should lead to more insightful interpretations of research results and fewer misconceptions . We are studying CIs because they are the focus of reformers who are keen to reduce emphasis on NHST. Difficulties in teaching NHST are well known, and misconceptions about