The opinions of the US Supreme Court are central to volumes of research on law, courts, and politics. To understand these complex and often-lengthy documents, scholars frequently rely on dichotomous indicators of opinion content. While sometimes appropriate, for many research settings this simplification of opinion content systematically omits important information. Using all US Supreme Court opinions from 1803 to 2010 in association with structural topic models, I instead demonstrate the value of representing the Court’s attention in opinions in terms of topic proportions.