We study misunderstandings of a widely used research graph that visualizes how parenthood affects gender inequality in the labor market. The child penalty graph typically displays overlapping trend lines for women and men before parenthood and a sharp divergence thereafter. A large survey experiment asks participants about the gender gap in earnings before and after random assignment across three alternative graph designs. Approximately one-third of respondents misinterpret the distance between the trend lines in the child penalty graph as the gender gap in earnings. After viewing the graph, nearly 50% of respondents (and 60% of those with a PhD) report an earnings gap of 0–5% before parenthood, far below the true gap of about 20%. We examine two consequences of this misunderstanding and assess whether alternative graph designs improve comprehension.