Research suggests that there is, at least, a risk that problematic viewpoints about rape (and other serious sexual offences) (rape myths) are influencing decisions made by jurors in cases involving alleged sexual offending. Research has examined whether any impact of such viewpoints can be minimised through providing correct information to jurors, either through judicial directions or expert testimony. However, findings in this research appear inconsistent, and while existing evidence syntheses have noted this inconsistency, they have not explained it. In this paper, we adopt a new approach to evidence synthesis, integrating evidence relating to mechanism (e.g., mechanisms of belief updating) with literature directly examining relationships between interventions and decisions in the mock jury context, in order to explain apparent inconsistency in the existing literature and draw robust conclusions in relation to when interventions are likely to be effective. We find that short and superficial statements of fact are unlikely to influence juror decision making, and, as such, potential interventions tested in much of the existing associative literature are impoverished versions of the optimal intervention. However, we also find that interventions have the potential to be reliably effective when they include more detailed information, particularly when they target broad schemas of rape myths rather than individual beliefs. These findings have the potential to inform more effective intervention in the trial context to reduce the likelihood of rape myths feeding through into legal decision making. As such they may be helpful in improving the effective prosecution of rape and serious sexual offending.