Sensitivity analysis for causal inference
Understanding how to conduct sensitivity analysis properly is essential for those aiming to strengthen their research findings. Let’s discuss this further.
How do we use sensitivity analysis?
Sensitivity analysis is pivotal for evaluating how findings from randomized trials hold up when the assumptions of the primary analysis are challenged [5]. However, there’s been a noticeable gap in guidance on selecting appropriate sensitivity analyses, resulting in their sporadic use despite their critical role in establishing causal inferences.
Morris, Kahan, and White (2014) [2] tackle this issue by outlining a principled framework for identifying meaningful sensitivity analyses through three essential inquiries:
- Does the sensitivity analysis explore the same question as the primary analysis?
- Could the sensitivity analysis potentially lead to a conclusion that diverges from that of the primary analysis?
- If differing...
