KonFound-It!

Sensitivity analyses that quantify the robustness of inferences to concerns about omitted variables and other sources of bias.

Questions? Issues? Suggestions? Reach out through the KounFound-It! Google Group.

Latest News

Time to look ahead to the next workshop: “What would it take to change your inference? Quantifying the Discourse about Causal Inferences in Epidemiology” (Modality: virtual on July 24, 2024) as part of the Society for Epidemiological Research annual meeting.

Description:

Causal inferences are often challenged because of uncontrolled bias. We will turn concerns about potential bias into questions about how much bias there must be to invalidate an inference. For example, challenges such as ‘But the inference of an exposure might not be valid because of pre-existing differences between the treatment groups’ are transformed to questions such as ‘How much bias must there have been due to uncontrolled pre-existing differences to make the inference invalid?’ By reframing challenges about bias in terms of specific quantities, this workshop will contribute to scientific discourse about uncertainty of causal inferences. Critically, the approaches presented in this workshop based on correlations of omitted variables and the replacement of cases have strong intuitive appeal.

In part I, we use Rubin’s causal model to interpret how much bias there must be to invalidate an inference in terms of replacing observed cases with counterfactual cases or cases from an unsampled population. This is extended to logistic regression and hazard models. In part II, we quantify the robustness of inferences in terms of correlations associated with unobserved variables. Calculations will be presented using the open-source app https://konfound-project.shinyapps.io/konfound-it/ with links to R and Stata modules. The format will be a mixture of presentation and individual hands-on exploration as well as group work.

Browse through a variety of resources and guides.