Ninth, by restricting our effect modification analyses to features that were significantly associated with exposure, we missed the opportunity to investigate features that did not reach statistical significance and we may overestimate the effect of features selected given that we had previously analyzed the feature in order to decide on its inclusion for further analysis.
← all excerpts
Linking Air Pollution Exposure to Blood-Based Metabolic Features in a Community-Based Aging Cohort with and without Dementia.
1
—
—