Step 1: Run regression of [scary Y] on [trendy X] using observational data
Step 2: Magically "adjust" for the 20 bajillion imperfectly-measured confounders
Step 3: Wait for a credulous journalist to write it up as causal
www.theguardian.com/science/2026...
Sophie E. Hill
Research from UCL suggests visiting art galleries or museums, singing and painting can help improve health outcomes
