Why Most Health Studies are Wrong
Questions to Ask:
1. What type of study was it?
a. Observational: ecologic/ecological studies (Cheaper & Quicker but
b. Experiment: Expensive & slower but not confounded
2. What group does the result apply to?
a. For example, if you test a new drug, it is best to test it on people that, if
deemed successful, would take it.
b. People take results from a controlled subset of people and assume it
applies to everyone
● For example, in 2012 there were articles published that egg yolks were as bad as
○ The test group had an average BMI that was equivalent to being
overweight, so these people were already at risk of disease due to their
3. Did they measure the outcome or a proxy?
4. What is the probability this result occurred by chance?
a. A p-value expresses the probability that an event occurred by chance.
● Rank counties by kidney cancer rate
● Counties with the highest rates are:
○ Sparsely populat