Before any compound, a method. These five chapters give you one.
The chapters
The Question Nobody Will Answer
From Dish to Mouse to Person
What a Real Trial Looks Like
The Six Ways a Cancer Claim Goes Wrong
The Evidence Grade
The method, as something you can run
These five chapters are the method the rest of the book applies. Two of
them are on this site as tools rather than as reading, which is the only way a
method is any use at two in the morning on a claim nobody has graded.
The six-question procedure from chapters four and five, as a decision tree. You walk a claim you met online through it and it returns the highest grade that claim could earn on the evidence behind it. It cannot tell you to take anything; the only thing it can do is take confidence away.
Chapter three, as a lookup. What kind of study is it, and what is the most that kind of study could ever show? A mouse study returns D however dramatic the result, and a consensus statement returns B however many trials stand behind it, because the trials are not the thing in front of you.
