Hello, I’m Pang Ying.

This is our first Q&A. We will answer the three questions about correlation and causation left at the end of the previous episode. We will probably follow this pattern: introduce a concept in the main episode, leave questions for reflection and discussion, then explain them in the Q&A.

Our first question concerns estrogen levels and heart-disease incidence. Is their relationship causal, or merely correlational?

Scientists initially found data showing a positive correlation between low estrogen and high incidence of heart disease, which was first interpreted as causation.

If causal, low estrogen would cause higher risk, so changing hormone levels should lower it. Estrogen therapy was therefore used in hopes of reducing disease. It did not lower high heart-disease incidence, revealing that the two outcomes were not causal but arose from a third factor, C.

C was menopause, which both lowers women’s estrogen levels and raises heart-disease incidence. Estrogen level and heart disease were correlated, not causally linked.

Our second question asked whether five minutes of daily running can help us live longer.

Data show that people who run five minutes daily live longer. That sounds plausible, but interpreting it causally means that anyone seeking longevity need only run at least five minutes a day to gain a definite benefit.

In fact, this was only an observational study. It established no biological, chemical, molecular, or hormonal mechanism connecting five minutes of running to longevity. People active enough to run daily probably differ in many other ways from inactive people. Were those differences controlled? Was this truly an all-else-equal comparison with only one variable?

A common factor C may be strong concern for health and self-discipline. These traits help people run daily, but also help them sleep and eat well, avoid smoking, and limit alcohol. Running and longevity may merely occur together; the qualitative improvement may come from all those dietary, sleeping, and lifestyle habits.

If you do none of those things and only run for five minutes, will it effectively prolong life? That is doubtful. The study did not control these variables.