Twenty percent of households in a Louisiana town of 11,000 are randomly selected, then people in these households were interviewed and their blood serum tested. From those selected, women aged 45 and over are categorized as having high cholesterol versus normal or low cholesterol results. It was found that 17% of the women in the high cholesterol category have coronary heart disease as well, as compared to 5% of those in the normal or low cholesterol category.

1. What type of study design is this?
2. What is the risk ratio of heart disease with high cholesterol comparted to normal/low cholesterol?
3. What possible selection bias may exist?
4. Would your type of selection bias over-estimate or under-estimate your answer to part B?