EPI 320 Lecture Notes - Lecture 14: Ecological Fallacy
Document Summary
W8d1: case-control studies part 2 & ecological studies. A: sweden is case control b/c grouped on outcome, finish is retrospective cohort b/c grouped on exposure. Case-control definitions: cases, precise disease / diagnostic criteria, time of disease onset, population cases came from, no reference to exposure, controls, population controls came from / selection procedures / no reference to exposure. Multiple control groups: in some cases, studies include one case group but more than one control group, second control group may: answer slightly diff questions, strengthen study conclusion. Ecological studies: not a single study design, but family of studies, experimental, cross-sectional, case-control, cohort, unit of observation is the group, not the individual, can be of any size (households, states, cities, countries, in practice, often larger. When: only have summary data on exposures or outcomes, may have a question explicitly about a group-level e/o.
