Ecological Studies Units of analysis are usually geographically defined populations example: the correlation between alcohol outlets and violent crimes) Example: the association between “broken windows” (a measure of social decay) and sexually transmitted disease rates)
Examples Aggregate measure – summarizes characteristics of individuals within a group. (eg. mean income, percent male, education level from census track data). Environmental measure – physical characteristics of the geographic (e.g. location percentage of homes that are blighted, percentage of of schools without after-school programs). Global measure – characteristics that are not reducible to the individual level Aggregate – Environmental – (e.g. drinking age laws, social security benefits)
Difficulties with interpretation of ecological studies No temporality Can’t really adjust for confounders because the analysis is on the aggregate level Susceptible to ecological fallacy or aggregate bias – when the association observed on the aggregate level does not necessarily represent the association that exists at the individual level.
Cohort Study A defined population is identified, persons are categorized by exposure then followed to an outcome of interest. There is a comparison of the rate of the outcome among those who are exposed compared to those who are not exposed. (measure of association – relative risk) Persons who are lost to follow-up are called censored or withdrawls