Ecological fallacy and Wikipedia

I really do like Wikipedia. Following up on my post about correction for attenuation, Wikipedia is also quite good on ecological fallacy (that you can’t make inference at the individual level from data observed at the population level).

Their example at the moment is nice too. Rich states voted for Kerry in the 2004 US presidential election – if we had just looked at this population level we might have concluded that richer individuals voted for Kerry (and this would be an example of ecological fallacy – where we make incorrect inference at the individual level from population level data). In fact, when explored at the individual level poorer people voted for him.

A hypothetical example from health services research could be a finding that at the hospital level a lower proportion of the overall hospital budget spent on cleaners was found in hospitals with lower patient experience scores – potentially leading us to the conclusion that patients have a better experience if their hospital is cleaner. However if we explore this further we might find that patients who are more unwell in general have a worse patient experience; and that in hospitals where sicker patients are cared for a larger proportion of the budget is spent on clinical care as the cleaning costs do not rise in proportion (and indeed these hospitals may still spend more in cleaning overall). Again it is incorrect to interpret a finding seen on the organisation level as a causal association.

For further reading any basic epidemiology text book gives a good introduction to the strengths and limitations of ecological studies.

This entry was posted in Stats and tagged . Group: . Bookmark the permalink. Both comments and trackbacks are currently closed.
  • The Cambridge Centre for Health Services Research (CCHSR) is a thriving collaboration between the University of Cambridge and RAND Europe. We aim to inform health policy and practice by conducting research and evaluation studies of organisation and delivery of healthcare, including safety, effectiveness, efficiency and patient experience.