Re: a study from Canadian health researchers
"Until every single person on the planet has everything they do recorded and every health issue recorded this kind of statistical "study" is just a waste of money" --- Lusty
If that were true, and every datum needed to be collected, opinion polling, a significant amount of quality control and a huge amount of science would be invalid --- we'd still be unsure as to whether or not cigarettes were harmful.
IS2R that a randomly chosen sample of 1000 from a population of millions would give you a worst case 95% confidence interval of about 3 percentage points on an either/or survey question. This is obviously a more complex case, but if sampling is good enough, it is certainly possible to draw conclusions about the population with a high degree of confidence.