A new US survey shows shocking numbers of people reporting side effects from the Covid gene therapy. How might these data be interpreted, and what could it mean for the future of public health policy?
Hi Alan - it is counterintuitive. But probability theory indicates that yes, a sample of 1,292 can reasonably represent the adult US population, producing a margin of error of about ±3% at 95% confidence. Beyond a certain point, the required sample size doesn’t scale with population size. So surveys with ~1,000 respondents can accurately represent huge populations.
(N.B. Sample size isn’t the only factor; good sampling methods and appropriate weighting to correct for demographic imbalances are just as important.)
Thanks for the reassurance. I'm just getting over a nasty bout of computer modelling and not fully recovered. Yes, sampling is extremely important and how the respondents were chosen could affect the outcomes.
Not that it seems to matter. There's been a tsunami of evidence of adverse outcomes but the jabbing never stopped. The article by Paula Jardine in yesterday's TCW shows the build up to bird flu which will be coming to greet us as soon as all the plans are in place.
“…getting over a nasty bout of computer modelling and not fully recovered”. I’m sure many can relate to that!
Thanks for raising the question of statistical extrapolation and large populations. It’s the kind of thing people do wonder about, so it’s good we could address it here.
"The Rasmussen data, collected in November 2025* via a national US survey of 1,292 individuals".
Is it reasonable and mathematically valid to sample just 1,292 out of 200+ million people and multiply the results accordingly ?
Hi Alan - it is counterintuitive. But probability theory indicates that yes, a sample of 1,292 can reasonably represent the adult US population, producing a margin of error of about ±3% at 95% confidence. Beyond a certain point, the required sample size doesn’t scale with population size. So surveys with ~1,000 respondents can accurately represent huge populations.
(N.B. Sample size isn’t the only factor; good sampling methods and appropriate weighting to correct for demographic imbalances are just as important.)
Thanks for the reassurance. I'm just getting over a nasty bout of computer modelling and not fully recovered. Yes, sampling is extremely important and how the respondents were chosen could affect the outcomes.
Not that it seems to matter. There's been a tsunami of evidence of adverse outcomes but the jabbing never stopped. The article by Paula Jardine in yesterday's TCW shows the build up to bird flu which will be coming to greet us as soon as all the plans are in place.
“…getting over a nasty bout of computer modelling and not fully recovered”. I’m sure many can relate to that!
Thanks for raising the question of statistical extrapolation and large populations. It’s the kind of thing people do wonder about, so it’s good we could address it here.