1
$\begingroup$

This question already has an answer here:

Assuming that you hear that X affects 1% of the global population, is it safe to assume that it should affect 1% of the population of your country/city/subgroup?

Given that there is nothing in the subgroup to affect the distribution of X. (Assume that the distribution of X is not mentioned, so should we assume normal/Gaussian?*)

I'm asking because, often, I hear X disease is very rare, it only affects 1% of the global population, which sounds HUGE to me.

*if so, you could say that your country is an outlier and it's much less that 1%, but on average, you should be correct, right?

$\endgroup$

marked as duplicate by Community Sep 18 '16 at 17:37

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

  • $\begingroup$ Let a be 'Each individual has a 1% chance of X' and b be 'X affects 1% of the global population'. a implies b, but b does not imply a. Example: pick two equally sized groups, raise each members' chance of X in the first group by the same amount you lower it in the second. Now a is false, but b is true. $\endgroup$ – conjugateprior Sep 18 '16 at 15:58
4
$\begingroup$

Very few things are independent of world location. In particular in disease, if someone says that it affects x% of the world population, it is very likely localized. For example: most Malaria is confined to a latitude range (there isn't much in northern Europe for example). And many genetic diseases fall within populations with identical genetics - therefore they affect certain locations more than others. So, no.

$\endgroup$
  • $\begingroup$ Correct, but if you assume a disease that lacks a particular pathogenic factor, like a mental disease, could you say that it's true? Also, since most stuff are indeed not distributed evenly at all, isn't it a useless statistic most of the time? $\endgroup$ – K. Gkinis Sep 18 '16 at 15:47
  • 1
    $\begingroup$ If it lacks a genetic factor, it most likely has an environmental one (malaria example). Useless most of the time..? Probably. It's a useful summary (as opposed to saying x% for every country or region), but clearly an incomplete one. $\endgroup$ – Filipe Sep 18 '16 at 15:53
3
$\begingroup$

The accepted answer is good, but let me generalize the concept to make it clearer.

In order to assume that probability of a characteristic in a sample (e.g. country) is the same as probability in the whole population (e.g. world), the sample needs to be representative of the whole. Would it be valid to say "X% of country A -> therefore X% of global population"? (For some characteristics, that might be true, but not for most things.) If not, then it's also not valid to go the other direction and say "X% of global population -> therefore X% of country A".

$\endgroup$
1
$\begingroup$

There are too many counter examples that make your premise false. Asians have a far lower rate of diabetes than Americans, for example. Yet, a second generation Asian in US has similar rates. Not genetic, then, but habit, diet. It would surprise me if any disease had a mathematically 'nice' spread across the world.

To offer a money counter example, US median income is equal to about top 1% number for world income.

$\endgroup$
  • $\begingroup$ The question was more general than diseases (for example wealth), but the point about large differences in concentration, still stands. $\endgroup$ – K. Gkinis Sep 18 '16 at 17:37
  • $\begingroup$ I added a different example $\endgroup$ – JoeTaxpayer Sep 18 '16 at 17:43

Not the answer you're looking for? Browse other questions tagged or ask your own question.