Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[HateLAPELoveSTEM]

Title.
And I heard that the heights above ±3SD deviate from the normal distribution. I dunno why.
I used the height percentile calculator to count the percentile of 185cm in American white males and the result showed '86%', which means 185cm in America is taller than 86% or the American white males. Does it stick with the fact?

[Exxon]

Title.
And I heard that the heights above ±3SD deviate from the normal distribution. I dunno why.
I used the height percentile calculator to count the percentile of 185cm in American white males and the result showed '86%', which means 185cm in America is taller than 86% or the American white males. Does it stick with the fact?

I guess it's true, I've been to the US and around 85% of the men I saw were less than 185cm. Around 25% of the men I saw while walking was below 170cm, 35% below 173cm and 50% below 175cm. This however varies between regions because the Midwest probably has an average of 179cm as it has high percentage of Scandinavian DNA in their local population so being 170cm there would probably mean you're only taller than 10% of men whereas in California the average is probably 172cm-173cm as there is a lot of Asian and Latino population as well as people from other immigrant backgrounds.

Edit: Also a lot of white Americans have Greek and Italian DNA and they tend to be shorter on average than other Europeans so it's not surprising to see that 85% of white men are below 185cm. Median white man (50th percentile) is probably 178cm but as soon as you add 1cm the percentiles increase like a rocket hence why 185cm is such a high percentile.

[Medium Drink Of Water]

Title.
And I heard that the heights above ±3SD deviate from the normal distribution. I dunno why.

Human height is subject to the normal distribution.  3SD is not the limit of the normal distribution.

[zaozari]

Human height is subject to the normal distribution.  3SD is not the limit of the normal distribution.

It doesn't deviate. The problem with more than 3 SD is that because the variable (individuals height) is so rare above that, that it is subject to an enormous error margin; the values are not to trust. If you express in percentage of population you can  have for example either 0,1 % or the double, 0,2%...
For example if you have 150 cm (3,5 standard deviation), male, USA,  you are theoretically (mathematically) in the 0.024 Percentile but in reallity this value is most probably not exactly true, far from that (calculator in Tall.com).

[HateLAPELoveSTEM]

Human height is subject to the normal distribution.  3SD is not the limit of the normal distribution.

Normal distribution can eradicate some girls' illusions that there are no men under 185cm or they are not men. Those are btches beyond crazy TBH

[HateLAPELoveSTEM]

I guess it's true, I've been to the US and around 85% of the men I saw were less than 185cm. Around 25% of the men I saw while walking was below 170cm, 35% below 173cm and 50% below 175cm. This however varies between regions because the Midwest probably has an average of 179cm as it has high percentage of Scandinavian DNA in their local population so being 170cm there would probably mean you're only taller than 10% of men whereas in California the average is probably 172cm-173cm as there is a lot of Asian and Latino population as well as people from other immigrant backgrounds.

Edit: Also a lot of white Americans have Greek and Italian DNA and they tend to be shorter on average than other Europeans so it's not surprising to see that 85% of white men are below 185cm. Median white man (50th percentile) is probably 178cm but as soon as you add 1cm the percentiles increase like a rocket hence why 185cm is such a high percentile.

But I dunno why, for instance, one place's average height is 173cm, but if you are 172cm then you are short in that place while if you are 174cm then you are not tall, very absurd. Seems like the standard of tall is higher implicitly than that of short.

[HateLAPELoveSTEM]

It doesn't deviate. The problem with more than 3 SD is that because the variable (individuals height) is so rare above that, that it is subject to an enormous error margin; the values are not to trust. If you express in percentage of population you can  have for example either 0,1 % or the double, 0,2%...
For example if you have 150 cm (3,5 standard deviation), male, USA,  you are theoretically (mathematically) in the 0.024 Percentile but in reallity this value is most probably not exactly true, far from that (calculator in Tall.com).

Is there any scientific paper that has proven the normal distribution will go out of line with the actual situations when the variable goes beyond ±3SD? I think Gauss mentioned this strange phenomena in the statistics in his writings or scripts or something.

[zaozari]

Look, contrary to others, I always put things here only after checking (evaluated subjectively sometimes, is impossible in another way) and of course I am far from reading only scientific magazines. But I studied Statistics among other things in my degree.  A normal distribution will allways be a normal distribution with its laws. But in some matters like biology, things, specially in extremes, don't allways follow Gaussian distribution exactly. Imagine a disease in which for example brain size after a certain value, is lethal. Maybe, just maybe,  there are not people above 2 or 3 SD in the left extreme only,  because that is incompatible with life.
The problem with stature is that samples never approach all population, so, final results for very minority groups have artificial, random deviations. Imagine 0,01% of the population.  This is allways mathematically correspondent to 1 in 10.000 individuals. If the sample is 1000, you have 90% chance of NOT finding any people shorter than those 0,01 % population.
That's one of the reasons also why more than 2 SD ("officially" "very very" "short" people in either sexes) is usually less than 2,3% of population in some countries and for example 2,5% in others (a different reason is simply that a population may be taller than the other). And it may be OR NOT an exact normal distribution in the extremes. For biological and even random reasons in a poor country for example, "very, very" tall people may be less than "very, very" short people in the other extreme.

[zaozari]

By "officially short" I mean the allmost unanimous concept according to scientists and doctors for being called "constitutional short stature" (it's not a disease), and that's why it is different between different populations.
There isn't a standard correspondent for tall people agreed unanimously, so it's more subjective to call someone "tall", altough by analogy you can say also being around average within 1 SD, as having height "around average",  "normal", to those with more than 1 SD as "slightly tall" to some and to others "tall ", and those with more than 2  SD as "very tall" or "extremely tall" (It would mean around more than 190,74 cm in the USA=6,26 feet). I know it may come has a surprise for many of you wanting LL but of course our feelings and personal views may be different than statistics. (Altough I would never, ever, risk LL if I measured the average value).
In parallel and subjectively, you can call "normal stature" values within 1 SD, "slightly short" for some and  "short" for others if within 2 SD, and "extremely short" (Constitutional Short Stature) those with more than 2 SD (but allways healthy).
Then, there is dwarfism and gigantism but these are statures with associated diseases.
Dwarfism is itself a disease, in some countries criteria include being less than 142 cm for girls for example.
But diseases causing these statures maybe several.

[zaozari]

Title.
And I heard that the heights above ±3SD deviate from the normal distribution. I dunno why.
I used the height percentile calculator to count the percentile of 185cm in American white males and the result showed '86%', which means 185cm America is taller than 86% or the American white males. Does it stick with the fact?

If your numbers are correct, 86% of "white american males" are 185 cm or LESS indeed. 
But it seems to me that 185 cm are in the 89 percentile and slightly above 1 SD (1,2 SD).

It's logic if you check the US average (175,9 cm in males). (Don't know if just "white". In my country people are fortunately very "mixed" and rarely we do statistics based on skin color (sometimes medical research does it by geographical origin for genetic studies because otherwise people with Japanese parents living in the US are taller than those born in Japan for example). On the other hand they are more prone to heart disease maybe because they eat less fish, etc.))

[zaozari]

And yes, 1 cm makes a lot of difference statistically the taller you are. If you are 191 cm ("vary tall") you are in the 97,8 percentil (wich is allmost again, like for short people (2 3%), only 2,2% (100-97,8) of population. The 0,1 difference may be either because I am not allways including millimeters or the gaussian distributation starts the deviation in the extremes.

[Elhemioe]

Nvm I read it wrong, you're right. Most American men are somewhere around 180.

[Medium Drink Of Water]

Your objection to SD's larger than three seems to be based on the increasing likelihood that sufficient sampling is not practical past them.

The normal distribution still exists even if you have sampling difficulties.  There are maximum and minimum limits to pretty much everything so the normal distribution will break down in the extremes eventually almost every time you apply it.

I think you're being too demanding in your standards.  It's an appropriate tool for conceptualizing human height.

[zaozari]

You are right, (I talked about "mathematical" gaussean and "biological" gaussean) and I believe too it's a great tool for studies,  research, medical and anthropological evaluation, etc, in fact maybe the single one for its purpose. I don't see much difference between your post and what I tried to explain. I don't criticise the technique even past 3 SD. We just have to know how to fully interpretate it. Often in some areas, the number of outliers may be exaggerated.
I am not being demanding, for me in practical terms this is irrelevant, if the 3 SD "delivers" a true or false value  (comparing to reality) of 0,01 % or 0,02 % percentiles. It is a technical aspect of something that is much more practical, psychological, sociologic, personal, etc.

[zaozari]

BTW, that "wording" for less or more than 1  2 or 3 SD is not mine, it's the usual approximate terminology used in science for fuman height.
Some countries prefer to consider for example 160 cm or less for adults without any concomitant disease (even if not corresponding to 3 SD) to define "Constitutional Short Stature".  This has practical implications in growing charts in treating or not treating the trend while still child/adolescent and it wouldn't be practical the reference to 2
SDs in hospitals for each child (updated when ? Based in what study or sample's composition?...)

[HateLAPELoveSTEM]

BTW, that "wording" for less or more than 1  2 or 3 SD is not mine, it's the usual approximate terminology used in science for fuman height.
Some countries prefer to consider for example 160 cm or less for adults without any concomitant disease (even if not corresponding to 3 SD) to define "Constitutional Short Stature".  This has practical implications in growing charts in treating or not treating the trend while still child/adolescent and it wouldn't be practical the reference to 2
SDs in hospitals for each child (updated when ? Based in what study or sample's composition?...)

I think the another reason why the sample over ±3SD doesn't subject to the normal distribution is that either they are dwarfisms or gigantisms, excluded by authoritative anthropometric researches and people tend to mistake 198, 199cm tall men for ≥200cm tall men and likewise mistake 150, 151cm tall men for 148, 149cm tall men and yeah you got it right for you mentioned subjective factors.
That's also why I'm all the time thinking 'men under 180cm are not men', 'There has been no men under 180cm' and similar bs told by idiots and retards as just laughing stocks.