Drexel dragonThe Math ForumDonate to the Math Forum



Search All of the Math Forum:

Views expressed in these public forums are not endorsed by Drexel University or The Math Forum.


Math Forum » Discussions » sci.math.* » sci.math.independent

Topic: Height differences -- statistical anomaly
Replies: 5   Last Post: Jan 1, 2014 9:18 PM

Advanced Search

Back to Topic List Back to Topic List Jump to Tree View Jump to Tree View   Messages: [ Previous | Next ]
Virgil

Posts: 9,012
Registered: 1/6/11
Re: Height differences -- statistical anomaly
Posted: Jan 1, 2014 9:18 PM
  Click to see the message monospaced in plain text Plain Text   Click to reply to this topic Reply

In article <la2gck$nkq$1@reader1.panix.com>,
ime@panix.com (Randy Hudson) wrote:

> In article <8761q6mxu6.fsf@bazspaz.fatphil.org>,
> Phil Carmody <thefatphil_demunged@yahoo.co.uk> wrote:
>

> > ime@panix.com (Randy Hudson) cites:
>
> >> http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0054186
> >
> > Statements like this worry me:
> >
> > For every cm increase in female height, partner height on average increased
> > with 0.19 cm [...]
> > Similarly, for every cm increase in male height, the female partner is
> > predicted to increase with 0.17 cm
> >
> > The two regression lines those statements are based on are 70 degrees
> > apart.
> > I'm not saying they're wrong, but are such statements actually useful?

>
> They seem to be pertinent to the question you asked.
>

> > If the two quantities are related to each other linearly, surely there is
> > one best relation between them, not two different ones, depending on which
> > axis you pick. Viewed another way, shouldn't the affine transform of a set
> > of data have a best-fit regression line which is the same affine transform
> > of the original data's regression line? Viewed another way - remove the
> > axes,
> > just show the points - now draw the line that best fits the data.

>
> That's not linear regression. The lack of perfect correlation attenuates
> the accuracy of predictions of the dependent variable based on the
> independent variable. At the extreme, zero correlation implies that the
> height of either partner tells nothing about the height of the other
> (zero inches of female height are implied by each inch of male height, and
> simultaneously, zero inches of male height are implied by each inch of
> female height.)
>

> >> >b) What is the probability under the assumption that the height of the
> >> >man is independent of the height of the woman?

> >>
> >> Indistinguishable from zero.

> >
> > Doesn't seem likely. I just paired random tourists outside my window, and
> > it only took me a quarter of an hour before "the next marriable female to
> > walk past" was taller than "the next marriable male to walk past from the
> > next group of people that walks past".

>
> I'm not sure what you are testing, but if you weren't examining married
> couples, I don't understand the relevance.
>

> > And that paper even says:
> > "The male-taller norm was thus violated in 10.2% of the couples when mating
> > was random"

>
> Yes, that's the null hypothesis. Based on their data, they rejected it.
> They observed far less violation of the "male-taller" condition than would
> be expected under random mating. They thus concluded that some couples
> practice assortative mating based on similarity of height.


I, for one, would certainly expect that exremely tall or exremely short
people of either gender would tend to mate with others fairly close to
their own height.
--





Point your RSS reader here for a feed of the latest messages in this topic.

[Privacy Policy] [Terms of Use]

© Drexel University 1994-2014. All Rights Reserved.
The Math Forum is a research and educational enterprise of the Drexel University School of Education.