oss
Posts: 27
Joined: Sat Feb 10, 2007 11:53 am

### The effect of heredity on person's height??

I just figured out this topic. But still don't know yet the procedure to collect the data. Can you help me?
Is this a possible topic??? Or should I find another one?
Thanks a lot for all your help =^.^=

MelissaB
Moderator
Posts: 1055
Joined: Mon Oct 16, 2006 11:47 am
You could certainly measure the heredity of height in humans. Keep in mind that you'll want to have a very large sample size and that you will have a lot of potentially confounding variables you simply won't be able to control. What grade are you in?

The way biologists typically measure (narrow-sense) heredity in populations is to measure the trait in both parents and offspring. Then we do a regression of the average parental values onto the average offspring values (a fancy way of saying we make a graph of parent values vs. offspring values and draw a best-fit line). The slope of that line is the narrow-sense heredity.

One thing you'll want to consider is that you can only measure (or ask for responses from) adults that have finished growing. If you try to look at it in children, they haven't finished growing so you'll probably get more variation than you would otherwise!

On the bright side, most adults know at least approximately how tall their parents were, so you may just be able to survey adults, ask them how tall their parents were/are, and use their responses. You would want to tell them not to write anything down if they're not absolutely sure, though, because if they guess it might bias your results.

Hope that helps some!

oss
Posts: 27
Joined: Sat Feb 10, 2007 11:53 am
thru this project, I wanna show that heredity isn't the only factor affect person's height. For instance, playing sports or eating properly can improve one's height. So how should I change the topic?
And to make the data more reliable, I think I need to do sth more than just ask and write down the data right??? Cause then I don't need any material to have this project done.
Thanks a lot for all your help =^.^=

MelissaB
Moderator
Posts: 1055
Joined: Mon Oct 16, 2006 11:47 am
That is a lot more complicated, and my gut feeling is that you'll need an extremely large sample size in order to see any results. What you can do is do exactly what I said before only look at the residuals of the regression--the deviation from that best-fit line--to see if that correlates with things like sports playing or eating right (note that 'eating right' is going to be EXTREMELY difficult to define and/or measure--many people will not be truthful about their food intake on a survey).

If you can actually measure the heights yourself, that would be great--but I think it will be easier for you to get a larger sample size (especially since you need parents' heights) if you just ask people to write down their heights. As long as you explain to them what you're doing, you probably won't have too many people lie to you.