Question

John's parents recorded his height at various ages up to 66 months. Below is a record of the results.
Age (months) 36 48 54 60 66
Height (inches) 35 38 41 43 45

Answer 1

John's parents decide to compute the least-squares regression line of John's height on age based on this data, and use it to predict his height at age at 21 years (252 months). Then
A. John's height, in inches, should be about half his age, in months.
B. the parents will get a fairly accurate estimate of his height at age 21 years because the data are clearly correlated.
C. such a prediction could be misleading because it involves extrapolation.
D. all of the above.

I say B the rest really don't make sense at all.

Answer 2

@CliffSedge

Answer 3

E. John's parents are nerds.

Answer 4

AHHAHA.

Answer 5

(and not very smart nerds if they think that'll work . . .)

Answer 6

My parents used to mark the door frame during each of our birthdays :P

Answer 7

actually it could be C, since I just learned what that word means... LOL

Answer 8

Mine did the same.
I guess their not really nerds then (John's parents, not yours or mine), they're just dorks.

Answer 9

I doubt it will be a good prediction at alll....

Answer 10

Just for fun, let's see what that prediction would be..

Answer 11

Predictor equation is y=0.3423x+22.32 (with an impressive correlation of r=0.9940!)

At 252 months, John would be over 9 feet tall!

Answer 12

HAHAHAHHAHAHA

Answer 13

Moral of the story: Don't extrapolate too much.

Answer 14

How do we figure out those numbers of y= again? I believe one is the slope, and the y intercept is the value at where y is intercepted on the y axis?

Answer 15

so that means C? :)

Answer 16

C. Definitely.
Tip for John's parents: Human growth is non-linear.

Answer 17

LOL :)

Answer 18

To get the slope and y-intercept for the predictor equation requires linear regression, which isn't too hard to do by hand for small data sets, but anything past 10 data points (or even 5 if I'm just not in the mood), I use my TI-84 or a computer spreadsheet program.

Answer 19

Wolfram ftw :)

http://openstudy.com/study#/updates/506f9953e4b060a360ffc484 next Q :)

Answer 20

Thanks for everything btw.

Answer 21

I still haven't tried Wolfram yet. I poked around at it a little bit, but everything I need to do, I can either do by hand, or pass on to my trusty TI-84.

To see the annoying process of doing it by hand, scope this:
http://stattrek.com/regression/linear-regression.aspx?Tutorial=Stat
and http://stattrek.com/regression/regression-example.aspx?tutorial=stat