Do you have a sibling? I could use your help

[pv=nrt]

make it up

[eehh]

Originally posted by Sleepy Weasel
Anyone else want to chime in?

no can do. dont have SAT's in Norway, so it would not help..

[Verthanthi]

From a friend of mine


1. 23/M
2. 20/F
3. 18/M
4. 16/F
5. 12/M
6. 11/F
______
1. #4
2. #6
3. 1370 (#3)
4. #5
5. smartest
6. #2

He only knew his own sat score, but :/

[Sleepy Weasel]

Thanks for all the help. Project is finished now.

For the record I came up with a slight correlation between older siblings scoring higher on the SAT than younger siblings. For those of you that know statistics that p-value was .09. The results also showed that the correlation was much stronger when the siblings were of the same gender than if they were different genders, and also that the extent of age difference between siblings had no effect.

So what does this all mean? Probably absolutely nothing since my sample was biased and my data largely innaccurate, but it got the paper written (12 pages of glorious bullshit. I also had to include a lot of background information on other studies that have been done on this in the past) so that's good enough for me.

[PaulOakenfold]

A: 33/F
B: 26/M
C: 21/M

A: 1450
B: -140
C: -0

EDIT: Oh crap, didn't read all the posts, sorry, if you still need info though here it is :/

[TelC@t]

Originally posted by Sleepy Weasel
For those of you that know statistics that p-value was .09.
...
Probably absolutely nothing since my sample was biased and my data largely innaccurate.

p-value is 0.09? dude, thats stuck in between the statistically significant and statistically insignicant. Did you do a 95% confidence interval thingy as well?

And I hope you didn't forget to mention that you sample size was small so the result does not necessarily reflect the true characteristics of the population.

But then again, you said it's for your psychology class, rigorousness in maths is probably not what they are looking for.

[Sleepy Weasel]

Yes I'm aware of all those things, and didn't do a confidence interval because that wouldn't have revealed any useful information. I used a paired difference z-test for the .09 result (n was 53), paired difference t-tests for the smaller sample size tests for the gender/age related tests, and two sample t-tests with pooling to compare the two samples for noticeable differences between them. But as you said it was for a psych class, so I didn't find it necessary to add any more statistics beyond those basic tests.

For the record I just didn't bother to define and alpha of .05 or .1 and just left it as statistically insignificant but close enough to suggest a possible correlation