MegaSack DRAW - This year's winner is user - rgwb
We will be in touch
I told my boffin daughter that STW has all the answers, can anyone help me win the internetz? 🙂
An excavation of an early Upper Paleolithic cave site in France gave two samples of hominid skeletons: a sample of Homo Sapiens Sapiens (n1 = 20) and a sample of Homo Sapiens Neandertalis (n2 = 20).
We are interested in establishing whether there is any significant difference between the cranial capacities of these two hominid species from this cave site.
Let X1 = 1514.6 the sample mean of cranial capacities (cm3) for the Homo Sapiens Sapiens specimens, and X2 = 1553.4 for the Homo Sapiens Neadnertalis.
Also, the sample standard deviation is s1 = 78.4 for the first and .s2 = 74.7 for the second group, respectively.
Q.
If one rejected the null hypothesis when the absolute difference of the two is 45cm³ what is the probability of a type 1 error?
She's been working at this for 5 hours apparently, and neither her or her friends have any idea.
And it needs to be for a 't' test not 'z' test...............
What language even is this?
standard errors 'n'shit, innit ?
(that's as far as my massive knowledge goes though)
5%
If we want to be 95% confidentr of rejecting the null hypothesis then there is a 5% chance we might make an error
Means & SDs are irrelevant
Oh, maybe they're supposed to shove the figures in using various "T" factors and see which one just gets it right ?
If the absolute difference is changed to 45, then at least one of the samples must have changed and so you couldn't rely on the SDs or means
(I iz not a sassystishan)
donncha - they'd only know that it was (less than) 5% if they used the values of T corresponding to that confidence limit*. I'm sure it's possible to calculate the actual CL, but not by me
*(he speculated)
What about 45cm cubed bit, was the reply I just got, she managed the CL
from an on-line calculator:
mean1 eq: 1514.6 (variance= 6146.56) (se= 17.5308)
mean2 eq: 1553.4 (variance= 5580.09) (se= 16.7034)
Probability that var1<var2
p=0.41764 (left: 0.5824; double: 0.8352)
Difference between means:
M1-M2=1514.6-1553.4=-38.8
sd=149.27; se=24.214
95% CI of difference:
-86.2593 <-38.8< 8.6593 (Wald)
t-difference: -1.602
df-t: 37.5; p= 0.05868
(left p: 0.9413; two sided: 0.1174)
Equal sd assumed
Difference not significant at 5%
By jove, I think you've done it, my Daughter now believes I am a genius and managed all this over a cup of tea and a digestive.
I will of course own up and give due credit after I have finished basking in the sunshine
Thanks guys
Dunstick
What class is this for?
So what was the answer?
Oh hang on, she reckons that online calculator is for a z test not a t, Donncha, 2nd yr Maths degree statistics
She got the ci as -17.72 and 95.32
Does this make any sense?
[img] 
[/img]
Apparently 'Saying that the means and sds are irrelevant for a type 1 error' was the crucial bit she was short of