Your question, shown below in its original context, was in 2 parts: "Did you get to the same conclusion on the 2 situations?" Well, did you? You didn't say, nor did you say what that conclusion is (or those conclusions are, if you DIDN'T get to the same conclusion).
And "This will happen to any sample and the tests are equivalent??" It's unclear why you end this sentence with question marks. Is this an assertion of your instructor, or of the textbook, that you want someone to explain? (This is suggested by that emphatic "WHY??????".) It might help if you tried to imagine data for which the null hypothesis would be rejected by the first (chi-square) test, but not by the test on the total proportion. (I will agree that data that do this would be rather unlikely to arise by the coin-tossing method described. But the exercise might help you to perceive the rather subtle difference between the two null hypotheses being tested by the two procedures, and the sense(s) in which the two tests can be said to be equivalent.) I'll leave you with that hint. Feel free to ask further, but answer the direct questions (express and implied) in the first two paragraphs above, and attempt the exercise suggested in the paragraph just above (the one beginning "It might help..."). On 19 Aug 2003, Flavio wrote: > Hi! My class had this exercise to do but no one seams to know how to > do it... > So we need a hand! :) > > The exercise is the following: > A coin was launched 320 times. Grouped in 80 sequences of 4 launches > each in the same order they were "conceived". The following happened > > Quantity os heads 0 1 2 3 4 > Quantity of Sequences 6 22 35 11 6 > > Meaning: 6 sequences with no heads, 22 sequences with 1 head, 35 seq. > with 2 heads... etc > > Now: > If the coin is balanced so the quantity of heads in each seq. of 4 > lauches is a aleatory binomial variable with parameters n=4 e p=50%. > Test this hipotesis using the Chi-Square method with 5% significance > Test now, with 5% significance, now using the total proportion of > heads > > Those we were able to do.. now the hard question: > Did you get to the same conclusion on the 2 situations? This will > happen to any sample and the tests are equivalent?? WHY???????? ----------------------------------------------------------------------- Donald F. Burrill [EMAIL PROTECTED] 56 Sebbins Pond Drive, Bedford, NH 03110 (603) 626-0816 . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
