Question 523005: Hi,
I am having problems with a chi-square related question.
Here is the question:
(a) Write the null and alternative hypothesis for a one-sided test to determine whether men are, on average, heavier than men from 20 years ago.
These are the hypotheses I have come up with:
Null Hypothesis – There is no difference between the average weight of men now and 20 years ago.
Alternative Hypothesis – Men, on average, are heavier than men from 20 years ago.
(b) Carry out a test of the hypothesis in (a) at the 5% significance level, i.e. calculate the:
- Test statistic
- Critical value of the Normal Distribution for a one-sided test at the 5% significance level
This is where I am having problems. I know how to do chi-square (I'm assuming we have to use chi-square as that's what we have been using in the other questions) when I have the data, ie. observed and expected. What I am unsure about is what data to use as this is all I have been provided with:
From data collected 20 years ago, the average weight of an adult male was 71 kilograms and the standard deviation 7.2 kilograms.
Table: Weights (in kilograms) of a new random sample of 29 adult males
59 79 61 81 62
82 66 68 83 69
84 70 73 77 84
74 88 76 78 91
78 74 86 59 66
73 74 75 76
So far I have been using the averaging weight of the new sample of men which is 74.7 and the average weight from 20 years ago which is 71 and trying to use a one-way chi-square method. So basically my table looks like this:
Today 20 Years Ago
Weight 74.7 71
First of all is this the correct way of doing it? If so, my next dilemma is trying to figure out the expected frequency from the observed. I know how to calculate the expected from a contingency table of 2 or more rows, but when i try to apply the same formula to just one row, the expected ends up the same as the observed, which surely cant be right?
If I am going about it totally the wrong way, any advice would be appreciated. I am not a mathematician or even studying it ,so apologies for my lack of knowledge.
Hopefully someone can help me out.
Ryan
Answer by stanbon(75887)   (Show Source):
You can put this solution on YOUR website! That is not a Chi-Sq problem. It is a 2-sample Z or T test problem.
---
Also, a strict rule for hypothesist testing requires that
equality ALWAYS appears in the Null Hypothesis.
----
I think if you use Google you will be able to find examples
online of a 2-Sample test.
Cheers,
Stan H.
|
|
|
