# Homework Help: P-value calculation

1. Jul 3, 2013

### lavster

1. The problem statement, all variables and given/known data

This isnt a homework question but figured this is the best place to ask :)

what I have done is measure doses from an x-ray tube at various positions using two different methods, method1 and method2. I would like to say if they are statistically significant or not by using a p-value test. I have no software except excel... I have so many results to do this analysis on. If you could help me with the first one that would be really appreciated. The points are meant to be (slightly) different so its not just a repetition of the same thing. Ie ive not just repeated it 7 times. Its method 1 and method 2 id like to compare

point method1 method2 % diff
1 203 ~~ 206.2~~ 1.57
2 202.2~~ 204.6~~ 1.17
3 202.9~~ 203.4~~ 0.27
4 202.4~~ 205.9~~ 1.73
5 202.6~~206.3~~1.86
6 202.3~~ 210.1~~3.83
7 202.4~~ 204.1~~ 0.92

2. Relevant equations

3. The attempt at a solution

so im assuming my null hypothesis is they are not different.

But is my degrees of freedom 1 or 6?

Also which one of the p-value tests do I use in this situation? most examples I can find is when its a group A vs group B, where the results from group A and group B are not linked. But in mine it is point 1 is point 1 in both and they should be compared together...

Thanks

2. Jul 3, 2013

### Ray Vickson

Just look at differences, not percentage differences. Why? Well, percentages are quite nonlinear functions of the data (as they are of the form 100*(x-y)/y or 100*(x-y)/x, which are not linear in x and/or y). If you look at the differences, you will have 6 degrees of freedom, because you have 7 differences.

Another issue that may be cause for worry is whether or not the data are sufficiently close to being normally distributed so as to allow for usable results from standard statistical tests. If the data are far from normal, the test results could be meaningless, and you might need to fall back on non-parametric tests.