Which statistical test to use?

Question

Karl 2013-9-5

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/86443-which-statistical-test-to-use

Hi, I have a quick question. Basically I have measured a set of quantities in two different ways (way A and way B), and have also measured the error (the values measured are normally distributed). I want to show that way A and B are equivalent. Now this is not the exact same as measuring say some value 50 times one way and some value 50 times another way, which I can then calculate the mean and standard deviations, and do a Z or t-test. I have measured a set of values which (should be) consistent between methods, but are not all the same values. Basically I'm measuring two (very strange) curves and have the mean and error at each point for both methods. I then want to see if the two curves are consistent (or not). I would think of doing a chi-square goodness of fit test, but, to my understanding, ths is for frequency data. Mine is not frequency data.

 One way I thought of doing this was a z-test on a point-by-point method. And I found about 92% of the values were within +/- 2 std. But i'm not certain how informative this number exactly is, I'm sure there's a better statistical method but I'm not too familiar with stats.
My question is what is the best way to do this? Does MATLAB have the built in test? Thanks very much for any help