Two outputs when calculating standard error of mean, WHY?

Hi,

I am calculating the standard error of the mean and I am getting two ouput. I do not understand why. The code is:

SEM_A = std(A,[])./ sqrt(size(A))

And my output is:

SEM_A =

0.0048 0.1290

Do You have any ideas?

:)

Answer 1

[Daniel M]

It's because size(A) returns two outputs. If you want to use the size of A in a particular dimension, use size(A,dim). (Since size operates columnwise, it's likely only dim=2 would make sense). If you want to divide by the total number of elements in A, use numel(A). If A is a vector then numel(A) is the same as length(A), which is the same as max(size(A)).

[Vlatko Milic]

Thank you! One question, why do the two values differ? To have a standard error of 0.0048 does not seem reasonable since my mean i around 1. The data is quite stable but to 0.0048 seems extreme?

[Vlatko Milic]

Now I think I solved it. My code is now:

STE = std(stat_W_per_K)./ sqrt(size(stat_W_per_K,2));

However, when it was:

STE = std(stat_W_per_K)./ sqrt(length(stat_W_per_K));

I got the super-low values. What is the reason for this?

Lastly, do you know how I can connect the standard error to a confidence interval of 95%? Is it my average +- my standard error and that's it? :)

Regards, Vlatko

[Daniel M]

Again it is the difference between the output of length(A) and size(A,2). I can't tell you which one is correct.

I also can't tell you how else to analyze/report your data, because I don't know what they are, how they were measured, underlying assumptions, statistical distributions, systematic errors and uncertainty, etc.