Reputation: 5579
I have 2 vectors
Upper_bound = [421.1706;418.7937;425.9144;431.7096];
Lower_bound = [376.0487;395.4193;402.7083;419.0457];
Representing the 95% confidence intervals of 4 measurements (A,B,C,D
).
How can I calculate in an automatic way whether there are significant differences between measures (i.e. 95% confidence intervals do not overlap).
my preferred output would be:
sign_diff = [0 0 0 0;
0 0 0 1;
0 0 0 0;
0 1 0 0];
Indicating that A does not differ from A,B,C,D.
B does not differ from A,B,C but it differs from D.
etc.
Thanks
Upvotes: 0
Views: 75
Reputation: 769
Taking Shai's answer:
sign_diff = ~(bsxfun( @le, Lower_bound, Upper_bound' ) &...
bsxfun( @ge, Upper_bound, Lower_bound' ))
I can't comment yet, so I post it as answer.
Upvotes: 0