Test for different means
Student's t-based significance test
This test examines whether the means and of two distributions (here: signals) are significantly different and displays the results in the message window.
It takes one parameter,
- diffvar, indicating whether the variances of the two
signals are assumed to be
- "identical" (0) or
- "different" (1)
and returns two parameters:
- t, which is the t-value resulting from Student's
t-test (see below), and
- signif, which is the significance, i.e. the probability [0..1] of randomly equal means, where small values (say 0.05 or 0.01) indicate significantly different means.
The t statistics is computed by the so-called 'pooled variance'.
If the variances are identical, t equals
where the standard error is
If the variances are different, t is computed by
where and denote the numbers of points in the first and second signal, respectively.
BDS test, Kolmogorov-Smirnov test,
Test for different histograms, Test for different variances,
Test for non-normal-distributed data,
Test for non-white-noise,
Test for the mean of a Gaussian, Test for the variance of a Gaussian,
Hollander/Wolfe , Lehmann