MATLAB: How to change the significance level of either kruskalwallis or ranksum test

alphakruskalwallisranksumsignificance levelwilcoxon

Hi, this is probably an easy question for rehular matlab users but leaves me standing infront of a problem: the default significance level of kruskalwallis is 0.01 and the obe for ranksum 0.05…I would like to use the same for both and found that you can change it just not "how". Where in the code do you put the alpha value in?

I typed it in like you would for a t-test but that should change the p-value depending on your significance level i.e. if the null hypothesis gets just about accepted with 0.05 it should be rejected with 0.01. However, the p-value remains the same in both cases…

Thanks!

Best Answer

For ranksum, you have to specify it as a name-value pair as 'alpha' (link).

The kruskalwallis doesn’t have that option. However, it outputs the p-value, so you can use it to make your own decisions.

Related Solutions

MATLAB: What is the output of kruskalwallis

I am not sure what "set of values" you are looking for. The output from the call:

p = kruskalwallis(X)

is simply the p value for the hypothesis test with null hypothesis H0: The columns of X all come from the same distribution and alternative hypothesis H1: At least one column comes from a different distribution. The table and stats variables simply give you more information regarding this test.

To interpret p, choose a significance level alpha (common choices are 0.01, 0.05, 0.1) and if the p value is less than alpha you would reject the null in favor of the alternative. Otherwise you would conclude there is statistical evidence to support the null hypothesis.

MATLAB: Does KSTEST not reject the null hypothesis when the P-value is less than the significance level in Statistics Toolbox 6.2 (R2008a)

The function KSTEST by default, performs a two-tailed test. In case of such tests, if the significance level is alpha (0.05 by default), the null hypothesis is rejected if the P-value is less than alpha/2. To account for the tail in the computations, you have to specify the 4th input to the KSTEST function ('tail').

Best Answer

Related Solutions

MATLAB: What is the output of kruskalwallis

MATLAB: Does KSTEST not reject the null hypothesis when the P-value is less than the significance level in Statistics Toolbox 6.2 (R2008a)

Related Question