How to get real p value in the 'multicompare' output table instead of rounded number

Dear

I would like to get real p value in the 'multicompare' output table instead of rounded number.

For example,

 [p,table,stats] = kruskalwallis(data,groups);

From my data, I got 2.478874278408479e-29 for p-value. (which is good.)

Then, I used multicompare function.

[c,m,h,nms] = multcompare(stats,'alpha',.05,'ctype','dunn-sidak');

1	2	107.358736884486	149.713883016209	192.069029147931	0
1	3	120.836759163490	153.690374387319	186.543989611148	0
2	3	-32.2920513577527	3.97649137111020	40.2450340999731	0.991192034848183

here is the result for c, and 6th column would be p value.

For group1-group2 comparison, I got '0' for p-value.

I wonder if I could get real value(ex) format like 2.54 x 10-29 ) instead of just 0 (rounded number)

I really appreciate it if someone could help.

Answer 1

[dpb]

As Walter points out beflow, that it prints an integer zero indicates the calculation of the p-value underflowed and the returned value was identically zero.

If it's a case of lack of precision on the command window, you can do something like:

fmt=['%5d%5d' repmat('%12.4f',1,3) '%12.3e\n'];
fprintf(fmt,c.')

In this case the values would be displayed 0.0000 if in format short mode, format long would give 15 digits but E-29 would still have underflowed.

[Walter Roberson]

No, for the output format being used, 0 with no decimal point indicates exact zero as far as the calculation was concerned

[dpb]

+1, Walter. He's got p-values small enough to have underflowed the calculation engine, you're right.

I just took a sample from one of the Examples but it isn't identically zero, correct.

I'll make note to the Answer.

[Sang Hoon Kim]

Got it~! Thanks a lot~!