Is it possible to have a significance rating of 0.00?

#1
Hi, I am currently working on my dissertation. I am measuring a muscular soreness rating /100 in 3 conditions after supplements, and at 3 time points, immediately after exercise, 24 hours post, and 48 hours post!I have tried and tried on spss and I keep coming out with a significance value of 0.00 in my test of between subjects, test of within subjects and friedmans test, I am getting pretty desperate as I need to write up my discussion and I dont even know if this is possible?any help greatly appreciated
 

TheEcologist

Global Moderator
#2
Hi kevin,

Its possible. It could be that SPSS is rounding your answer to fit within its tabled output (double click on the values to find out) or it could be that the p-value is actually zero - or the practical equivalent of zero for your computer. This can happen when you are testing against zero and your values are very high and not very variable - it also depends on the test distribution. You havent given us enough details to narrow it down.
 

CB

Super Moderator
#3
I think the convention would be to describe a p value that SPSS has rounded to zero as "p < .001" or similar when actually writing up your analysis. Maybe change the decimal points output to show 3 dp though - the p value may not round to zero when you have 3 dp displayed.
 

trinker

ggplot2orBust
#4
In theory it could be possible to have p=0 but highly unlikely. That would mean absolutely no difference between groups, measures, etc. Just do to chance, error or what ever there's be some sig. As others have pointed out SPSS rounds after a few digits, likely the zeroness you experienced is a result of SPSS.