Close Home Forum Sign up / Log in

Do I understand this correctly?

W

I've done some sums on SPSS and I have this as a p-value: -3.403E38

Can I presume that that is a very, very small p-value below p= 0.01

I just want to make sure I've got it right. Thanks.

A

hmmm, I would have thought it's actually a really high number, but now I'm not so sure....
If for example 1.5E5 = 150000, then your -3.403E38 should = -340300000000...

I have been doing SPSS all day myself however, so my brain is frying as I type so I could be very wrong!

D

Do you mean 3.403E-38?

W

That's precisely the output for the P-value I got. Thinking about it, it has to be the power of minus 10 to the power of 38 because you can't get a p-value of above 1.0 I don't think. I really hope I haven't done it wrong now, but most of the others look fine.

A

lol that would make more sense Wally! :) Good catch DanB, I was seriously getting worried about myself there!

D

I don't know which is less sensible - a p-value greater than one or a negative one! Although I do remember a strange conversation once at Uni about negative probability... I think I was drunk at the time.

C

I'm useless at stats. If it's less than 0.05 that's good, right? :p

D

Well that's often the case but it depends what confidence you are looking at, a p-value of 0.01 might be appropriate, 0.05 is quite arbitrary... yada yada yada... Claudia, come and see Dr DanB for Statistics 101 lessons :-x

C

I was just joking, but that's what it's like in my field :p

However, you can teach me statistics anytime, Dan :-x :$

W

I know why I was getting a negative p-value now. I used the wrong test. It's not my fault though; it was SPSS offering me all these fancy additional tests as an add on to the original stats test I was running. For paired ordinal test-retest data, my good friend Wilcoxon has shown my the way. I've now got p-values that look more normal. None of that minus p-value to infinity and beyond business.

D

I have to say I've never used SPSS so cannot comment on how easy (or difficult) it is to use.

However, Claudia, I shall see you at 9am sharp in my office. Come armed with a Normal distribution.

P

Quote From DanB:


However, Claudia, I shall see you at 9am sharp in my office. Come armed with a Normal distribution.


No extreme kurtosis allowed!

On the SPSS front I've found that versions 15 and below do just fine. But the newer ones are like a statistical assault course.

A

Tell me about it, I'm currently battling my way through version 17! We are forced to upgrade every time at our uni. Consequently I have turned to Simca P+ for PCAs (they scare me in SPSS) and even braved R for CA and CCA! I have a serious new found respect for R, if I don;t get a job when I'm finished and have about 3 months with nothing to do, I'm going to sit and properly learn it... :p

C

I quite like SPSS. Mainly cos i've never used anything else so don't know what I'm missing! Hate the weird bugs it gets. Thinking of upgrading from15....is 17 much different? Never seen a negative p value though!

Everyone I know who uses r raves about it, esapecially it's ability to draw better graphs, but don't you need programming skills? I'm sadly laking any of these!

ooh I love a bit of SPSS :-x

15540