Thanks for your reply .... It turned out that in my path there was
a version of erfcore.m (not really sure how it came to be there) that
gets called by erf.m and betainc which gets called by spm_t2z.m
Mystery solved!
-Luis
Begin forwarded message:
> From: Ged Ridgway <[log in to unmask]>
> Date: November 14, 2006 6:17:28 AM GMT-05:00
> To: Luis Hernandez <[log in to unmask]>
> Subject: Re: [SPM] spm_t2z function
>
>
> Hi Luis,
>
> I just copy-pasted your code and it looked fine (all positive, no
> gaps), I've tried it on both spm5 and spm2 and it seems okay, so
> which version are you using and do you have all the updates?
>
> Best,
> Ged.
>
>
> Luis Hernandez wrote:
>> Hi,
>> I'm getting some unexpected results using the spm_t2z function....
>> say we get some t-score between 0 and 5 and convert them to z
>> scores with 8 degrees of freedom;
>> t = rand(1000,1)*5;
>> z = spm_t2z(t,8);
>> plot(t,z,'.')
>> I get a gap in the plot in values around t=0,65, and the z scores
>> are all negative!
>> The distribution looks like the negative of what it should look
>> like...
>> could someone shed some light on this?
>> thanks
>> -Luis
>
>
--
( http://www.eecs.umich.edu/~hernan )
|