Hello! I am new to SDM and would like to do a meta-analysis using fMRI data. 
Some papers reported their stats with z score so I converted them to t using the converter on SDM. However, I got a lot of  high t scores. For example, a paper had z score of 8.72 (N=8) and the t score i got was infinite. 
Anther study had a t value as high as 323.34 after conversion. 
I am not sure if I should trust these numbers and how to deal with "inf"in my analysis. 
Thank you very much.

Hi. Converters from z to t may return infinity if the value is too high, though this might be overcome using logarithms; z=8.72 would correspond to t=714 with 7 df.

However, I would be very cautious with such extremely high t-values. Are the values reported in the study really z-values? I strongly suggest to check it, as they could be t-values incorrectly labeled as z-values.

Hello!

I too am facing the same problem except my t-values are in the range of 30. While most of the studies in my meta-analysis report a t values of range 1 to 5.
Do you think these studies have incorrectly labled t-values as z-values
If not, would these variations not create a bias?

Thanks

Navkiran


Originally posted by Joaquim Radua:

Dear Navkiran,

these high z-values seem very unlikely, so I would double check.

If they are correct, you might interpret that some few samples have extreme effects while most not, and you may be interested in studying the source of such heterogeneity.

I hope this helps,

Joaquim

Thanks a lot for the prompt reply!

Dear, Joaquim
   
      If the Z value is infinity, how to deal with it?

Hello! Have you solved this problem? When I was extracting coordinates, there was also a coordinate with a z value greater than 9.99, and after converting it to a t value, it displayed inf. How should I solve this situation?