[Eeglablist] ERP grand average does not reflect actual latency

[Daniele Marinazzo]

Dear Alvin

focusing only on a tiny part of the erp, and only on the average, is not a
good way to analyze the data, since it makes too many assumptions on the
underlying distribution of data, individual variability, and on the ERP
time course as a whole.

Guillaume Rousselet, Cyril Pernet and others have explained this in great
detail, with code and examples, see for example these papers

http://onlinelibrary.wiley.com/doi/10.1111/ejn.13400/full
http://biorxiv.org/content/early/2017/03/27/121079

and in these blog posts with code

https://garstats.wordpress.com/2016/07/28/neuroscience-group-results/
https://garstats.wordpress.com/2016/04/02/simple-steps-for-more-informative-erp-figures/

https://garstats.wordpress.com/2016/07/28/neuroscience-group-results/

hope this helps

Daniele

On Tue, Mar 28, 2017 at 3:30 AM, Shih-Chun Kao <shihchunkao at gmail.com>
wrote:

> Dear list,
>
> If I found a statistically significant P3 latency effect (~50ms) between
> two conditions in a within-subject design study but my grand average plot
> does not show such difference, what would be the cause for inconsistent
> results and what does such inconsistency suggest? I understand grand
> average is not the best illustration but I thought 50ms is a pretty big
> difference. Thank you!
>
> Alvin
>
> _______________________________________________
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to eeglablist-unsubscribe at sccn.
> ucsd.edu
> For digest mode, send an email with the subject "set digest mime" to
> eeglablist-request at sccn.ucsd.edu
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20170328/bd4f04e4/attachment.html>