Talk:Sobel test

Calculation
I find it strange that the calculation method states that you should compare the t-statistic with the normal distribution. While I know that this is the case when N is large, shouldn't the more proper thing to do be to compare a t-statistic with the t-distribution? Then you are always safe? Comments? — Preceding unsigned comment added by Justacec (talk • contribs) 18:49, 10 November 2013 (UTC)

SEM notation
Using arrow diagrams assumes knowledge of SEM on the part of the readers, but the diagram actualy just show a regression model. Therefore this should be simplified and the SEM model should be annotated and linked to the SEM article. Cato The Censor (talk) 13:42, 21 November 2013 (UTC)

true p-value
There is no such thing as a "true p-value" in inferential statistics. --176.199.208.44 (talk) 14:37, 25 March 2020 (UTC)