I have seen, in the Q&A section of Andrew Hayes website, that he doesn't use BCa bootstrap (with his macro PROCESS) to estimate the indirect effect of a mediation analysis, but uses percentile bootstrap. But, in Andy Field's book (Discovering Statistics Using IBM SPSS Statistics - North American 5th Edition), on page 375 and 376, Field wrote "... BCa CI ..." to report the indirect effect, even if he is using PROCESS. So... is this a mistake by Field? Or is Hayes Q&A section of his website outdated (because he also wrote that "Bias-corrected bootstrap confidence intervals will again be available in a future release of PROCESS"... so maybe BCa CI are now used?)? Here is a fictional example to illustrate my question - I used PROCESS v3.4 (see attachments).
Haroon R Abbu Thank you! Do you know where we can find the information about the fact that PROCESS uses BCa bootstrap CI? Because as I said, Andrew Hayes says the opposite in his website, and in his latest book, he puts emphasize on percentile bootstrap CI instead of BCa bootstrap CI. :)
So, I asked the question directly to mister Hayes and he wrote back to me saying that PROCESS does not uses BCa bootstrap intervals. He also said that
the next release of PROCESS will allow BC bootstrap intervals as an option (BC and BCa are different things).
- Badges
- Science topic
- Similar topics
- Books