Parallel Analysis for EFA with paran (Dinno)

[Theresia Kuepper]

Dear colleagues,

I'm performing an exploratory factor analysis and tried to figure out how many factors to extract by using the paran command in Stata which is an alternative command for parallel analysis. Paran is a user srcipted code.
My code is:

paran, mat(C) n(236) centile(95) iteration(100) graph seed(6327456) all

By using this code I'm getting a type mismatch r(109).
Does anyone know how to solve this problem?
Thank you in advance!

[Joseph Coveney]

Make sure that you're using the current version of the user-written command, which seems to be Version 1.5.3 2015-12-16 from the author's website and not from SSC.

Using that version and your exact line of code verbatim doesn't give any error: the following code runs without any problem (do-file and output attached below).

version 17.0

clear *

log using paran.smcl, nomsg name(lo)

foreach file in "paran.ado" "paran.hlp" "simdata.dta" {
    copy "https://alexisdinno.com/stata/`file'" `c(pwd)'/`file'
}

*
* Begin here
*
which paran.ado

quietly use simdata

quietly factor var*

matrix define C = e(C)

// Copied and pasted verbatim from the Statalist post:
paran, mat(C) n(236) centile(95) iteration(100) graph seed(6327456) all

quietly graph export Paran.png

quietly log close lo

exit

[ericmelse]

For those visitors who want to download the current version - *! version 1.5.4 18mar2025 - you can find it here as well as some other interesting packages created by Alexis Dinno.
.