And related, is there a particular reason why SPSS doesn’t show pooled results of PCA for multiple-imputation data sets, both with and without bootstrap confidence intervals, while there are solutions for this at hand? See, the following references:
Van Ginkel, J.R. & Kiers, H.A.L. (2011). Constructing bootstrap confidence intervals for Principal Component loadings in the presence of missing data: a multiple-imputation approach. British Journal of Mathematical and Statistical Psychology, 64,
498-515.
Van Ginkel, J.R. & Kroonenberg, P.M. (2014b). Using Generalized Procrustes analysis for multiple imputation in Principal component analysis. Journal of Classification, 31, 242-261.
Best regards,
Joost van Ginkel
-----Original Message-----
From: SPSSX(r) Discussion [
[hidden email]] On Behalf Of Ian Martin
Sent: woensdag 26 november 2014 15:51
To:
[hidden email]
Subject: bootstrapped PCA?
I was wondering if there is some particular reason for the scattered implementation of the bootstrap option in SPSS? While I understand there are some unique considerations in the implementation of this option for factor analyses (flipping polarity of
variable loadings on subsamples, leading to overestimation of variance), I think there are solutions that have been published.
Is there any intent, or timeline, at SPSS, for making bootstrap available for PCA and similar analyses, or does anyone know of a scripting solution for bootstrapped eigenvalues and variable loadings?
cheers,
Ian
Ian D. Martin, Ph.D.
Environmental Analysis & Consulting
=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of
commands to manage subscriptions, send the command INFO REFCARD