Hi SPSSers,
I have a friend who's looking for either conceptual or empirical references for the following situation:
He would like to run a Confirmatory Factor Analysis (CFA) for Q1 (below), then use an Exploratory Factor Analysis (EFA) data analysis strategy for Q2. He has a large enough sample to be able to
randomly split our sample to do the CFA on the first half, then EFA on the second half...
Q1: can we first replicate the current state of
knowledge, factors A vs B using CFA, with a pool of items,
several are deleted based on poor loadings. The CFA
will give more confidence that the subsequent EFA
findings have a foundational grounding in the basic
factors identified in previous research, hopefully
allowing for easier future replication.
Q2: Now that Q1 is a 'yes' are there further
gradations here?, A1, A2 vs B1, B2, B3. This is the EFA part,
again a couple of items are deleted but A1,A2, B1, B2
B3 emerge...
Can anyone please give me conceptual and/or empirical study
references to support using this data analysis strategy,
randomly splitting a sample then doing CFA followed by EFA?
(I have found studies randomly splitting but doing EFA first
then CFA, but I would like to do CFA then EFA, using the
logic above)
Thanks in advance.
Best,
Lisa
Lisa T. Stickney, Ph.D.
Assistant Professor
Merrick School of Business
University of Baltimore
Ph: 410-837-6607
Fax: 410-837-5675
[hidden email]
====================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
Locked
