|
|
 |
| Bifactor EFA taking a very long time ... |
 
|
| Message/Author |
|
|
| LSH posted on Monday, March 06, 2017 - 6:14 am
|
|
|
Good morning,
I have had a bifactor EFA that has been running for about four days now. It's a large analysis (n = 2067, 210 variables, 2 - 12 proposed factors) with 10,000 iterations permitted (since I had trouble getting an earlier version of the model to converge); however, four days seems long even for an analysis this size. My syntax is below; I would appreciate any guidance you can provide.
TITLE: Bifactor EFA;
DATA: FILE IS [trimmed for space]
VARIABLE: NAMES ARE [trimmed for space]
CATEGORICAL ARE ALL; MISSING ARE ALL (999);
ANALYSIS: TYPE = EFA 2 12;
ROTATION = BI-GEOMIN;
ITERATIONS = 10000;
OUTPUT: TECH1 TECH5 modindices (all); |
|
|
|
|
| This is a huge problem especially if the items are polytomous. It may not be realistic. Also, asking for 10,000 iterations instead of the default of 1,000 is going to increase the time for the analysis. You may need to reformulate the analysis, for example, extract just one factor solution as a first step to see how long that take. |
|
|
| LSH posted on Monday, March 06, 2017 - 2:49 pm
|
|
|
| Thank you; I'll try that. |
|
|
| Back to top |
|
|
