BIFACTOR
bifactor.zip (121 KB)

The BIFACTOR program estimates the bifactor model for ordinal and dichotomous data. The bifactor model can be used to represent factorial structures for many types of psychological and educational tests, since they can exhibit a general factor and one or more group or method factors. The bifactor structure results from the constraint that each item has a nonzero loading on the primary dimension and, at most, one of the group factors.
 
Authors list:
Robert D. Gibbons and Donald Hedeker
 

In Full-Information Item Bifactor Analysis of Graded Response Data View PDF, Gibbons et al. develop estimation procedures for fitting the graded response model when the data follow the bifactor structure. A limited-information solution based on polychoric correlations using robust WLS, which is available in Mplus software for the full-information approach, is presented as an alternate estimation method for the bifactor model. Read a comparison of estimates obtained from Mplus and BIFACTOR. View PDF
 

The downloaded bifactor.zip contains the following files:
POLYBIF.EXE - the BIFACTOR program
README.TXT - important notes on the BIFACTOR program
TEST1.DEF through TEST5.DEF - input files for POLYBIF.EXE
bi-factor-rating scale-theta-y1at0.002.inp - Mplus program setup file
bi-factor-rating scale-theta-y1at0.002.out - Mplus program output file
 
1. Please fill in your name and e-mail address.
  By providing this information you are consenting to be the potential recipient of an email questionnaire about the program. If you receive the questionnaire, you are under no obligation to fill it out and return it, however, your response will be appreciated.
   
  Name  
       
  E-mail  
   
2. Press the Download button. If prompted whether to open or save the file, choose to save it to your computer.
   
   
 
Last updated February 8, 2008. Copyright © 2003-2008 Center for Health Statistics. All Rights Reserved.
Center for Health Statistics Home UIC Home