This message is printed where many scores are not represented (for your
4 items, certainly you have few individuals with a score of 1 or 2 or 3
). If the -autogroup- option don't "repair" this problem, I suppose that
all the individuals of your sample have a positive or a negative
response to all the five items (and so all the scores are 0 or 4). I
think that none fit tests can be used in this case and that the fit can
not be good (because this is a very deterministic framework to use the
Rasch model).
I hope this helps !
Best
Jean-Benoit
Ousmane FAYE a écrit :
Dear all,
I am running the "Raschtest" command on four dichotomous variables
to analyze fit for an index variable. I used the following
command: "raschtest Q5_111 Q5_114 Q5_113 Q5_112, id(hhid)".
However, I received the following error message: "(Files with residence episodes and individual characteristics) At least one group of scores concerns none individuals. Tests will be not computed. Use the group or the autogroup options."
I tried both options (group or autogroup) but I still receive the same error message. Is there any way to fix this and compute the tests I
need? Your help would be very much appreciated.
Thanks in advance,
Ousmane
Nairobi, Kenya
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
--
Jean-Benoit Hardouin, PhD
Maitre de Conférences - Assistant Professor
EA 4275 "Biostatistics, Clinical Research and Subjective Measures in Health Sciences"
http://www.sante.univ-nantes.fr/biostat/
Departement of Biomathematics and Biostatistics
Faculty of Pharmaceutical Sciences
University of Nantes
1, rue Gaston Veil - BP 53508
44035 Nantes Cedex 1 - FRANCE
Email : [email protected]
Personal website : http://www.anaqol.org
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
