Raju's commands: To run DFIT, there are several steps that you will have to do. 1. get referent group item parameters via Multilog. 2. get focal group item parameters via Multilog. 3. get focal group theta estimates via Multilog. 4. link parameters via Equate (transforming "from" the referent group "to" the focal). 5. Create DFIT command file 6. run DFIT and answer questions regarding files. -------------------MULTILOG RUNS---------------------------------------------- get referent group item parameters via Multilog Created on: 13 February 2003, 14:08:45 >PROBLEM RANDOM, INDIVIDUAL, DATA = 'C:\Temp\ref.dat', NITEMS = 8, NGROUPS = 1, NEXAMINEES = 1000; >TEST ALL, GRADED, NC = (5(0)8); >SAVE ; >END ; 5 12345 11111111 22222222 33333333 44444444 55555555 (21x,8A1) get focal group item parameters via Multilog Created on: 13 February 2003, 14:08:45 >PROBLEM RANDOM, INDIVIDUAL, DATA = 'C:\Temp\foc.dat', NITEMS = 8, NGROUPS = 1, NEXAMINEES = 500; >TEST ALL, GRADED, NC = (5(0)8); >SAVE ; >END ; 5 12345 11111111 22222222 33333333 44444444 55555555 (21x,8A1) get focal group theta estimates via Multilog. Created on: 13 February 2003, 14:08:45 >PROBLEM SCORE, *notice "Random" was changed to "score" INDIVIDUAL, DATA = 'C:\Temp\foc.dat', NITEMS = 8, NGROUPS = 1, NEXAMINEES = 500; >TEST ALL, GRADED, NC = (5(0)8); >SAVE ; >END ; 5 12345 11111111 22222222 33333333 44444444 55555555 (5x,8A1) ------------------------------------EQUATE COMMANDS----------------------------- Equate commands - Note. Check format of of .par files from multilog. -- equating referent (to) and focal (from) groups 21 *quad points g *model used ref.par *referent group item param file created by Multilog (5F12.5) *format of this file 8 *# items l *logistic model foc.par *focal group item param file created by Multilog (5F12.5) *format of this file 8 *#items l *model refeq.par *equated referent group item parameter file to be written by equate. You may or may not use this n *no equating of thetas y *specs are correct 1-8: *anchor items (all items are anchors here) 5 5 *# response options is first 5, which represents a greater theta is second. 5 5 *there is one line of this for each item in referent group 5 5 5 5 5 5 5 5 5 5 5 5 1-8: 5 5 *ditto for focal group 5 5 5 5 5 5 5 5 5 5 5 5 5 5 *this ends the equate program ------------------------------------DFIT COMMAND FILE---------------------------- Note on DFIT - you can either use as input the transformed parameters from equate for time 2 (refeq.par) and put scalers to 1.0 and 0, or you can read in original from multilog and put scaling values into Raju program as this program does. Notice that if you use the 1.0 and 0.0 scaling values that you will have to change the format of the input .par file. (DFIT command file) Comparing referent and focal group data *Title line 8 1.000 .096 1.000 0.000 *# items, multiplicative term, NCDIF cutoff, 2 scaling constants 11111111 *lowest reponse option for each item 55555555 *highest response option for each item (5F12.5) *format of focal group item param file (/5F10.3) *format of equated referent group item param file (F10.3) *format of focal group theta file 12345 *response options, 1 of these for each item 12345 12345 12345 12345 12345 12345 12345 *end file (here it is saved as "dfit.inp") -----------------------------------------filenames file-------------------------- *you need to either create a .txt file with all of the file names, or just type Dfitp6 and follow instructions. Below are the contents for the .txt file.: dfit.inp * file that contains the above foc.par * focal group item params (from initial multilog run) refeq.par * referent group item params (I'm using the equated file from equate output) foc2.sco * focal group thetas dfit.out * dfit output file * this file was saved as 'dfitfiles.txt'. * you can then goto DOS, cd until you get to the folder with all the data, specify the path and type "dfitps6