The OUTMODEL= data set contains the model specification, the computed parameter estimates, and the standard error estimates. This data set is intended to be reused as an INMODEL= data set in a subsequent analysis by PROC IRT.
The OUTMODEL= data set contains the following variables:
BY variables, if any
_GPNUM_
variable for group numbers, if used
_TYPE_
, a character variable that takes various values that indicate the type of model specification
_SUBTYP_
, a character variable that takes various values that indicate the type of model parameters
_NAME_
, a character variable that indicates the model type, parameter name, or variable name
_NUM_
, a numeric variable that takes various values that indicate model specifications such as the number of items, number of factors,
number of groups, and number of levels for each item
_VAR1_
, a character variable that is the name or number of the first variable in the specification
_VAR2_,
a character variable that is the name or number of the second variable in the specification
_ESTIM_
, a numeric variable that is the final estimate of the parameter
_STDERR_
, a numeric variable that is the standard error estimate of the parameter
Each observation (record) of the OUTMODEL= data set contains a piece of information about the model specification. Depending
on the type of specification that the value of the _TYPE_
variable indicates, the meanings of the _SUBTYP_
, _NAME_
, _NUM_
_VAR1_
, and _VAR2_
variables differ. Table 65.3 summarizes the meanings of these variables for each value of the _TYPE_
variable.
Table 65.3: Meaning of Variables in the OUTMODEL=Data Set
|
Description |
|
|
|
|
|
---|---|---|---|---|---|---|
MODEL |
Model Info |
NGROUP |
Number of groups |
|||
MODEL |
Model Info |
NITEM |
Number of items |
|||
MODEL |
Model Info |
NFACTOR |
Number of factors |
|||
MODEL |
Model Info |
LINK |
Link function |
|||
MODEL |
Model Info |
RORDER |
RORDER option |
|||
MODEL |
Model Info |
DESC |
DESC option |
|||
VAR |
Variable |
Name |
Number of levels |
Response function |
||
FACTOR |
Factor |
Name |
||||
GPVAR |
Group |
Group var name |
||||
PARM |
Parameter |
CEILING |
Name |
Number |
Variable |
|
PARM |
Parameter |
GUESSING |
Name |
Number |
Variable |
|
PARM |
Parameter |
INTERCEPT |
Name |
Number |
Variable |
Category |
PARM |
Parameter |
SLOPE |
Name |
Number |
Variable |
Factor |
PARM |
Parameter |
COV |
Name |
Number |
1st variable |
2nd variable |
PARM |
Parameter |
MEAN |
Name |
Number |
Variable |
For computational convenience, the intercept parameters rather than the threshold (or difficulty) parameters are saved in the OUTMODEL= data set. For multidimensional exploratory analysis, the OUTMODEL= data set includes the unrotated factor loading matrix. In the OUTMODEL= data set, fixed parameters do not have names but only parameter numbers. For free parameters, the OUTMODEL= data set includes the parameter name and the parameter number. When this data set is used for the INMODEL= option, PROC IRT sets an equality constraint for parameters that have the same name.