Table 33 supports the investigation of item bias, Differential Group Functioning (DGF), i.e., interactions between classes of items and types of persons. Specify DIF= for person classifying indicators in person labels, and DPF= for item classifying indicators in the item labels.

 

Example output:

You want to examine item bias (DIF) between Females and Males in Exam1.txt. You need a column in your Winsteps person label that has two (or more) demographic codes, say "F" for female and "M" for male (or "0" and "1" if you like dummy variables) in column 9.

 

Table 33.1 is best for pairwise comparisons, e.g., Females vs. Males. Use Table 33.1 if you have two classes of persons, and Table 33.2 if you have two classes of items.

 

Table 33.3 or Table 33.4 are best for multiple comparisons, e.g., regions against the national average. Table 33.3 sorts by item class then person class. Table 33.4 sorts by person class then item class.

 

Table 33.4

-----------------------------------------------------------------------------------

| ITEM      OBSERVATIONS    BASELINE       DGF     DGF   DGF   DGF         PERSON |

| CLASS    COUNT AVERAGE EXPECT           SCORE    SIZE  S.E.   t   Prob.  CLASS  |

|---------------------------------------------------------------------------------|

| 1          234     .46    .46             .00     .06   .28  -.22 .8294  F      |

| 1          221     .51    .50             .01    -.16   .27   .59 .5552  M      |

| 2           54     .85    .84             .01    -.18   .50   .36 .7224  F      |

| 2           51     .86    .86             .00     .00   .61   .00 1.000  M      |

| 3           18     .89    .85             .04    -.50   .83   .60 .5541  F      |

| 3           17     .82    .87            -.04     .76  1.04  -.74 .4734  M      |

| 4           18     .00    .01            -.01     .00< 3.00   .00 1.000  F      |

| 4           17     .00    .01            -.01     .00< 2.40   .00 1.000  M      |

-----------------------------------------------------------------------------------

 

This displays a list of the local difficulty/ability estimates underlying the paired DGF analysis. These can be plotted directly from the Plots menu.

 

DGF class specification identifies the person-label columns containing DIF classifications, with DIF= set to @GENDER using the selection rules. The item-label columns for item classes are specified by DPF=.

 

The DGF effects are shown ordered by Person CLASS within Item CLASS.

 

ITEM CLASS identifies the CLASS of items.

OBSERVATIONS are what are seen in the data

COUNT is the number of observations of the classification used for DGF estimation, e.g., 18 F persons responded to TAP item 1.

AVERAGE is the average observation on the classification, e.g., 0.89 is the proportion-correct-value of item 4 for F persons.

COUNT * AVERAGE = total score of person class on the item

BASELINE is the prediction without DIF

EXPECT is the expected value of the average observation when there is no DGF, e.g., 0.92 is the expected proportion-correct-value for F without DIF.

DGF: Differential Group Functioning

DGF SCORE is the difference between the observed and the expected average observations, e.g., 0.92 - 0.89= -0.03

DGF SIZE is the relative difficulty for this class, e.g., person CLASS F has a relative difficulty of .07 for iten CLASS 1-. ">" (maximum score), "<" (minimum score) indicate measures corresponding to extreme scores.

DGF S.E. is the approximate standard error of the difference, e.g., 0.89 logits

DGF t is an approximate Student's t-statistic test, estimated as DGF SIZE divided by the DGF S.E. with a little less than (COUNT-2) degrees of freedom.

Prob. is the probability of the t-value. This is approximate because of dependencies between the statistics underlying the computation.

KID CLASS identifies the CLASS of persons. KID is specified with PERSON=, e.g., the first CLASS is "F"