﻿ Table 33.1, 33.2 Differential group functioning DGF pairwise

# Table 33.1, 33.2 Differential group functioning DGF pairwise

This Table identifies Differential Group Functioning DGF interactions between classification-groups of persons (identified by DIF=, referencing the person labels) and classification-groups of items (identified by DPF=, referencing the item labels) using the column selection rules. Differential average classification-group performance (DGF) is powerful when looking for latent classes among the persons. Differential bundle functioning (DBF) is powerful when looking for local dependence among the items. For more details, see Table 30 (DIF) and Table 31 (DPF). A graphing technique can be used to display DIF item characteristic curves for non-uniform DIF.

Table

33.1 DGF report (paired person classes on each item class)

33.2 DGF report (paired item classes on each person class)

33.3 DGF report (list of person classes within item class)

33.4 DPF report (list of item classes within person class)

We put a code in the person label of every person indicating whether the person belongs to the Control Group, "C", or the Treatment Group, "T". The column for this code is the Differential Item Functioning DIF= column. Then, if we want a report for every item, it is a DIF analysis, Winsteps Table 30.

If we want to group items, then we put a code in the item label of every item which indicates the item group to which the item belongs. The column for this code is the Differential Person Functioning, DPF= column. Then, if we want a report for every person, it is a DPF analysis, Winsteps Table 31.

If we want to do "person group" with "item group" . Then this is DGF (Differential Group Functioning). Person DIF= column and Item DPF= column, Winsteps Table 33.

DGF analysis: the log-odds model for an individual person in the group and item in the group is:

log(Pni1/Pni0) = Bgn - Dhi - Mgh

where

Bgn is the overall estimate from the main analysis of the ability of person n (who is in group DIF g)

Dhi is the overall estimate from the main analysis of the difficulty of item i (which is in DPF group h)

Mgh is the interaction (bias, DGF) for person group g on item group h. This is estimated from all persons in group g combined with all items in group h.

The DGF dialog displays when Table 33 is called from the Output Tables or Plots menus.

Table 33.1

DGF CLASS-LEVEL BIAS/INTERACTIONS FOR DIF=@GENDER AND DPF=\$S1W1

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

| PERSON   DGF    DGF   DGF  PERSON   DGF    DGF   DGF    DGF    JOINT  Rasch-Welch    ITEM  |

| CLASS  SCORE   SIZE  S.E.  CLASS  SCORE   SIZE  S.E. CONTRAST  S.E.   t  d.f. Prob.  CLASS |

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

| F        .00    .06   .28  M        .01   -.16   .27      .22   .39   .57 452 .5708   1    |

| F        .01   -.18   .50  M        .00    .00   .61     -.18   .79  -.23 101 .8208   2    |

| F        .04   -.50   .83  M       -.04    .76  1.04    -1.26  1.33  -.95  32 .3484   3    |

| F       -.01    .00< 3.00  M       -.01    .00< 2.40      .00  3.84   .00  32 1.0000  4    |

| M        .01   -.16   .27  F        .00    .06   .28     -.22   .39  -.57 452 .5708   1    |

| M        .00    .00   .61  F        .01   -.18   .50      .18   .79   .23 101 .8208   2    |

| M       -.04    .76  1.04  F        .04   -.50   .83     1.26  1.33   .95  32 .3484   3    |

| M       -.01    .00< 2.40  F       -.01    .00< 3.00      .00  3.84   .00  32 1.0000  4    |

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

The most important numbers in Table 33.1: The DGF CONTRAST is the difference in difficulty of the item between the two groups. This should be at least 0.5 logits for DGF to be noticeable. "Prob." shows the probability of observing this amount of contrast by chance, when there is no systematic item bias effect. For statistically significance DGF on an item, Prob. ≤ .05.

Table 33.2

DGF CLASS-LEVEL BIAS/INTERACTIONS FOR DIF=@GENDER AND DPF=\$S1W1

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

| ITEM    DGF    DGF   DGF  ITEM    DGF    DGF   DGF    DGF    JOINT  Rasch-Welch    PERSON |

| CLASS SCORE   SIZE  S.E.  CLASS SCORE   SIZE  S.E. CONTRAST  S.E.   t  d.f. Prob.  CLASS  |

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

| 1       .00    .06   .28  2       .01   -.18   .50      .24   .57   .42 119 .6774   F     |

| 1       .00    .06   .28  3       .04   -.50   .83      .56   .87   .64  30 .5261   F     |

| 1       .01   -.16   .27  2       .00    .00   .61     -.16   .67  -.24 100 .8097   M     |

| 1       .01   -.16   .27  3      -.04    .76  1.04     -.92  1.07  -.86  25 .3973   M     |

| 2       .01   -.18   .50  1       .00    .06   .28     -.24   .57  -.42 119 .6774   F     |

| 2       .01   -.18   .50  3       .04   -.50   .83      .32   .97   .33  39 .7420   F     |

This Table contrasts, for each item class, the size and significance of the Differential Item Functioning for pairs of person classifications.

DGF class specification defines the columns used to identify DGF classifications, using DIF= and DPF=, see the selection rules.

Reading across the Table 33.1 columns:

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

DGF estimates with the  the iterative-logit (Rasch-Welch) method:

DGF SCORE is the average response score-point difference between the observed and the expected scores for this PERSON CLASS on this ITEM CLASS. Higher scores mean locally higher ability or locally lower difficulty.

DGF SIZE is the differential difficulty of this item (scaled by USCALE=) for this class, with all else held constant, e.g., .07 is the relative difficulty for Kid Class F on Item Class1. The more difficult, the higher the DGF measure.
-.52> reports that this measure corresponds to an extreme maximum person-class score. EXTRSCORE= controls extreme score estimate.
1.97< reports that this measure corresponds to an extreme minimum person-class score. EXTRSCORE= controls extreme score estimate.
-6.91E reports that this measure corresponds to an item with an extreme score, which cannot exhibit DIF

DGF S.E. is the standard error of the DGF SIZE (scaled by USCALE=).

PERSON CLASS identifies the CLASS of persons, e.g., the second CLASS is "M".

DGF SCORE is the average response score-point difference between the observed and the expected scores for this PERSON CLASS on this ITEM CLASS.

DGF SIZE is the differential difficulty of this item for this class, with all else held constant, e.g., -.15 is the relative difficulty for Kid Class M on Item Class1. The more difficult, the higher the DGF measure.

DGF S.E. is the standard error of the second DGF SIZE.

DGF CONTRAST is the difference between the two DGF SIZE, i.e., size of the DGF across the two classifications of persons, e.g., .07 - -.15 = .23 (usually in logits). A positive DGF contrast indicates that the item is more difficult for the first, left-hand-listed CLASS. See details in Table 33.3.

JOINT S.E. is the standard error of the DGF CONTRAST = sqrt(first DIF S.E.² + second DIF S.E.²), e.g., .38 = sqrt(.27² + .27²)

t gives the DGF significance as a Student's t-statistic = DGF CONTRAST / JOINT S.E. The t-test is a two-sided test for the difference between two means (i.e., the estimates) based on the standard error of the means (i.e., the standard error of the estimates). The null hypothesis is that the two estimates are the same, except for measurement error.

d.f. is the joint degrees of freedom. This is shown as the sum of the sizes of two classifications (see Table 33.3 less 2 for the two measure estimates, but this estimate of d.f. is somewhat high, so interpret the t-test conservatively, e.g., d.f. = (426 A + 1 D - 2) = 425. When the d.f. are large, the t statistic can be interpreted as a unit-normal deviate, i.e., z-score.

INF means "the degrees of freedom are so large they can be treated as infinite", i.e., the reported t-value is a unit normal deviate.

Prob. is the two-sided probability of Student's t. See t-statistics.

ITEM CLASS is the item classification specified by DPF=. Here the first ITEM CLASS is "1"

Each line in the Table is repeated with the PERSON CLASSes in reversed order.

Help for Winsteps Rasch Measurement Software: www.winsteps.com. Author: John Michael Linacre

 Forum Rasch Measurement Forum to discuss any Rasch-related topic

Rasch Publications
Rasch Measurement Transactions (free, online) Rasch Measurement research papers (free, online) Probabilistic Models for Some Intelligence and Attainment Tests, Georg Rasch Applying the Rasch Model 3rd. Ed., Bond & Fox Best Test Design, Wright & Stone
Rating Scale Analysis, Wright & Masters Introduction to Rasch Measurement, E. Smith & R. Smith Introduction to Many-Facet Rasch Measurement, Thomas Eckes Invariant Measurement with Raters and Rating Scales: Rasch Models for Rater-Mediated Assessments, George Engelhard, Jr. & Stefanie Wind Statistical Analyses for Language Testers, Rita Green
Rasch Models: Foundations, Recent Developments, and Applications, Fischer & Molenaar Journal of Applied Measurement Rasch models for measurement, David Andrich Constructing Measures, Mark Wilson Rasch Analysis in the Human Sciences, Boone, Stave, Yale
in Spanish: Análisis de Rasch para todos, Agustín Tristán Mediciones, Posicionamientos y Diagnósticos Competitivos, Juan Ramón Oreja Rodríguez
Winsteps Tutorials Facets Tutorials Rasch Discussion Groups

Coming Rasch-related Events
Jan. 5 - Feb. 2, 2018, Fri.-Fri. On-line workshop: Practical Rasch Measurement - Core Topics (E. Smith, Winsteps), www.statistics.com
Jan. 10-16, 2018, Wed.-Tues. In-person workshop: Advanced Course in Rasch Measurement Theory and the application of RUMM2030, Perth, Australia (D. Andrich), Announcement
Jan. 17-19, 2018, Wed.-Fri. Rasch Conference: Seventh International Conference on Probabilistic Models for Measurement, Matilda Bay Club, Perth, Australia, Website
Jan. 22-24, 2018, Mon-Wed. In-person workshop: Rasch Measurement for Everybody en español (A. Tristan, Winsteps), San Luis Potosi, Mexico. www.ieia.com.mx
April 10-12, 2018, Tues.-Thurs. Rasch Conference: IOMW, New York, NY, www.iomw.org
April 13-17, 2018, Fri.-Tues. AERA, New York, NY, www.aera.net
May 22 - 24, 2018, Tues.-Thur. EALTA 2018 pre-conference workshop (Introduction to Rasch measurement using WINSTEPS and FACETS, Thomas Eckes & Frank Weiss-Motz), https://ealta2018.testdaf.de
May 25 - June 22, 2018, Fri.-Fri. On-line workshop: Practical Rasch Measurement - Core Topics (E. Smith, Winsteps), www.statistics.com
June 27 - 29, 2018, Wed.-Fri. Measurement at the Crossroads: History, philosophy and sociology of measurement, Paris, France., https://measurement2018.sciencesconf.org
June 29 - July 27, 2018, Fri.-Fri. On-line workshop: Practical Rasch Measurement - Further Topics (E. Smith, Winsteps), www.statistics.com
July 25 - July 27, 2018, Wed.-Fri. Pacific-Rim Objective Measurement Symposium (PROMS), (Preconference workshops July 23-24, 2018) Fudan University, Shanghai, China "Applying Rasch Measurement in Language Assessment and across the Human Sciences" www.promsociety.org
Aug. 10 - Sept. 7, 2018, Fri.-Fri. On-line workshop: Many-Facet Rasch Measurement (E. Smith, Facets), www.statistics.com
Sept. 3 - 6, 2018, Mon.-Thurs. IMEKO World Congress, Belfast, Northern Ireland www.imeko2018.org
Oct. 12 - Nov. 9, 2018, Fri.-Fri. On-line workshop: Practical Rasch Measurement - Core Topics (E. Smith, Winsteps), www.statistics.com

Our current URL is www.winsteps.com