1996
DOI: 10.1111/j.1745-3984.1996.tb00491.x
|View full text |Cite
|
Sign up to set email alerts
|

Applying the Mantel‐Haenszel Procedure to Complex Samples of Items

Abstract: This Monte Carlo study examined the effect of complex sampling of items on the measurement of differential item functioning (DIF) using the Mantel‐Haenszel procedure. Data were generated using a 3‐parameter logistic item response theory model according to the balanced incomplete block (BIB) design used in the National Assessment of Educational Progress (NAEP). The length of each block of items and the number of DIF items in the matching variable were varied, as was the difficulty, discrimination, and presence … Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

2
44
1

Year Published

1996
1996
2019
2019

Publication Types

Select...
8
1

Relationship

0
9

Authors

Journals

citations
Cited by 22 publications
(47 citation statements)
references
References 14 publications
2
44
1
Order By: Relevance
“…These results are consis tent with the findings of Allen and Donoghue (1994), who have shown in certain cases extremely large Type I errors (as high as 0.8 for a nominal level of 0.05) for simulation models designed to realistically model the American NAEP assessment examination.…”
Section: Extension Of Sibtest To Ordered Polytomous Itemssupporting
confidence: 88%
“…These results are consis tent with the findings of Allen and Donoghue (1994), who have shown in certain cases extremely large Type I errors (as high as 0.8 for a nominal level of 0.05) for simulation models designed to realistically model the American NAEP assessment examination.…”
Section: Extension Of Sibtest To Ordered Polytomous Itemssupporting
confidence: 88%
“…Similar assumptions have been made in the Allen and Donoghue (1996) study. Because Block A is present in all three booklets, it is said that these booklets are ''anchored'' by a common-item Block A (see Table 2).…”
Section: New Methodsmentioning
confidence: 67%
“…When tests are long or when the variation of discrimination across test items is not large, these procedures can be effective (see for example Shealy & Stout, 1993a, or Zwick et al). However for short or moderate length tests and large but realistic levels of variation in item discrimination, the Type I error rate can become substantially larger than the nominal level (see Allen & Donoghue, 1996;Chang, Mazzeo, & Roussos, 1996; for simulated DIF studies using IRT models designed by the authors to model actual test settings. Including the studied item in the matching score X does nothing to prevent the stochastic ordering of the distributions of 9 R | X R = x and 9 F | X F = x.…”
Section: Notes On the Mantel-haenszelmentioning
confidence: 99%