خلاصة:
The present study aimed at investigating DIF sources on an EFL reading comprehension
test. Accordingly, 2 DIF detection methods, logistic regression (LR) and item response
theory (IRT), were used to flag emergent DIF of 203 (110 females & 93 males) Iranian
EFL examinees’ performance on a reading comprehension test. Seven hypothetical DIF
sources were examined in this regard: text familiarity, gender, topic/text interest, guessing,
and the social variables of location, income, and educational status. Only LR, for gender
and text familiarity, could preempt DIF with gender supporting the gendered-text effect
while text familiarity benefiting, inversely, the participants with low level of text
familiarity. For interest in topic, LR found a single item favoring the group with higher
levels of interest and the IRT model detected DIF in either extreme. Regarding guessing
and income, the LR indicated DIF supporting the low guessers and high-income group
whereas IRT, conversely, showed DIF favoring the high guessers and low-income group.
For location and education both methods, correspondingly, demonstrated DIF for the
expensive location and educated groups. Finally, the differential test functioning result
made it clear that only three sources of DIF (gender, income, & interests) were transferred
to the test level. The findings could support a proportional effect of DIF sources.
ملخص الجهاز:
"Such studies have indicated that for example east Asian test takers outperform test takers from European-language backgrounds on Michigan English Language Assessment Battery (Li & Suen, 2012); that females tend to outperform males in tests of verbal and written abilities, especially if constructed response items are included (Willingham & Cole, 1997); they also outperform males on aesthetics, human affairs, and contextualized reading items (Carlton & Harris, 1992) on the mood, impression, or tone items of a reading passage (Pae, 2012), on items related to human relations, human rights, aesthetic and on items referring to stereotypical female activities (O'Neill & McPeek, 1993; Sadker & Sadker, 1994), whereas males outperform females on antonyms, and analogies (Carlton & Harris, 1992), on reading items of inferencing type (Carlton & Harris, 1992; Lawrence & Curley, 1989), on items involving visualization and items calling upon knowledge and experiences acquired outside of school (Hamilton & Snow, 1998), and on items related to science and on items referring to stereotypical male activities (O'Neill & McPeek, 1993; Sadker & Sadker, 1994).
Can text familiarity, gender, interest, guessing, and contextual variables (location, income, and educational status) function as sources of DIF on EFL reading comprehension items?
Method Design of the study Unlike the DIF studies in the literature which select an exploratory approach toward the analysis of DIF sources, the present study benefited from a mixed exploratory-confirmatory approach in this regard; that is, in dealing with social sources of DIF, an exploratory approach was adopted in which the test was administered to the test takers who were divided into groups based on the social variables (location, income, & educational level) and then their performance on the test was analyzed for instances of DIF.
The results of DIF analysis in this regard indicated that only one item (inference-type item) presented this contextual factor as a significant predictor of DIF in favor of the test takers coming from educated families."