Differential Item Functioning: The Consequence of Language, Curriculum, or Culture?
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
In recent decades, the use of large-scale standardized international assessments has increased drastically as a way to evaluate and compare the quality of education across countries. In order to make valid international comparisons, the primary requirement is to ensure the measurement equivalence between the different language versions of these assessments due to their multilingual and cross-cultural nature. In this study, we investigated the measurement equivalence of one of the most popular international assessments, PISA (Programme for International Student Assessment), between U.S. and Canadian, Hong Kong and mainland Chinese, and U.S. and mainland Chinese students. Both unidimensional and multidimensional random coefficient multinomial logit model (RCML) were applied to detect differential item functioning (DIF). Furthermore, we exerted great efforts to identify possible explanations of DIF via detailed content analyses. The results showed that the number of DIF items is the smallest between Canadian and U.S. students and the largest between U.S. and Chinese students. We also noticed that for all three comparisons the number of DIF items reduced significantly when we analyzed the data using the multidimensional approach. Our content analysis revealed that language difference only accounted for a small proportion of DIF between U.S. and Chinese students, whereas differential curriculum coverage was found to be the most serious cause of DIF in both the Hong Kong-Mainland and the U.S.-Chinese comparisons. In addition, we found that differential content familiarity is also a potential cause of DIF. Further investigations of more potential sources of item bias require the collection of additional data.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.046 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.003 | 0.002 |
| Open science | 0.003 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.012 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it