Investigating and optimizing score dependability of a local ITA speaking test across language groups: A generalizability theory approach
Bibliographic record
Abstract
With the present study I investigated the sources of score variance and dependability in a local oral English proficiency test for potential international teaching assistants (ITAs) across four first language (L1) groups, and suggested alternative test designs. Using generalizability theory, I examined the relative importance of L1s (i.e., Indian, Korean, Mandarin, and Spanish), examinees, tasks, and ratings to score variability, and estimated dependability across the L1s. The analyses identified examinees as the largest contributor, which is important for high dependability and validity arguments for test scores. Effects of ratings and tasks were small, but L1 effects on score variance were considerable, with the Indian group’s dependability lowest. Unlike previous generalizability theory studies on L1 effects, however, further analyses revealed that the L1 effects highly likely reflect proficiency differences rather than strong bias when comparing the percent agreement of the ratings, external criteria of examinee English proficiency, and underlying score distributions. I discuss the proficiency differences related to varied socio-linguistic contexts of using and learning English. Lastly, I suggest an alternative design with fewer items and one additional rating for improved dependability. Considering multiple test purposes specific to ITA testing (i.e., efficiency, construct representation, formative advantages), I propose a flexible approach.
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.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.083 | 0.248 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.003 |
| Bibliometrics | 0.004 | 0.003 |
| Science and technology studies | 0.001 | 0.003 |
| Scholarly communication | 0.003 | 0.003 |
| Open science | 0.002 | 0.004 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 0.000 |
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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".