The internal consistency and accuracy of automatically scored written receptive meaning-recall data: A preliminary study
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
Vocableveltest.org is a testing platform on which users can create online self-marking meaning-recall (reading or listening) and form-recall (typing) tests that address a number of limitations of the existing vocabulary level tests and vocabulary size tests. A major limitation of many existing vocabulary tests is the written receptive meaning-recognition (multiple-choice or matching) format which is associated with increased error due to guessing and decreased power to measure the type of vocabulary knowledge suitable for reading practice (McLean et al., 2020, Stewart et al., 2021a, Stoeckel et al., 2021), despite being designed for this purpose (Nation, 2012, Schmitt et al., 2020, Webb et al., 2017). Conversely, scoring meaning-recall tests by hand is labour-intensive, and the internal consistency and accuracy of automatically marked data are unknown. Thus, this study investigated the internal consistency and accuracy of automatically marked responses of 98 words from the fifth 100 most frequent words of English. This study tested for knowledge of high-frequency words as a more robust test of the marking system, as these words possess multiple-meaning senses, making their automatic marking problematic. Furthermore, the predicted limited range of learners’ knowledge of these 98 words was expected to result in data of a low internal consistency. However, the automatically marked data had a high internal consistency (Cronbach’s α = 0.868) and was 98% similar to human marked meaning-recall responses.
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 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 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 it