IAT faking indices revisited: Aspects of replicability and differential validity
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
Research demonstrates that IATs are fakeable. Several indices [either slowing down or speeding up, and increasing errors or reducing errors in congruent and incongruent blocks; Combined Task Slowing (CTS); Ratio 150-10000] have been developed to detect faking. Findings on these are inconclusive, but previous studies have used small samples, suggesting they were statistically underpowered. Further, the stability of the results, the unique predictivity of the indices, the advantage of combining indices, and the dependency on how faking success is computed have yet to be examined. Therefore, we reanalyzed a large data set (N = 750) of fakers and non-fakers who completed an extraversion IAT. Results showed that faking strategies depend on the direction of faking. It was possible to detect faking of low scores due to slowing down on the congruent block, and somewhat less with CTS-both strategies led to faking success. In contrast, the strategy of increasing errors on the congruent block was observed but was not successful in altering the IAT effect in the desired direction. Fakers of high scores could be detected due to slowing down on the incongruent block, increasing errors on the incongruent block, and with CTS-all three strategies led to faking success. The results proved stable in subsamples and generally across different computations of faking success. Using regression analyses and machine learning, increasing errors had the strongest impact on the classification. Apparently, fakers use various goal-dependent strategies and not all are successful. To detect faking, we recommend combining indices depending on the context (and examining convergence).
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 | 0.010 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.012 | 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