Strong Preferences or Simplifying Heuristics? Using Internal Validity Tests and Latent Class Analysis to Better Understand Stated Preference Survey Results. A Case Example in Health Preferences Research
Bibliographic record
Abstract
OBJECTIVES: Internal-validity tests (IVTs) are used in discrete choice experiments (DCEs) to check decision heuristics, choice logic, response consistency, and tradeoffs. There is no standard for how many IVT failures classify respondents as having unacceptable data quality or how to account for failures in choice models. We assessed IVT failures and used latent class analysis to identify choice patterns consistent with statistically informative DCE data. METHODS: We conducted a DCE with 4 attributes (3 ordered), 12 experimental choice tasks, and 2 constructed IVT choice tasks. Respondents with IVT failures were asked questions about their choices. We evaluated preference heterogeneity controlling for attribute dominance using a 4-class latent class model with attribute-specific alternative-specific constants and compared with a 1-class model without attribute-specific alternative-specific constants. RESULTS: Of the 201 respondents, 34 had IVT failures of which 38% to 42% provided reasons other than nonattendance or simplifying heuristics. Comparing the 4-class latent class model no-dominance class with the 1-class model, the coefficients of 2 ordered attributes were significantly different, illustrating potential bias due to simplifying heuristics. Attribute-specific dominance class probability varied by number of choice tasks respondents exhibited attribute dominance on, ranging from 8 to 10 for a class-membership probability of 50%. CONCLUSIONS: IVT "failures" should be interpreted as unexpected responses warranting further inquiry. Including understanding questions could yield insights about stated preferences; however, these increase respondent burden and may not explain simplifying heuristics. Single subjective "rules of thumb" for attribute dominance thresholds may not be adequate. Latent class models controlling for attribute dominance are a data-driven approach that should be considered to assess simplifying heuristics and attribute dominance thresholds.
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 | 0.170 | 0.549 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.003 |
| Bibliometrics | 0.004 | 0.005 |
| Science and technology studies | 0.001 | 0.007 |
| Scholarly communication | 0.007 | 0.014 |
| Open science | 0.002 | 0.003 |
| Research integrity | 0.002 | 0.006 |
| Insufficient payload (model declined to judge) | 0.004 | 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".