Multi‐cohort evaluation of “Don't know” responders to self‐report oral health questions: Implications for etiologic research
Bibliographic record
Abstract
BACKGROUND: Self-reported data can extend the reach of oral health research, but "Don't know" responses may threaten validity. We explored characteristics of participants who responded "Don't know" to a periodontal health question across three distinct cohorts. METHODS: We used data from three questionnaire-based observational studies, namely, the Pregnancy Study Online (PRESTO) (N = 10,996), the Black Women's Health Study (BWHS) (N = 479), and the National Health and Nutrition Examination Survey (NHANES) (N = 15,502), to evaluate responses to questionnaire items related to periodontal health (e.g., "Has a dentist or dental hygienist ever told you that you have periodontitis or gum disease?"). We compared sociodemographic and behavioral factors across each response category ("Yes," "No," "Don't know"). We used Monte Carlo simulation to create multiple datasets of 100,000 participants under different scenarios to calculate the percent change in observed effect estimates in analyses using the full cohort compared to analytic cohorts excluding "Don't know" respondents. RESULTS: "Don't know" prevalences ranged from 1.6% to 4.1%. We observed differences between "Don't know" responders and those who answered "Yes" or "No" across all three cohorts. "Don't know" responders were more likely to have lower educational attainment, lower income, and reduced engagement with oral healthcare services. We observed substantial bias in complete-case effect estimates in some simulated scenarios. Bias was larger when the underlying population prevalence of "Don't know" responses was higher. CONCLUSIONS: "Don't know" responders had distinct patterns of sociodemographic characteristics and oral healthcare engagement. The degree of bias in complete-case analysis was dependent on simulated factors. PLAIN LANGUAGE SUMMARY: Research about oral health often asks people to answer questions about their teeth and gums. Sometimes people respond that they "Don't know" the answer to these questions, which can make data challenging for researchers to analyze. In this study, we used three different data sources to look at whether there were particular characteristics that were more common among people who said they "Don't know" in response to a question about their gum health. "Don't know" responses were not very common in any of the three groups, ranging from 1.6% in a representative survey of people in the United States to 4.1% in a group of women in the United States and Canada trying to become pregnant. In all three groups, people who said "Don't know" had a lower household income, less education, and were less likely to have seen a dentist recently. We also used simulated datasets to evaluate when excluding people who responded "Don't know" would be expected to cause the most bias in analyses. The expected bias increased with the number of "Don't know" responses in the 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.
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 | 0.306 | 0.347 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.004 |
| Bibliometrics | 0.003 | 0.003 |
| Science and technology studies | 0.003 | 0.003 |
| Scholarly communication | 0.003 | 0.003 |
| Open science | 0.004 | 0.003 |
| Research integrity | 0.003 | 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; the direct Gemma label and the distilled Codex classifier agree on what is shown here.
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".