Reconciling disparate data to determine the <i>right</i> answer: A grounded theory of meta analysts' reasoning in meta‐analysis
Bibliographic record
Abstract
While the systematic review process is intended to maximize objectivity and limit researchers' biases, examples remain of discordant recommendations from meta-analyses. Current guidelines to explore discrepancies assume the variation is produced by methodological differences and thus focus only on the study process. Because heterogeneity of interpretation also occurs when experts examine the same data, our purpose was to examine if there are reasoning differences, ie, in how information is processed and valued. We created simulated meta-analyses based on idealized randomized studies (ie, perfect studies with no bias) to ensure differences in interpretations could only be due to reasoning. We recruited published meta-analysts using purposeful variables. We conducted 3 audio-recorded interviews per participant using structured and semi-structured interviews, with paraphrasing and reflective listening to enhance and verify responses. Recruitment and analysis of transcripts and field notes followed the principles of grounded theory (eg, theoretical saturation, constant comparative analysis). Results show the complexity of meta-analytic reasoning. At each step of the process, participants attempted to reconcile disparate forms of knowledge to determine a right answer (moral concern) and accurately draw a treatment effect (epistemological concern). The reasoning processes often shifted between considering the meta-analysis as if the data were whole, and as if the data were discrete components (individual studies). These findings highlight paradigmatic tensions regarding the epistemological premises of meta-analysis, resembling previous historical investigations of the functioning of scientific communities. In understanding why different meta-analysts interpret data differently, it may be unrealistic to expect objective homogenous recommendations based on meta-analyses.
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 | 0.796 | 0.820 |
| Meta-epidemiology (narrow) | 0.006 | 0.008 |
| Meta-epidemiology (broad) | 0.018 | 0.023 |
| Bibliometrics | 0.044 | 0.023 |
| Science and technology studies | 0.008 | 0.032 |
| Scholarly communication | 0.027 | 0.019 |
| Open science | 0.021 | 0.019 |
| Research integrity | 0.014 | 0.024 |
| Insufficient payload (model declined to judge) | 0.004 | 0.001 |
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".