Bibliographic record
Abstract
In a recent paper published in Addiction, Rogeberg & Elvik 1 replicated and criticized two meta-analyses synthesizing the epidemiological evidence for marijuana use and crash risk in drivers 2, 3. The two meta-analyses were performed independently by two research teams (ours at Columbia University and Asbridge et al.'s [3] at Dalhousie University) and were published a few months apart in 2012 in two highly regarded academic journals. While we appreciate Rogeberg & Elvik's interest in our work, we feel it is incumbent upon us to respond to their criticisms of our study. Rogeberg & Elvik's critique of our study centers on three main issues: study selection, estimated effect size and data extraction. First, Rogeberg & Elvik state that our ‘selection criteria are unclear and hard to rationalize’. Study selection is an important step in any meta-analysis. In the Methods section, we stated explicitly that we followed ‘the guidelines put forth in the PRISMA statement and MOOSE guidelines for reporting systematic reviews and meta-analyses of observational studies in Epidemiology’ and spelled out the precise search terms as well as eligibility criteria, including a detailed 11-level graph illustrating the study selection process. We chose to cast a wide net in terms of exposure to cannabis use within studies, entering results as they were reported by the original investigators. We made clear that our estimated summary odds ratio was based on empirical data reported in the source studies and was not adjusted for any confounding factors. Albeit imperfect, our approach reflected accurately the state of the evidence as supported by the epidemiological literature. Secondly, Rogeberg & Elvik assert that the comparable odds ratios reported by us and Asbridge et al. overestimate substantially the true odds ratio of crashes ascribed to marijuana-induced impairment. This criticism, based on results of simulation exercises with sparse data and unsubstantiated assumptions, is unwarranted and misguided. The studies included in our meta-analysis assessed the associations of crash risk with marijuana use as indicated by drug testing or self-reporting, not assumed marijuana-induced impairment. Simulation models are no substitute for empirical evidence. Finally, Rogeberg & Elvik state that they found a data abstraction error regarding the Fergusson & Horwood study 4 in our meta-analysis, ‘where the correct OR was 1.4 rather than the 2.4 extracted by Li et al.’ There is no error in our data abstraction. The Fergusson & Horwood study 4 is the only study included in our meta-analysis that is based on a cohort design. Accordingly, the authors reported their results in rate ratios based on person-years. To include the data from this cohort study in our meta-analysis, we reconstructed the rate ratio data presented in Table 1 of the Fergusson & Horwood paper and converted them into odds ratios. This conversion is necessary, because other studies included in our meta-analysis were based on case–control and cross-sectional designs and reported their findings in odds ratios. Rogeberg & Elvik do not seem to understand the differences between these epidemiological study designs. It is apparent that they mistake the rate ratio reported by Fergusson & Horwood for odds ratio. None.
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 | 0.020 | 0.095 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.003 | 0.012 |
| Bibliometrics | 0.005 | 0.004 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.004 | 0.001 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.002 | 0.003 |
| 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".