Intra- and inter-rater reliability of a manual codification system for footwear impressions: first lessons learned from the development of a footwear database for forensic intelligence purposes
Bibliographic record
Abstract
To generate forensic intelligence from footwear impressions and link crime scenes, most law enforcement agencies and forensic laboratories rely on a manual codification system based on pattern recognition and classification by human analysts. However, although they are commonly used in practice, to date we still know little about the reliability of such systems. Taking advantage of the development of a footwear database for forensic intelligence purposes at the Laboratoire de sciences judiciaires et de médecine légale in Quebec (Canada), this study aims to make a preliminary assessment of the intra- and inter-rater reliability (i.e., the level of repeatability over time and the level of consensus between analysts) of the proposed codification system. To do so, three forensic intelligence analysts classified a set of 27 crime scene impressions and test impressions at two different times (two weeks apart). Percent agreement, Cohen’s Kappa, and Light’s Kappa were then calculated. Results show that two out of three analysts have reached an almost perfect level of intra-rater agreement, while the other have achieved a substantial level of intra-agreement, and that all analysts have reached a substantial level of inter-rater agreement. Findings suggest that, although a few patterns may have lower levels of agreement, overall, the developed codification system presents a satisfactory level of reliability. This preliminary study thus suggests that contrary to what advocates of fully automated systems may sometimes imply, manual codification of footwear impressions may be fairly appropriate for intelligence purposes. It calls for further evaluative research in the field.
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 | 0.142 | 0.233 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.003 | 0.002 |
| Open science | 0.002 | 0.003 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 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; 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".