A preliminary study investigating the overlay method in forensic podiatry for comparison of insole footprints. Issue 5 (September 2022)

• Record Type: Journal Article
• Authors: Howsam, Nicolas; Reel, Sarah; Killey, Jenny
• Is Part Of: Science & justice. Volume 62:Issue 5(2022)
• Page Start: 494
• Page End: 505
• Publication Date: 2022-09
• Subjects: Overlay -- Insole -- Footprint -- Validity -- Forensic -- Podiatry; Forensic sciences -- Periodicals; Criminal investigation -- Periodicals; Forensic Medicine -- Periodicals; Jurisprudence -- Periodicals; Criminalistique -- Périodiques; Enquêtes criminelles -- Périodiques; Criminal investigation; Forensic sciences; Electronic journals; Periodicals; 363.2505
• DOI: 10.1016/j.scijus.2022.07.009
• Languages: English
• ISSNs: 1355-0306

Abstract

Highlights: Findings support advice that the overlay method should not be used in isolation. Agreement between two examiners' comparisons showed high inter-rater consistency. One examiner showed poor agreement with two other examiners' overlay comparisons. Whilst a lack of consistency was found, this is ameliorated at verification stage. Validity of method: Sensitivity: 77.8%; Specificity: 61.9%; FPR: 38.1%; FNR: 22.2%. Abstract: The overlay method is used in forensic podiatry to compare the shape, position, and overall fit of features between questioned and reference footprint evidence. However, the scientific foundation underpinning its validity in the comparison of insole footprints is not fully understood because of a lack of published data defining its statistical operating parameters. A review of literature revealed its subjective nature with little scientific validation, limited only to face validity. The aim of this study investigated strength of agreement between three expert footprint examiners' overlay comparisons of ten reference insole footprints using Cohen's weighted kappa (K w ). Validity of method was explored using measures of sensitivity, specificity, false positive rate (FPR), false negative rate (FNR), receiver operating characteristic (ROC) curve and area under the ROC curve (AUC). Results showed two examiners demonstrated high inter-rater consistency between their overlay comparisons (K w : 0.981; 95 %CI: 0.943 to 1.020 for Rater1 v Rater2) whilst one examiner did not (K w : 0.340; 95 %CI: 0.148 to 0.532 for Rater1 v Rater3; K w : 0.310; 95 %CI: 0.100 to 0.519 for Rater2 v Rater3). Validity was investigated using a conclusion scale in a verbal expressions table to indicate support or rejection of compatibility of footprints between one questioned insole and ten reference insoles. Data analysis indicated validity as follows: Sensitivity: 77.8 %, Specificity: 61.9 %, FPR: 38.1 % and FNR: 22.2 %. ROC analysis corroborated this finding. AUC was calculated at 0.762 or 76.2 % indicating an 'acceptable' measure of overall accuracy of the overlay method for insole footprint comparison. Findings provide novel data supporting previous suggestions that the overlay method should not be used in isolation to compare insole footprints. Data also offers insight into the scientific foundation of this method, whilst highlighting its limitations and providing some implications and recommendations for forensic podiatry practice.