Homicide Resources, Solvability and Detection

Abstract

The key objectives in this chapter are, first, to understand how homicide investigations are resourced, and the ways in which investigative costs, investigative actions and exhibits (output evidence items) are related to detection outcomes. A second important objective is to identify solvability factors which distinguish between detected and undetected cases, and to identify factors that help predict differences in solvability among the detected cases. Third, the accuracy of detectives’ homicide classification for predicting case costs in England and Wales is also assessed, and fourth differences between the two police jurisdictions in the study are examined. The study is based on a population of 290 homicides committed in Hampshire and Thames Valley Police jurisdictions, which cover shire counties in lowland England and Wales. There is a fuller set of homicide characteristics for Thames Valley Police data which enables a more comprehensive solvability factor analysis. Nine-tenths of cases were detected. Solvability factors explained over a third of detection outcomes whereas investigative costs did not help explain whether or not cases were detected. Solvability factors were identified which included suspect numbers, homicide type (murder/manslaughter/infanticide), offender–victim relationship (stranger/friend/relation) and suspect age. In contrast to existing research, whether the homicide occurred in a private or public place and whether knives or firearms were used had no effect on detections. Murders were less solvable than family-related incidents while homicides where hammers, golf clubs, spades, axes and blunt instruments were used as weapons reduced the prospects of a detected outcome. At odds with other studies, stranger homicides were linked with higher not lower odds of detection, while offences involving drugs tended to improve detection odds, a probable reflection of how solvability factors differ under circumstances of virtually unlimited resourcing and very high detection rates. Given exceptionally high detection rates, alternative indicators of solvability were evaluated that could be used to predict solvability differences among the detected cases. These indicators included number of investigative actions, number of exhibits per action and costs per action, all of which differed between detected and undetected incidents, plus time taken to charge suspects. Higher numbers of actions, less costly actions, fewer exhibits per action and greater time taken to charge suspects indicate lower solvability and characterise difficult-to-solve cases for which additional, often unproductive, investigative activities are undertaken in exhausting possible investigative avenues. Homicide characteristics predicted 28% of the variation in number of investigative actions which was the best alternative solvability indicator. It is notable that resources had little, if any, statistical effect on detection outcomes. This is because these homicide investigations were fully funded so that investigative effort continued to be expended until all investigative options were tried. This resulted in more actions, albeit lower costs and less productive ones, as case solvability fell, so that earlier cases involved fewer more costly and productive actions, while later cases involved more less costly actions which produced fewer exhibits. This tends to equalise investigative costs and make undetected cases almost as costly. Hence resources tended to reflect case solvability and add little statistical explanation of detection outcomes. Police homicide grades based on detectives’ experience were good predictors of homicide costs, though it appears there is some potential to improve these predictions by drawing on statistical analysis of solvability factors. That detectives’ grades proved to be better predictors of investigative costs than variables measuring homicide characteristics is undoubtedly affected by the limited number and range of homicide characteristics in the data set, whereas detectives have all factors available to them on which to base their case cost estimates. Investigative costs were higher in Thames Valley Police jurisdiction, the principal effect of which was to accelerate the investigation so that suspects were charged earlier than in Hampshire Police Force jurisdiction. There is a need for more detailed examination of investigative costs and the type—witness, forensic, electronic—and quality of evidence—which results from various investigative actions for different types of homicide to understand their role in detecting cases. This would provide insights into how to make homicide investigation more cost-effective by tailoring investigative actions to particular sorts of incidents.