The Danish Crime Harm Index: How It Works and Why It Matters

How can Danish Police calculate a single metric, similar to the Cambridge Crime Harm Index (Sherman, Neyroud and Neyroud 2016 ), that allows the total harm from all victim-reported crimes to be compared over time and across different areas, offenders, gangs, and victims, and what difference can it make in analyzing crime? This analysis examines 2,129,550 reported criminal incidents in the national police crime reports spanning 16 years (2011–2016) and 93% of all of the crime categories in the Danish Criminal Code, excluding those that are primarily detected through proactive policing. It uses the officially recommended days of imprisonment for each crime type to generate the weight of punishment that would be applied if one offender were convicted of each crime, by standards independent of prior criminal history of the offender. The analysis coded the recommended number of days in prison for each offense type based on guidelines set out by the Danish Office of Public Prosecutions. The sentencing value from the prosecutor guidelines was reviewed by five prosecutors. The reliability between the prosecutor ratings and the prosecutor guidelines’ sentencing value was calculated by using Cronbach’s alpha ( α = 0.93). While the count of all victim-reported criminal events in Denmark in 2016 had dropped by 14% (54,000 fewer crimes) from 393,000 in 2011, the Danish Crime Harm Index value (after excluding adult rapes due to a 2015 change in reporting rules) for all victim-reported crimes rose by 0.5% or 46,640 days of recommended imprisonment, from 9,714,057 recommended days of imprisonment in 2011 to 9,760,697 days in 2016. Including rapes, the Danish CHI rose by 6% from 2011 through 2016. Because the Danish Crime Harm Index (DCHI) can lend a completely different perspective (and opposite direction) for interpreting crime trends, it can do the same across individuals and areas within Denmark. The value of adding the DCHI totals to the historic reporting of crime counts would seem to be substantial, at all levels of analysis.