Recorded Crime Q4 2017

Introduction

The publication of Recorded Crime Statistics was suspended following the release of the Garda Inspectorate Report in 2014. CSO published a Review of the Quality of Crime Statistics in 2015. Following this review CSO resumed publishing crime statistics but included a note to indicate the limitations of the data to users. Since then there has been further accumulating evidence that suggests the quality of the data recorded on PULSE may not be reliable.

Recorded Crime statistics were suspended ahead of the scheduled 2017 Q1 release in June 2017 due to concerns regarding the accuracy of PULSE data relating to homicide incidents. An Garda Síochána (AGS) initiated an internal review of records relating to 41 suspicious deaths in the period 2013 – 2015. Separately, the CSO had made a number of observations relating to the accuracy of PULSE records relating to homicides over the period 2003 - 2016.

The issues identified both by the internal Garda review and in the CSO observations have required the rectification of PULSE incident records by An Garda Síochána. The changes to the underlying data mean that there are considerable revisions to Recorded Crime statistics from those last published by the CSO in March 2017. These changes are most apparent in Homicides (Group 01).

Homicide refers to all unlawful killings. It is comprised of four subcategories:

• Murder 
• Manslaughter
• Infanticide
• Dangerous driving leading to death

Impact of the revisions

There has been considerable change in the number of recorded homicide incidents during the period 2003 - 2016 from figures previously published by the CSO. The total net increase in the number of homicide incidents for that period is 234, or 18.3%. This is in large part due to the number of incidents of Dangerous Driving Leading to Death, which has been revised upward by 196, or 38.9%.

The revised and previously published data on homicide incidents is shown in tables 3.1 and 3.2, with the difference for each year shown in table 3.3. A comparison for total recorded homicides and total recorded incidents of Dangerous Driving Leading to Death by year are shown in figures 3.1 and 3.2 respectively.

The data has been revised in three of the four homicide categories; Murder, Manslaughter and Dangerous Driving Leading to Death. The majority of revisions occur in relation to incidents of Dangerous Driving Leading to Death with increases ranging from 13.2% or 9 incidents in 2006 to 84% or 21 incidents in 2014. The number of murders has decreased by 3 in the revised data.

Reasons for the revisions

Firstly the Garda review into the classification of suspicious deaths on PULSE has resulted in a number of incident reclassifications.

Secondly, following bilateral engagement between CSO and AGS a number of additional recording errors were identified, including:

1.    Incidents  marked with the incorrect statistical flag.

Where there are multiple crimes at the same time, the most serious one is marked as the primary incident and the others as secondary. Only primary incidents are included in recorded crime statistics. In some instances homicide incidents were incorrectly flagged as secondary incidents.

 2.    Instances where homicide incidents were recoded as non-homicide incidents.

When crime incidents are recorded they are assigned a classification. In some situations it becomes necessary to reclassify the incident as the investigation progresses. An example of this would be where an incident of dangerous driving is reclassified to dangerous driving causing death and in some instances this was not done. 

 3.    Instances where there were multiple unlawful killings in the same incident but only one crime incident was recorded.

 This typically occurred in incidents of Dangerous Driving Causing Death where multiple fatalities occurred but only one crime incident was recorded, rather than one per victim as specified in Crime Counting Rules for homicide. 

 4.    Instances where Homicide incidents were recorded in the incorrect period.

In most instances this only impacts the distribution of the incidents across the period 2003 – 2016 and there is no impact on the total number of incidents for the period. In a small number of instances (less than three) the incidents are no longer included in the data as they occurred prior to 2003. 

 5.    Instances where incidents were misclassified within homicides.

According to the Crime Counting Rules, a reclassification within the homicide group should occur when, for example, a charge of Murder results in a conviction for Manslaughter. This was not done in some instances. This would have no effect on the total number of homicides.

Finally, it was discovered that, as a result of a technical error, CSO were not receiving information relating to incidents updated more than 2 years after first being created in PULSE. This error has been corrected but there will be some, albeit minor, changes to all crime groups periods prior to 2016 as a result.  One of the effects of this in relation to homicide was that on occasions where AGS reclassified incidents from murder to manslaughter this would not have been reflected in the data. 

It should be noted that in the case of some incidents there were multiple data quality issues. 

Further revisions to the data

The period of the Garda review into the classification of suspicious deaths on PULSE has since been extended and this review is ongoing. There are also other outstanding PULSE data quality issues which remain unresolved. As a result it is very likely that there will be further revisions to PULSE records and, accordingly, to recorded crime statistics in both the homicide and other crime groups.

In addition it is likely that further revisions will occur as a result of the complex and lengthy nature of homicide investigations. This complexity and time lag sometimes requires in reclassifications as incidents progress through the criminal justice system.

The Statbank Tables are available here