Introduction
Crime statistics are a vital information source on the prevalence and type of crime in Ireland. They are of significant public interest. Trust in these statistics is important and high-quality information is a critical element of evidence-based policy-making.
The CSO publish Recorded Crime statistics in accordance with Section 10 of the Statistics Act 1993 and Section 47 of the Garda Síochána Act 2005. Statistics produced by the Central Statistics Office (CSO) are compiled in accordance with the European and the Irish Statistical Codes of Practice (ESSCOP and ISSCOP). Records of crime incidents recorded in PULSE by An Garda Síochána (AGS) are the only source of data available for Recorded Crime statistics.
In March 2018 the CSO resumed publication of Recorded Crime statistics using the category of Statistics Under Reservation. This allowed CSO to provide the best available measure of police-recorded crime in Ireland while simultaneously informing users of their concerns regarding the quality of the underlying data.
The Statistics Under Reservation categorisation indicates that the quality of Recorded Crime statistics does not meet the standard required of official statistics published by CSO and is in line with principle 15 of ESSCOP, whereby users should be informed of the quality status of the statistics they are using.
This document identifies the actions which the CSO see as critical to improving the quality of PULSE data. It delineates the criteria which must be satisfied by AGS for the categorisation as Statistics Under Reservation to be removed.
The actions in this document are in line with the AGS Policing Plan for 2018. The Policing Plan identifies actions to be undertaken under the general heading of “Improved Data Quality.” The Policing Plan outlines a commitment to ensure “that all incidents are correctly classified” and that the “data quality on our PULSE system [should be enhanced]”. These actions are a fundamental starting point for addressing data quality, and show a shared understanding in CSO and AGS of the need for data quality improvement.
The CSO will continue to assess data quality and to address data quality concerns while improvements are ongoing and will assess progress made on the actions listed in the document. The CSO are committed to assisting AGS in improving data quality wherever possible. The decision to re-categorise Recorded Crime statistics, or otherwise, lies with the Director General of the CSO.
Action Area 1: All data governance activities should be led by, and overall responsibility for data quality should be assigned to, one senior manager.
Effective data governance is essential for assuring the quality of crime data. Roles and responsibilities in respect of the recording, validation, review and management of data need to be identified and assigned to appropriate personnel. The assignment of overall responsibility for data quality to one individual is an important step in ensuring that data quality improvement initiatives are focused and coherent. It will make it easier to resolve data quality issues and to drive positive change.
Data quality should be monitored on a continuous basis and audited regularly to ensure that both the procedures and outputs are fit for purpose.
A data risk register should be established and maintained outlining the risks associated with the recording and use of crime data. Risk management is an essential element of all data production processes. These risks include processing risks such as recording and coding errors and risks around data analysis. It is important is that a risk assessment of the current crime data production process be conducted. This assessment should identify the controls in place to mitigate and manage data quality risks.
Training must be fit-for-purpose to ensure that those with responsibility for incident and data recording are cognisant of the value and relevance of data, and are up-to-date on policy and procedural best-practice.
Action Area 2: A data quality management framework must be devised which clearly outlines what constitutes good quality data, what data quality control mechanisms are in place, and how data quality is to be measured, monitored and independently audited to ensure fit-for-purpose crime data.
It will be important for AGS to identify and specify what is meant by good quality data and how it is to be measured. A shared understanding of what constitutes “fit for purpose” crime data is a fundamental starting point in improving data quality.
In order for the CSO to be able to compile statistics in which users can have trust, sound data quality principles and methodologies must be employed by AGS at the recording stage so as to ensure a high standard of quality in the underlying administrative data source. These principles (with illustrative, not exhaustive, examples) include:
Are user needs around recorded crime statistics are being met? For example, the need for good quality data on incidents with domestic abuse or discriminatory motives etc.
Does the data recorded on the PULSE system accurately reflect the information known? Are the data fields recorded correctly (e.g. dates, incident classifications, locations, names, dates of birth)? Are the data fields amended if relevant information changes?
Are data being entered onto the recording system in a timely manner? Is there a standard or time limit for when an incident must be recorded, since excessive delay can undermine data quality?
Are data in different parts of the PULSE system logically consistent with each other (for example the offender detection status and the recorded investigation outcome)?
Is the way an incident is recorded objective and consistent over time and across different regions?
An important principle of quality management is that any errors are detected and corrected early in the process. By implementing a data quality management framework for PULSE and other AGS recording systems, the capacity to improve the quality of Recorded Crime statistics will be much enhanced.
Specific concerns in relation to the quality of data currently held on the PULSE system have been identified, including by the CSO. This plan is forward-looking and is focused on addressing deficiencies in the data recording process. The critical objective is to improve data quality and ensure fit-for-purpose crime data for future statistical outputs.
Action Area 3: A comprehensive (publicly available) Crime Recording Rules document must be produced which explains the current rules and procedures for the recording of crime incidents and investigation outcomes. In addition, a framework for maintaining the Crime Recording Rules document must be agreed with key stakeholders.
The requirement for a comprehensive manual informing users and the public about crime recording, classification, detections and crime incident outcomes, and counting principles has been outlined by both the Garda Inspectorate and the Expert Group on Crime Statistics, and an undertaking has been made to produce an up-to-date and fit-for-purpose document. It is important that this document meets these requirements in a manner that does not compromise operational activity.
A comprehensive publicly-available document explaining how and why crime is recorded by An Garda Síochána is a key, not only for accountability in how crime is recorded, but also in allowing users to understand and utilise Recorded Crime statistics.
Consideration should be given to the establishment of a formal group of key stakeholders charged with ensuring that the Crime Recording Rules document is maintained up to date and in line with policing policy and legislative developments. An annual process of review and sign-off for this important document is necessary.
Appendix I: Illustrative sample actions
The following sample actions are included for illustrative purposes to indicate the measures which are necessary to address the Action Areas outlined in the plan. This is not intended as an exhaustive or finalised list of Actions.
Action Area 1
1.1 A senior manager with adequate authority and resource to effect positive change in data governance and data quality to be appointed. The appointee will assume responsibility for the implementation of this Plan.
1.2 Clearly defined structures of accountability and responsibility in respect of recording, validation, review and management of data to be drawn up.
1.3 A data recording risk register to be created and signed off on an at-least annual basis.
Action Area 2
2.1 A review of incident types and valid relevant investigation outcomes to be carried out and the system rationalised where appropriate.
2.2 The timeliness of incident recording on AGS information systems to be monitored and reported.
2.3 The completeness of crime incident records (i.e. that crime events reported to AGS are represented as crime incident records on PULSE), the use of non-crime incident types (e.g. Attention and Complaints) to record crime events, and the accuracy of PULSE crime incident type classifications to be independently audited on an annual basis.
2.4 The accuracy in key data fields to be assessed on a continuous basis and errors to be detected and corrected in a timely manner, and data quality to be measured and regularly reported on. For example,
2.5 The rationale for PULSE incident type re-classifications and incident invalidations to be recorded in an accountable and consistent manner.
2.6 Crime recording rules such as the Primary Incident counting rule, the One Incident per Victim countingrule, and the Series of incidents, same victim, same offender counting rule to be implemented automatically (e.g. the Primary Incident counting rule determines that, where two or more incidents occur at the same time, only the more serious incident should be counted for statistical purposes).
2.7 Documented procedures on the outcomes that can be used to mark an incident detected for each crime type and the charges that relevant to each incident type need to be in place (e.g. an incident should not be marked as Detected until a relevant and verifiable investigation outcome is recorded).
2.8 Measures to ensure the logical coherence of data on the PULSE system should be devised and regularly reported on (e.g. the incident detection status, offender role, offender detection status and investigation outcome must be logically consistent with the Crime Recording Rules).
2.9 The needs of internal and external users for accurate data on, for example, domestic abuse, discriminatory motives, and the use of weapons in crime incidents to be examined and incident recording procedures and standards updated if required.
2.10 Implementation of rules regarding mandatory recording of victims (e.g. for crimes against the person) or offenders (e.g. drug possession) depending on the incident type.
2.11 Implementation of rules regarding mandatory recording of victim-offender relationship where appropriate (e.g. in domestic abuse or sexual offence incidents).
2.12 Victim and offender profiles to be monitored with the aim of reducing person profile duplication.
2.13 Date of birth and gender to be recorded accurately for victims and offenders (e.g. to allow analysis of the age and gender profiles affected by different crime types).
2.14 The system of transfer of court outcome data from the Courts Service to PULSE to be audited for completeness.
Action Area 3
3.1 A Crime Recording Rules document to be produced as soon as possible.
3.2 A framework for maintaining the Crime Recording Rules document to be agreed.
