Using data to design and monitor sentencing guidelines: The case of Ireland

There is a growing trend across common law jurisdictions towards more structured sentencing processes. Many have implemented some form of sentencing guidelines in recent years; others are designing or considering introducing them. This article focuses on Ireland, a jurisdiction that has recently legislated for sentencing guidelines but does not have sufficient sentencing data available to inform key decisions on guideline design and implementation. It highlights the importance of sentencing data for guideline development and outlines how the lack of data on current sentencing trends and outcomes means that there is no frame of reference for current practice against which to design guidelines, nor assess their likely impact or gauge their effectiveness once introduced. The article examines the types of data that are needed to develop and monitor sentencing guidelines and sets out how they might be collected and analysed in Ireland. The Irish example serves as a case study for other jurisdictions with respect to the importance of sentencing data and the need to ensure that such data are available prior to the introduction of sentencing guidelines.


Introduction
As in many common law jurisdictions, the Irish sentencing process remains largely unstructured. Judges in Ireland have a high degree of discretion when it comes to sentencing decisions, which are made within broad statutory, constitutional and common law parameters.
This discretion has its benefits and limitations. On the one hand, it enables judges to individualise sentences and respond to the unique circumstances of each case. On the other hand, it means that there is little to prevent disproportionate, inconsistent or arbitrary sentencing, 1 which can have serious implications for victims, offenders and public confidence in sentencing. Domestic law and policy must seek to structure judicial discretion in a manner that balances individualised sentencing with consistency, in that it must allow judges to respond to the unique characteristics of each case, while minimising unwarranted disparities between cases. 2 Recent public commentary in Ireland tends to argue that the process is too discretionary and warrants a greater emphasis on consistency. Academics 3 and the media 4 lament perceived inconsistencies in sentencing, while several high-profile sentences have been the subject of public criticism. 5 As a result, both the sentencing process and the need to address the alleged inconsistencies therein have become popular political issues.
In recent years, Irish opposition parties Fianna Fáil and Sinn Féin proposed Bills that would introduce sentencing guidelines. Fine Gael (the Governing party at that time) were originally against the idea, citing constitutional arguments relating to the separation of powers and judicial independence, 6 as well as a 2014 penal policy review report 7 that recommended against the introduction of statutory sentencing guidelines, as the basis for their opposition. 8 The Government acknowledged the need to address inconsistencies in sentencing but thought that this could be done by providing sentencing information to judges, without the need for sentencing guidelines. In 2017, the Government introduced a Judicial Council Bill which included the creation of a subcommittee of the Judicial Council called the Sentencing Information Committee, their role being to collect, analyse and disseminate sentencing information. 9 However, following discussions with opposition parties to secure support for the passing of another Bill (the Judicial Appointments Bill), the Government agreed to amend the Judicial Council Bill to include sentencing guidelines. The proposed Sentencing Information Committee was renamed the Sentencing Guideline and Information Committee and given the additional role of developing sentencing guidelines. The amended Bill received widespread support across the political spectrum and was passed and signed into law on 23 July 2019.
Parliamentarians from across the political spectrum expressed the view that sentencing guidelines would improve Ireland's sentencing process by enhancing consistency and transparency. 10 Absent from this discourse, however, was a meaningful analysis of the extent to which sentencing in Ireland is inconsistent, the causes of any inconsistencies, how sentencing guidelines can be designed to address these inconsistencies 11 and how the significant obstacles to introducing sentencing guidelines will be overcome (most notably, the lack of sentencing data currently available in Ireland).
This article considers these issues and discusses the implications of Ireland's experience for guideline design in other common law jurisdictions. It begins by highlighting the current dearth of sentencing data in Ireland and showing how this has affected the development of sentencing policy to date. It then turns its focus to the future. With the introduction of sentencing guidelines now in statute, the article discusses the vital role of sentencing data in developing and monitoring guidelines. It explores the types of data collected in other jurisdictions, before setting out how the Irish Courts Service and the Sentencing Guidelines and Information Committee might seek to amass and analyse data in Ireland. Finally, we conclude that there is considerable work to be done in collecting the necessary data and that a sentencing census would be the most comprehensive approach to realising this goal.

Limits to the available sentencing data
Researchers often lament the lack of sentencing data that most jurisdictions collect and publish, noting that this represents the main barrier to empirical sentencing research. 12 13 Yet, policymakers, too, suffer from a dearth of data with which to inform an evidence-based approach to sentencing policy. Consider the question of consistency: if we assume that policymakers are responsible for reducing unwarranted disparities in sentencing, then they must begin by investigating the extent, nature and causes of these inconsistencies, requiring a substantial amount of data on current practices and trends to do so with any accuracy or precision.
This also causes problems for other stakeholders. In legal and judicial training, for example, a lack of sentencing information makes it difficult to educate prospective judges and other court actors about how and why certain sentences are used in practice, the impact of different sentences and how sentencing has changed over time. For journalists, penal reformers and others who seek to hold judges to account, the lack of aggregate information means that they have no frame of reference against which to compare outcomes in individual cases. Likewise, it is difficult to identify and challenge myths and exaggerations within the media and political discourse without knowing the full picture, while the general public are left to develop perspectives about the reality and effectiveness of sentencing processes based on incomplete information.
A small number of jurisdictions produce, or have produced, some useful information. In the United States, for example, many sentencing commissions publish rates of judicial compliance with the specified ranges within sentencing guidelines. 14 However, very few make the raw data available to researchers for more advanced analysis-the United States Sentencing Commission, the Minnesota Sentencing Guidelines Commission and the Pennsylvania Commission on Sentencing are some examples that do. 15 In England and Wales, the Ministry of Justice collects various data sets and makes these available to researchers. Most notably, academics have used their quarterly statistical releases and the (now defunct) Crown Court Sentencing Survey (CCSS) to examine judicial decision-making under sentencing guidelines and assess empirically other questions around sentencing practices and outcomes. 16 Importantly, these data also informed the development of the sentencing guidelines in England and Wales. 17 In Ireland, in contrast, there are just three limited sources of sentencing data, of which only two are publicly available. Firstly, the Court Service Statistics provides basic administrative statistics on sentencing. They set out the number of times each type of sentence (imprisonment, community service etc.) was imposed for different offences. However, they provide no data on sentence length or the factors underpinning sentencing decisions. 18 Secondly, the Irish Sentencing Information System (ISIS) was set up in 2007 by the then Chief Justice, the Hon. Ms Justice Susan Denham. Its purpose was to provide sentencing information to judges, legal practitioners and the public. Researchers undertook an initial data collection phase between 2007 and 2009 and a second in 2013. For both, researchers collected data by attending the two criminal courts with the highest caseloads (namely, the District and Circuit Courts) in selected areas, using a template to record information on the sentences imposed. In some cases, the template captured the comments made by the judge prior to imposing the sentence. In other instances, it was more difficult for the researchers to determine what to record. For example, while the researchers recorded the mitigating and aggravating factors when stated by judges in sentencing hearings, some judges did not articulate these in detail. 19 The data collected were then anonymised, inputted into a database and published on the ISIS website, irishsentencing.ie (currently inactive). This information could be searched by offence type and using keywords.
While ISIS provided valuable insights into sentencing in Ireland, it had significant limitations. Aside from the methodological issues noted above, the majority of the data are now over 10 years old and reflect a small, unrepresentative sample of sentences imposed in Ireland: less than 1000 sentences were collected from three locations (the urban areas of Dublin, Cork and Limerick). In essence, ISIS data represented a snapshot of how a limited number of judges sentenced particular offences over a short period. As such, neither the Court Service Statistics nor ISIS allows for any meaningful analysis of sentencing trends in Ireland, nor of how specific offences are currently sentenced. They also provide no insight into the issue of inconsistency in sentencing.
The third source of sentencing data and information in Ireland, accessible only by members of the judiciary, is the Judicial Researchers' Office. 20 Under the guidance of the Hon. Mr Justice Peter Charleton, the office produces sentencing reports and publishes these on a service-wide intranet. These focus on selected serious offences (e.g. manslaughter, rape, sexual assault, robbery, possession or importation of drugs for sale or supply and child pornography). 21 For a time, anonymised versions of some of these reports were also publicly available on the ISIS website. While the reports themselves do not set out the exact methodology used, an Irish Times article on the Judicial Researchers' Office described this as follows: The Judicial Researchers' Office have developed their own, improvised system [ . . . ]. First they scour relevant judgments on specific offences from the Court of Criminal Appeal, placing each one in a batch corresponding to sentence bands. Who is getting suspended sentences or up to three years? Who is getting 5-10 years? And so on. Next they turn to the online archive of The Irish Times and carry out a trawl for additional cases. Taking the names and dates from these, they listen back to the original sentence hearing using the digital audio recording system that has been in the This methodology allows for some useful analysis of sentencing. For example, its inclusion of offences sentenced in the Central Criminal Court (which sits above the Circuit Court in the Irish hierarchy and tries only the most serious offences) would permit analysis of sentence severity among the most serious offences, including, perhaps, the use of life sentences. However, this process, too, has its limitations, particularly in relation to its sample size, its reliance on media outlets for data and its lack of information about the lower courts where most offences are processed and where judges generally do not provide detailed sentencing rationales.
These judge-led initiatives represent an important first step towards addressing the dearth of sentencing knowledge in Ireland. Yet, as in many common law jurisdictions, chasms remain between what happens across the system and what data are ultimately collected and published, particularly regarding volume offences in the most active criminal courts. Consequently, we lack the ability to analyse sentencing practices and trends in a manner that could inform evidence-based developments in penal policy.

The importance of sentencing data for policy development
In Ireland, the lack of sentencing data has demonstrably affected the development of judicial and legislative sentencing guidance and policies. While the appellate courts have developed a range of sentencing principles overtime to guide judges when sentencing, 23 they have been reluctant to issue guideline judgments which include indicative sentencing bands. In DPP v Tiernan, 24 for example, the Supreme Court explicitly refused to give a guideline judgment for sentencing rape, noting that the absence of reliable sentencing information on rape cases was a key barrier to doing so.
The first case to provide this form of guideline judgment was DPP v W.D. 25 Here, Charleton J in the Central Criminal Court collaborated with Aoife Marie Farrelly from the Judicial Researchers' Office to produce data on which to base such a judgment. They collected and analysed 'dozens of rape sentences' to 'show the circumstances that might guide a mild response, an ordinary response, an exceptional response and, finally, a sentence tending 22. Ruadhan Mac Cormaic, 'Using Complex Data to Answer a Simple Query: How Do Judges Decide on Sentences?' Irish Times (4 June 2013) <https://www.irishtimes.com/news/crime-and-law/using-complex-data-to-answer-asimple-query-how-do-judges-decide-on-sentences-1.1416005> accessed 21 July 2020. 23. For example, an overarching principle in sentencing is the principle of proportionality. In  towards life imprisonment'. 26 The approach taken in W.D. 27 was cautiously approved by the Court of Criminal Appeal (CCA) 28 in DPP v Adam Keane. 29 As more sentencing information has become available to judges, in particular the internal sentencing reports from the Judicial Researchers' Office, we have seen other examples of guideline judgments being delivered by the appellate courts. In DPP v Ryan 30 and DPP v Fitzgibbons, 31 the CCA produced guideline judgments with indicative sentencing bands for assault causing serious harm and firearms offences, respectively. In late 2014, the CCA was effectively replaced with the Court of Appeal and while it was initially slow to issue guideline judgments, 32 it eventually did so for residential burglary in the case of DPP v Casey. 33 It also revised the sentencing bands for assault causing serious harm in DPP v O'Sullivan. 34 O' Malley notes that this initial delay may well have been due to the enormous backlog of cases that the Court had to deal with when first established. 35 This point was also made by the Hon. Mr Justice John Edwards, Judge of the Court of Appeal, writing in the Irish Judicial Studies Journal. He stated that the backlog of cases, combined with the substantial amount of additional internal research required to produce a guideline judgment, delayed the continuation of the experiment commenced by its predecessor in 2014. 36 Finally, the Supreme Court issued its first guideline judgment of this nature in DPP v Mahon, 37 the focus of which was the offence of manslaughter. This was followed by DPP v FE, 38 where the Supreme Court updated the indicative sentencing bands for rape.
These cases clearly show a willingness among the appellate courts to offer guidance and to issue guideline judgments. In Tiernan, the Court refused to issue a guideline judgment in part due to the lack of sentencing information. More recently, however, the Courts have attempted to overcome this obstacle by using the resources available to them (Judicial Assistants and the Judicial Researchers' Office) to produce their own research on sentencing and develop guideline judgments. In FE, Charleton J noted that the Courts have conducted a substantial amount of research since Tiernan and that reliance on the Courts' own research and on judgments on precedent are both useful and an aid to practice in the administration of justice. 39 However, the appellate courts' willingness to provide guidance has and continues to be been tempered by the need to rely almost exclusively on their own limited internal resources and on 26 the lack comprehensive sentencing data for all criminal offences. This is evident from the comments of Justice Edwards noted above and from the judgment in Ryan, in which the Court stated that it would only consider providing guidance where 'it has available to it detailed information of sufficient quality on the type of sentence typically imposed by sentencing judges'. 40 As for legislative sentencing guidance, Ireland has taken a minimalist approach so far. For most offences, the only statutory guidance is a prescribed maximum sentence. A very limited number of offences have a presumptive minimum sentence, 41 and there is a mandatory life sentence for adults convicted of murder. There are statutory provisions providing some guidance on sentencing children 42 and people who have committed an offence while on bail. 43 There is limited guidance regarding the effect of a guilty plea 44 and a requirement that the relationship between a defendant and victim be an aggravating factor in sentencing for certain offences. 45 There is also legislation which encourages (but does not require) judges to use community service orders as an alternative to short-term prison sentences. 46 It is within this context that legislators have decided to take action to structure sentencing, in the form of sentencing guidelines. Yet, the associated parliamentary debates did not involve much discussion about the lack of sentencing data and the key role of these data in developing and monitoring guidelines.
Indeed, a recognition of this challenge has long been conspicuously absent from the political discourse. A prior opposition effort to introduce a Judicial Sentencing Commission in 2013 would have required the Commission to create numerous guidelines within the first year of its establishment. 47 By way of comparison, the Scottish Sentencing Council took over 3 years to publish its first guideline: a five-page document attending only to the general principles and purposes of sentencing. 48 Almost 2 years later, 49 their second guideline, on the sentencing process, is yet to be finalised, while the jurisdiction's first offence-specific guideline (on death by driving) is still (as of July 2020) undergoing an initial drafting process, to be followed by lengthy judicial and public consultations before being finalised. 50 This illustrates the size of the gap between the political expectations of what such a body might achieve in Ireland and what is realistic and achievable in practice. It also indicates a lack of understanding about the process of creating guidelines, of the vital role of sentencing data therein and of the many challenges that must be overcome in guideline development.

The introduction of sentencing guidelines in Ireland
Before exploring which data sets and methods can support guideline design and monitoring in Ireland, we can briefly outline the relevant legislation. As noted earlier, the Judicial Council Act 2019 provides for the creation of a Judicial Council. This Council, once established, will be composed of all current members of the judiciary, with its primary functions carried out by a Board consisting of ex officio and internally elected members of the Council. The Act requires that the Council establish several committees with responsibility for a range of matters, including a new Sentencing Guidelines and Information Committee. This will have 13 members: 8 judges appointed by the Chief Justice and 5 'lay members' appointed by the Government. The Committee's functions include preparing and submitting to the Board, for its review, draft sentencing guidelines and draft amendments to adopted guidelines. The Act explains that, as in other jurisdictions, these guidelines may relate to sentencing generally (generic guidelines) or to particular offences or categories of offence or offender (offence-or offender-specific guidelines). 51 The Committee will also be required to 'monitor the operation of sentencing guidelines', 52 although the statute does not explain precisely what the goals of this will be, nor the activities that the Committee must undertake to perform this duty. Importantly, the statute envisages that the Committee should play a role in data collection, two of its functions being to 'collate in such manner as it considers appropriate, information on sentences imposed by the Courts, and disseminate that information from time to time to judges and persons other than judges'.
It is important that the statute explicitly recognises the need for data collection. Without information on the current practices, the Sentencing Guidelines and Information Committee will not be able to perform its other key statutory functions of developing, monitoring and revising sentencing guidelines.

Developing and monitoring sentencing guidelines
The first step to creating sentencing guidelines is to understand current sentencing practices and trends. This is true for both offence-specific and generic guidelines. When developing an offence-specific guideline, it is essential to understand how and why judges currently sentence that offence, to have a frame of reference against which to propose sentence ranges and weightings to apply to different factors. Similarly, when creating a generic guideline, drafters need to have some knowledge of current practices in relation to the sentencing principle covered by the guideline. To introduce new sentencing guidelines without this knowledge and understanding would jeopardise their quality and integrity by opening them up to accusations of arbitrariness or politicisation.
In Ireland, the primary reason given for introducing new guidelines is to improve consistency in sentencing-not to alter it fundamentally. 53 If there is no information available on current practices, the Sentencing Guidelines and Information Committee will not be able to assess the likely impact-in terms of consistency, severity and other factors-which a draft guideline will have on sentencing outcomes. It would essentially mean that a group of 13 51 people (i.e. the 8 judges and 5 'lay members' provided for in the Judicial Council Act) would be tasked with establishing sentence ranges and setting out sentencing principles without any ability to explore the current ranges, trends and practices. Of course, there may be circumstances in which a sentencing guideline will seek to do more than enhance consistency. In England and Wales, for example, new drug offence guidelines were designed to encourage more proportionate sentencing of 'drug mules', relative to other drug traffickers who had greater culpability. 54 Those who design guidelines might also wish to alter sentencing practice if public or judicial consultation, or relevant research, indicates this to be necessary. For example, the Scottish Sentencing Council's draft guideline on sentencing young people encourages judges to consider young adults' maturity when sentencing, based on a literature review the Council commissioned which indicates that such considerations are warranted. 55 In these situations, sentencing data are equally crucial to provide a baseline from which to encourage deviation and, as importantly, to enable us to measure the extent to which practices do change over time, if at all. This brings us to the need for sentencing data to help monitor sentencing guidelines. Recent research indicates that sentencing guidelines may influence sentencing in a variety of ways, including individualisation 56 and severity. 57 Once a guideline is adopted, therefore, sentencing practices must be monitored to discover the extent of its impact on sentencing processes and outcomes. This would enable any issues arising to be pinpointed and addressed at the earliest opportunity. It would also enable successful approaches to be identified and replicated.
In England and Wales, for example, an impact assessment of the assault guideline found that, for two assault offences (s 18 causing grievous bodily harm/wounding with intent and s 47 assault occasioning actual bodily harm), there were unexpected increases in severity following its introduction. For s 18, the guideline increased sentences in excess of estimations. For s 47, sentences became more severe, despite predictions that the opposite would occur. 58 Another recent empirical study in England and Wales found that the introduction of guidelines might have increased sentence severity for some offences, but not others. 59 If this 'sentence inflation' were to occur in Ireland as a result of introducing guidelines, it would be a damaging, regressive step, with vast implications for probation and prisons services, for offenders subject to more severe sentences and for their families and communities more broadly-especially at a time when prisons are overcrowded and the prison population is trending upwards. 60 Such an outcome would also run counter to the penal policy aims and objectives that Irish policymakers have recently articulated. 61 The central role of sentencing data in creating and monitoring sentencing guidelines is clear. Yet, Ireland, like many countries, lacks the infrastructure to collect and analyse the type and quality of data needed to perform these tasks effectively. Overcoming this obstacle requires us to consider what types of data are needed and how they might be collected.

Types of sentencing data collected in common law jurisdictions
The best way to collect reliable, comprehensive sentencing data is to establish a sentencing census. Sentencing censuses seek to collect detailed data on all sentences imposed within a jurisdiction, on a continuous basis. They produce up-to-date, nationally representative data, permitting sophisticated analyses of sentencing processes, outcomes and trends.
Examples of this approach have existed in England and Wales and in the United States. In England and Wales, the CCSS operated between 1 October 2010 and 31 March 2015 and acted as a census of all cases sentenced in the higher court. The CCSS was a paper-based survey, completed by the sentencing judge when passing sentence in the Crown Court. It asked both for the factors the judge took into account when determining the appropriate sentence and for the final sentence given. 62 The survey had a response rate of around 60% for each year it operated. 63 While this may seem low for a 'census', it is quite high for a survey and provided unprecedented amounts of valuable data on sentencing in the Crown Court. These data have been central to the Sentencing Council's efforts to develop new guidelines and monitor the impact of adopted guidelines. Academics have also used this data set to gain empirical insights into the impact of sentencing guidelines, 64 as have other state bodies conducting research on criminal justice matters. 65 The CCSS was discontinued in 2015 and replaced with a strategy of bespoke data collection. The Sentencing Council portrayed this decision as an evolution in their analytical approach to develop more targeted, 'guideline-specific' data collection in the criminal courts. 66 However, an independent review of the Sentencing Council's methodology clarified that the decision to discontinue the survey was based primarily on resource considerations. The report went on to state that, while the reduced resources available to the Council meant that this decision was inevitable, the replacement strategy reduces the Council's ability to carry out its statutory functions. 67 In other words, the decision reflected the Council's efforts to adapt to reduced resources, rather than a strategic decision to improve data collection.
In the United States, the United States Sentencing Commission, an independent agency within the judicial branch of the federal government, runs a sentencing census at the federal level. The Commission's primary mission is to promulgate and amend the federal sentencing guidelines, but it is also responsible for the collection, analysis and dissemination of sentencing data. 68 To perform these duties, the Commission collects data from sentencing records sent directly from the federal courts. Within 30 days of the entry of judgment in a criminal case, the Chief Judge of each sentencing court must submit several documents to the Commission, including the Judgment and Commitment Order, the Statement of Reasons, any plea agreement, the charging documents, the pre-sentence report and any other information the Commission finds appropriate. 69 Some data are extracted from these documents using optimal character recognition technology, while Commission staff collate other data manually. The information obtained by analysing these data then informs the Commission in performing its mission. The Commission publishes a Sourcebook of Federal Sentencing Statistics each year, which displays the data in various tables and charts. 70 The Commission also makes the raw data available to the public for research purposes. 71 In the absence of a sentencing census, another approach used is to combine administrative data with bespoke sentencing research. This is now the approach used in England and Wales, following the discontinuation of the CCSS. It is also the approach used by Scotland's newly established Sentencing Council. Administrative data usually include some information about all sentences imposed but tend to be limited in detail. As such, they might help analyse broad trends but are less useful for engaging in more in-depth analysis of sentencing practices and the reasons why sentences are imposed or when trying to examine issues such as consistency in sentencing. 72 In England and Wales, court administrative systems (LIBRA and CREST) are used to collect data in the magistrates' courts and the Crown Court. These systems produce data on sentencing for specific offences, broken down by age group, gender and ethnicity. The Ministry of Justice uses these to develop its quarterly statistics bulletin that, in turn, the Sentencing Council uses to understand broad sentencing trends.
The Sentencing Council then supplements this with various types of bespoke research. For example, it recently requested that magistrates and District Judges from 81 courts fill out a survey 73 every time they sentenced an offender for theft from a shop or stall. They were asked to do this for 3 months before the Theft Offences Guideline was introduced and for a 3-month period thereafter. The response rates stood at 30% and 26%, respectively. 74 Similarly, when assessing the impact of the Robbery Guideline, 75 the Council requested that all Crown Court Judges fill out a form for every offender they sentenced for robbery, again for a 6-month period.
The response rate was 48%. 76 As part of the same study, the Council also conducted interviews, mainly over the phone, with 26 Crown Court Judges. In advance of the interview, judges were given a case scenario to consider and asked to complete a sentencing form to facilitate further discussion about their sentencing practices. 77 The Scottish Sentencing Council, in setting out its methodological approach to collecting data, has indicated that it will first look to use administrative data to gain insight into the sentencing of different offences. 78 It will obtain these from the Scottish Government and criminal justice organisations like the Scottish Courts and Tribunals Service, the Crown Office and Procurator Fiscal Service, the Scottish Prison Service and Police Scotland. The limitations of this type of data have, however, been duly noted in a literature review conducted for the Council as part of the process of developing the Causing Death by Dangerous Driving Guideline. 79 In addition to using administrative data, they have stated that they will review other relevant research and, if needed, conduct their own research to understand current sentencing decisions for the offence that the guideline will cover. This may include research with judges in the form of focus groups, formal discussion papers, in-depth interviews and larger group discussions. 80 There are also jurisdictions that do not have sentencing guidelines but still collect and analyse sentencing data to inform sentencing practices and the development of sentencing policies. For example, in the Australian state of New South Wales, the Judicial Information Research System contains legal and statistical information about sentencing intended to assist judicial officers in their decision-making. It is made up of a number of databases, the most relevant of which are the Sentencing Statistics Database and the Judgment Database. The Sentencing Statistics Database uses administrative data sourced from the New South Wales Bureau of Crime Statistics and Research to set out the range of penalties imposed for each type of offence. 81 The data are more detailed than the administrative data in England and Wales or in Scotland. They allow the statistical profile to be refined based on the offender's plea, age, prior record, number of counts of the principle offence and whether the sentence imposed took into account other offences. 82 The Judgment Database contains judgments of the CCA that are relevant to sentencing. These data sets provide a baseline understanding of broad sentencing trends and judicial directions relating to sentencing and could be used in combination with other methods to provide a holistic understanding of sentencing. On their own, however, they are clearly insufficient for the development and monitoring of sentencing guidelines.

Suggested approaches for Ireland
Having highlighted the lack of sentencing data available in Ireland and detailed the collection and use of different types of sentencing data elsewhere, the final section of this article seeks to 76. Note the response rates are much lower than the Crown Court Sentencing Survey. As has been noted by others, the Crown Courts Survey was ongoing and so it likely become habitual for many judges and part of their regular process. recommend what Ireland should do next. The most comprehensive approach to collecting sentencing data and producing sentencing information in Ireland would be simultaneously to operate an ongoing sentencing census in the District and Circuit Courts, to collect and analyse relevant judgments from the appellate courts and to conduct or commission bespoke studies and systematic evidence reviews on specific sentencing issues on an ad hoc basis. This would produce reliable, comprehensive, representative and up-to-date sentencing data. It would also provide detailed information on certain offences and themes and on the impact of proposed or adopted guidelines. A sentencing census must have buy-in from judges to achieve a high response rate and ensure that survey forms are filled out frankly and comprehensively. Getting this buy-in to participate in an ongoing survey may be difficult in the District Court in particular: while there is a duty on District Court Judges in Ireland to give reasons for their decisions, this duty is limited 83 and does not include a duty to give written reasons for all sentences imposed. In 2003, the Law Reform Commission recommended introducing such a requirement, 84 but the Working Group on the Jurisdiction of the Courts rejected this soon afterwards. 85 In rejecting the Law Reform Commission's recommendation, they placed significant weight on representations made to the Working Group on behalf of District Court Judges that any obligation to record reasoning in writing would be onerous, time-consuming and impossible within existing workloads. While this was over 15 years ago, it nevertheless highlights the challenges in convincing busy judges to complete a survey every time they impose a sentence.
If judges were reluctant to support a census due to time constraints, a compromise might be to run a more limited census in the District Court, alongside a full census in the Circuit Court. The limited approach might include a shorter survey or one focused only on certain offences. As for filling out the survey forms frankly and comprehensively, this may be an issue in both courts. For a census to have its full value, judges need to be willing to identify and record the actual factors they considered and their reasons for imposing a sentence and not just the factors they feel the need to state to defend or justify their sentence. 86 All efforts should be made to attain buy-in from the judiciary, including by communicating why it is necessary and facilitating judicial participation in the survey's design to maximise its clarity, brevity, usability and legitimacy. Although this may seem to be a momentous task, it has been done successfully in larger jurisdictions than Ireland and is by far the best way for the Sentencing Guidelines and Information Committee to perform its statutory functions. Efforts should also be made to explore ways in which new technologies could be used to collect the necessary data while limiting the impact on judges and their workload. 87 Data quality is of paramount importance and Ireland must seek to achieve a 'gold standard' of data collection in this field, working backwards from this goal if absolutely necessary. 83 As noted, one of the responsible bodies (or, perhaps, the responsible body) for determining what sentencing information will now be collected in Ireland-the Sentencing Guidelines and Information Committee-is a subcommittee of the Judicial Council. A majority of its members are judges, which may have some benefits and present certain challenges. The benefits relate to its legitimacy among judges: if the subcommittee were to determine that a census is necessary, then one might expect the chances of judicial participation and acceptance to be greater than if the request came from another body. On the other hand, judicial committee members might not wish to promote methods of data collection that, while comprehensive, would be (or would be seen to be) onerous for their colleagues.
Even if the Sentencing Guidelines and Information Committee were inclined to establish a census, and sentencing judges were willing to participate, there may still be an issue of resources. To operate a census in the District and Circuit Courts would involve ongoing costs. This money would need to come from central government, meaning that they, too, must be convinced of this approach. 88 There are signs that the value of good quality data is now being recognised in government. The Department of Justice and Equality, for example, recently published a new data and research strategy which acknowledges the necessity of data for evidence-based policy and the need for Ireland to enhance its own data collection capacities and practices to support this goal. 89 If a sentencing census is not possible, another approach is to combine different methods of collecting less comprehensive sentencing data. This would include the use of administrative data, the conducing of bespoke research in the District and Circuit Courts and the ongoing analysis of relevant Court of Appeal and Supreme Court judgments. This approach still requires many improvements to the administrative data that are currently collected. At a minimum, there would need to be data collected on the sentences imposed for every offence. Additional non-subjective data should also be collected (age, gender, ethnicity etc.), allowing for some limited analyses of broad sentencing trends. When making improvements to the administrative data, there should be some consideration of issues raised in Scotland by McPherson and Tata about how to represent and distinguish between single conviction and multi-conviction cases. 90 Bespoke research could be conducted on particular offences or on sentencing issues in the District and Circuit Courts. This would neither be ongoing nor for every offence (like a census). Still, it would need to happen regularly to gather enough data to develop and monitor guidelines. Again, judicial buy-in and willingness to participate would be crucial. While this approach would be less resource intensive than a census (at least, according to the Sentencing Council of England and Wales), it would still involve some costs, with the committee hiring research staff or having funding available to commission research. Likewise, it would require time and expertise for someone to analyse Court of Appeal and Supreme Court Judgments.
The Irish Government must not underestimate the level of resources required to move from the current system, to one in which sufficient data are collected to design and monitor guidelines effectively. This will require significant and ongoing investment. We have not even spoken about the public and other stakeholder consultations that the Sentencing Guidelines and Information Committee will need to conduct, along with their other statutory duties.
The Government, in its original impact analysis of the Judicial Council Bill, suggested that the 'Sentencing Information Committee' would have an annual budget of €200,000. 91 There has been no assessment published since amending the Bill to assign this committee the additional responsibility of developing sentencing guidelines. By comparison, the Sentencing Council in England and Wales has an annual budget of over €2,000,000 92 and a recent review noted that their research teams are stretched. 93 The Scottish Sentencing Council has a budget of over €600,000 and receives additional support from the Scottish Courts and Tribunals Service. 94 The legal information section of the Judicial Commission of New South Wales has annual expenditure of over €1,300,000. 95 Ireland will need to increase its original estimated budget and find new and innovative ways of collecting data and conducting the necessary research. Otherwise, the potential benefits of sentencing guidelines are unlikely to be realised, and an arbitrary or neglectful approach to their design and implementation will push us further away from the goal of evidence-based penal policy.

Conclusion
Sentencing guidelines appear to be gaining favour among policymakers in many common law jurisdictions. Recent research suggests that guidelines can influence sentencing practice in a variety of ways. 96 It is vital, therefore, that much care is taken as guidelines are introduced and implemented. We have argued that the first step to creating sentencing guidelines must be to collect and analyse sentencing data to understand current sentencing practices and to provide a frame of reference against which sentencing guidelines can be developed. Data are also needed to facilitate the monitoring of guidelines and to ensure that their impact on sentencing practices and outcomes is understood.
The passage of Judicial Council Act 2019 indicates the inevitability of sentencing guidelines in Ireland. As it currently stands, however, there is a severe dearth of sentencing data and a lack of infrastructure with which to collect and analyse the data that are required to develop and monitor sentencing guidelines. Drastic improvements are urgently required. Ideally, this would have been addressed in advance of legislation, but it is not too late to make the appropriate investments at this stage, while the Committee is being established and relevant stakeholders are giving initial consideration to the purpose and content of guidelines. In this article, we have outlined the types of data that are needed and made some suggestions as to how these might be collected. In particular, we have argued that a sentencing census should be established to provide the most comprehensive data set possible, giving policymakers the best opportunity of designing effective guidelines. This can be supplemented with bespoke research on specific themes and offences. Ultimately, the further Ireland departs from this 'gold standard' of data collection, the more difficult it will be to design guidelines that achieve their intended goals. Guidelines are not cheap to produce, and resources must be invested in maximising their quality and credibility.
While there is still be a degree of uncertainty as to whether sentencing guidelines can achieve everything that their proponents advocate, there is little doubt about the need for sentencing data to inform their content and monitoring. To introduce sentencing guidelines without first ensuring the availability of comprehensive sentencing data would be to set them up for failure from the outset.