OxRec (Oxford Risk of Recidivism Tool)

The Salvation Army Youth Protection & Rehabilitation Foundation

The OxRec is a scientifically based questionnaire and tool to estimate the likelihood of recidivism. The probation service uses the tool as part of the broad investigation it conducts into a suspect's or convict's personal circumstances for judicial authorities. It helps to arrive at an informed recidivism risk assessment in a structured way.

Last change on 15th of December 2025, at 11:13 (CET) | Publication Standard 1.0

• Publication category: Impactful algorithms
• Impact assessment: Field not filled in.
• Status: In use

General information

Theme

Law

Begin date

2018-04

Contact information

[e-mail]

Responsible use

Goal and impact

One of the tasks of the probation service, is to issue an advice. In this advice, the probation worker describes the factors associated with the delinquent behaviour and estimates, among other things, the risk of recidivism. The advice states whether, according to the probation worker, interventions are needed to influence the recidivism risk, to control it, and to contribute to the client's resocialisation and/or protection of society/victim. For example, a (partly) suspended sentence may be advised with the special condition of treatment by a healthcare provider, following a behavioural intervention and/or an alcohol ban. The probation service must monitor compliance with these conditions. The proposed interventions match the recidivism risk identified by the probation service and the interests of the victim, the suspect or offender and society.

To assess the risk of recidivism, the probation service examines what risk factors and protective factors are present. Scientific literature shows that these risk and protective factors either reinforce or reduce risky (criminal) behaviour. To this end, the probation service conducts extensive research into the client and his environment. It does this by, for example, talking to the client, consulting references and retrieving information from judicial data. The probation officer works within a digital working environment with several components and questionnaires that structure the probation officer's judgement (a structured professional judgement). This environment guides the probation worker in the process leading to a risk assessment and supports in substantiating advice to partners in the criminal justice chain. Within this environment, the probation worker also completes the OxRec's questions as a component. The aim of this is to arrive in a structured way at a substantiated judgement in, for example, the advice. The probation officer writes this down in a report explaining the outcomes of the advice and any instruments used, including those of the OxRec. This advisory report is given to the client, prosecutor and judge.

The probation organisations in the Netherlands all use the OxRec. These are the institutions of Stichting Verslavingsreclassering GGZ, Reclassering Nederland and Stichting Leger des Heils Jeugdbescherming & Reclassering.

Considerations

Research shows that an assessment based on the results of a scientific instrument together with the professional's assessment is better than an assessment based solely on the professional's experience. The calculation model behind the recidivism rates is based on scientific research into influencing factors on recidivism. The outcome of the OxRec does not guide the final risk assessment and advice given by the probation service, it is ultimately the probation worker who reaches a judgement in the individual case. The model supports professional judgement and does not replace it. The outcome of the structured professional judgement, which considers a multitude of factors, including the OxRec's assessment, leads to the final advice.

Human intervention

There is always meaningful human intervention when using the OxRec. The model was developed to be used alongside professional judgement, as a support tool for structured decision-making, not as a substitute for professional judgement. After a professional has conducted the broad investigation, the OxRec is completed. The estimation on recidivism that emerges from this compares the professional with their own judgement. The professional always makes the final choice for the individual risk assessment and the most appropriate advice for the client. This includes consultation with colleagues (such as the four-eyes principle, but also consider case histories). The choice is substantiated whereby the various considerations are discussed and described (including the risk assessment model used as well as any other relevant information that has emerged in the investigation).

Risk management

The OxRec is only used by probation workers. They are trained to work with a structured professional judgement (with the OxRec as part of it) and therefore weigh up all information before arriving at a final judgement themselves, so not just the risk assessment tool used such as the OxRec. The probation service has commissioned research into whether this way of working is carried out as intended. This shows that it is.

In training for probation workers, it is discussed that a risk assessment is an estimate in which one compares characteristics of an individual against a group average of a group of persons with similar characteristics. Estimating the recidivism risk as low does not mean that someone will definitely not recidivate and estimating the risk as high does not mean that recidivism in an individual can be determined with certainty. This is why the probation service arrives at a weighted and substantiated advice in a structured manner. This includes explanations including how the final judgement deviates or matches outcomes of the instrument used (such as the OxRec).

The OxRec was developed in Northern Europe (Sweden) on a very large and pure dataset, after which it was externally validated in several countries. So too in the Netherlands on Dutch data, before it was put into use by the probation service. This validation is currently being repeated.

The OxRec results are shown in the appendix of the counselling report that goes to the client (the suspect or convict), the prosecutor and judge. In this way, it is clear and transparent to everyone which instruments the probation service has worked with and on which the probation service's advice is based. To prevent other people from misinterpreting the results of the OxRec, the report contains the description of how the advice was created, a clear rationale for the advice and also how the outcome of instruments such as the OxRec was taken into account in it.

Legal basis

The basis for the probation service's use of risk assessment instruments lies in their legal duties. The legal tasks of advice and supervision are relevant here and are laid down in various articles in the Penal Code and the Code of Criminal Procedure, among others. A list of the relevant legal articles for each task is given below. A link to the main articles is included under reference legal basis.

Opinion:

The basis for processing the personal data of clients (suspects and convicted persons) for the purpose of the counselling task lies in a legal obligation (Article 6(1)(c) AVG) and/or the performance of a public law task (Article 6(1)(e) AVG) in conjunction with, among others, the following articles

Reclasseringsregeling 1995: articles 8 paragraph 1 sub a, 9 and 10;

Code of Criminal Procedure: articles 63 paragraph 6, 147, 177 paragraph 2, 310, 6:2:12 paragraph 2, 6:6:23a paragraph 3, 6:6:23c paragraph 3;

Penal Code: articles 38m paragraph 4, 38z paragraph 2;

Transfer of Enforcement of Criminal Judgements Act: section 19(1);

Penitentiary Institutions Act: article 18a

Penitentiary Measure, article 7 paragraph 2

Supervision:

The basis for processing the personal data of clients (suspects and convicts) for the purpose of supervision and guidance by the probation service is a legal obligation (Article 6(1)(c) AVG) and/or the performance of a public law task (Article 6(1)(e) AVG) in conjunction with

Reclasseringsregeling 1995: articles 8 paragraph 1 sub b, 11 and 12;

Penal Code: articles 14c paragraph 3 under b jo paragraph 6, 38 paragraph 2, 77z paragraph 3 under b, 77aa paragraph 3;

Code of Criminal Procedure: articles 80, 167 paragraph 2, 6:2:11 paragraph 2 sub b, 6:3:12 paragraph 3, 6:3:14, 6:6:10, 6:6:19 and 6:6:23b Code of Criminal Procedure;

Pardons Act: article 15 paragraph 1;

Penitentiary Institutions Act: articles 4 and 43 paragraph 4;

Beginswet verpleging ter beschikking gestelden: article 51 paragraph 2;

Regulations for treatment of persons placed at risk: Articles 59 to 61, 68 to 70;

Decree on forensic care: article 2.7;

Decree on enforcement of criminal decisions: articles 3:1 to 3:5.

Below is a reference to some of the most important articles of law:

Links to legal bases

• Probation Regulation 1995: https://wetten.overheid.nl/BWBR0007120/2019-06-26#Hoofdstuk3
• Art. 147 WvSv: https://wetten.overheid.nl/BWBR0001903/2025-07-01#BoekTweede
• Article 6:2:12 WvSv: https://wetten.overheid.nl/jci1.3:c:BWBR0001903&boek=6&hoofdstuk=2&titeldeel=Tweede&artikel=6:2:12&z=2025-07-01&g=2025-07-01
• Article 14c WvSr: https://wetten.overheid.nl/jci1.3:c:BWBR0001854&boek=Eerste&titeldeel=II&artikel=14c&z=2025-07-01&g=2025-07-01

Elaboration on impact assessments

DPIA (Data Protection Impact Assessment ) is carried out

Operations

Data

The OxRec consists of 13 multiple-choice questions answered in the probation officer's digital work environment by the probation officer. For this, they can click a bullet at the correct answer. These are questions related to health, gender, age, imprisonment, previous offences, education and social environment. All the questions and explanation of the questionnaire can be found on the developer's site via the link to OxRisk below. As the developer points out, these are for information, not for professional use.

All questions are variables whose likelihood of association with recidivism is known from empirical (scientific) research.

Links to data sources

• Validation of the OxRec in the Netherlands is publicly available and freely downloadable.: https://www.nature.com/articles/s41598-018-37539-x
• Information with examples of tools by Oxford University developers.: https://oxrisk.com/

Technical design

In the OxRec, questions are completed by clicking a bullet. Each question has its own weight. This means that not every question carries the same weight in determining the instrument's final risk estimate. The choice of these weights is scientifically based.

The OxRec adds up all the outcomes of the questions. This gives as outcome a probability of recidivism. This means that the probability of recidivism is indicated from 0 to 100% for both general recidivism and violent recidivism within 1 and 2 years. These percentages are visible.

In addition, it is indicated whether the percentage of recidivism within two years falls into a risk category. These categories are low risk, medium risk and high risk. This category is also visible. The cut-off values for risk for these categories are determined as follows:

- Probability of recidivism overall: 0%-29% = Low; 30%-49% = Medium; 50% or more is High

- Probability of violent recidivism: 0%-9% = Low; 10%-29% = medium; 30% or more is high

The choice of sums and weights is scientifically based. Extensive analyses have determined what these should be. This also applies to the categories and their cut-off values.

The OxRec is built into the probation service's digital work environment and not available or removable on paper. Explanations and explanations are readily available for each question via so-called i-texts. The model was developed to support professional judgement, not to replace it. The use of the OxRec supports structured professional judgement.

External provider

The tool was developed by Oxford University.