Select Committee on Public Accounts Eleventh Report


Review of National Audit Office Report HC548 Session 2001-02

HM Prison Service: Reducing Prisoner Re-offending


  1.  This report examines the role and performance of the Prison Service, other related criminal justice bodies, and education or health services in reducing the rate of prisoner re-offending in England and Wales. It deals specifically with prisoner programmes in developing thinking skills, basic national curriculum skills, and drug rehabilitation and detoxification. The analysis covers a range of issues, such as access to programmes, the success of programmes, matching prisoners' needs to programmes, and the availability and use of data for informing programme innovation.

Administrative and management context

  2.  The administrative and managerial context is well conveyed in Part one of the report. The two large Figures two and three help to illustrate the text. Several committees and agencies have been set up over the past decade to address the concerns of the government and the public about the high levels of re-offending by ex-prisoners within two years of release, reconviction rates averaging 56 per cent. The aims and methods employed by such key units as the Offending Behaviour Programmes Unit and the Drug Strategy Unit, their responsibilities and lines of accountability, are summarised effectively. The study team provide some useful metric data on prison populations and costs in a clearly understandable form. For example, the average cost of a prison sentence imposed at the Crown Court is £30,500, and the average cost of a sentence to the Prison Service is £22,900 per year (paragraph 1.2). There is a good summary of the target mechanisms used in the prison sector and of links to other criminal justice organisations. There are some useful data on major constraints on effective rehabilitation of prisoners, for instance, the fact that 80 per cent of prisoners admit to drug use in the year before prison. However, it is unclear what kinds of drugs are included here: data for drugs more serious than cannabis would have been useful. The study team also give some statistical information: for example, an 18 per cent rise in the number of prisoners employed in workshops is noted, but prison populations have risen by 45 per cent in the last decade.

Structure and presentation

  3.  The report is clearly and accessibly structured. It begins with a "Foreword by the Comptroller and Auditor General", which distils eight main points for action. This section is a bit hard to interpret as a one-off. Does it reflect the fact that the Executive Summary is rather long at eight pages? Part one sets out the background to programme development. It focuses well on the detailed structure of the system and the main programme delivery units. Part two addresses programme accreditation, and the important problem of scope and reliability of data gathering. Part three covers the arrangements for assessing the needs of individual prisoners, and trends in access to programmes, which vary significantly by the type of prison and region of the country. Part four examines the reliability of information about resettlement arrangements and activities, and reveals an array of shortcomings in both knowledge and performance in resettling prisoners after release. Five appendices cover methods and provide useful data, some of which should have been included in the main text.

  4.  The report uses descriptive headings for the Parts, which work well, but two orders of long narrative sub-headings within Parts. On several occasions the second and third level headings occur together, without any intervening text (see pages 19, 21, 27, 29 and 37 (twice). In additional cases narrative sub-headings of two sentences length are used (see pages 22, 24, 27, 28, 39 and 41). Some of the narrative sub-heads are also long-winded. All three practices create over-long and complex "heading spaces" in the text, and should be avoided. The key rules should be: never have two headings directly adjacent to each other without intervening text: never have two sentence headings; keep even narrative headings as punchy as feasible. The report is otherwise clearly written, tackling a difficult topic in an accessible way. There are some bits of awkward wording or phrasing, (such as the over-long redescriptive sentence in paragraph 1.9 relating to improving thinking skills), but these are rare.

Graphics and statistics

  5.  Those graphs and statistics included are relevant and well designed. The report includes a useful amount of data and some more innovative graphs are used, notably the box and whisker plots showing the range of spending per prisoner per year on education in Figure 18, although the columns here should have been arranged in a numerical progression. Figure 15 also shows nicely the extent of variation in access to Enhanced Thinking Skills courses (page 30), although a horizontal box plot should have been used allowing regions' names to be read more easily on the Y-axis. There were a few other small defects (see Minor points below).

  6.  Other data in the report are not well presented or sufficiently brought into the main text. Appendix two showing the results of the study team's survey of prisons is not well presented. For example, on page 45 the basic data on programmes is not arranged in a numerical progression and percentages are needlessly computed correct to one decimal point—the overall result being a jumbled table with 80 cells. To make matters worse each of the columns is heading "number of prisoners" when it should be "prisons". No real analysis is given of this table, and it does not emerge clearly enough in the report that there are three classes of programmes. Two are provided almost universally (Basic Skills Education and CARATS). Three more are extensively available in a majority of prisons (Other Programmes that are mainly locally driven, Enhanced Thinking Skills and Drugs Rehabilitation). And four are available in a fifth or fewer prisons (Sex Offender Treatment Programme, Reasoning and Rehabilitation, the CALM Programme, Therapeutic Communities, and Problem Solving). All of the data in Appendix two should have been properly processed and analysed in the main text. The data in Appendix four on hours of purposeful activity should also have been analysed and presented there. The scores for individual prisons are just listed here broken up by type of prison. The key summary statistics below for prisons as a whole are not given:

24 hours

Upper quartile
Lower quartile
Upper limit
Lower limit
Mid range

  The Prison Service target for purposeful activity was 24 hours per week, and 58 prisons failed to meet this level while a further 28 prisons achieved scores of 24 or 25 hours. Effectively nearly two thirds of prisons were failing to meet or only just meeting the target.


  7.  The report is well-focused on a defined topic with clear policy effectiveness implications and significant financial costs (accounting for between 0.5 of 1 per cent and 11 per cent of the average annual cost of a prisoner, according to Figure 18). The report highlights the lack of reliable data or research to inform Prison Service policy and working. At times, however, the study team do seem to skirt over potentially serious issues. For example, paragraph 2.25 states that unaccredited programmes play an important part in the service delivery without any real analysis of why this might be the case. One or two links out to other topics might have been pursued. "Prison" in some ways extends into the community via the early home release of prisoners who have agreed to be subject to electronic surveillance or "tagging", and so is relevant to the coverage of resettlement programmes.


  8.  The study methodology is well designed in terms of generating data. Questionnaires were sent to all prisons and virtually all replied on most topics. Interviews were undertaken with staff and prisoners in 10 prisons of varied type, as well as with officials and the inspectorate, providing reasonably full and triangulated coverage. Twenty prisoner files were examined in each of these locations, 200 in all. The methods Appendix one is reasonably informative, but might helpfully have listed the key studies in the literature review; and said more about the nature of consultations with NGOs and why parole Board personnel were not interviewed (see Minor Points). "Targets", such as the aim of doubling the number of prisoners released into work or training by 2004, are also cited without much or any exploration of the underlying rationale for the target itself. A drop in the target level for "average number of prisoner hours spent on purposeful activities" after 1996-97 is shown in Figure 20 but not explained. It would have been interesting to know whether this was policy change under the current government or whether it reflected only some definitional change in the meaning of "purposeful activity".

  9.  The study team seemed too willing to rely on data provided by either Home Office or the Prison Service. Most figures are sourced to one of these two, and not much use is made of independent data collection and analysis. Above all the team does not fully exploit the data, which it has generated. For instance, there are evidently problems in the extent to which moves of prisoners disrupt their access to programmes, and one aspect of this is the time it takes for prisoner records to be transferred to a new prison after a move. Question nine in the survey of prisoners asked how long it took each prison to send on records, and 93 per cent claimed to accomplish this within seven days (it was not clear if these are office working days of just any days). But when asked how long it took for new records to be transferred in from other prisons, only 63 per cent said it was accomplished within one to seven days, 30 per cent said eight to 30 days, and four per cent over 30 days. This pattern clearly suggests overly favourable reporting on the question of delays in sending on records. Neither question is referred to elsewhere in the report, along with most of Appendix two. Figure 19 does refer to data drawn from question 11 on page 50, but it is clearly calculated slightly differently and so the numbers do not match up consistently.

Conclusions and recommendations

  10.  There are 10 sets of recommendations overall, presented in the Executive Summary and re-distilled in the Foreword. They all seem sensible and timely in a rather general way but we were a little sceptical that the study team has necessarily got a grip on possible intractable barriers to achieving improvement. At some points the study team come across as seeing information technology change as a universal solution to deep-seated problems of poor information collection and systems. Phrases such as "the Prison Service expects that a new IT system, due to be introduced under the Quantum project, will provide improved cost information" (paragraph 2.10) are pretty vague. They do not address the likely problem of changing the culture of an organisation that has not collected so much crucial management information for such a long period of time. It should be clear by now to any NAO study team, that new IT systems often create new problems or build in new limits as well as solve old ones. The study team might have been able to take a better view of the scope and limits of changes from improving data collection and analysis had the study included comparisons with other countries. The New Zealand Treasury, for example, has developed very detailed costings of prison regimes and penal alternatives and has linked them effectively to success rates in preventing re-offending, albeit in a much smaller country with a far lower prison population.

  11.  There is also a constant and relatively well-attested risk that targets and key performance indicators come to be more important for audited agencies than the activities themselves. One possible area is the recommendation, given the imprimatur of the Comptroller and Auditor General in the Foreword, that "management information systems [be developed] that would enable an assessment to be made of [...] the success of individual prisons in reducing re-offending". Yet the report itself at several points emphasises reasons why the Prison Service has seen this task as difficult. Prisoners transfer from one prison to another for a host of reasons in the course of their imprisonment. It would be perverse to introduce an incentive for prisons to maximise their KPI score for prisoner programme completion by, for example, unwontedly hanging onto prisoners who are good risks and shedding poor risks.

  12.  There are also some rather strange conclusions drawn throughout the report that seem to contradict the analysis in places. For example, paragraph 2.12 says that "the Prison Service has made a determined effort to ensure that the design of its offending behaviour programmes conforms to the best available evidence of what works". This is strange given that the overriding message of the report is that there is scant data around anyway to form a foundation for any analysis. Presumably what works is being determined by the quasi-experimental studies reported in Appendix three. But this is a single small study and may be vulnerable to Hawthorne effects and to the quality shading of programmes, which move from experimental or national implementation. Perhaps the key report message is that the Prison Service should spend less time being "determined" to use such studies to inform choices and priorities, and more time collecting and using wide-scale data to inform programme improvement. Although the report deals with resourcing of prison programmes it does not draw inferences from the clear difficulties facing probation staff in fulfilling their prisoner-related tasks. And it would have been useful to compare performance across publicly-run prisons and privately-contracted providers, assessing relative strengths and weaknesses with both. Paragraph 2.20 says that "contracts let to suppliers since 1999-2000 have stipulated that their drug treatment programmes should gain accreditation by March 2002. The Drug Strategy Unit recognises, however, that they are unlikely to do so [...]". This nugget of information is not really developed in terms of private prison providers and their management of rehabilitation programmes.


  14.  The report provides a very useful and quite comprehensive picture of the current state of prisoner re-education programmes. More use should have been make of the data collected. A more evidence-based approach and more specific and independent analysis and recommendations could have been pursued.

Minor Points

  Some tiny aspects of presentation might have been better done. In our copy of the report, the four page sheet for pages 17-18 and pages 35-6 was unbound within the report. Part 1 involves quite a lot of repetition of points already made in the Executive Summary. In Part 2 "data" is used as singular, and an overly small font is employed for the case study panels. The number of acronyms included in the report means a glossary would be useful. Some of these acronymic inventions seem quite bizarre (not, of course, the fault of the study team), such as OASys or CARAT. Presumably OASys stands for Offender Audit System, but this is not clear from the report.

  The repeated use of the cover photograph throughout the report (on pages 2, 4-5, 6-7, 18, 26 and 36) considerably detracts from its overall appearance. The shot shows a female prison officer talking to a male prisoner with a female social worker taking notes (no description of the scene is given at any point). All the people involved are white, and a black prisoner is shown in only one small photograph (Figure 6). Over-use of one image may inadvertently lead readers to underestimate the problems of ethnic minorities in the prison system. Matching different images to the stage of the training processes described in each part of the report would have been more helpful.

  One apparent omission in the methods was the absence of any interviews with or evidence from Parole Board staff and panel members. Their key task is to evaluate prisoners for release on license in the light of reductions in their risk of re-offending during imprisonment. Part of that process is the evaluation of how well prisoners have taken up and completed opportunities to engage in prisoner programmes, and paragraph 3.22 mentions this criterion for parole.

previous page contents next page

House of Commons home page Parliament home page House of Lords home page search page enquiries index

© Parliamentary copyright 2002
Prepared 7 November 2002