Deradicalisation and Disengagement Programming in Prisons and Rehabilitation Centres

Abstract

The concepts of deradicalisation and disengagement and are often applied in parallel, with the former generally interpreted in terms of attitudinal change (i.e. reducing sympathy for such violence), and the latter relating instead to behavioural change (i.e. no longer directly contributing to violence). The number of programmes aiming to achieve these objectives has increased substantially over recent years, including in locations such as Indonesia, Nigeria, the Philippines, Saudi Arabia, Singapore, Somalia, and Yemen. While the effects of Preventing / Countering Violent Extremism (P/CVE) programmes is notoriously difficult to measure, there is a strong case that such initiatives have the potential to deliver greater 'bang per buck' than other intervention types, for the reasons discussed in this paper. Of course, optimising efforts to counter this violence should clearly be of importance to the EU both on moral grounds (in relation to addressing violence in the countries in which interventions occur), and as a matter of 'enlightened self-interest' (as there is a continued threat of this violence being imported to Europe). Of course, the risks associated to such programmes are also prominent, precisely because they provide an existential threat to violent extremist organisations. This paper covers a range of themes relating to disengagement and deradicalisation programmes, including their activities, staffing considerations, facility upgrades, segregation policies, how to determine their success (or failure), and criteria through which to assess suitable locations for such initiatives.

Full text

1
Deradicalisation and Disengagement Programming in Prisons
and Rehabilitation Centres
By: James Khalil
1
, PhD
Executive Summary
The concepts of deradicalisation and disengagement and are often applied in parallel, with the former
generally interpreted in terms of attitudinal change (i.e. reducing sympathy for such violence), and the
latter relating instead to behavioural change (i.e. no longer directly contributing to violence). The
number of programmes aiming to achieve these objectives has increased substantially over recent
years, including in locations such as Indonesia, Nigeria, the Philippines, Saudi Arabia, Singapore,
Somalia, and Yemen. While the effects of Preventing / Countering Violent Extremism (P/CVE)
programmes is notoriously difficult to measure, there is a strong case that such initiatives have the
potential to deliver greater ‘bang per buck’ than other intervention types, for the reasons discussed
in this paper. Of course, optimising efforts to counter this violence should clearly be of importance to
the EU both on moral grounds (in relation to addressing violence in the countries in which
interventions occur), and as a matter of ‘enlightened self-interest’ (as there is a continued threat of
this violence being imported to Europe). Of course, the risks associated to such programmes are also
prominent, precisely because they provide an existential threat to violent extremist organisations.
This paper covers a range of themes relating to disengagement and deradicalisation programmes,
including their activities, staffing considerations, facility upgrades, segregation policies, how to
determine their success (or failure), and criteria through which to assess suitable locations for such
initiatives. The key recommendations are as follows:
• Tailored interventions: Most of these programmes converge around a similar range of
services, including basic education, vocational training, religious education, civic education,
psychological (or psychosocial) support, family initiatives, sports and recreational activities
and reintegration support. Individuals may be drawn to violent extremism for a broad range
of reasons and have distinct personal aspirations. As such, these interventions should ideally
be tailored to individual circumstances and needs, rather than one-size-fits-all. Of course, the
programme in its entirety will also have to comply with national, regional, and international
rights standards.
1
The author of this paper was responsible for research and monitoring and evaluation (M&E) at the Serendi rehabilitation
centre for former members of al-Shabaab in Mogadishu between 2015 and 2019. He also designed the M&E system for a
prison deradicalisation programme in Baidoa, Somalia, and provided technical assistance for a similar programme in Abuja,
Nigeria, with the latter supported by the EU’s Instrument contributing to Stability and Peace (IcSP).

2
• Sequential onset of programme activities: Programme designers should also consider
adopting a ‘lowest hanging fruit’ approach in which individual components of the broader
initiative are sequentially added over time. This way, activities that do not require extensive
specialist input during their design phase and that are likely to generate immediate buy-in
from beneficiaries (e.g. sports and recreational activities), can be prioritized over those that
are more complex and have the potential to provoke suspicion or hostility simply because of
their sensitive nature (e.g. psychological support and religious education).
• Staff suitability: Many of the above activities require specialist staff, including vocational
instructors, teachers, imams, social workers, and sports coaches. Irrespective of whether
these roles are filled by existing facility staff, external implementing partners, or a
combination of both, it is often difficult to identify suitably experienced and qualified
professionals who are willing to work in such facilities. This is particularly problematic given
that is also difficult to overstate the importance of personal relations between centre staff
and programme beneficiaries.
• Staff training: Such programmes should generally include training within their budgets, both
for the specialist staff and facility guards. This may include standard modules for such facilities
(e.g. administrative procedures, human rights, detention centre security, vulnerable groups,
complaints and procedures, and so on), as well as ones relating specifically to violent
extremism (e.g. recognizing signs of radicalisation, assessing violent extremist prisoners, and
so on).
• Facility upgrades: There is a general consensus that interventions of this nature are less likely
to succeed in centres that fail to meet adequate standards in terms of physical facilities,
including accommodation, educational areas, kitchen facilities, and so on. As such, there is
also a strong argument for those tasked with designing these programmes to consider
addressing such issues through the programme budget.
• Segregation policies: To the extent that those already involved in violent extremism have the
potential to influence others in such facilities, there is certainly a strong argument for
segregation policies that limit or prevent such interactions. Such policies also enable the
authorities to concentrate scarce resources, including staff with specialist skills and training
relating to violent extremism. Yet, there are also strong arguments against such policies,
including that they may reinforce the notion that such organisations hold a political status,
and that they may inadvertently help such groups maintain their internal structure and
discipline. Such policies also have the potential to undermine the process of deradicalisation
as this can occur through interactions between violent extremists and their non-ideological
counterparts if the two groups are able to mix. The bottom line is that there is no ‘correct’
approach, and the preferred strategy will depend on a range of contextual factors.

3
• Benefits to other detainees: Irrespective of the above, programme designers must also
consider that tertiary P/CVE programmes may potentially deliver adverse consequences if
those detained on grounds other than violent extremism are excluded from the benefits. In
particular, they may generate jealousies and hostilities that may undermine the general
functioning of these facilities.
• Programme objectives: Debates among the community of experts often revolve around
whether the objectives of such programmes should relate to the concepts of deradicalisation,
disengagement or reintegration, and whether such programmes should also adopt more
elevated overarching statements of purpose relating to violent extremism in the location in
question more broadly. This is something that should be considered at the outset of each
initiative as it is likely to influence programme design decisions.
• Suitable locations for such programmes: While there is no simple formula to determine
environments that may be suitable for this form of initiative, at the very least policymakers
should consider the extent to which key government stakeholders are likely to be willing and
able to provide their support, security considerations, existing levels of community
acceptance for such programmes, and (for centres designed to encourage voluntary exits from
this violence) the current rates of disengagement. In the current context, decisions regarding
suitable locations may also have to consider local COVID-19 situations.

4
1. INTRODUCTION
This policy paper provides information about deradicalisation and disengagement programmes in
prisons and rehabilitation centres. To place its content in context it is worth first elaborating on public
health interpretations of the Preventing / Countering Violent Extremism (P/CVE) framework,
2
which
categorise initiatives in the following manner:
• Primary P/CVE: These interventions correspond to primary prevention initiatives in health
care that aim to prevent disease or disorder before it occurs. In the case of P/CVE, they aim
to prevent broad communities from sympathising with violent extremism, including through
communications campaigns, interfaith dialogues, cultural events with a message of tolerance,
and so on.
• Secondary P/CVE: These initiatives correspond to health initiatives targeted at individuals
deemed ‘at risk’ of specific diseases and disorders. In the context of P/CVE, this relates to
those ‘at risk’ of becoming involved in violent extremism (according to their age, gender,
affiliations, behaviours, etc.), with initiatives attempting to channel them in alternative
directions through mentorship, vocational training and a range of other tailored interventions.
• Tertiary P/CVE: This corresponds to health initiatives that aim to lessen the impact of existing
diseases and disorders. In the case of P/CVE, this refers to initiatives that aim to positively
influence the attitudes and/or behaviours of individuals currently or previously involved in
violent extremism through basic education, vocational training, religious education, and so
on.
The disengagement and deradicalisation programmes considered in this paper fall under the third of
these categories. The number of such programmes has increased substantially over recent years, with
prison-based initiatives in the global south occurring in countries such as Indonesia, Nigeria, the
Philippines, Saudi Arabia, Singapore, Somalia, and Yemen. While each of these programmes is unique,
many converge around a similar range of services, including basic education, vocational training,
religious education, civic education, psychological (or psychosocial) support, family initiatives, sports
and recreational activities and reintegration support. Semi-open rehabilitation centres for former
violent extremists deemed to be at ‘low risk’ of returning to violence, including those in Nigeria and
Somalia, also qualify as tertiary P/CVE. Aside from offering a similar range of basic services, these
initiatives also attempt to directly encourage disengagement through outreach initiatives and other
‘upstream’ activities.
Given that they target individuals with existing connections to violent extremism, tertiary initiatives
occur at the ‘sharp end’ of the P/CVE framework, and this is reflected by greater potential rewards
and risks. Considering these in turn, it is worth first observing that those tasked with designing P/CVE
programmes must answer a variety of key questions, including the following:
2
On this framework, see for instance Shandon Harris-Hogan, Kate Barrelle, and Andrew Zammit, “What is Countering Violent
Extremism? Exploring CVE Policy and Practice in Australia”, Behavioural Sciences of Terrorism and Political Aggression 8/1
(2016), 6-24.

5
• Who should the beneficiaries be? Programme designers must consider whether their
initiatives should target entire communities or specific population subsets, and in the latter
case they must establish suitable criteria (location, age, gender, and so on) through which to
determine the beneficiaries.
• What form should the interventions take? P/CVE initiatives take a multitude of forms (as
indicated above), and programme designers must consider which of these are most likely to
deliver the desired objectives, financial costs of different intervention types, whether they are
likely to generate ‘buy-in’ from local officials, and so on.
Such questions tend to be more easily answerable for those tasked with designing tertiary
programmes, and as such there is a strong case that they have greater ‘bang per buck’ potential
than other P/CVE intervention types. Put simply, the beneficiaries are identified as those with a
background of involvement in this violence, and this helps avoid inadvertently targeting individuals
never genuinely likely to become involved (as occurs with secondary initiatives). In addition, the most
suitable forms of intervention can be determined through simply interviewing these beneficiaries to
determine their prior motives for involvement in violence, future aspirations, and so on. Of course,
optimising efforts to counter this violence should clearly be of importance to the EU both on moral
grounds (i.e. in relation to addressing violence in the countries in which interventions occur), and as a
matter of ‘enlightened self-interest’ (i.e. as there is a continued threat of this violence being imported
to Europe).
Of course, the risks associated to such programmes are also prominent, precisely because these
interventions provide an existential threat to violent extremist organisations. On this basis,
policymakers should be prepared to allocate a substantial proportion of their budget to security
measures, at least in the case of rehabilitation centres. Given the sensitive nature of this programming,
issues with local stakeholders are also likely to be more prominent. One consequence may be that
they limit access to these beneficiaries, which raises obvious questions about transparency. It may
also be that key local stakeholders fail to ‘buy-in’ to the notion that those previously involved in this
violence can genuinely reform. It is worth adding that tertiary programmes are multidisciplinary in
nature, and this also may be problematic in certain locations. For instance, it may be problematic if
there are gaps in local capacity in certain programmatic areas, or if cooperation between the different
departments proves to be challenging for whatever reason.
2. KEY CONCEPTS
Before elaborating on the dual concepts of deradicalisation and disengagement, it is first necessary to
briefly discuss the prominent disconnect between sympathy for violent extremism and actual
involvement in this violence.
3
On the one hand, individuals who sympathize with this violence often
3
This section draws heavily from James Khalil, John Horgan and Martine Zeuthen, “The Attitudes-Behaviours Corrective
(ABC) Model of Violent Extremism”, Terrorism and Political Violence,
https://www.tandfonline.com/doi/abs/10.1080/09546553.2019.1699793

6
remain uninvolved in its production (as represented by Individual E in Figure 1). Conversely, and more
importantly for the purposes of this paper, many of those actually involved in this violence are
unsympathetic or indifferent to its ideology and objectives. Instead, they often act in pursuit of a
salary, vengeance, adventure, status, and other personal incentives, or as they are simply coerced into
involvement (as represented by Individuals A, B and C in Figure 1). As discussed in the subsequent
sections, tertiary initiatives must incorporate a range of different services in an attempt to address
these varied motives.
Figure 1: The Partial Disconnect between Attitudes and Behaviours
4
4
The figures in this paper are adapted from James Khalil, Rory Brown, Chris Chant, Peter Olowo and Nick Wood,
“Deradicalisation and Disengagement in Somalia: Evidence from a Rehabilitation Programme with Former Members of al-
Shabaab” (Whitehall Report 4-18, Royal United Services Institute, UK, 2019), 9, 11.
Individual C: Motivated to
involvement in violence largely by
the fear of repercussions by a violent
extremist entity
Behaviours: Extent of involvement in
ideologically justified violence (y-axis)
Individual D: Motivated to involvement
in violence largely by a sense of purpose
gained through acting in accordance
with ideological beliefs
Individual A: Motivated to
involvement in violence largely
by economic gain
Individual B: Motivated to
involvement in violence largely by
the pursuit of status and
adventure-seeking
Attitudes: Extent of sympathy for
ideologically justified violence (x-axis)
Attitudes: Extent of opposition to
ideologically justified violence (x-axis)
Individual E: Sympathetic
towards violence, but not
involved in its creation

7
With this critical disconnect in mind, the term deradicalisation is most often applied to refer to
positive attitudinal change in relation to violent extremism (as demonstrated schematically in Figure
2).
5
By contrast, the concept of disengagement is invariably defined in behavioural terms, either in
relation to individuals exiting violent extremist organisation, or to them ending their involvement
in this violence. It is important to consider these to be distinct (although obviously interrelated)
processes as attitudinal change often occurs without a corresponding behavioural change (as
represented by Individual X), and vice versa (Individual Z). For instance, the former transpires in cases
where individuals are prevented from exiting by the threat of retaliation by the violent extremist group
in question. The latter often occurs where individuals elect to dedicate more time to their families
rather than violence, if they ‘burn out’ due to their clandestine and dangerous lifestyle, and so on.
Figure 2: Deradicalization and Disengagement Pathways
5
For more on these concepts, see for instance Mary Beth Altier, Christian Thoroughgood and John Horgan, “Turning away
from Terrorism: Lessons from Psychology, Sociology and Criminology,” Journal of Peace Studies 51, no. 5 (2014): 648; Shane
Bryans, Handbook on the Management of Violent Extremist Prisoners and the Prevention of Radicalization to Violence in
Prisons (United Nations Office on Drugs and Crime, New York, 2016), 141; Sarah Marsden, Reintegrating Extremists:
Deradicalisation and Desistance (Lancaster: Palgrave Macmillan, 2017), 8; and Bart Schuurman and Edwin Bakker,
“Reintegrating Jihadist Extremists: Evaluating a Dutch Initiative, 2013-2014,” Behavioural Sciences of Terrorism and Political
Aggression 8, No. 1, (2015), 3.
Attitudes: Extent of opposition to
ideologically justified violence (x-axis)
Attitudes: Extent of sympathy for
ideologically justified violence (x-axis)
Behaviours: Extent of involvement in
ideologically justified violence (y-axis)
Individual Z:
Disengaged but not
deradicalized
Individual X:
Deradicalized but not
disengaged
Individual Y:
Deradicalized and
disengaged
Deradicalization
Disengagement

8
The third key concept of relevance to this paper is reintegration, which is often broken down into
social, economic, and political components.
6
The first of these is usually interpreted in terms of
former violent extremists ending their association with violent extremists and enhancing their
connections to ‘positive influences’ among their family and community members. It is important to
recall that tribal and clan authorities can help facilitate this process in many parts of the global south.
Without overlooking that unemployment and underemployment levels are also frequently high in
locations that experience this violence, economic reintegration is usually interpreted in terms of these
individuals obtaining livelihoods opportunities that meet their financial needs (and those of their
dependants, where relevant), or in certain cases pursuing additional education. These first two
components are often closely interlinked, in that supportive family or community members are
frequently the gatekeepers to work. As expressed by Stina Torjesen, the political component of
reintegration ‘involves ending efforts to achieve political goals through violent means.’
7
Although
widely applied, this third element is problematic as many do not become involved in this violence in
pursuit of political goals (as discussed in relation to Figure 1), and as those that did pursue such
objectives may simply decide to forfeit these aims for a low-key life following disengagement.
3. PROGRAMME ACTIVITIES
As discussed in the previous section, individuals become involved in violent extremism for a variety of
different reasons, and to address this tertiary P/CVE programmes tend to provide a broad range of
services, including: basic education, vocational training, religious education, civic education,
psychological (or psychosocial) support, family initiatives, sports and recreational activities and
reintegration support.
8
Before discussing these in turn, it is worth first highlighting the following four
overarching points:
• Gap analysis and local stakeholder priorities: Some of the initiatives listed below are likely to
already be provided in certain facilities, and it is necessary to ensure that the support is not
redundant in such cases. Not all tertiary programmes include all of these elements, and the
support provided will also often have to align with the priorities of local stakeholders.
6
See, for instance, Marsden, Reintegrating Extremists; and Stina Torjesen, “Towards a Theory of Ex-Combatant
Reintegration”, Stability: International Journal of Security & Development 2, No. 3, (2013), 1-13.
7
Torjesen, “Towards a Theory”, 4.
8
For more on these activities in a variety of different contexts, see for instance Atta Barkindo and Shane Bryans, “De-
Radicalising Prisoners in Nigeria: Developing a Basic Prison Based De-Radicalisation Programme,” Journal of Deradicalization
7, (2016); Adrian Cherney, ‘Evaluating Interventions to Disengage Extremist Offenders: A Study of the Proactive Integrated
Support Model (PRISM), Behavioural Sciences of Terrorism and Political Aggression, in print; Hamed El-Said, New Approaches
to Countering Terrorism: Designing and Evaluating Counter Radicalization and De-Radicalization Programs (Hampshire:
Palgrave Macmillan, 2015); Hamed El-Said and Jane Harrigan, Deradicalizing Violent Extremists: Counter-Radicalization and
Deradicalization Programmes and their Impact in Muslim Majority States (Oxon; Routledge, 2013); and Malkanthi
Hettiarachchi, ‘Sri Lanka’s Rehabilitation Program: A New Frontier in Counter Terrorism and Counter Insurgency’, PRISM:
Journal for the Center of Complex Operations 4/2 (2013); Khalil et al, “Deradicalization and Disengagement”.

9
• Tailored to the needs of beneficiaries: The provisions should be tailored to the needs of each
beneficiary, rather than one-size-fits-all, to the extent that this is possible in each location
given the resource constraints.
9
• Human rights compliant: The programme in its entirety will have to comply with national,
regional, and international human rights standards. While the specifics will reflect the nature
of each initiative in question, it is worth adding that in the case of the Serendi centre in
Mogadishu this covered the adequacy of living and learning facilities, disciplinary procedures,
complaints mechanisms, entrance and exit processes, and so on.
• COVID-19 implications: While the impact of COVID-19 on future tertiary programmes is
difficult to predict at this stage, it may also influence the nature of the activities. For instance,
certain activities (perhaps sports) may be deemed excessively risky, it may impact on the
availability of suitably experienced and qualified professionals, and so on.
It is also worth observing that in many contexts there is a strong case for adopting a ‘lowest hanging
fruit’ approach in which implementers sequentially add new components of the programme over
time.
10
Through this approach, elements that do not require extensive specialist input for design (e.g.
sports and recreational activities) can be prioritised over those that do (e.g. religious education and
psychological support). Additionally, implementers can prioritise elements that are most likely to
generate ‘buy-in’ from beneficiaries (e.g. again, sports and recreational activities) over those that may
provoke suspicion simply because the programme is new and being implemented by an external
agency (e.g. psychological support and religious education).
Basic education
Basic education provisions are common to many tertiary P/CVE programmes, including those in
Malaysia, Nigeria, Saudi Arabia, Singapore, Somalia, and Sri Lanka. For instance, at the Serendi
rehabilitation centre in Mogadishu, the education provisions cover ‘standard school subjects such as
maths, Somali, Arabic, English, science, history and social studies,’ with the classes aligned to the
national curriculum.
11
Of course, such classes should ideally cover a range of levels to ensure that they
are relevant to each individual, and to allow beneficiaries to progress. Aside from offering a key life
skill, part of the logic of providing basic education is that it may help beneficiaries find work at a later
date, thereby encouraging them away from violent extremism. As with all of the interventions listed
in this section, it also provides an important communication effect by demonstrating that the state is
genuinely interested in the wellbeing of the beneficiaries, in contradiction to the message of many
violent extremist groups.
9
See, for instance, Christopher Dean, ‘The Healthy Identity Intervention: The UK’s Development of a Psychologically
Informed Intervention to Address Extremist Offending’, in Silke (ed.), Prisons, Terrorism and Extremism, 94; and Khalil et al,
“Deradicalization and Disengagement”, 23-25.
10
This would replicate the approach adopted by an UN-funded prison deradicalisation initiative in Somalia, as described in
James Khalil, ‘Pilot Rehabilitation Project for High Risk Prisoners in the Baidoa Prison’, 2017, at:
https://www.un.org/peacebuilding/sites/www.un.org.peacebuilding/files/documents/somalia_january_2017_-
_evaluation_of_rehabilitation_project_in_baidoa_prison.pdf
11
Khalil et al, “Deradicalization and Disengagement”, 23.

10
Vocational training
Vocational training is also common to many programmes, including those in Nigeria, Saudi Arabia,
Singapore, Somalia, and Sri Lanka. Even more so than basic education, vocational training is provided
specifically with future livelihoods prospects in mind, thus relating primarily to disengagement (rather
than deradicalisation) aims. To place this service in context, it is worth observing that the promise of
a salary and other material benefits are often key motives for involvement in violent extremism in
many locations.
12
Of course, the vocational options must be appropriate to local contexts, as ideally
revealed by recent livelihoods surveys, with these potentially including trades such as carpentry,
metalwork, printing, electronics, hairdressing and painting.
13
Given that the prospects for employment
are poor in many locations experiencing violent extremism, implementers should certainly also
consider supplementing these provisions with ‘build your own business’ and entrepreneurship
training.
Religious education
Religious education is also regularly provided through such programmes, including in Australia,
Malaysia, Nigeria, Saudi Arabia, Singapore, and Somalia. With many motivated into involvement in
violence by religious and political ideologies, this intervention relates specifically to deradicalisation
(rather than disengagement) aims. Unsurprisingly, religious elements are invariably identified as being
among the most problematic component of tertiary interventions. While perhaps generally more
applicable in ‘the West’, a first key issue relates to the right of freedom of religion, and concerns that
the state may be acting as ‘thought police’.
14
A second is that religious leaders involved in such
initiatives are often viewed as being little more than ‘captured’ state agents, with this often reportedly
impacting of their ability to influence the beneficiaries.
15
Nevertheless, there remain strong arguments
in favour of religious education initiatives given the central role that ideological motives often play.
Civic education
Civic education is seemingly less frequently provided through such programmes, and as with religious
education is principally associated with deradicalisation (rather than disengagement) objectives. As
observed by Shane Bryans, ‘space should be provided within the wider education curriculum for
learning about citizenship, reinforcing the values, rights, duties and responsibilities of individuals
towards each other and in relation to the State.’
16
Bryans adds that ‘learning about law, justice,
fairness, human rights and ethics in public life, democracy, the role of government, critical thinking
12
See, for instance, Khalil et al “The Attitudes-Behaviours Corrective (ABC) Model”
13
Bryans, “Handbook on the Management”, 83.
14
See, for instance, Daniel Koehler, Understanding Deradicalization: Methods, Tools and Programs for Countering Violent
Extremism (Abingdon: Routledge, 2017) 201.
15
Although relating specifically to the UK, this point is perhaps best expressed by Douglas Weeks, ‘Doing Derad: An Analysis
of the UK System’, Studies in Conflict and Terrorism 41, No. 7 (2018), 529-30.
16
Bryans, “Handbook on the Management”, 80.

11
and constructive debate is essential in countering subjective interpretations of the world propagated
by violent extremists’, and that ‘civic responsibility and citizenship may also build a shared sense of
culture within a society that consists of a diverse mix of ethnic, religious and cultural backgrounds.’
17
In the case of the Serendi centre in Somalia, such efforts contributed to positive attitudinal change
regarding political leaders, the Somali National Army (SNA), and elections and broader democratic
principles.
18
Psychological (or psychosocial) support
Psychological (or psychosocial) support takes many different forms and is implemented with many
differing objectives in mind. For instance, in the case of the Serendi centre in Mogadishu, the
provisions include both psychosocial support (involving solution-focused counselling and group-work
to address conditions such as anxiety and PTSD) and psychiatric treatment, with cases that require
institutional care referred to external mental-health service providers.
19
This support is provided
primarily on ethical grounds, rather than to encourage individuals away from violent extremism. By
contrast, psychological interventions through the Healthy Identity Intervention (HII) in the UK do
specifically aim to influence attitudes and behaviours. Specifically, the purpose of these efforts is to
encourage beneficiaries to fulfil their needs legitimately, increase their emotional tolerance and
acceptance, reduce their offense-supportive attitudes, beliefs and thinking, increase their personal
agency, and express their values and pursue their goals legitimately.
20
Also under the umbrella of psychology, creative activities (including painting, poetry, and so on) are
also often provided for therapeutic value, as a means of self-expression, and to provide a safe avenue
for beneficiaries to learn about their own motivations for involvement in violent extremism, social
issues, and so on. This umbrella also incorporates cognitive training, which often involve approaches
that are ‘rooted in addressing the structure of thinking rather than the content of the ideas believed
to support violence.’
21
More than the other forms of intervention, psychological support represents a
specialist thematic area, and so it is critical that such initiatives are designed by suitably qualified and
experienced individuals, ideally with prior experience in the country in question. While many of the
interventions discussed in this section should be tailored to individual needs, this is particularly the
case with psychological initiatives.
Family support
Efforts to reconnect beneficiaries to their families are also often incorporated into tertiary initiatives,
including in Australia, Malaysia, Saudi Arabia, Singapore, Somalia, and Sri Lanka. While families are
most often viewed in relation to their potential role during aftercare and reintegration (as discussed
17
Bryans, “Handbook on the Management”, 80.
18
Khalil et al, “Deradicalization and Disengagement”, 24.
19
Khalil et al, “Deradicalization and Disengagement”, 24.
20
Dean, ‘The Healthy Identity Intervention’, 98-100.
21
As discussed in Marsden, Reintegrating Extremists, 75.

12
shortly), they can also have an important influence over violent extremists while they are still in
facilities.
22
This issue is discussed at length by Shane Bryans:
Violent extremist prisoners retain, within certain limits, their right to family life and not to be
totally isolated from society, even though they may be prevented from physical interaction with
the outside world. Prisoners’ ability to maintain contacts with the outside world, in particular their
families, is also fundamental to their prospects of successful social reintegration … As appropriate
prison facilities may be located far from prisoners’ families in some countries, prison
administrations should make special efforts to prevent the breakdown of family ties, with visits,
letters and phone calls being allowed to the maximum possible extent, while ensuring that the
requisite security precautions are in place.
23
Sports and recreational activities
Sports and recreational activities are also often included in tertiary programmes, including in locations
such as Nigeria, Saudi Arabia, Somalia, and Sri Lanka. While certain commentators primarily emphasise
the physical benefits of such initiatives, they are also important in that they may encourage pro-social
thinking and behaviour.
24
As already observed, sports and recreational activities may provide a ‘low-
hanging fruit’ option for implementers keen to generate buy-in to such programmes during their initial
phases. In the words of Shane Bryans, ‘sport will often be the first point of entry into the prison regime
– an initial engagement in an activity which is likely to be regarded as providing fun and escape from
everyday routine and less likely to be a public statement of conformity.’
25
Reintegration support
Reintegration support refers to aftercare services offered to beneficiaries, with provisions tending to
include family support, assistance with connecting to the community, various forms of financial
support (including relocation support in certain cases), and the facilitation of work and education
opportunities. As with psychological support, these provisions tend to vary notably between contexts.
For instance, in the case of Saudi Arabia:
Officials consider three elements for re-socialisation and successful reintegration: a detainees’
willingness to get married and re-join his wife and family; his participation in university or other
continuing education; and his re-joining the workforce. With this in mind, the Ministry of Interior
activity helps former detainees on all these fronts by providing financial assistance, helping him
find a job, or providing other necessary social services. Prince Mohammed bin Nayef has even
attended ex-detainees’ weddings in order to show support for their post-release life and to
encourage their settling down with a family.
26
22
Khalil et al, “Deradicalization and Disengagement”, 25.
23
Bryans, “Handbook on the Management”, 13.
24
For instance, as observed by Barkindo and Bryans, ‘De-Radicalising Prisons in Nigeria’, 16.
25
Bryans, “Handbook on the Management”, 93.
26
Marisa Porges, ‘Saudi Arabia’s “Soft” Approach to Terrorist Prisoners: A Model for Others?’, in Andrew Silke, (ed.) Prisons,
Terrorism and Extremism, 176.

13
By contrast, the Serendi programme in Mogadishu offers more limited support,
27
with this partly
reflecting the fact that resources are more constrained in this case. However, a conscious decision was
also made by stakeholders to channel more resources to other aspects of the initiative on the grounds
that the objectives of the Serendi programme are to incentivise voluntarily disengagements from al-
Shabaab, rather than to help beneficiaries reintegrate per se.
4. STAFFING CONSIDERATIONS
Many of the programme activities listed in the previous section require specialist staff, including
vocational instructors, teachers, imams, social workers, and sports coaches. Irrespective of whether
these roles are filled by existing facility staff (as occurs in the Nigerian prison programme), external
implementing partners (as is the case in the Somali prison programme), or a combination of both (as
occurs in the Serendi centre), it is often difficult to identify suitably experienced and qualified
professionals who are willing to work in such environments (a problem potentially exacerbated by
COVID-19, if a critical mass of qualified professionals choose to avoid working in crowded locations
such as these). As expressed by Shane Bryans:
Not every faith professional will be well suited for a therapeutic function. Faith professionals who
are incredibly knowledgeable and experienced in their pastoral work may find it difficult to
operate in the prison environment, as the rehabilitation of violent extremists requires a different
set of skills and aptitudes. Assessing the faith professional’s willingness and ability to work in the
therapeutic environment should therefore be a key factor in the selection process.
28
This is of key importance as it is difficult to overstate the relevance of personal relations between
centre staff and programme beneficiaries. Research conducted at the Serendi centre in Mogadishu
revealed that the attitudes of staff was a key driver of deradicalisation in many cases.
29
Focusing on
the Nigerian prison programme, Atta Barkindo and Shane Bryans similarly observed that:
When Treatment Team members work with violent extremists with transparency, empathy and
decency, this can have a positive impact on the way the prisoners view government and its
institutions. In the Nigerian de-radicalisation programme, Team members had very few resources
to attend to the immediate needs of prisoners (such as health, legal issues, family contact, diet
etc.), but they worked with the prisoners and the prison authorities to try and resolve such issues
when possible. Beyond meeting immediate needs, the Treatment Team gave time; they
empathised with prisoners and were neither harsh nor judgemental in their interaction and
treatment. The prisoners stated to independent observers that this was the first time in their
direct experience that government had showed an interest and they were treated with dignity and
humanely.
30
27
Khalil et al, “Deradicalization and Disengagement”, 27-31.
28
Bryans, “Handbook on the Management”, 35.
29
Khalil et al, “Deradicalization and Disengagement”, 25.
30
Barkindo and Bryans, ‘De-Radicalising Prisons in Nigeria’, 21.

14
In any case, the designers of these programmes should certainly consider the potential need for
training within their budget, both for specialist staff and facility guards. This may include standard
modules for such facilities (e.g. administrative procedures, human rights, detention centre security,
vulnerable groups, complaints and procedures, and so on), as well as ones relating specifically to
violent extremism (e.g. recognising signs of radicalization, assessing violent extremist prisoners, and
so on).
5. FACILITY UPGRADES
There is a general consensus that tertiary interventions are less likely to succeed in centres that fail
to meet adequate standards in terms of physical facilities, including accommodation, educational
areas, kitchen facilities, and so on. As such, there is a strong argument for those tasked with designing
such initiatives to consider addressing applicable issues of this nature through the programme budget.
To place this suggestion in context, it is worth highlighting the following upgrades that began at the
Serendi centre in Mogadishu in 2016:
Having previously been furnished with only basic mattresses, sheets and mosquito nets, the
dormitories were equipped with new bunk beds and lockable personal storage boxes. The
vocational skills workshops also underwent comprehensive renovations, and a barbershop was
constructed to establish an additional livelihood option. The residents additionally benefited from
a refurbished medical centre, a new canteen, kitchen block, games and television room, and
football and basketball courts.
31
With this in mind, it is worth observing that several Serendi beneficiaries reported that they were
surprised by the quality of conditions on arrival, and that this provoked them to reconsider prior
negative opinions of the Somali government, which in turn contributed to their deradicalisation.
32
6. OTHER DETAINEES
Regarding prison contexts in particular, it is worth briefly considering the debate surrounding policies
that segregate violent extremist inmates from those detained on other grounds, even though
international stakeholders may ultimately have limited influence over such decisions. While the core
premise of tertiary programmes is that they can encourage certain individuals to deradicalise and
disengage, prisons are also widely considered to be ‘breeding grounds’ of violent extremism. It is
certainly not implausible that these opposing processes occur simultaneously to different individuals
within the same facility. While several notable commentators have cautioned against an uncritical
acceptance of the supposed ‘breeding ground’ phenomenon,
33
to the extent that those already
31
Khalil et al, “Deradicalization and Disengagement”, 20.
32
Khalil et al, “Deradicalization and Disengagement”, 25.
33
See, for instance, Peter Neumann, “Prisons and Terrorism: Radicalisation and De-radicalisation in Fifteen Countries”
(International Centre for the Study of Radicalisation, London, 2018), 25-27; Andrew Silke and Tinka Veldhuis, ‘Countering

15
involved in violent extremism have the potential to influence others, there is certainly a strong
argument for segregation policies that limit or prevent such interactions. Variants of this approach
having been adopted within prison contexts in locations such as Australia, Malaysia, Nigeria, the
Netherlands and the UK. Such policies also enable the authorities to concentrate scarce resources,
including staff with specialist skills and training relating to violent extremism.
However, there are also strong arguments against such policies, including firstly that they may
reinforce the notion that such organisations hold a political status, as was notoriously the case in
Northern Ireland. Secondly, they may also inadvertently help such groups maintain their internal
structure and discipline, as well as plan and orchestrate activities. In the words of Andrew Silke and
Tinka Veldhuis, these may include ‘mass escapes, riots, weapon smuggling, coordinated hunger
strikes, the intimidation and assassination of prison staff, and sophisticated campaigns of political
protest.’
34
Third, such policies have the potential to undermine the process of deradicalisation as this
process can occur through interactions between violent extremists and their non-ideological
counterparts if the two groups are able to mix. Finally, under certain circumstances such policies may
exacerbate overcrowding issues. The bottom line is that there is no ‘correct’ approach, and to again
quote Silke and Veldhuis ‘the preferred strategy is likely to depend on a range of contextual factors,
such as background and size of the inmate population, available resources and staff, levels of violence
and gang activity, and inmate culture.’
35
Irrespective of the above, programme designers must also consider that tertiary initiatives may
inadvertently deliver deleterious consequences if those detained on grounds other than violent
extremism are excluded from the benefits. To be more specific, they may generate jealousies and
hostilities that may undermine the general functioning of these facilities. While the literature is
relatively silent regarding solutions to this problem, the below quote from a prison programme in
Baidoa, Somalia, is worth considering:
To complicate matters further, there is seemingly a threat that the benefits received by the former
al-Shabaab prisoners may provoke hostility the others –which in turn may undermine the
functioning or security of the prisons. While the Prison Commander downplayed his concerns
about this issue, the project team highlighted their apprehensions over such possible negative
effects. Indeed, the provision of goats to the entire prison population was specifically designed to
address this threat.
36
7. DETERMINING PROGRAMME SUCCESS (OR FAILURE)
While it is notoriously difficult to determine the success of P/CVE programmes generally, this issue is
arguably somewhat less pronounced for tertiary interventions simply because of their proximity to
Violent Extremism in Prisons: A Review of Key Recent Research and Critical Research Gaps’, Perspectives on Terrorism 11,
No. 5 (2017), 1-10.
34
Silke and Veldhuis, ‘Countering Violent Extremism in Prisons’, 4.
35
Silke and Veldhuis, ‘Countering Violent Extremism in Prisons’, 5.
36
Khalil, ‘Pilot Rehabilitation Project’, 5.

16
beneficiaries. In any case, this section sequentially considers programme objectives, Theories of
Change (TOC), and indicators, relying on the terminology outlined in Box 1. As is always the case,
determining programme success (or failure) requires the team to articulate suitable objectives at the
outset, with budgetary and time constraints in mind.
Programme Objectives
37
Debates among the community of experts often revolve around whether the objectives of tertiary
programmes should relate to attitudinal (i.e. deradicalisation) or behavioural (i.e. disengagement)
change. Some favour the former, simply as they interpret the ultimate purpose of such initiatives to
relate to a reduction in the rate of recidivism.
38
Viewed in this manner, an appropriate objective
statement may be articulated along the following lines:
• To prevent beneficiaries returning to involvement in violent extremism.
Other experts instead promote the notion of deradicalisation on the grounds that individuals are often
motivated to exit violence in the first place by attitudinal change, and as sustained disengagement
may also be more likely if underpinned by deradicalisation. In other words, there are strong
counterarguments for treating deradicalisation as being a parallel program objective, at least on a par
with disengagement. This is the approach explicitly adopted by the Nigerian Prison Service and at the
Serendi program in Somalia. Depending on circumstances, a potential objective statement along these
lines may be:
• To prevent beneficiaries returning to involvement in violent extremism, and to help them
become (or remain) unsympathetic to such violence.
Of course, it is worth recalling that not all of those involved in this violence are sympathetic to its
ideology (as discussed in relation to Figure 1), and thus the deradicalisation objective is not necessarily
applicable to all beneficiaries. This aside, empirical evidence regarding the relationship between
deradicalisation and disengagement (see Figure 2) in the context of such programmes remains scarce,
and this is perhaps an additional argument for explicitly aiming for both, at least until more evidence
becomes available.
Shifting attention to the final concept discussed in Section 2 of this paper, programme designers may
choose instead to view their objectives in terms of reintegration,
39
with a suitable statement perhaps
expressed as follows:
37
This section draws heavily from Khalil et al, “The Attitudes-Behaviours Corrective (ABC) Model”.
38
For instance, this stance is advocated by Andrew Silke, “Disengagement or Deradicalization: A Look at Prison Programs for
Jailed Terrorists,” CTC Sentinel 4, no. 1 (2011), 20.
39
An increasingly number of experts are advocating for this approach, including Marsden, Reintegrating Extremists, 3, 10-
11.

17
• To help enable and facilitate beneficiaries reintegrate (socially, economically, and politically)
into communities.
Of course, certain donors may argue that while these three statements are perhaps adequate at the
programme ‘outcome’ level, they are insufficiently elevated to qualify as ‘impact’ statements in the
context of broader campaigns of violent extremism. As such, the following uppermost objective
statements may be more suitable:
• To contribute to peace and stability in [Location X]
• To undermine [Violent Extremist Group X]
Theory of change (TOC)
40
Figure 3: Template Results Framework
40
This section draws heavily from James Khalil and Martine Zeuthen, “Countering Violent Extremism and Risk Reduction: A
Guide to Programme Design and Evaluation” (Whitehall Report 2–16, Royal United Services Institute, UK, 2016).
IMPACT
OUTCOMES
OUTPUTS
ACTIVITIES
2.2: Beneficiaries participate in basic
education, personal development
courses, VET, religious, and
psychosocial support, tailored to
individual needs and ambitions
2 – Rehabilitation: Beneficiaries’
knowledge, skills and attitudes
are enhanced / improved in an
environment conducive to
positive change
2.1: The physical environment
in which the programme
occurs (i.e. in terms of
facilities) is conducive to
beneficiary rehabilitation
To contribute to peace and
stability in [Location X]
3 – Reintegration Support:
Beneficiaries’ prospects for
sustained socio-economic
reintegration are
enhanced
3.1: Beneficiaries receive integrated
support to social and economic
reintegration as permitted by the
security context
1 – Support to National
Stakeholders: National
stakeholder capacity to deliver
tertiary P/CVE initiatives is
enhanced
1.1: National stakeholders (ministry
officials, prison staff, etc.) receive
contextualised technical assistance,
guidance, and training

18
TOCs serve primarily to articulate the intended pathways from activities to the desired impact of
programmes, and to identify the key assumptions associated with these pathways. To do so, it is
necessary to establish a results framework (or equivalent), such as Figure 3. The purpose of listing
assumptions is to help identify aspects of programme design that may be problematic. For instance,
these may relate to the security context remaining suitable for such an initiative, key national
stakeholders buying-in to the programme and offering genuine support, beneficiaries being willing
and able to participate, all materials and equipment (e.g. for vocational training) being available at a
suitable cost, local communities being sufficiently sympathetic towards reintegration, and so on. Of
course, the point is that entire ‘branches’ of the programme may entirely fail to contribute to the
intended impact if even only one assumption is misguided. In this scenario it may be necessary to
redesign certain aspects of the initiative, with this potentially involving anything from subtle
alterations to the removal of entire programme branches.
Candidate Indicators
In line with best practice, suitable indicators should be linked to each Impact, Outcome and Output
statement of the results framework. Potential indicators for the intentionally ‘elevated’ impact
statement in Figure 3 may include:
• Violent incidents in Location X during Period Y (#)
• Population in Location X who claim that violent extremism is justifiable (%)
Of course, there is a strong case that such indicators would fail to provide meaningful information
about specific tertiary programmes given the so-called ‘attribution problem’. The point is that while a
reduction in violent extremism or sympathy for such acts in any given location may be driven by a
specific P/CVE intervention, such a change may alternatively be a consequence of lower levels of
violence in neighbouring locations, political accommodations, an economic boost offering increased
employment, a decrease in the availability of arms and other equipment, and so on. Such issues may
essentially be insurmountable, but elevated impact statements are nevertheless useful at the very
least as they remind stakeholders and implementers of the intended purpose of the programme,
helping to ensure that all parties are aligned.
The following top-level indicators are worth considering for rehabilitation facilities where the
overarching objective is understood in terms of incentivising voluntary disengagements:
• New arrivals at the centre (#)
• Proportion of new arrivals claiming they were partly incentivised to disengage by a prior
knowledge of the available rehabilitation and reintegration support (%)
The former again does not fare well against the attribution problem as it is subject to changes beyond
the control of the programme team, including battlefield momentum, and so on. By contrast, the
latter seemingly offers greater potential as a means to determine whether the programme is achieving

19
success. Increases in the percentage figure would suggest that knowledge of the programme is
spreading, and that its reputation is sufficient to incentivise at least certain individuals to exit violence.
Shifting attention to the outcome level of Figure 3, much of the measurement can revolve around the
three core concepts discussed in Section 2 of this paper. While recidivism rates should in principle
provide an ideal measure for sustained disengagement, in practice this metric is notoriously
problematic.
41
Among the various issues often identified, the relevant security agencies may simply
be unaware of former beneficiaries returning to violence, or may elect not pass this sensitive
information on to the implementing agencies. These rates may also be misleadingly deflated in cases
where former residents return to violence only after the relevant reporting periods. With this in mind,
and in the context of Islamist violent extremism specifically, it may be helpful instead to consider the
following proxy measures relating to the beneficiaries’ period of residence:
• Beneficiaries attending a ‘moderate’ prison mosque (%)
• Beneficiaries interacting with ‘moderate’ imams (%)
• Beneficiaries interacting with female staff (%)
• Beneficiaries interacting with other residents / staff of different faiths (%)
• Beneficiaries taking part in art / sport (where this may be deemed haram) (%)
Periodic questionnaires may also provide an avenue to directly measure deradicalisation, although
the reliability of such data is open to debate. These may include questions that deliver data for the
following indicators:
• Beneficiaries claiming that they would vote at the next elections (if applicable) (%)
• Beneficiaries claiming to be willing to befriend individuals from other religions (%)
It is also worth considering data from risk assessments, which typically aggregate metrics relating to
both disengagement and deradicalisation.
42
Shifting attention to reintegration, one option may be
again to rely on questionnaires, in this case administered after residency, to deliver data for the
following indicators:
• Beneficiaries (excluding those who have chosen to return to education) claiming to 'generally'
earn an income sufficient to cover their basic needs and (where applicable) that of their
dependents twelve months after exit (%)
41
On this issue, see for instance Barkindo and Bryans, “De-Radicalising Prisoners in Nigeria”, 19–20; Khalil et al
“Deradicalization and Disengagement”, 5; Porges, “Saudi Arabia’s ‘Soft’ Approach”, and Michael Williams and Steven
Kleinman, “A Utilization-Focused Guide for Conducting Terrorism Risk Reduction Program Evaluations”, Behavioural Sciences
of Terrorism and Political Aggression 6, No. 2 (2014), 3.
42
On the subject of risk assessments, see for instance Elaine Pressman and John Flockton, ‘Violent Extremist Risk Assessment:
Issues and Applications of the VERA-2 in a High-Security Correctional Facility’, in Silke, ed., Prisons, Terrorism and Extremism;
Joanne Richards, ‘High Risk or Low Risk: Screening for Violent Extremists in DDR Programmes’, International Peacekeeping
25, No. 3 (2018); Kiran M Sarma, ‘Risk Assessment and the Prevention of Radicalization from Nonviolence into Terrorism’,
American Psychologist 72, No. 3 (2017), 278-288.

20
• Beneficiaries claiming to 'have been accepted' by their family one year after exit (%)
• Beneficiaries claiming to 'have been accepted' by their community one year after exit (%)
8. INTERVENTION LOCATIONS
43
While there is no simple formula to determine environments that may be suitable for the initiatives
discussed in this paper, the following should be considered:
• State support: Key government stakeholders may be unwilling or unable to support such
programmes, for instance as they oppose ‘rewarding’ the perpetrators of violence, as they are
sceptical about the prospects of genuine rehabilitation and reintegration, as they need to
satisfy public opposition to such interventions, as they lack the technical capacity, and so on.
Of course, donors should favour locations where these issues are limited, as this will greatly
enhance the prospects of programme success.
• Security considerations: The beneficiaries and staff associated with such programmes may be
vulnerable to intentional targeting by violent extremist groups, and the facilities in which
these initiatives occur may also be attacked. Indeed, the more successful the programme, the
more likely it is that such groups will be incentivised to target what could amount to an
existential threat to their survival. With this in mind, policymakers must consider whether it
is possible to identify locations that are sufficiently secure for such programmes.
• Community acceptance: Many communities are the immediate victims of violent extremism,
and thus certain members outright reject the idea of living alongside those who were
previously involved in such violence. While government stakeholders may have some
influence over the level of support for the tertiary programmes (for instance, via community
consultations, messaging campaigns, and so on), donors should consider that public
opposition may be sufficiently strong to make such initiatives unfeasible in certain locations.
• Disengagement rates: In certain locations there may be factors that limit the rate of
disengagement to an extent that initiatives designed to incentivise such acts would be
essentially be destined to fail. For instance, this may be the case where the relevant violent
extremist groups maintain battlefield momentum and/or elevated levels of territorial control.
• COVID-19 considerations: It is difficult to predict the extent to which COVID-19 will continue
to influence everyday life over the coming years, and which specific geographies may be most
affected. Nevertheless, it seems reasonable to assume that the virus may render certain
locations unsuitable, at least temporarily, either as the risk of such programmes exacerbating
the spread of the virus is excessively high, or as it is not possible to identity suitably qualified
professionals willing to work in such crowded environments.
Disclaimer - This publication was produced with the financial support of the European Union. Its contents are the sole
responsibility of the author and do not necessarily reflect the views of the European Union.
43
This section draws heavily on Khalil et al “Deradicalization and Disengagement”, 33-35.