Research and analysis

Evaluation of the Safer Streets Fund round 2, year ending March 2022

Published 10 April 2024

Applies to England and Wales

Executive summary

This report presents the findings from an independent evaluation of the Safer Streets Fund round 2 (SSF2). In January 2021, the Home Office announced SSF2 which represented a £20 million investment to help reduce neighbourhood crime in hotspot areas through the delivery of situational crime prevention interventions such as CCTV, street lighting and target hardening measures. Local authorities (LAs) and Police and Crime Commissioners (PCCs) across England and Wales were eligible to apply for up to £432,000 per bid to fund their proposed interventions in residential and rural hotspots and including commercial hotspot areas in this round of funding. Fifty bids secured SSF2 funding and implemented their interventions in residential, rural and commercial areas between June 2021 and March 2022. SSF2 builds and expands on the learnings from the earlier Safer Streets Fund round 1 (SSF1) (Home Office, 2023), which focused on residential and rural hotspot areas only.

Verian (formerly known as Kantar Public) conducted a process and impact evaluation of SSF2 between June 2021 and November 2022. The process evaluation explored how local bid teams implemented and delivered interventions. The impact evaluation measured the impact of SSF2 on police recorded crime (PRC) and public perceptions of crime and safety (through a series of quantitative resident surveys).

Process evaluation findings

The process evaluation comprised 5 stages of qualitative research, beginning with a review of all 50 SSF2 bids. Other key steps in the process evaluation included a pre-implementation qualitative survey of all bid areas, in-depth case study research amongst 10 selected areas and a post-implementation survey of all bid areas.

The evaluation found bid development and delivery to be highly collaborative processes led by the LA or PCC’s office. Multiple local organisations such as community safety teams, resident groups and local charities played a key role in both phases.

A key point of feedback highlighted by several teams was the limited window of time available for both bid development and delivery (again, consistent with SSF1 findings). Several teams felt that this had shaped the choice of interventions and also limited the scope for community engagement at the time of bid design and development.

Local stakeholders confirmed that implementation went largely to plan and that it was a positive experience overall. While coronavirus (COVID-19) did cause some delays and challenges in resource availability, these factors did not have a major impact as many teams felt that their experience of dealing with similar issues on SSF1 had enabled them to adapt and deliver successfully.

Most of the key enablers of delivery success at a programme level were similar to those highlighted on SSF1. They included established relationships between different stakeholders, engagement of partners early in the delivery process and effective project management.

A key factor of success highlighted in this SSF2 round was the experience and expertise of the delivery team in the interventions they were involved in. This was especially important for interventions that had complex legal and governance requirements, such as closed-circuit television (CCTV), automated number plate Recognition (ANPR) and video doorbells.

Stakeholders also recognised that building awareness of the SSF programme and engaging the community well before implementation was important for delivery success.

Feedback on delivery in rural and commercial areas was largely aligned to residential areas. Thus, learnings can be shared across these areas. Stakeholders also fed back that business engagement was largely positive and was facilitated through the engagement of local commercial networks such as Bid Improvement Districts (BID).

Most bid teams, at the time of the research, felt it was too early to evaluate the longer-term impacts of the interventions. However, some areas anecdotally reported examples of early successes, including improved engagement with local communities and businesses with positive feedback on the interventions being delivered.

Impact evaluation findings

Verian measured the impact of SSF2 on neighbourhood crime and residents’ perceptions and self-reported experiences of crime and safety. The team did this by analysing PRC data about neighbourhood crime (domestic burglary, robbery, theft from the person and vehicle crime) and comparing resident survey responses (both within SSF2 intervention and comparison areas) before and towards the end of SSF2 implementation. Verian also combined the SSF1 and SSF2 resident survey data to conduct a pooled analysis (for variables included in both surveys).

Overall, there is evidence of statistically significant positive impacts of SSF2 on several measures (at the 95% confidence interval, which in social research is the typical way of conveying how precise a survey measurement is)^{[footnote 1]}. The survey-based analysis detected improved overall understanding of crime prevention measures (+10 percentage points, or pp) and increased use of property marking equipment (+3pp). There were also reductions in the proportion of residents reporting having experienced or witnessed shoplifting (-6pp) and robbery (-2pp) in SSF2 bid areas (versus comparison groups).

Regarding police recorded incidents of neighbourhood crime, there was no statistically significant reduction at the overall fund-level. Verian could not identify any conclusive explanations from the process evaluation data to explain why or how SSF2 did not impact these outcomes. There are various possible explanations, including intervention dosage (such as, whether the interventions delivered were of sufficient intensity or scale to have a measurable impact). There could also be several broader national trends unrelated to SSF2 (for example, socio-economic changes), as well as very localised events (for example, the arrest or release of a prolific local criminal, or increased gang activity in an area), which could have affected PRC trends over a short period.

There were also some unexpected negative findings in the impact analysis with a statistically significant increase in vehicle crime at the fund-level (in all bid areas that indicated targeting this crime). In the survey-based analysis, a greater proportion of residents reported acquisitive crime (+9pp) and anti-social behaviour (ASB; +11pp)^{[footnote 2]} as a problem in their local area. More residents reported experiencing or witnessing ASB (+9pp), particularly relating to witnessing the use or dealing of drugs (+8pp). One possible explanation is that communications around the Safer Streets Fund raised residents’ awareness of target crimes, leading to increased perception of crime and its reporting.

In the pooled SSF1 and SSF2 analysis, there was a statistically significant impact on understanding and use of crime prevention measures and awareness of crime prevention activities. Specifically, there was an increase in the proportion of residents saying they understand the measures they can take to prevent crime on their property (+7pp). Also, there was an increase in awareness of some crime prevention activities – partnerships to prevent crime (+2pp), the Safer Streets project as a whole (+4pp) and activities to improve public spaces, such as building car parks (+3pp).

Conclusions and recommendations

The insights generated through this evaluation highlighted some clear learnings for the implementation teams that they can apply to bid development and delivery in future rounds. Furthermore, the quantitative and qualitative work as part of this evaluation also helped to identify some key learnings to be considered by the Home Office when commissioning future rounds of the SSF programme.

For local area delivery teams, the report identified the following key recommendations:

• early engagement of delivery partners, stakeholders and the community (residents and businesses) is essential to ensure prioritisation of the programme across delivery team stakeholders and strong take-up of interventions amongst residents and businesses
• prior knowledge of intervention implementation processes is key to success so must be considered when allocating responsibilities for intervention lead and delivery at the time of programme set-up; special efforts are required to build awareness of legal and governance requirements associated with interventions
• specific audience groups, such as landlords and students are harder to reach and engage, as experienced by the SSF1 stakeholders; thus, developing customised engagement plans for these groups is important, which may include leveraging networks specific to them (for example, universities) rather than engaging directly; in addition, it is also important to develop a bank of learnings on what has worked for these audiences in the past to guide future work

For the Home Office, feedback received from stakeholders informed the following set of considerations that it can apply to future rounds:

• consideration of the time given to local areas for bid development and programme delivery is key for future rounds; most implementation teams felt that the current windows for bid development and implementation were very limited and had impacted the bid developed, and intervention selection and their timely delivery
• there is an opportunity for the Home Office to create and manage a platform that enables the exchange of learnings and best practice across bid areas; this would also enable the creation of an evolving bank of best practice that delivery teams could leverage for all future rounds; as a note, the Police Crime Prevention Initiative (PCPI) launched a new website in February 2023, which is open to all SSF practitioners and supports local projects with a variety of resources to assist with effective delivery of SSF interventions
• stakeholders expressed the need for additional guidance on specific elements of programme delivery, such as legal and governance processes, project management, intervention implementation and running of community surveys
• most interventions selected for implementation in commercial and residential areas were the same; to drive the implementation of interventions that are more suited/specifically for commercial areas/properties, it will be important to make available inputs and guidance on this to local delivery teams
• to enhance the effectiveness of future evaluations, the Home Office must consider the following recommendations:
  • a longer scoping phase to allow for assessing various impact evaluation options
  • ensure bid areas implement the interventions at a scale (such as, large intervention dosage) significant enough for residents to notice them and/or be reflected in the data
  • collect management information data at regular points (including about how many interventions have been implemented/provided to residents in that time period) in the delivery period to enable accurate assessment of the dosage and take-up of each intervention

Acknowledgements

Verian relied on the contributions from several parties to evaluate the Safer Streets Fund round 2 and report on our findings. First, our thanks go to Professors Nick Tilley and Shane Johnson of University College London for their invaluable advice on methodology, fieldwork, analysis and reporting. We are also grateful to Home Office analysts and policy colleagues for their regular guidance and feedback throughout the evaluation. Our thanks also go to the anonymous peer reviewers who provided us with comments during the reporting stage and finally, to the evaluated project staff and their local residents who gave their time generously to the evaluation process.

1. Introduction

The Home Office commissioned Verian to independently evaluate the Safer Streets Fund round 2 (SSF2). This report presents the findings from the impact and process evaluation carried out between June 2021 and November 2022.

1.1 Policy context and background of the Safer Streets Fund

Neighbourhood crime (domestic burglary, robbery, theft from the person, vehicle crime)^{[footnote 3]} can have devastating impacts on the livelihoods and safety of individuals and communities, alongside significant social and economic costs (Home Office, 2021). Neighbourhood crime comprises approximately a third of all crimes, excluding fraud and computer misuse. Neighbourhood crime had been falling before COVID-19 impacted levels of crime, with a 12% decrease in the year ending March 2020 compared to the previous year (1.96 million incidents down from 2.24 million in year ending March 2019) (ONS, 2022).

Although COVID-19 reduced the opportunities for theft, crime levels were expected to revert back to previous levels once government restrictions were removed and the economy started to open up again (Langton et al., 2021). When the Home Office announced SSF2, there was already some evidence indicating an increase in crime levels but neighbourhood crime remains below pre-pandemic levels (Home Office, 2023). Investing in crime prevention solutions was therefore of importance going forward. Crime prevention measures such as alleygating, CCTV, street lighting and home security have been proven to prevent neighbourhood crimes within certain individual, household and area contexts as they remove opportunities to commit crime or act as a deterrent (Piza et al., 2019; Welsh & Farrington, 2008). The Home Office introduced SSF2 to tackle neighbourhood crime by implementing these crime prevention interventions (College of Policing, 2022a).

1.2 SSF2 overview

SSF2 is a £20 million government fund which aims to reduce neighbourhood crime in key residential and commercial hotspot areas. Local authorities (LAs) and Police and Crime Commissioners (PCCs) across England and Wales were eligible to apply for up to £432,000 per bid to fund their proposed interventions. As outlined in the SSF2 prospectus (Home Office, 2021), the key objectives of the Fund are to:

• reduce neighbourhood crime in areas that receive funding – making local areas safer and reducing demand on the police allowing them to focus on higher harm crimes
• continue building evidence about the impact of targeted investment in situational prevention^{[footnote 4]} in high-crime areas to strengthen the case for future investment, both at a local and national level
• grow local capability to undertake data-driven problem solving and capture evidence and practical learning about how best to implement situational interventions to prevent crime

SSF2 builds and expands on the Safer Streets Fund round 1 (SSF1), with a few minor changes to address the feedback and lessons learned through the first round of the Fund’s delivery. These changes include:^{[footnote 5]}

• focusing the Fund onto the 4 neighbourhood crime types
• allowing investment into commercial and non-residential areas
• increasing flexibility on the defined geography of eligible areas, bringing rural areas into scope
• allowing LA-led (local authority) bids

1.2.1 Overview of the SSF2 bids and interventions

Overall, there were 50 bids that secured SSF2 funding (out of 97 bids received by the Home Office).

On SSF2, driven by the bidding requirements, there was a mix of both PCC-led (24%) and LA-led bids (76%). While previously on SSF1, all bids were led by PCCs while LAs involvement was specific to individual interventions. In SSF2, coverage of acquisitive crimes has increased, with more bids focusing on crimes such as robbery (62% in SSF2 versus 21% in SSF1) and car theft (70% in SSF2 versus 29% in SSF1).

Similar to SSF1, close circuit television (CCTV) (92% of bids), street and other lighting (68% of bids), and alleygates (38% of bids) were the most prevalent interventions. However, the number of bids that included video doorbells in the proposed designs in this round was slightly higher (22% of bids in SSF2 versus 13% in SSF1). A higher number of bids on SSF2 included ANPR (32%) compared to on SSF1 (13%). Notably, as in SSF1, more bids focused on home target hardening than vehicle or business target hardening.

Figure 1 outlines the 11 categories of interventions and the number of bid areas that planned to deliver them (in grey) and those that were actually implemented (in green). Overall, while the SSF2 interventions were largely delivered as planned in the initial bid applications, there were some discrepancies between what was planned at the outset and what was delivered by the end of the programme. Section 3, ‘Process evaluation findings’, captures the implementation process and experience and, thus, outlines factors that may have caused these discrepancies.

A detailed breakdown of the interventions and bid areas that delivered them is captured in Appendix A, with the amount of funding each bid area received in Appendix B.

Figure 1: Planned and delivered interventions

Bids in the SSF2 programme targeted both residential and commercial areas, with 24 SSF2 bids included commercial area interventions in their bid design. Most of these tended to be a mix of interventions focused on the residential and business communities. Interventions included, amongst others, CCTV and lighting, gating, target hardening of public spaces such as improvements to subways, graffiti removal, and high police visibility in hotspot areas. The commercial areas tended to include a mix of commercial establishments and residences. This was a change from SSF1, which focused on residential areas only.

Despite the expansion of eligibility of the Fund, only 3 bid areas in SSF2 were defined as rural towns or rural villages. These bids delivered interventions that were largely similar to the non-rural areas such as fixed and re-deployable/transportable ANPR and CCTV (monitoring the entrances to sheds), property marking kits and community engagement activities, Crime Security Assessments to repeat victims of crime, as well as portable cameras, video doorbells, crime prevention kits and perimeter intruder alarms for residential and farm properties.

1.2.2 Overarching objectives of the interventions

Aligned with the Fund’s requirement, bid teams reported that the focus of their SSF2 programmes was to reduce acquisitive and neighbourhood crime. Similar to SSF1, most bid teams also targeted a set of secondary crimes that included anti-social behaviour (ASB), violent crime, drug-related offences, and environmental crimes such as fly tipping and dog fouling. Less frequently targeted crimes included domestic abuse and gang activities, catalytic convertor theft and HGV-related crimes.

As on SSF1, several bid teams also highlighted several broader objectives they felt the programme could help achieve, such as fostering a sense of pride and ownership of the area, improving feelings of safety amongst the community and building confidence of the community in the local police force. Some stakeholders also reported using the programme to promote the attractiveness of city and commercial centres, thus accelerating economic benefits to the area.

“Engaging with local communities to build resilience beyond the time scales of the project through targeted and environmental activities and improvements. To improve the look and feel of the targeted areas which will contribute to improvements in resident perceptions of crime and safety.”
(Local authority, Residential)

2. Evaluation design and methods

This section outlines the evaluation objectives and the impact and process evaluation methodology. Appendix C provides further technical details on the evaluation design and methodology.

2.1 Evaluation objectives

SSF2 evaluation aimed to provide both a robust impact assessment and an in-depth exploration of implementation learnings through a combination of bid-level case studies and programme-level evidence synthesis. The core objectives of this evaluation align with the stated key outcome measures that the Home Office specified from the outset, including:

1. Explore whether, and how, funded projects help to reduce neighbourhood crime (as the primary focus of SSF2 interventions) and other crime and disorder types (defined as secondary objectives / outcome measures, such as wider acquisitive crime types or ASB) and produce other benefits (for example, improved public perceptions of safety and policing) as a result of implementing local crime prevention plans (see sections 3.3, 3.4, 3.5, 4.1 and 4.2).
2. Produce evidence that can be useful to practitioners and policymakers interested in reducing neighbourhood crimes (sections 3 and 4).
3. Bring together findings across areas through a common outcomes framework and project clustering by intended effects, intervention approach and local contexts (see section 4).

2.2 Evaluation approach and methods

The SSF2 evaluation triangulates findings from multiple data sources and research methods, as outlined below, to provide an account of the programme implementation and its impact.

If the impact analysis had indicated that SSF2 led to a reduction in crime, then the intention was to conduct an economic evaluation of SSF2 (using the Home Office economic and social costs of crime framework (Heeks et al., 2018)), in conjunction with the Manning cost-benefit tool (Manning et al., 2016). However, it was not possible to monetise the overall benefits of the Fund, given the findings of the impact analysis.

2.2.1 Impact evaluation methodology

This section outlines the methodology Verian used for the impact evaluation (section 4 covers the findings). The impact evaluation used 2 quasi-experimental approaches:

a) The synthetic control methodology to measure bid-level changes in neighbourhood crime using historic police recorded crime (PRC) data. Verian constructed synthetic control areas by taking a weighted average of non-SSF funded areas similar to the bid areas in terms of crime trends and other characteristics. The team generated the estimates of the impact of the Fund in each area by comparing the post-intervention trends of the bid areas to their synthetic control.

b) A difference-in-difference (DiD)^{[footnote 6]} analysis to measure changes in public perceptions of safety, crime and policing, using repeated cross-sectional surveys of residents in treatment and control areas, at the start and end of the SSF2 intervention period. The team analysed the differences in the changes in aggregated outcomes between the bid areas and the control areas using regression modelling^{[footnote 7]} to generate estimates of the impact of the Fund.

a) Impact analysis of PRC data

Verian used a time series of administrative crime data, applied to the synthetic control method, to analyse the impact of the Fund – and of each individual local scheme – on all neighbourhood crime or one or more specific neighbourhood crime types.

The team used the number of reported crimes recorded in either the Home Office’s DataHub or police.uk for the analysis. DataHub was the preferred source of data, but the team used police.uk data for bid areas in police force areas (PFAs) that have historically not returned or returned incomplete data to the DataHub^{[footnote 8]}.

The follow-up period for the impact analysis was April to June 2022 (quarter 1) (after all the bid areas were expected to finish implementing their interventions). At the point of the analysis, PRC data was available until June 2022. The team applied the synthetic controls method to generate a control group for each of the 50 SSF2 bid areas, matching on population and PRC statistics for the five years before the start of the intervention (quarter 1, 2016 to 2017 (April to June 2016) to quarter 1, 2021 to 2022 (April to June 2021)).

Population characteristics included demographics as recorded in the census, housing density statistics, and information about the number of COVID-19 cases recorded in an area (see Appendix C for further information).

This method used a weighted average of multiple control areas (census Lower Layer Super Output Areas or LSOAs) to create an artificial control group whose trend in PRC before introducing the Fund was closely matched to the trend in the SSF2 bid areas.

To ensure the team compared bid areas to similar areas, it drew the synthetic control group from a pool of potential comparison LSOAs with similar population characteristics, as opposed to considering the entire pool of available comparison LSOAs. The team excluded areas from the pool where the first round of SSF was implemented.

It was not possible to control for the presence of other crime reduction interventions (outside SSF2) within both the test and control areas. Similarly, this analysis did not control for different types of SSF2 interventions that were implemented in different SSF2 bid areas.

The team measured the impact of the Fund and bids by comparing the level of crime in the synthetic control areas to those in the SSF2 bid areas, after administering the interventions. The impact estimates represent the difference between the number of crimes reported in a bid area and the synthetic control estimate for that bid area. As a note, there are some limitations to this approach – including controlling for external factors such as COVID-19 and data-related dynamics including statistical floor effects – which are discussed in more detail in Appendix C (see section C2.1.1).

b) Survey-based DiD impact analysis

To assess the impact of the SSF2 interventions on public perceptions of safety relating to crime and policing, the team conducted repeat cross-sectional surveys at pre- and post-implementation intervals in each bid area, and in 100 control areas (2 control LSOAs for each bid area). The control areas were selected using Coarsened Exact Matching (CEM) to ensure similarity to bid areas across a range of characteristics (more detail in Appendix C).

The survey data was collected using an Address-Based Online Surveying (ABOS) methodology, with local addresses being randomly sampled from Royal Mail’s Postcode Address File. In summary, the baseline and follow-up survey achieved the following:

• baseline survey (9 August to 12 September 2021)
  • 3,327 achieved individual responses comprising 1,198 paper completes and 2,129 online completes
  • 74% of completes (2,456) were from SSF2 bid areas
  • the household response rate was 21.9%
• follow-up survey (6 May to 19 June 2022)
  • 3,311 achieved individual responses comprising 971 paper completes and 2,340 online completes
  • 74% of completes (2,455) were from SSF2 bid areas
  • the household response rate was 19.1%

See Appendix C for the method used for calculating these response rates.

After collecting, processing and weighting the survey data (more detail in Appendix C), the team conducted the DiD analysis to assess the impact that SSF2 interventions had on a range of observed outcomes. This analysis measured the change in outcomes for treatment areas (SSF2 area residents) at 2 points in time (pre- and post-programme) and compared this change with the trend in outcomes experienced by non-SSF2 residents in the control group areas over the same time period. The difference between these 2 levels of change is the impact attributed to the SSF2 interventions, as illustrated in figure 2.

Figure 2: Difference-in-difference analysis

All social survey data is subject to sampling error, which may have caused changes in the same measures between collection points. Statistical significance testing helped to determine which changes could be attributed to the SSF2 programme, rather than to any sampling or measurement error. The DiD analysis focused on impact findings that were statistically significant at the 5% level. This means that if the surveys were to be conducted 100 times, we would expect a finding of the same nature on at least 95 occasions.

2.2.2 Process evaluation methodology

The process evaluation aimed to explore the set-up, delivery and implementation experience of local SSF2 projects. It aimed to identify the key enablers and barriers to successful delivery, and capture any lessons learned by the teams on the ground, to inform best practice for future rounds of funding.

The process evaluation comprised 5 tasks, with an overview provided in figure 3 below.

Figure 3: Process evaluation stages

Scoping phase: Bid review (June to July 2021)

Verian reviewed all 50 successful SSF2 bids to get a better understanding of the interventions implemented as part of SSF2. The team developed a bid review framework for summarising the key information from the bids.

Task 1: Process evaluation priorities workshop (August 2021)

Verian organised a half-day workshop with Home Office SSF2 team members, academic partners and College of Policing stakeholders. The aim of the workshop was to determine which interventions were the priority for analysis in the process evaluation case studies, as well as to identify information gaps and develop the key research questions the evaluation should aim to address.

Task 2 : Pre-implementation questionnaire (September to October 2021)

The team distributed the pre-implementation questionnaire among all 50 successful bid applications. The questionnaire was qualitative in approach and was designed to develop an understanding of each bid prior to intervention delivery.

Task 3: Intervention case studies (December 2021 to February 2022)

An in-depth analysis of 10 case study areas was conducted via interviews with stakeholders and, where possible, resident or business organisations. The objective of this exercise was to develop bid-level insights into delivery approach. See Appendix C 1.4, Task 3 for detail on how the team selected case study areas.

Task 4 : Post-implementation questionnaire (May to June 2022)

All 50 successful bid teams completed the post-implementation questionnaire. This was a qualitative questionnaire and administered after the core implementation period. It focused on bid teams’ experiences of implementation, early signs of impact and their assessments of future programme sustainability.

The process evaluation findings (section 3) provide detailed analysis, learnings and implications for practitioners implementing crime prevention measures. Specific learnings and implications are provided on all stages of the project process, from bid development through to project delivery and intervention-specific learnings.

3. Process evaluation findings

This section outlines the process evaluation findings, highlighting the experience of local SSF2 delivery teams throughout the programme journey. It captures the experience and learnings in the bid development and programme set-up phases, before moving on to the project implementation phase. Broadly speaking, these findings are consistent with the earlier SSF1 evaluation, with some important differences, for example, the SSF2 implementation was less affected by the COVID-19 disruption and it had different policy priorities. As a note, Verian also delivered a separate ‘Process Implementation Checklist’ for the Home Office to promote amongst local SSF teams and to support them with their bid development and delivery efforts.

To deliver insight and learnings for subsequent rounds of the SSF programme, this section covers the multiple enablers and barriers to SSF2 implementation according to the stakeholders interviewed. It documents these findings from an overarching programme point of view and at the intervention level. It also covers some early outcomes and successes reported by stakeholders and their approach to ensuring sustainability of the interventions beyond the programme period.

As mentioned earlier (in section 2.2.2), the process evaluation draws on multiple sources of insight, including: a review of all the SSF2 bids; a pre-implementation survey of all bid areas; in-depth case study analysis of 10 local areas; and a post-implementation survey.

3.1 Bid development and programme set-up

This section outlines key insights into the approach adopted by teams and their experience of the bid development and programme set-up phases.

3.1.1 Approach to bid development and programme set-up

The research findings suggest that bid development was a highly collaborative process. While efforts tended to be coordinated and led by LAs or PCCs, several other organisations – such as local charities, resident and business groups, academic partners, Designing Out Crime Officer teams – were also engaged and input from them sought through surveys and consultations.

Besides the above, bid teams also reported taking a data-led approach by tapping into a range of other sources as key inputs into bid development, such as:

• College of Police and Safer Streets toolkits to draw on the existing evidence base around intervention choice and delivery
• Environmental Visual Audits (EVAs), property reviews and walkabouts to understand needs of the area and opportunities for change
• feasibility studies in collaboration with partners (for example, partnering with housing associations on home security improvements)
• triangulation of different data sources – such as police crime data, publicly available data on demographics and levels of deprivation in the area, and MOSAIC data^{[footnote 9]} – as an input into hotspot mapping and geographical decisions on intervention installation

Bid teams’ initial project set-up included establishing governance arrangements, including project management, and operating structures. Areas varied in how complex their governance structures were, but usually these included steering groups formed of LAs, police forces, PCCs, charities, businesses and/or residents’ groups.

Following this, bid teams proceeded to verify resource availability and initiated procurement and staff recruitment, where required. Bid teams selected intervention leads based on their practical understanding and prior experience of implementation.

In this phase, some teams also reported adapting the bid design. For example, one bid area identified locations for CCTV camera installations, which were deemed unsuitable during the set-up phase and new locations were assessed and selected. Another team reported reducing the number of video doorbells to be installed and increased the CCTV provision due to GDPR requirements around video capture in private spaces.

3.1.2 Feedback on the bid development and project set-up experience

Overall, bid teams found the bid development process to be a positive learning journey, with clarity on what was expected of them through the process.

“We felt the bid process was clear and precise and wouldn’t change it.”
(PCC, Residential and Commercial)

Bid teams generally considered the guidance on bid development to be helpful and complete, but some gaps highlighted by stakeholders were:

• Safer Streets Fund (SSF) toolkits to incorporate more case studies to support the planning of interventions and highlighting synergies between different interventions
• guidance on how to manage Safer Streets projects
• guidance to support the completion of EVAs
• guidance on how to run community surveys and engage the community
• guidance on effectively implementing neighbourhood crime interventions
• guidance on regulatory and legal compliance (related to specific interventions, for example, GDPR).
• information and guidance on governance processes (related to specific interventions, for example, planning permission)

For approximately half of the teams, the key area for improvement was bid timeframes. These teams felt the tight turnaround for bid submission had restricted the opportunity for stakeholder engagement and contributed to inadequate groundwork when preparing for delivery. It also, they reported, additionally pressured their teams.

“The time frames are unrealistic, as we have had to find resources in already stretched teams to try to pull together the data and information required.”
(PCC, Residential and Commercial)

Similar to SSF1, there was feedback from a few areas identifying Lower Layer Super Output Areas (LSOAs) as challenging as crime was not reported at this level, making analysis more complicated.

A few stakeholders also felt there were some gaps in the breakdown of the data, making analysis more difficult. For example, the feedback suggested that burglary and robbery data available to them does not break down to residential or personal crime, thus requiring further analysis.

A few practical challenges were also highlighted by some teams that included difficulties accessing the Jaggaer portal^{[footnote 10]} and communication channels (especially for some LA systems), and navigating the high number of separate submission documents.

A few teams could not complete this scoping work before bid submission owing to the lack of time. A few also felt that the limited time for bid development prevented more in-depth analysis to better identify the right set of interventions for their local areas. They also felt that additional time for bid development would have enabled more community engagement, and thus better understanding of interventions required for the area.

“With more submission time we could have conducted more in-depth analysis of the data, including more interaction with local residents and a better analysis of the best interventions for the area. We also would have been able to include more actual intervention work. Most of the recommendations from our Environmental Visual Audit that we did not include in the bid were due to delivery time scales.”
(Local Authority, Residential)

3.2 Programme delivery experience

3.2.1 An overview of what has worked well at the programme level

Overall, bid areas were positive about their delivery experiences and proud of what had been achieved, despite some barriers and initial setbacks. The following emerged as key elements that had gone well on implementation at the programme level.

Leveraging existing frameworks and partnerships at the local level contributed to partnership working and smooth bid plan delivery

Established team relationships, from SSF1 and other ongoing collaborations, were reported to be crucial for mitigating risks and challenges.

Bid teams also consistently highlighted that leveraging existing frameworks and contracts in place was a significant support in timely bid plan delivery. Working with contractors and suppliers already in place and using existing procurement frameworks led to considerable savings in time and resource compared to bid teams that were required to develop new relationships, set up new contracts or initiate a lengthy procurement process.

“One of my recommendations to people would be, if you’re going to do something like that [procuring], have a look and see if there’s any existing frameworks you could use, because you could potentially identify that upfront, and have that ready to go.”
(PCC, Residential and Commercial)

Experienced and knowledgeable team members resulted in positive intervention implementation experiences

Bid teams consistently reported that team members’ prior experience of interventions, such as using and installing the technology/equipment (for example, cameras, lighting, target hardening kits) played a central role in smooth and timely implementation. Additionally, team members with knowledge of the area (for example, of the community, the physical environment and the types of crime) and familiarity with contractor teams were also key to facilitating a positive delivery experience.

“Sometimes the implementation doesn’t go to plan, so we have to be creative to look for solutions. This is where having a good network of contacts with experience in a variety of subjects is very helpful. We try to build our positive experiences and solutions to problems from previous projects into the design and implementation of our new projects.”
(PCC, Residential)

Effective project management and leadership was crucial to driving successful implementation and a positive experience for teams

As highlighted in the SSF1 report, several bid teams reported that having a team member dedicated to project management was core to successful delivery.

“People all have their day jobs, it can become a big beast to manage. If I was to say anything, I would say get yourself a good project manager who can oversee it.”
(PCC, Residential and Commercial)

A minority of areas reported they faced challenges when the project manager had no project management experience or was not adept at using project management tools.

Regular and frequent communication across delivery teams played a pivotal role in a positive and successful delivery experience

All bid teams felt they had successfully ensured regular and frequent communication (meetings, written updates and reports) regarding delivery, including the challenges, mitigations and learnings, as well as regular updates on progress. This included providing summary updates to community safety partnerships/ community safety strategic groups and local community meetings, in addition to social media updates and press releases.

While the findings above closely align to the experience on SSF1, including commercial areas was specific to SSF2. While the experience of teams in these areas did not differ from those in residential areas, the stakeholder feedback specific to implementation in these contexts is captured below.

Delivery in rural areas was smooth and efficient despite minor delays in procurement and, for some, a slightly under target take-up of residential interventions

Stakeholders in rural areas felt that implementation had been smooth, despite challenges reported across the board such as procurement delays and slightly lower than planned take-up of residential interventions. No significant differences were reported in the implementation approaches or experience between rural and non-rural areas.

Interventions in commercial areas were delivered effectively and the experience was aligned with that of teams implementing the programme in residential areas

Bid teams reported working with local commercial networks such as Business Improvement District^{[footnote 11]}, and local Chamber of Commerce to engage businesses and understand their needs.

Case study 1: Successful business engagement

One bid area reported that local businesses were very supportive of their team’s intervention delivery plan and were generally more on board with the programme, as a whole, than local residents. A major facilitator of success in this area was the programme dovetailing on the city’s Business Improvement District and Business Watch initiatives. Through these channels the delivery team could respond to business requests for taller alleygates, necessary in commercial areas, and adapted plans accordingly. “Both of those organisations have been key in getting especially small businesses to engage with us and tell us exactly what the issues are and how we can help them. If we haven’t had got that, we could be going in installing all things in the wrong places, not really tackling the problem.” (OPFCC, Commercial)

Business response to SSF2 was largely positive

The business response to the programme and the interventions was positive. Business groups were involved in regular project management discussions and offered opportunities to feedback on delivery of the programme. Groups saw being involved, particularly in early stages, as a positive because they could establish links/networks, understand benefits of the intervention for them, and overall felt their concerns were listened to and addressed.

“Businesses engaged well and really saw the benefit in the aims of the work being undertaken. The guardianship elements were well received, and positive feedback was received. Some challenges to get started but once those conversations had occurred, they were really receptive.”
(PCC, Residential and Commercial)

3.2.2 An overview of the challenges to delivery at the programme level

This section captures a summary of what bid teams felt could have worked better at the programme level, meaning they were not intervention-specific. Barriers relevant to specific interventions have been captured in the intervention-specific section later in this section.

Supply chain issues hindered implementation across bid teams

Similar to the SSF1 experience, supply chain issues and product shortages caused some delays. Specific issues highlighted included global shortages of products like microchips and slowing down of import processes driven by Brexit, which impacted interventions such as CCTV, video doorbells, ANPR, alleygating, environmental improvements and target hardening. However, teams reported responding to this by adapting their plans (for example, through the use of existing equipment already owned by local organisations) or delivering on a smaller scale.

Staffing challenges exacerbated by COVID-19-related staff absences impacted some bids

While not a key challenge, like in SSF1, a few bid teams reported facing resourcing challenges internally and amongst partners, causing some delays in delivery.

“There were some challenges around timescales for delivery which were incurred internally due to staff sickness and loss of momentum.”
(PCC, Residential)

Lengthy and complicated procurement processes delayed delivery for some bid teams

Several bid teams highlighted delays in implementation due to their internal processes for procurement of equipment and contractors at the local level. This was relevant for interventions such as home target hardening and street lighting. For some teams, particularly those that had not implemented video doorbells previously, the absence of pre-approved suppliers required them to go through lengthy processes to set up new contractors in the system. A few teams reported that even when there were approved service providers, they were required to obtain quotes from others before committing to the contract.

“[The LA partner] doesn’t have single suppliers, which slows down procurement and processes. Tendering processes can slow things down by a couple months.”
(PCC, Residential).

Conflicting agendas and priorities and lack of coordination sometimes acted as a barrier to effective collaboration, and thus, delivery

As on SSF1, some SSF2 teams reported challenges in engaging contractors and getting them to prioritise delivery on SSF within the timelines. They felt that this could have been avoided by engaging them earlier in the project planning phase and making everyone aware of agendas and timelines.

“Our experience of working with a contractor was good, but we are aware of their many conflicting priorities.”
(PCC, Residential)

This also was experienced amongst teams delivering different interventions on SSF2, leading to lack of coordination in their implementation, as illustrated in the case study below.

Case study 2: Coordination of complementary interventions

Coordination between teams delivering different interventions proved challenging in one area. ANPR, CCTV and street lighting were viewed as complementary interventions, but were installed by different delivery teams. Lack of coordination between delivery teams across this bid meant that the installation of each intervention was staggered, thus limiting their efficacy. Here project management was hindered by lack of understanding of the different organisational procedures followed by each partner organisation. These different approaches to delivery were, thus, not built into project plans. “When the camera goes in, the camera is one part, but all of the partners are also doing their part in it as well.” (PCC, Residential and Commercial)

Difficulties in engaging the community or key stakeholders and lack of buy-in was a challenge for some interventions

A few bid teams delivering street and other lighting, alleygating, video doorbells and environmental improvements reported challenges in engaging the local community, including landlords and landowners, especially in commercial areas where property owners were difficult to identify. This reflects the experience of implementation teams on SSF1 too.

Lower take-up rate than planned on certain interventions acted as a key barrier to implementation

While most teams reported a positive response from the community on interventions, a few teams also fed back on the lower than anticipated response from the public on interventions such as target hardening of homes, property marking and video doorbells. Stakeholders felt that several factors had affected take-up rates across SSF2 target areas. Firstly, perceived low levels of trust in police/public authority figures were seen to lead to lack of community confidence in the efficacy of interventions. Second, COVID-19 restrictions had limited opportunities for face-to-face engagement, leading to lower awareness of the interventions and understanding of how they would work. Additionally, specific targets groups such as transient populations and landlords were identified as difficult to engage, leading to lower take-up rates amongst them.

Case study 3: Low take-up rate amongst hard-to-engage groups

One bid reported facing difficulties engaging the area’s highly transient student population, who were less responsive to community engagement strategies and often unwilling to use target-hardening equipment distributed by the OPCC. This lack of engagement impacted the efficacy of strategies targeting burglary. “It can be quite difficult for us to deliver with individuals that aren’t long-term residents of the community, because they don’t have the same home feeling that they would have in their hometown – that has been a challenge because we have a lot of members of our community who are just passing members.” (PCC, Residential and Commercial) Similarly, the large student population meant delivery teams had difficulties identifying and contacting their landlords to provide them with safety equipment, and in receiving their permission to install video doorbells, leading to a lower-than-expected take-up rate in some cases.

Knowledge sharing within and across SSF2 bid teams was not frequently reported

However, those teams that did engage in knowledge-sharing activities reported benefitting greatly from doing so. They reported sharing:

• good practice (specifically of best procurement processes)
• technology and equipment procurement and installation advice (from specialist teams)
• recommended providers of services (best contractor relationships)
• feedback from engagement with the community (shared with local police and councillors)

One stakeholder fed back that there would be value in sharing knowledge and information across different bid teams in the future rounds.

Some teams noted that the delay in notifying local areas of their successful bid application contributed to limiting the programme delivery period. This delay also meant that it was not possible for staff working on SSF1 to continue on to SSF2, thus limiting the opportunity to ensure consistency in staff knowledge and engagement.

Several teams said they would like to see changes to the time available for implementation

Besides finding the bid development window limited, as reported in section 3.1.2 many teams also felt the implementation window was inadequate for the installation of several interventions, including alleygates and video doorbells. Teams reported not being able to implement interventions optimally due to the funding deadlines.

“The biggest issue with all of these grants, and not just Safer Streets, is it is too short, it’s too short funding and too short turnaround – it’s not practical. So it puts this ridiculous pressure on you all the time, and you end up doing what’s easy, rather than what’s best.”
(PCC, Residential)

3.3 Intervention-specific delivery experience

This section presents the key enablers and barriers specific to the implementation and delivery of each intervention. These findings come primarily from the case study analysis and the post-implementation qualitative questionnaire completed by bid areas.

3.3.1 CCTV

CCTV was a planned and delivered intervention for most bid areas (43). Implementation included new fixed and re-deployable cameras, and network upgrades (replacing existing cameras with new installations). The scale also varied greatly between bid areas, with installation ranging from over 70 cameras to just three across a breadth of sites through rural, commercial and residential areas. In several cases, bid teams ultimately delivered more CCTV than planned, primarily driven by underspend in other areas or requests from the community. CCTV was most frequently accompanied by street and other lighting and community engagement.

While CCTV supports operational intelligence by identifying offenders, it was primarily seen by bid teams as a key deterrent measure. As in SSF1, it was also highlighted that CCTV is often requested by the community, and inclusion in the programme helps to engage residents and businesses.

Factors that enabled effective implementation of CCTV
Bid teams were very positive about their implementation experience. This positive experience on SSF2 differed from SSF1 where some bid teams had highlighted key challenges owing to the complicated governance requirements. Besides the programme-level enablers, bid teams felt there was one key factor that also contributed to effective implementation of CCTV.

Previous experience in procurement and installation of CCTV was key to smooth delivery
Driven by the complex governance processes and technical knowledge requirements involved in CCTV implementation, it was felt that a working and practical understanding of CCTV technology and having experienced third-party contractors enabled both effective procurement and maximised intervention efficacy.

Barriers to effective delivery of CCTV
No specific barriers were reported for CCTV, besides those experienced at a programme level – procurement, supply chain and COVID-19-related resourcing shortages.

For information on the key steps in the CCTV implementation process, PCCs should refer to the ‘Process Implementation Checklist’^{[footnote 12]} created as part of the SSF2 evaluation.

3.3.2 Automatic Number Plate Recognition (ANPR)

ANPR was included in the initial bids of 16 areas and ultimately delivered in 14. Installation of ANPR cameras provided police with intelligence to aid the investigation of crimes, support the recovery of stolen cars and, to a lesser extent, discourage criminality by working as a visible deterrent. ANPR cameras were usually deployed on key routes into busy population centres, allowing law enforcement to track vehicles of interest. ANPR was generally part of a package of complementary measures, alongside street lighting and CCTV. Furthermore, ANPR was generally included in PCC-led bids, where existing contracts were in place for delivery. The scale of ANPR implementation varied considerably across bid areas, ranging from 2 to 20 cameras.

Factors that enabled effective implementation of ANPR delivery
Bid teams were mostly positive about their experience of ANPR implementation. Stakeholders highlighted multiple factors as key to effective delivery of the intervention.

A dedicated lead responsible for ANPR delivery was key to success
Several bid teams reported that implementation progressed smoothly partly due to a dedicated lead for ANPR in place. They felt it enabled a focus on delivery of this intervention, which is heavy on governance process requirements.

“Having a dedicated lead officer for delivery and accountability [facilitated ANPR delivery].”
(PCC, Residential and commercial)

Given the specialist nature of this intervention, bid teams emphasised that knowledge of ANPR technology was also vital to delivery
ANPR implementation knowledge, understanding of the technology and an insight into crime patterns in the local area were reported as critical factors for lead teams as well as delivery partners. Detailed understanding of how the technology worked contributed to effective decision-making on key matters such as camera locations, thus maximising effectiveness. For instance, a stakeholder in the case study analysis attributed the success in their area to the early inclusion of delivery partners with ANPR expertise and previous experience of the issues that arose in SSF1.

“Local knowledge of crime routes helped with placement of cameras. Experienced contractors [were] used [which facilitated delivery and intervention effectiveness].”
(PCC, Residential)

“I’ve had less stress over this ANPR implementation than anything else because the person who’s running it, he knows the system, he has a complete understanding of the data protection issues. In truth, it’s been a joy.”
(PCC, Residential)

Barriers to effective delivery of ANPR
No specific barriers were reported on ANPR, besides those experienced at a programme level – supply chain issues and COVID-19-related resourcing shortages.

For information on the key steps in the ANPR implementation process, PCCs should refer to the ‘Process Implementation Checklist’^{[footnote 12]} created as part of the SSF2 evaluation.

3.3.3 Street and other lighting

Street and other lighting was included in the bids of 34 areas and ultimately delivered in 27. The scale of delivery varied in scope, with teams installing both new lighting structures (for example, lighting columns and lanterns) and/or making improvements to existing ones (for example, replacing with brighter, more cost-effective lights or repositioning columns to improve coverage). New lighting, or repairs to existing infrastructure, was generally implemented behind shopping centres, near subways and bridges, and on main access routes, as well as among some residential premises.

Lighting was often included in a package of measures alongside CCTV and ANPR. By improving visibility, lighting was seen to increase the surveillance powers of these complementary interventions, aiding in the detection of criminals. Similarly, stakeholders also explained that lighting was a key intervention in deterring ASB.

Factors that enabled effective implementation of delivery of street and other lighting
Most stakeholders felt that implementation had been largely straightforward. Enablers of effective delivery included factors also highlighted in other interventions, such as collaborative working between the council and PCC/police, the use of experienced contractors and leveraging of existing contracts for the supply of relevant equipment.

Barriers to effective delivery

Coordinated delivery of complementary interventions was challenging and could be improved
A few bid teams reported poorly managed delivery of complementary interventions, such as CCTV. This was driven by teams responsible for these interventions failing to communicate and coordinate implementations, undermining the efficacy of the lighting plans.

Other barriers that teams reported in the context of street and other lighting were those experienced across other interventions – some minor delays and challenges with procurement, as well as some difficulties engaging the community and a lack of buy-in.

3.3.4 Alleygating

Alleygating was a relatively common intervention that was included in 19 bids and delivered in 14. Its selection may be explained by it being perceived as a ‘quick win’ by bid teams, along with other hard infrastructure interventions such as CCTV, which offer visual evidence of progress/action. Alleygating varied considerably in scale across bid areas, ranging from the installation of 41 new gates and refurbishment of 200 existing gates, to the installation of one gate and fencing along a single residential alleyway. There was also variation in equipment installed in bid areas – almost all bid areas installed gates designed for passageways to limit pedestrian access. Several areas incorporated other equipment into their bids, such as vehicle gates, fencing and roller shutters.

Factors that enabled effective delivery of alleygating

Successful community engagement was essential to effective implementation
Most bid teams reported successfully implementing alleygates once they had engaged the community. They felt that securing resident views and approval, legally required in some instances, enabled them to implement the intervention in ways that worked for the community.

Implementation was customised to suit the needs of businesses
One bid team reported that engaging businesses impacted by the alleygates enabled implementation to be customised to their needs. For example, the height of the gates required was greater than standard alleygates, so amends were made to the plan.

“We had face-to-face discussions with businesses impacted, to discuss whether businesses used alleygate for something operational. Having that communication allowed us to understand that they had deliveries, so instead of putting a vehicle gate…we went to extended pedestrian gate, which meant that they could still have a lorry or a vehicle come up the alleyway and easily deliver their supplies.”
(OPFCC, Commercial)

Barriers to effective delivery of alleygates

Lack of understanding of legal and governance requirements caused delays for some teams
Bid teams highlighted that alleygate installation required strict adherence to governance processes. In a few areas, where they had not anticipated these challenges, it had delayed delivery and sometimes even led to the dropping of the intervention. For example, Public Spaces Protection Order (PSPO) applications were required in several locations where the public had right of way. This proved to be a lengthy process, which many teams had not planned for. In one area, it meant that a gate could not be installed at the target location.

“[A] Location identified within the bid was identified to have public rights of way and therefore a PSPO would be required, which would have taken 12 months to complete.”
(Local authority, Residential)

Barriers experienced at the programme level, specifically supply chain issues and lack of buy-in from the community and from landowners who are the key stakeholders here, were also experienced by teams implementing alleygates.

3.3.5 Environmental improvements

Environmental interventions were implemented across rural, commercial and residential bids and included measures such as graffiti removal, general street cleaning, fixing/maintenance work, tree planting and installation of new street furniture. Thirty-four bids planned to implement environmental interventions, of which 21 finally implemented them. This variation, according to bid areas, was driven mostly by resident feedback, difficulty in securing permission from the local council or property owners, and delays in delivery on account of weather conditions or COVID-19 restrictions. Environmental interventions were generally in LA-led programme areas, with existing roles within the council. However, local community groups and stakeholders supported these interventions, which were generally included in bid design to drive the broader social outcomes, such as increased pride in the area and perceptions of safety.

Enablers of effective delivery of environmental improvements
Most teams reported a positive experience in the delivery of environmental interventions. Below are the key enablers of delivery for these interventions.

Delivery of environmental interventions was by and large smooth due to the bid teams’ ability to adapt implementation approaches when required
In some cases, bid teams reported having to adapt their original plans due to the reallocation of budgets or in response to community feedback. For example, in one bid area, initial plans involved structural improvements to fencing and alleyways but this was adapted in response to community requests for aesthetic improvements to the area, such as planting.

Support and engagement from local community groups was a key facilitator
Bid teams overwhelmingly highlighted the importance of partnerships with local networks such as local community centres, community officers, environmental community groups and council teams to deliver these interventions effectively. For example, one team reported that identifying key collaborative opportunities with a local housing organisation enabled them to successfully fix broken fencing and trim back overgrowth.

Barriers to effective delivery of environmental interventions

Poor weather in the winter months hampered delivery for some teams
By their nature, environmental interventions are impacted by poor weather conditions. For example, poor weather conditions were reported to have disrupted the application of anti-graffiti paint on street art installations in one bid area.

Programme-level barriers such as supply chain issues and lack of community buy-in also impacted the delivery of environmental interventions.

3.3.6 Video doorbells

Video doorbells, including Cube (a wireless security camera^{[footnote 13]}), were implemented in 9 SSF2 bid areas, with implementation frequently led by PCCs and occasionally supported by contracted partners.

Enablers of effective delivery of video doorbells
Teams reported relatively smooth delivery of video doorbells. Stakeholders highlighted 2 important enablers:

Using experienced contractors enabled quick and effective implementation
A few stakeholders felt that using experienced external contractors to install the doorbells was key to successful delivery, as it alleviated police/LA team pressures.

Legal agreements and data privacy forms used previously (on SSF1) simplified implementation in SSF2
Teams that had implemented video doorbells on SSF1 could leverage documentation that was developed then. For example, teams reported using the same supplier, which streamlined the purchase process. Similarly, GDPR forms (required to be signed by the resident receiving the doorbell) developed for SSF1 were used for SSF2 too, simplifying the process further.

Barriers to effective delivery of video doorbells
A few teams highlighted some challenges, besides supply chain issues experienced by other intervention delivery teams.

Concerns regarding privacy rules governing the use of video doorbells slowed delivery for many bid teams and, in some cases, led to adaptations in the bid plan
Recent rule changes on the use of video doorbells led to uncertainty amongst bid teams. GDPR and other privacy laws governing the legality of filming public spaces were seen as barriers, with some bids scaling back or adapting their implementation plans as a result. One area, for instance, opted to only distribute doorbells to those properties where it would capture private spaces, greatly reducing the number implemented (the underspend was diverted to CCTV cameras). In another area, the team chose to adapt their plan, as illustrated in the case study below.

Case study 4: Adapting plans to mitigate legal issues

To circumvent privacy concerns, one bid area adapted their plans by opting to implement another style of security camera, the Cube, in some locations. The Cube is a portable motion-activated camera which can be deployed within private spaces, such as sheds and garages, therefore avoiding the issues around filming public spaces associated with video doorbells. In this case, bid teams also worked with residents to agree responsibility and liability to mitigate risks to the delivery team if the Cube was used by residents in ways that contravened GDPR or similar laws. “They’re [criminals] going to have to do a lot more walking, scanning and everything else, before they can break in, with the risk of getting caught. It sends a live time photograph notification having been broken into.” (PCC, Residential)

Installation was a time and resource intensive process, especially for those not using external contractors
Since delivery of the equipment involved visits to each property, it sometimes necessitated multiple trips to a property before they could get the GDPR form signed and deliver the doorbell. In direct contrast, teams that had external contractors supporting implementation reported not experiencing any challenges.

Some bid teams experienced a lack of buy-in from some resident groups
Stakeholders felt that lack of trust in the police (or other authority figures) contributed to the lower take-up rates in some areas.

“In a very deprived area, where lot of residents don’t like the council or authority, (there was) lot of animosity from residents. When we first did the door knocks, we got so much animosity – you know, why are you interested now? We rang the police three years ago – that kind of stuff.”
(PCC, Residential)

For information on the key steps in the video doorbell implementation process, PCCs should refer to the ‘Process Implementation Checklist’^{[footnote 12]} created as part of the SSF2 evaluation.

3.3.7 Property marking

Twenty-five bid teams reported delivering property marking interventions. Property marking kits were distributed to residents through door-knocking initiatives or distributed at local community engagement events. SmartWater and Selecta DNA^{[footnote 14]} were the primary property marking kit brands used.

Enablers of effective delivery of property marking

Property marking kits are a familiar intervention to bid teams, and most reported delivery experience as straightforward
Furthermore, most bid teams were familiar with the kits (as also found in SSF1) and highlighted the relative ease with which they were obtained and distributed to residents. Unlike other interventions, teams did not experience any supply chain issues, as they often had standard and approved suppliers already in place.

The community engagement approach to distributing kits enabled teams to deliver broader crime prevention advice and other information to residents
Bid teams that distributed the kits via door-knocking initiatives could engage residents, brief them on how the kits work and also give them other crime prevention information and advice.

“Providing these kits free of charge encouraged residents to talk further to our team enabling us to offer crime prevention advice and signpost for further support where needed. The delivery team had the knowledge and the right personalities to engage with the community.’’
(PCC, Residential)

Training of delivery staff enhanced the effectiveness of the intervention and drove registrations of the kit
Several bid teams explained issues arose previously (on SSF1 and other similar programmes) due to inadequate understanding of property marking kits. This was mitigated in SSF2 by training delivery teams on how the kits worked. This helped them to share the correct information with the community, engage second-hand retailers to help identify stolen goods, and ensure kits were registered by residents on delivery.

Barriers to effective delivery of property marking

Tendency to not register kits impacted the effectiveness of the intervention
Some teams found that when the kits were delivered via community events and residents were not required to register them immediately, the kits were not registered. Police cannot identify stolen equipment without registration of marking equipment, making the intervention ineffective.

Besides the above, the barriers were similar to those experienced by teams implementing video doorbells – time consuming approach and some residents were unwilling to engage because of their distrust of authority figures.

As in the case with some other interventions that required residents to request/sign up (such as video doorbells and target hardening), the delivery of property marking was also impacted by a lower take-up rate than anticipated.

Case study 5: Property marking kit delivery

A bid area raised concerns that despite freely distributing property marking kits to residents, they were not being used, limiting the intervention’s possible impact in the local area. “We’ve handed them the kit, we’ve registered the kit for them, but we don’t necessarily know if they then mark all their property.” (PCC, Residential) Consequently, the PCC designed a door-to-door delivery plan that would mitigate this concern, maximising equipment take-up among residents and encouraging usage. Intervention dosage was high, with every house in the bid’s target areas offered property marking kits. The PCC delivery team opting to distribute the kit door-to-door was expected to increase take-up. This strategy also enabled officers to register marking equipment at the door to reduce burden on the resident and encourage use. “When you’ve got time to talk through, tell them it’s free, do the admin for them, take up is pretty good.” (PCC, Residential)

3.3.8 Target hardening (homes, businesses and vehicles)

Target hardening measures for homes, such as locks, window alarms and doors, were a popular intervention, implemented in 34 bid areas. Target hardening of vehicles was carried out in 16 areas, while target hardening of commercial properties was carried out in only 10 areas, with measures ranging from providing equipment, such as burglary packs, alarms and dummy cameras, to information and guidance (for example, advising a business on how to target harden premises through redesign).

Enablers of effective delivery of target hardening

Positive engagement with business and resident communities, in most cases, enabled effective delivery
Engaging businesses at an early stage and making use of established links and local networks encouraged take-up. Bid teams frequently reported utilising local facilities, such as leisure centres, for the collection of equipment and to hold community meetings where bespoke advice could be provided to residents. They felt this increased attendance and overall buy-in from the community.

Stakeholders also felt that having the right people on the team was essential for getting target groups to take-up the interventions.

“A dedicated Burglary Reduction officer (BRO) PCSO performed this role. The person was extremely good at this role and much more effective in securing the support from residents to have target hardening work undertaken.’’
(PCC, Residential)

“An enabler was definitely the appointment of coordinators. The fact that they were not police made a big difference to the trust they managed to gain from dealing with the community in the colliery. Their Safer Streets branded hoodies increased visibility and engagement.’’
(PCC, Residential and Commercial)

Barriers to effective delivery of target hardening

Legal processes were the only barrier identified for target hardening of vehicles
A few teams reported that legal processes involved in commissioning local garages to undertake work took longer than anticipated, causing some delays.

3.3.9 Community engagement

Thirty-seven bid teams implemented initiatives aimed at engaging the community. The delivery of community engagement activities was largely smooth and well received by local businesses and residents.

Community engagement activities varied widely across bid teams, with the most common activities including local resident and business surveys. Other activities included public events, community days of action, Neighbourhood Watch schemes, Community Champion initiatives, door-knocking initiatives, organised litter picks and employing community wardens. A few bid teams rolled out bespoke apps to deliver public safety messages, crime reduction reminders and alert messages.

Enablers of effective delivery of community engagement
Most bid teams reported that activities to engage residents and businesses in the community were delivered successfully and to schedule. In many cases, bid teams reported delivering on a larger scale than planned in response to positive community feedback.

Tapping into networks and organisations with established relationships with the community is a key enabler of successful engagement
Bid teams reported that engagement interventions were heavily facilitated by good working relationships with local organisations and community groups that could step in when there was distrust in the community of authority figures.

“Having external facilitators helped where there were tensions with statutory agencies.” (Local authority, Residential)

Feedback from teams engaging businesses also reflected this. Engagement initiatives were carried out through the local Business Improvement District or the Business Crime Reduction Partnership teams, which already had relationships with the businesses in the area.

Barriers to effective delivery of community engagement
COVID-19, poor weather in the winter and lower levels of trust amongst some resident groups were highlighted as key barriers to community engagement initiatives.

A few audience groups were identified as being more difficult to engage, such as students, landlords, residents who first language was not English. This was driven by lower levels of motivation to engage in the programme (specifically for landlords) and lower trust levels in authority. Some teams reported that landlords were often more difficult to contact as they were not at the properties in the target areas.

While most teams did not report any barriers to engaging businesses, one team noted they had found the larger business organisations more difficult to engage as they had established, national policies in place which sometimes made it difficult to deliver on some security advice given by the SSF teams.

3.4 Impact of interventions and early successes

This section summarises the approach that bid teams have taken to measure impact and any early impacts reported at an intervention level.

3.4.1 Impact of interventions

Approach to monitoring programme outcomes and impact

While most teams felt it was too early to report on the impacts of the programme, plans were in place to monitor outcomes and impacts.

Formal crime and ASB statistics were cited by most teams as the main form of evidence to measure the impact of SSF2 interventions. Aside from this, teams reported plans to collect resident and community feedback through surveys, as well as feedback from local business groups. Monitoring reports (for example, CCTV monitoring reports and incident logs) were another source of information that teams reported planning to use to establish the impact of the intervention.

Furthermore, many bid teams also reported early successes based on community surveys, observations from team members and qualitative feedback from residents/businesses.

3.4.2 Intervention-specific early successes

Captured below are some of the early successes that teams reported the programme had led to in their areas.

CCTV: Multiple teams reported early successes with CCTV interventions directly leading to crime prevention (for example, a reduction in burglary, graffiti, ASB), offender identification and arrests (as well as fixed penalty notices or warnings). Several bid teams also reported increased public perception of safety, evidenced by the reduced calls for service, their own post-programme surveys, and an observed increase in intended use of public spaces.

“79% of residents surveyed felt that the installation of additional CCTV was a good use of resources and 64% said it made them feel safer.”
(PCC, Residential)

“Reduced number of calls for service in each area [where] cameras have been installed. More people are using the park area as intended.”
(PCC, Commercial)

ANPR: Compared to other interventions, early successes for ANPR were the least frequently reported by bid teams. However, one interviewed stakeholder reported that data gathered through an operational ANPR camera had enabled police to identify and recover a stolen car.

Street and other lighting: Bid teams reported early indication of positive impacts of street and other lighting evidenced through resident survey results and anecdotal evidence. Bid teams reported that initial community responses outlined heightened feelings of safety in areas with improved lighting.

“87% of residents surveyed feel that improvements to street lighting is a good use of resources; 75% agree that this makes them feel safer.”
(PPC, Residential and Commercial)

“[There is] increased public confidence in routes which are now covered by enhance lighting and feeling of safety.”
(PCC, Residential and Commercial)

Alleygates: Most bid teams reported alleygates elicited an overwhelmingly positive community response, making residents feel safer. Resident feedback also suggested that alleygates had improved the local areas visually.

“The new gate that was fitted has completely changed the area – the road is cleaner, the residents feel safer, and we are once again a united community.”
(PCC, Residential)

Environmental improvements: Bid teams reported that the local community positively received these interventions. Anecdotal feedback tended to be overwhelmingly positive, with residents commenting that the streets looked cleaner. Multiple teams felt that the interventions had created a greater sense of pride within the community.

“It’s brought the community together; they’ve got a little bit of pride in their local community.”
(Local authority, Residential)

“Because people have seen some landscaping work going on, it’s encouraged local organisations to say oh that’s a good idea, let’s start doing some spring plantings.”
(Local authority, Residential)

Video doorbells: Bid teams felt that video doorbells had increased perceptions of safety in the community. Some also reported that video doorbells were helping to reduce ASB in the local community.

Property marking: Most teams expected property marking to drive a reduction in acquisitive crime. A few bid teams that delivered bike marking reported a reduction in bicycle thefts. Other bid teams that distributed property marking kits reported improved police-community engagement and a few bid teams stated the kits had led to increased prosecutions for thefts. Finally, several bid teams reported residents had confidence in the efficacy of the kits, driven by their understanding of the technology.

Target hardening homes: Anecdotal and local survey statistics in some areas highlighted that residents felt safer in their homes.

“Two gentlemen came and fixed the locks on our front and back doors yesterday afternoon. They were extremely professional and addressed the issues…Please share this feedback, great locksmiths. Myself and my husband feel a lot more secure now.”
(Local resident)

“I am so happy to benefit from this project and be able to sleep in the dark again, knowing I have a CCTV camera and window security locks. It will make me feel safer and I know it will be the same for my kids too.”
(Local resident)

Community engagement: Overwhelmingly, engagement activities were reported to have successfully contributed to community support for SSF2 interventions and the programme more widely.

“People feel reassured that funding is going into their area, they feel reassured that they can see the outcome of that.”
(PCC, Residential)

“[it was a] great success based on the feedback from all the partners and community.”
(PCC, Residential)

Business engagement: Feedback from businesses was highly positive. Many bid teams reported that businesses in their area felt listened to, were satisfied that their concerns were being addressed and that interventions were being delivered in a timely manner. Bid teams also felt the programme had contributed to new, or improved, working relationships between local businesses and delivery teams, which increased engagement in local business crime partnerships and Business Improvement Districts.

“Retailers who we have engaged through the Town Safe radio scheme commented that they appreciated support from the council and wanted to see more schemes like this develop, especially after COVID-19 impacts.”
(PCC, Residential and Commercial)

3.5 Sustainability of interventions

This section outlines the approaches that bid teams had taken, or planned to take, to ensure the sustainability of interventions beyond the funding period and the risks to sustainability in the longer term.

Most SSF2 bid teams reported having plans in place to continue funding most of their interventions in the longer term. Most bid teams had built intervention maintenance costs into future budgets. They reported that these funding plans would work to prevent equipment from falling into disrepair and ensure that interventions continue to be delivered successfully.

In some cases, the LA had agreed to fund maintenance costs of interventions for a minimum of 3 years

In most bid areas, the LA agreed to fund repair and maintenance costs. Interventions financially supported by LAs and PCCs in the longer term tended to be those with hard infrastructure situated on public property, including CCTV and ANPR.

Bid teams considered the sustainability of interventions by procuring durable intervention equipment

Bid teams reported considering the sustainability of intervention equipment at the planning stage and much of the equipment procured was said to be durable, resistant to damage through vandalism or use, and expected to last beyond the funding period.

There were also several cases where responsibility to maintain and fund intervention equipment rested with local residents

Interventions that involved residents directly, and those installed on privately owned land or property such as video doorbells and target hardening, were identified as being at risk in the longer term, given that ongoing repair, maintenance and replacement of the equipment relied on continued resident buy-in.

Future funding cuts were another commonly identified risk

Bid teams suggested that local authority spending cuts or reallocation of budgets could prevent the long-term delivery of interventions and maintenance of equipment, leaving it in disrepair. Bid teams also suggested that budget cuts put short-term schemes at risk, such as environmental improvement drives and community engagement events.

The experience of programme and intervention implementation covered in this section has led to multiple learnings amongst delivery teams. Verian have distilled these learnings into recommendations in section 5.

4. Impact evaluation findings

The impact evaluation findings presented in this section are based on the analysis of police recorded crime (PRC) data and the comparison of residents’ survey responses at 2 time points (before SSF2 was implemented and towards the end of SSF2 implementation). Further technical detail about the impact measurement approach can be found in section 2.2 and in Appendix C. In addition, Appendices D and E provide the detailed bid-level PRC data analysis results and DiD analysis results for each variable in the resident survey, respectively.

Overall, there is evidence of a small-to-moderate statistically significant positive impact of SSF2 on several measures. In summary:

(i) The survey-based analysis detected statistically significant changes on multiple measures. Specifically, improved understanding of property crime prevention measures (+10pp) and increased use of specific crime prevention measures (such as, property marking, +3pp), as well as reductions in the proportion of residents reporting having experienced or witnessed shoplifting (-6pp) and robbery (-2pp) in SSF2 bid areas (versus comparison groups). There were multiple variables for which no statistically significant impact was detected; a summary is included under each relevant subsection below.

(ii) While there was no statistically significant impact at the overall Fund-level on PRC incidents. Verian also conducted additional exploratory analysis by segmenting the bid-level impact estimates; the team did not find any statistically significant differences in impact by type of area (residential versus commercial), urban versus rural areas, or recent crime trajectory.

There were also some unexpected negative impact analysis findings. In terms of PRC data, there was an increase in vehicle crime (in the 7 bid areas that targeted vehicle crime). In the survey-based analysis, a greater proportion of residents reported that acquisitive crime (+9pp) and anti-social behaviour (ASB)^{[footnote 15]} are a problem in their local area (+11pp), and more residents reported experiencing or witnessing ASB (+9pp; particularly relating to the use or dealing of drugs (+8pp). One possible explanation for this is that the Safer Streets programme and related communications led to increased awareness of these issues among residents.

To better understand these findings, Verian further looked into the process evaluation data. While the team could not identify any conclusive explanations, they note that several impact measurements (see C2.1.1 Limitations in Appendix C for more details) and implementation-related factors (see section 3.2.2) for an overview of the Fund-level challenges to delivery that influenced whether the SSF2 evaluation could detect significant impacts within the timeframe and budgetary parameters.

These should be taken into consideration when interpreting the impact findings in the below sections. As a note, Verian cannot rank or put specific weights on any of these influencing factors and it is likely that they all have a bearing synergistically on how impact was being detected (or not).

Intervention dosage

Regarding the intervention dosage (the number (or magnitude, or intensity) of interventions implemented or distributed to residents) and take-up by residents varied across bid areas. Where interventions landed at a relatively small scale, it might have caused the impact of SSF2 to be diluted, which reduced the ability of this evaluation to observe major trends in the impact data. For example, CCTV implementation ranged from the installation of 70 cameras in one bid area to only 3 in another bid area. It is very hard for impact to be detected in the survey-based analysis of overall resident perceptions if only a small number of residents noticed the interventions in localised spaces, or if only a few received the interventions. In addition, even when residents were offered interventions directly (for example, video doorbells), several teams reported that some residents from hard-to-reach groups, such as students, were reticent to accept and use them.

Impact measurement

Regarding impact measurement, it was likely too early to measure the impact of the interventions given the timescales of SSF2 intervention implementation and evaluation timings. Specifically, implementing certain interventions was time intensive in some bid areas, particularly for more complex interventions such as ANPR. Some bid areas also experienced implementation delays due to difficulties in procuring the interventions and mobilising for the planned activities, such as due to global supply shortages in equipment (for example, video doorbells). Because of these factors, implementation was still ongoing in several bid areas while the follow-up survey data was being collected or completed shortly before this. This reduces the likelihood of the interventions having a noticeable and measurable impact, in terms of residents’ perception of crime and crime incidents. It is important to note that the additional SSF1 PRC data analysis with a longer reference period (12 months after interventions were delivered) did not find any significant impact of SSF1 on PRC incidents^{[footnote 16]}. Lastly, it is possible that some broader national trends (for example, socio-economic changes driving up acquisitive crime and the relaxation of COVID-19 restrictions providing criminals more opportunities for theft and burglary) and very localised events (for example, the arrest or release of a prolific local criminal) might have impacted PRC incidents.

The following sub-sections outline the impact findings from the SSF2 in more detail:

• 4.1: impact on PRC incidents
• 4.2: residents’ experiences and perceptions of crime and ASB
• 4.3: awareness and adoption of crime prevention measures
• 4.4: residents’ recollection of communications

Alongside the findings from the SSF2 evaluation, the sub-sections also describe the findings from the SSF1 PRC data analysis with a 12-month reference period and combined SSF1 and SSF2 survey responses.

4.1 Impact on Police Recorded Crime incidents

To estimate the impact of SSF2 interventions on numbers of crime incidents, police recorded crime (PRC) data from April to June 2022 was used to compare SSF2 areas with matched non-intervention areas. Specifically, for each bid area, Verian looked at the types of crime which were to be targeted as specified in the bids. In some cases, overall neighbourhood crime was targeted whereas in others one or more specific neighbourhood crime types were focused on (domestic burglary, robbery, theft from the person and/or vehicle crime).

At the Fund-level, the evidence suggests that SSF2 did not have a statistically significant reduction on PRC; the impact estimate was close to zero (see table 1 below). This applies to the analysis of impact on crime overall, as well as additional analysis breaking down the PRC data by crime type, type of area, urban/rural classification and recent crime trajectory. However, one exception was an unexpected finding that the number of vehicle crimes recorded was higher in SSF2 areas than would be expected without the Fund (58 more vehicle crimes across all SSF2 areas targeting this crime type in the 3-month observed period). Generally, while it was not statistically significant, the impact estimates for burglary and theft from person were moving in the right direction in terms of a reduction in crime incidents. For neighbourhood crime and robbery, the trend is indicating that there in a non-statistically significant increase in crimes.

Table 1: Fund-level impact estimates by crime type (April to June 2022)^{[footnote 17]}

Crime type	No. of SSF2 LSOAs	No. of crimes recorded across all SSF2 LSOAs	Estimated counter-factual	Impact estimate	Lower CI*	Upper CI	p-value
Neighbourhood crime (overall)	73	841	825	16	-131	163	0.830
Burglary	106	525	541	-16	-89	57	0.666
Robbery	90	156	138	18	-21	57	0.358
Theft from the person	78	138	168	-30	-65	5	0.096
Vehicle crime	89	442	384	58*	8	108	0.022

That said, at the bid-level, the impact analysis identified six local areas with a significant reduction in police recorded crime (than would have been expected without the Fund) and three areas where there was an increase in PRC incidents (See Appendix D for more detail).

There are a variety of possible explanations why there were statistically significant increases in PRC, which were discussed with academic experts. These range from factors directly related to the nature of the SSF2 interventions, their practical implementation, and the specific outcomes that are captured by this evaluation. For instance, in terms of SSF2 intervention aims, the process evaluation evidence shows that some interventions were designed to make crime more difficult (for example, target hardening equipment for homes), while others focused on increasing the risk of getting caught, or to help police detect and identify offenders (for example, ANPR). Thus, it is possible that more crimes were being detected and recorded than would have been without the SSF2 interventions. It is also likely that different types of interventions may have impacted on individual beneficiaries more so than achieving an overall impact across an intervention area, while other interventions could possibly have displaced some of the criminal activity into adjacent streets, within or outside the defined bid-level area. The team explored all these contextual and intervention-specific factors (excluding analysis of crime displacement), but it was not possible to draw out any firm conclusions from the available evidence.

Further, thinking about SSF2 implementation, the intervention dosage (that is, quantity of interventions implemented or given out to residents) was low in some areas and generally varied. Where the intervention dosage was small, it might have caused the impact of SSF2 to be diluted, which reduced the ability of this evaluation to observe major trends in the impact data. Also, some interventions took the full delivery period to be fully implemented. As a result, the evaluation may not have been able to capture their impact yet, due to the local project delivery timescales. As noted above, Verian also conducted exploratory analysis to understand whether there are any differences between bids targeting specific areas (rural versus urban, and commercial versus residential) or with different crime trajectories in PRC data prior to SSF2. There is not enough evidence to conclude that different levels of impact are driven by these area-level characteristics, given that the median impact estimates (and most of the individual impact estimates) are close to zero in each case.

In addition, several broader national trends (for example, socio-economic changes driving up acquisitive crime and the relaxation of COVID-19 restrictions offering criminals more opportunities for theft and burglary), as well as very localised events (for example, the arrest or release of a prolific local criminal, or increased gang activity in an area) are likely to have affected PRC trends over a short period, independently of any crime prevention measures.

4.1.1 Additional SSF1 PRC data analysis and SSF1 PRC data analysis with a 12-month reference period evaluation

This impact analysis of PRC data for SSF2 broadly mirrors the earlier findings that were reported in our evaluation of SSF1. The main SSF1 analysis was conducted using PRC data covering the first quarter (quarter 1 from April to June 2021) and it did not detect any statistically significant impact of SSF1 on PRC at the overall Fund-level. As with the SSF2 evaluation, the SSF1 evaluation 3-month data reference period also overlapped with some areas, which were still implementing their interventions – the degree of overlap with implementation activities was to a greater extent with the SSF1 compared to the SSF2 evaluation.

Verian subsequently carried out further analysis of PRC data relating to SSF1, this time using a longer data reference period (12 months between April 2021 and March 2022). They did this to explore the possibility that it may have taken some more time for impacts of SSF1 interventions to be reflected in the PRC data. Additionally, across many crime types, fewer crimes were recorded during the COVID-19 pandemic. This could have made it difficult to detect any impacts during the initial 3-month reference period. However, the impact estimates of this additional analysis were close to zero at the Fund-level, indicating that there is no significant evidence to support the hypothesis that SSF1 led to a substantial reduction in neighbourhood crime. Similarly, at bid-area level, the analysis suggested that there generally were no substantial reductions in the number of recorded crimes that could be attributed to SSF1. However, it is important to note that it is possible that a reference period longer than 12 months might be needed to detect an impact (where there is any); it is not possible to confidently say exactly how long it would take for these interventions to impact on PRC.

4.2 Residents’ experiences and perceptions of crime and ASB

The survey-based analysis detected some positive findings in terms of the impact of SSF2 on residents’ experiences of specific types of acquisitive crime, alongside some unexpected findings about crime and ASB perceptions and reported experiences. Specifically, there was a statistically significant reduction in the proportion of residents reporting that they had experienced or witnessed shoplifting (-6pp) and robbery (-2pp) in the last month, which could be attributed to SSF2 interventions.

However, contrary to expectations, there was a statistically significant increase identified in the proportion of residents who say that acquisitive crime is a problem in their area (+9 pp when compared to non-SSF2 areas). This aligned with a similar finding about ASB in terms of increases in the proportion of residents who say that ASB is a problem in their area (+11pp) and that they had experienced or witnessed ASB in their local area (+9pp). In particular, the proportion of residents reporting that they had witnessed people using or dealing drugs increased in the SSF2 intervention period (+8pp).

Based on the process evaluation findings, one possible explanation for these unexpected findings is that the communications about the SSF2 interventions increased residents’ awareness and perceptions about the prevalence of acquisitive crime and ASB in their local area. For example, some SSF2 areas used social media communications and community events to alert residents to the threat of burglary and other acquisitive crimes, potentially making residents more sensitive to signs of or the problem in general. Similarly, some delivery teams reported that interventions targeting ASB were often designed to improve visibility, for instance, street lighting and environmental improvements (for example, shrub clearance) in hotspots to improve surveillance and deter ASB activity. However, by improving visibility of locations where ASB is likely to occur, these interventions may have contributed to increased resident awareness of it in the short term.

Compared to SSF1, there were several outcome variables for which Verian found no statistically significant impact in SSF2, but for which significant change was detected in the SSF1 impact evaluation:

1. In SSF1 there was a significant increase in the proportion of residents saying that the police were doing a good or excellent job in their local area (+7pp), whereas no change was found in SSF2.
2. In SSF1, there was a significant decline in the proportion of residents reporting worry about being mugged or robbed (-9pp), whereas no change was found in SSF2.
3. The proportion of residents saying they had experienced begging, vagrancy or problems with the homeless in the last month saw a decrease in the SSF1 analysis (-6pp) – a finding that is likely to be an anomaly, reflecting government initiatives to reduce homelessness during the COVID-19 pandemic – whereas no change was found in SSF2.

Verian found no significant impact in either SSF1 or SSF2 in:

• residents’ perceptions about reported experiences of other types of acquisitive crime (apart from robbery and shoplifting)
• residents’ feelings of safety when walking alone in their area after dark
• impact of crime on residents’ quality of life in the last month (on a scale from 0 to 10)
• residents’ worry about theft of/from their cars
• residents’ perceptions of how pleasant or safe they considered their local area

Verian combined the SSF1 and SSF2 resident survey data to conduct a pooled analysis (for variables included in both surveys). They did this to exploit the potentially higher statistical power of a larger survey sample size. This additional impact analysis mostly aligns with the patterns reported so far. There were no further or additional impacts to report on residents’ experiences and perceptions of crime and ASB.

4.3 Awareness and adoption of crime prevention measures

The survey-based analysis found that residents’ understanding, and adoption of crime prevention measures were positively impacted by SSF2 interventions.

Specifically, SSF2 had a statistically significant impact on residents’ reporting that they understand the measures they can take to prevent crime on their property (+10pp compared to non-SSF2 areas). Also, a significant increase in the proportion of people reporting using property marking equipment (+3pp) can be attributed to SSF2.

The process evaluation suggests 2 possible interconnected explanations for these findings: intervention quantity (such as, dosage); and how security measures were distributed to residents. Specifically, the positive impact detected around the use of property marking equipment may be explained by its high dosage. Given the intervention’s low-cost, many SSF2 bid areas distributed equipment widely across the local community. As a result, many people would have received this intervention, which made its use more detectable in the survey-based analysis.

The process evaluation findings suggest some bid teams’ approaches to distributing target hardening and property marking equipment enabled more intensive engagement with residents. This included handing out equipment at community engagement events and distributing door-to-door. This gave delivery staff the opportunity to explain how to implement and use these interventions, thereby building residents’ understanding of crime prevention measures to protect their property (although correct use of the equipment could not be guaranteed). This explanation is reinforced by the finding from the follow-up survey that 31% of residents who had heard about what the police and council were doing to tackle crime in the local area, recalled information about home security measures (see section 4.4).

4.3.1 Comparing and pooling SSF1 and SSF2 findings

Comparing the SSF1 and SSF2 survey-based findings, there was a statistically significant increase in residents’ awareness of the Safer Streets project in their local area in the SSF1 evaluation (+5pp), but not in the SSF2 evaluation.

Analysing the pooled SSF1 and SSF2 survey data (covering variables surveyed in both rounds), a statistically significant impact was found for the:

• increase in the proportion of residents who say they understand the measures they can take to prevent crime on their property (+7pp)
• increase in awareness of some crime prevention activities, including partnerships to prevent crime (+2pp); the Safer Streets project as a whole (+4pp); and activities to improve public spaces, such as building car parks (+3pp)
• increase across bid areas in using double-glazed windows (+7pp), which was not an SSF-funded intervention so unlikely to be attributed to SSF2

Across both SSF1 and SSF2, there was no significant impact identified for:

• use of home or vehicle security measures
• awareness of community engagement measures (for example, Neighbourhood Watch)
• awareness of environmental improvement measures in public spaces

4.4 Residents’ recollection of communications

In the SSF2 follow-up survey, residents were asked to recall information about the actions that police and the local council had taken to tackle crime and ASB in their local area since August 2021. This section presents only the descriptive findings; no survey-based impact analysis to assess statistical significance was conducted because these questions were not included in the pre-SSF2 baseline survey. Please note that residents could select multiple options for the questions about the types of information and mode of receiving information; this means that response percentages do not add up to 100%, as with single-coded survey questions. The key findings are:

Awareness of crime prevention related information: Nearly a quarter of residents (24%) surveyed in areas receiving SSF2 funding reported that they had seen or heard information about what the police and local council were doing to tackle crime and ASB in their local area. This is lower compared to SSF1 in which almost a third of residents (32%) reported that they had seen or heard this information.

Types of information: Of the residents who were aware of this information, they had most commonly seen or heard about the protective measures in their local area (36%), home security measures (31%), and community engagement (22%). It’s important to note that 31% reported that they did not know which types of information they had seen or heard. Resident recall of information on crime prevention was broadly consistent across both SSF1 and SSF2.

Mode of receiving information: Residents who had seen or heard information reported that this was via Facebook (24%), in newsletters (20%), and in local newspapers (14%). This differed from SSF1 findings where newsletters (29%) were the most commonly reported engagement channel. Awareness of information through social media was broadly consistent across SSF1 and SSF2.

5. Summary of key findings and recommendations

This section provides a summative assessment of the evidence about the impact of SSF2 on neighbourhood crime and other key measures, as well as the learnings about intervention implementation.

Overall, the process evaluation found ample confirmation from local implementation teams that SSF2 was seen as a successful programme as a whole. Local delivery stakeholders generally wished for more time for their bid development and mobilisation efforts, a few struggled with implementation delays, and they had a variety of constructive suggestions about how the grant administration process could be improved. That said, on balance most areas were positive about the way the Home Office engaged local areas and about the difference their interventions were making to help reduce neighbourhood crime in their communities.

Thinking about the impact evaluation of SSF2 interventions, the evidence is fairly mixed and modest, very much in line with our earlier SSF1 evaluation (see section 3: Impact evaluation findings). Overall, there are indications that a number of the stated programme outcomes are moving in the right direction.

Overall, there is evidence of statistically significant^{[footnote 18]} positive impact of SSF2 on several measures. The survey-based analysis detected improved overall understanding of crime prevention measures (+10pp) and increased use of property marking equipment (+3pp). There were also reductions in the proportion of residents reporting having experienced or witnessed shoplifting (-6pp) and robbery (-2pp) in SSF2 bid areas (versus comparison groups). However, there were some unexpected outcomes, such as more people reporting having experienced or witnessed ASB (+9pp) and viewing ASB (+11pp) and acquisitive crime as a problem in their area (+9pp).

There were no statistically significant reductions detected from the PRC data analysis at the overall Fund level, which does not support the hypothesis that SSF2 has led to a substantial reduction in crime in the intervention areas.

At the same time, there were also some unexpected negative findings in the impact analysis, which showed a statistically significant increase in vehicle crime at the Fund level (in all those bid areas that indicated targeting this crime). In the survey-based analysis, a greater proportion of residents reported that acquisitive crime (+9pp) and ASB (+11pp) are a problem in their local area, and more residents reported experiencing or witnessing ASB (+9pp).

It is important to interpret these results in the context of several factors that can enable or inhibit the ability of a programme evaluation to detect any major impact. These typically include the contours of the original policy design, the dosage of interventions, implementation fidelity, and the methodological design of the impact measurement approach. Where a programme seeks to implement a broad range of relatively small (in terms of intervention dosage) and very localised interventions – as is the case here with SSF2 – it is generally difficult for an evaluation to observe a consistent trend of significant impacts.

We discuss the key implications of these findings in the following sections, which are relevant for both the Home Office in terms of future commissioning approaches, and local SSF teams who may be designing and delivering their interventions.

5.1 Recommendations for local implementation teams

5.1.1 Bid development

At the bid design stage, consultations with implementation teams of previous rounds are key to draw learnings and insights from their implementation experience. This is a key factor to help shape the choice of interventions and the development of a realistic time plan. It will also help teams to identify potential barriers and proactively plan the programme accordingly.

Early engagement of delivery partners and stakeholders at the bid development stage is important to ensure prioritisation of the programme across organisations, paving the way for smooth delivery. This should include engagement with residents and businesses in the community.

Adequate consideration of communication activity (through mainstream media) as well as direct engagement of the community (through events, door-knocking initiatives etc) must be a part of the delivery design to ensure take-up of interventions and overall engagement with the programme.

Building strong insight into the target area – physical structures and spaces, community/local population – is important and needs to guide bid development. For example, it is essential to know the physical state of properties in the target area to plan target hardening interventions. It is equally important to gain an insight into the community needs and how they feel in the area when planning interventions.

5.1.2 Programme set-up

There is value in building an internal, local bank of knowledge on the governance processes and legal requirements associated with interventions, specifically CCTV, ANPR, alleygates and video doorbells, early in the bid development phase. This will drive awareness and advance preparation for delivery. As part of this evaluation, Verian developed a Process Implementation Checklist to initiate a bank of best practice implementation processes for specific interventions. Learnings and insight from local teams would be of great value to build on and further develop this checklist.

It has been seen that prior knowledge and experience (of the intervention) is a factor driving successful delivery. This is an important consideration at bid design phase when identifying roles and responsibilities. Where possible, it is recommended that some continuity be maintained in delivery teams and partners across different rounds.

Where possible, it is recommended that delivery teams plan for a project manager dedicated to managing the delivery and tracking of this programme.

5.1.3 Programme delivery

Community and business engagement must start early on to build awareness of the programme and drive buy-in to the interventions. This is especially critical for interventions that depend on resident/business take-up for their success (for example, video doorbells, target hardening, alleygates).

Specific hard-to-reach targets groups, such as landlords and students, have been flagged as more difficult to engage in the SSF programmes so far. Thus, if they are part of the target group in the area, then it is important to build customised plans for engagement. Rather than directly engaging with them, reach out through networks associated with them. For example, for students, it may be more effective to communicate and engage with them through the university rather than as residents of a particular area. For landlords, identify associations or local authority teams that may have an established connection with them.

The above is also true for some audience groups that may have lower levels of trust in authority figures and may not be willing to engage with them. In these cases, it is important to establish these groups and then identify the ideal way to reach them. Alternative approaches such as reaching out through community leaders, local charities, religious organisations or schools can be explored. This is especially critical for interventions that may require a community engagement approach for implementation, such as video doorbells, target hardening, property marking.

5.2 Recommendations for the Home Office

5.2.1 Programme commissioning

Reviewing the time accorded to teams for bid development is an important consideration for future rounds. Ideally, this can be done through engaging local area bid teams and other relevant government stakeholders to understand the ideal development window that would enable teams to develop an effective bid, while also meeting government-specific timelines.

Reviewing the programme delivery period is also a key consideration for future rounds, based on feedback received from local teams. There is an opportunity to consider further flexibility or extension of timescales to enable effective delivery.

Driven by the requests in round 1 for additional guidance on specific elements of bid development, there is an opportunity to include the following in future Safer Street toolkit editions:

• SSF toolkits to incorporate more case studies to support the planning of interventions and highlighting synergies between different interventions
• guidance to support the completion of EVAs
• guidance on how to run community surveys
• guidance on implementing community engagement initiatives
• guidance on regulatory and legal compliance (related to specific interventions, for example, GDPR)

It is important for the Home Office to explore further (with bid teams) the need expressed by some for guidance on the balance between the different types of interventions (some stakeholders in SSF2 felt the guidance for SSF2 favoured the ‘harder’ physical interventions).

There is an opportunity for the Home Office to understand (from bid teams and other intervention experts) the average time required for implementation of different steps involved in the key interventions such as CCTV, ANPR, video doorbells, property marking or target hardening. This can then feed into the guidance on bid development that implementation teams can leverage for their bid design.

5.2.2 Supporting programme delivery

On the back of the SSF1 evaluation, the Home Office provided a number of bespoke resources to SSF3 bid areas, including a communications toolkit and implementation guidance.

Nevertheless, there was an absence of knowledge and information sharing between bid teams. Given the level of insight and experience being built within teams over subsequent rounds of the SSF programme, there is a strong opportunity for the Home Office to create and manage a platform for knowledge sharing across teams^{[footnote 19]}. Exchange of information on experiences, proven ways of resolving challenges and an opportunity to collaborate on innovative approaches would be valuable to both the bid development and implementation phases.

While most teams that had implemented interventions in commercial areas felt their programmes had gone smoothly, no significant differences were observed in the types of interventions targeting commercial and residential areas. Thus, there is an opportunity for the Home Office to provide guidance and evidence in the SSF toolkits on interventions customised for commercial areas/establishments and business audience groups.

The approach taken to project management varied across areas and depended largely on the experience/knowledge of the allocated project manager. Building in best practice approaches, tools and advice on project management into the SSF guidance/toolkits will enable a more consistent and efficient approach to project management.

5.3 Considerations for future SSF evaluations

Consider having a longer scoping phase during which feasibility of various impact evaluation options can be thoroughly assessed. This requires appropriate time for the evaluators to reach out to bid teams and confirm key details about locations, audience, timeline, types and quantities of interventions delivered (as these might have changed since the bid was submitted).

Consider enhancing the available management information data from bid areas by collecting it at a few time points (depending on the length of the delivery period) and making this a requirement as a condition of receiving funding. For instance, this could include collecting data about the number of interventions implemented/delivered, timings and spend for each type of intervention to enable accurate assessment of the dosage and take-up of each intervention.

For the evaluation to capture robust evidence about specific types of interventions (as opposed to aggregated Fund-level impacts), the area eligibility and bid selection criteria could be more targeted. This would be required to ensure the quantity of interventions (such as, dosage and intensity) implemented or delivered is significant enough to make a large enough difference for residents to notice interventions and/or to be reflected in official PRC data. An increase in the number of programme beneficiaries (for example, residents in receipt of any SSF intervention) could also help to increase the statistical power for detecting changes. This might also be enabled through the delivery of a smaller number of select pre-specified interventions in each area; the consistency in the types and quantities of interventions would allow for more robust comparison given some level of standardisation in terms of what interventions are being evaluated.

Develop a theory of change for the Fund and its interventions to identify the short- and medium-term outcomes, in addition to the anticipated longer-term impact of reducing crime and changing perceptions of crime. While it can take a long time to see change in the impact measures (and various external factors can influence this), any changes in short- and long-term outcomes should be noticeable sooner. Changes in outcomes would indicate whether the interventions are moving in the right direction towards achieving the anticipated impact^{[footnote 20]}.

Verian also re-iterate several general points that were presented in the conclusion section of the published SSF1 evaluation report, which generally reflect best practice in programme evaluation and will increase the likelihood of detecting statistically significant effects of interventions. Added to this, the team feel there is value in sharing the evaluation data with the wider research community, for example, by putting anonymised data sets into the public domain for other researchers or scholars to potentially conduct further analysis in the future.

Appendix A: Crime breakdowns

Table 2: Breakdown of Safer Streets Fund round 2 (SSF2) target crimes (from the bid analysis)

Crime type	Target crime	No. of bid areas	% of total bid areas
Acquisitive crime	Domestic burglary	41	82
	Theft of/from car/vehicle	35	70
	Robbery	31	62
	Theft from person	28	56
	Commercial burglary	15	30
	Bike theft	8	16
	‘Neighbourhood’ or ‘acquisitive’ crimes without specification	4	8
	Shoplifting	3	6
	Fraud	2	4
Anti-social behaviour	ASB (non-specified)	42	84
	Criminal damage/vandalism	10	20
	Alcohol misuse	4	8
	Neighbourhood disputes/neighbourhood nuisance	2	4
	Nuisance motorbike riding/joyriding/speeding	2	4
	Intimidation/harassment/threats	2	4
	Arson	1	2
	Noise nuisance	1	2
Drug-related activity	Drug misuse	8	16
	Drug dealing (excl. county lines)	8	16
	County Lines	6	12
	Drug offences or other drug-relates activities (non-specified)	6	12
	Cuckooing	2	4
Violent crime	Violence/assault	20	40
	Violence against women/domestic abuse	3	6
	Knife crime	2	4
	Unspecified crimes associated with the night-time economy	1	2
Sexual offences	Rape/sexual violence/sexual offences	7	14
	Child and sexual exploitation	2	4
	Sex-working	1	2
	Kerb crawling	1	2
Organised crime	Gang recruitment/ gang-related crime/ organised crime/ criminal exploitation	10	20
	Terrorism	1	2
Vagrancy	Rough sleeping	2	4
	Offences associated with begging	1	2
Environmental crime	Environmental crimes (non specified)	3	6
	Fly tipping/littering/dog fouling	8	16
	Hare coursing/poaching (rural)	1	2

Table 3: Breakdown of SSF2 intervention types (from the bid analysis)

Type of intervention	Intervention	No. of bid areas	% of total bid areas
Target hardening: public spaces	CCTV	46	92
	Street & other lighting (new or improved)	34	68
	Alleygating	19	38
	Public area fencing	7	14
	Anti-climb spikes/paint and defensive toppings/trellises	2	4
	Other restrictive measures, e.g. K barriers, bollards	2	4
	Knife-detection equipment	1	2
Target hardening: homes	Locks/chains	17	34
	Lights (inside/outside homes)	12	24
	Doorbells, e.g. video doorbells	11	22
	Home security improvement measures/ target hardening/packs (non-specified)	11	22
	Door/window alarms	9	18
	Surveillance cameras (excl. video doorbells)	8	16
	Door/communal entrance replacements/upgrades	6	12
	Letterbox guards/upgrades	3	6
	Window jammers	2	4
	Dummy cameras/locks	2	4
	Smoke alarms	1	2
	Anti-climb spikes/paint and defensive toppings/trellises	1	2
	Safes	1	2
	Window privacy film	1	2
	Phone blockers	1	2
Target hardening: vehicles	Bike Storage	4	8
	Garage defenders/ security upgrades	3	6
	Motorbike/‘vehicle’ chains/locks	3	6
	Bike Locks	2	4
	Steering wheel locks	2	4
	Capture trap bicycles	2	4
	Capture trap motor vehicles	2	4
	Number plate screws	1	2
	Other/unspecified car park target hardening	1	2
	Safer/Secure Parking Scheme	1	2
Target hardening: Commercial Properties	Commercial target hardening/pack distribution (non-specified)	2	4
	Commercial letterbox guards/upgrades	1	2
	Commercial locks/chains	1	2
	Commercial door/window alarms	1	2
	Commercial surveillance cameras	1	2
Identifying stolen property	Property marking incl. SmartWater	23	46
	ANPR	16	32
	Forensic (chemical-dousing) trap devices	1	2
	Bike registration	1	2
Equipment upgrades	Radio systems	5	10
Environmental interventions	Signage, e.g. Neighbourhood Watch, CCTV, property marking	28	56
	Cleaning/clearing waste	11	22
	Planting (also community based)	7	14
	Graffiti removal or replacement	6	12
	Painting surfaces/artwork	6	12
	Environmental improvements (non-specified)	5	10
	Pruning/clearing/cutting plants or trees	5	10
	Repairs/maintenance	4	8
	Screening/boarding empty properties	3	6
	Public use sheds, e.g. for bins, gardening, litter collection tools	2	4
	Landscaping	2	4
	Repurposing public spaces, e.g. creating gardens	1	2
	CCTV screen	1	2
	Public Wi-Fi	1	2
	Subway network development	1	2
Community engagement	Neighbourhood Watch / crime prevention groups	15	30
	Crime prevention and safety advice/training	14	28
	Community engagement events (open to all)	10	20
	Community consultation/ surveys	8	16
	Support of/ collaboration with community groups and initiatives	7	14
	Neighbourhood Alert/ Crimestoppers engagement or implementation	6	12
	Recruiting volunteers (aside from Neighbourhood Watch)	5	10
	PABCIS or business engagement	5	10
	Home visits	3	6
	Community engagement (non-specified)	2	4
	Education programmes, e.g. sports, art	2	4
	Victim/perpetrator support, incl. mediation	1	2
Other community interventions	Employment of dedicated engagement staff	9	18
	Interventions specified to be using nudge theory	3	6
	Police patrols	3	6
	Community engagement training to staff	2	4
	Stop and search	1	2
	Treatment, recovery and diversionary activities, e.g. with drug users	1	2
Partnerships for community engagement/communications	Multi-agency partnerships, e.g. council, police, fire services, youth services, NHS	7	14
	Partnership with experts, e.g. academics, specialists	3	6
Communications	Promotional communications/ campaigns incl. emails, texts, leaflets, billboards, post	37	74
	Social media	9	18
	Communication targeting hard-to-reach/transient groups, e.g. students, elderly residents, non-English speaking	8	16
	Digital screens/ signs in public	2	4
	Dedicated apps	1	2

Appendix B: Funding allocation (2021 to 2022)

Table 4: Funding distribution across bid areas (ranging from £73,786 to £482,000)

Police and Crime Commissioner (PCC)/Local authority (LA)	Area	Grant awarded
Birmingham City Council (West Midlands PCC)	Stockland Green, Birmingham (West Midlands PCC)	£417,852
Blackburn with Darwen Borough Council	Whalley Banks (Lancashire PCC)	£482,000
Cambridge and Peterborough PCC	Peterborough City Centre	£430,000
Cheshire PCC	Chester City Centre	£432,000
Chesterfield Borough Council	Chesterfield Town Centre (Derbyshire PCC)	£422,916
Great Yarmouth Borough Council	Great Yarmouth and Gorleston (Norfolk PCC)	£96,000
Ipswich Borough Council	Gipping & Westgate (Suffolk PCC)	£226,244
Lincolnshire PCC	Ingoldmells	£170,548
North Yorkshire OPFCC	South Selby Rural Borderland Parishes	£393,320
Northamptonshire PCC	Northampton	£432,000
Nottinghamshire PCC	Nottingham City Centre	£429,248
Nottinghamshire PCC (Secondary)	Mansfield	£442,763
South Yorkshire PCC	Sharrow, Sheffield	£97,931
South Yorkshire PCC (Secondary)	Dearne Valley	£396,960
Staffordshire OFPCC	Hanley	£426,305
Thames Valley PCC	Oxford Road East & Coley, Reading	£376,563
West Mercia PCC	Greenlands, Redditch	£346,616
West Mercia PCC (Secondary)	Sutton Hill	£432,000
West Yorkshire PCC	Armytage Lane, Calderdale, Carr Wood Road Industrial Estate, Wakefield and Dewsbury Town Centre/The Flatts	£303,965
Wolverhampton City Council (Secondary)	Bilston (West Midlands PCC)	£239,102
Bristol City council	Portland Square, Bristol (Avon & Somerset PCC)	£398,451
Essex PFCC	Grays, Thurrock	£432,000
Exeter City Council	Exeter City Centre (Devon & Cornwall PCC)	£482,000
Hampshire PCC	Portsmouth Town Centre & Landport	£408,933
Hampshire PCC (Secondary)	Eastrop, Basingstoke	£257,815
MOPAC	Noel Park, Haringey	£202,169
MOPAC (secondary)	Hackney	£312,280
Surrey PCC	Godstone and Bletchingley	£266,847
Wealden District Council	Wealden (Sussex PCC)	£162,797
Wiltshire PCC	Swindon Town Centre	£420,384
Bedfordshire PCC	Manshead	£346,516
Cleveland PCC	Stockton Town Centre	£327,808
Cumbria PCC	Workington St Michael’s	£333,621
Durham PCC	Easington	£427,561
Durham PCC (Secondary)	Darlington	£266,603
Dyfed Powys PCC	Llanelli	£73,786
Gwent PCC	Pillgwenlly	£395,924
Gwent PCC (Secondary)	Moriah, Pontlottyn, Twyn Carno	£282,821
Hertsmere Borough Council	Borehamwood (Hertfordshire PCC)	£123,997
Humberside PCC	West Marsh	£437,669
Kent PCC	Central Harbour and Eastcliff, Ramsgate	£267,921
Kent PCC (Secondary)	Westgate, Canterbury	£382,158
Leicestershire PCC	Leicestershire Town Centre	£438,425
Manchester City Council	Moss Side and Rusholme (Greater Manchester Combined Authority)	£395,182
Merseyside PCC	Birkenhead	£450,000
Newcastle City Council	Byker (Northumbria PCC)	£410,320
North Wales PCC	Deiniol, Hirael Garth, Bangor, Gwynedd	£238,913
North Wales PCC (Secondary)	Glyn and Rhiw, Conwy	£246,420
South Wales PCC	Grangetown and Butetown, Cardiff	£421,934

Source: Data supplied by the Home Office.

Appendix C: Evaluation technical methodology

The evaluation of the Safer Streets Fund round 2 (SSF2) made use of multiple evidence streams to provide a nuanced and rounded account of the process and its impact. It explored implementing the Fund, as well as measuring its impact.

This Appendix presents the technical details of Verian’s evaluation approach and associated methodology. It documents the team’s design choices and the practical procedures that they adopted iteratively, building on the original research proposal, the project initiation document and the evaluation design framework.

C1. Process evaluation

Verian designed the process evaluation aiming to:

• explore the set-up, delivery and implementation experience of successful SSF2 bid area teams
• identify the key enablers and barriers to a successful delivery
• capture lessons learned by teams on the ground, to inform best practice for future rounds of funding
• combine findings with those from the impact evaluation to deliver a comprehensive view of the SSF2 programme

The process evaluation approached research in 5 tasks, commencing with a review of the 50 successful bids. The main stage of the evaluation was then undertaken through 4 tasks: a process evaluation priorities workshop with stakeholders, a pre-implementation qualitative survey, priority intervention case studies and a post-implementation qualitative survey.

C1.1 Scoping phase: Bid review

The Home Office provided Verian with bid documentation for successful and unsuccessful bids. Verian used the information from the bids to create a tailored bid review framework, which gave background to each bid area, categorised interventions by type, outlined intervention numbers, target crimes and expected outcomes by crime type. Broadly, the team categorised interventions as either physical or community. The bid review exercise enabled Verian and the Home Office to identify bids to be analysed in more detail as case studies.

C1.2 Task 1: Process evaluation priorities workshop

In August 2021, Verian facilitated a half-day workshop, using information developed from the bid review, with Safer Streets programme team members, academics from University College London (UCL), and College of Policing stakeholders. The workshop aimed to determine the priority interventions throughout the process evaluation and to identify information gaps and develop the key research questions the evaluation should aim to address. This approach to the SSF2 evaluation was guided by the objective of enabling meaningful input into critical strategies and tools, such as the SSF toolkit (College of Policing, 2022c).

The workshop resulted in 7 priority interventions being identified for this process evaluation – Automatic Number Plate Recognition (ANPR), Close Circuit television (CCTV), video doorbells, property marking, environmental, community and anti-social behaviour (ASB) – as they were highlighted as those that were either key to the SSF programme or those that the Home Office were keen to develop an insight into.

C1.3 Task 2: Pre-implementation questionnaire

Verian administered the pre-implementation questionnaire between September and October 2021. The questionnaire was qualitative in approach and designed to develop an understanding of each bid prior to intervention delivery, specifically:

• area background
• the programme choices (package of interventions) and rationale for these, including if/how SSF1 or other local initiatives had impacted these choices, the role of central government resources and external experts in shaping the bid, and the communication and media plan

The questionnaire also allowed focused data collection on the 7 priority interventions and for bid teams to feedback on any challenges they faced in the bid process.

The self-completion questionnaire took the format of a digital form emailed to all 50 successful bid areas in September. The questionnaire was designed to take around 30 to 40 minutes to complete and was expected to be shared across team members. The teams were given a 3-week period to complete and return it. In total, 44 survey responses were completed by successful SSF bid areas by the end of October 2021.

Where SSF1 conducted telephone interviews, SSF2 pivoted to a pre-implementation questionnaire, aiming to reduce the burden on bid teams while maintaining quality and depth of data collected.

C1.4 Task 3: Priority intervention case studies

Verian and the Home Office identified 10 case study areas based on agreed criterion, which included priority intervention type, area characteristics, targeted crimes, delivery partnership structure, geographic location and previous experience of delivering SSF funded interventions.

Verian conducted an in-depth analysis of these 10 case study areas between December 2021 and February 2022. The central objective of this exercise was to develop insights at bid-level for a select number of areas on their approach to set-up, implementation and monitoring of interventions, and to provide rich examples of how delivery is working ‘on the ground’. The case studies also aimed to explore the barriers and challenges faced by delivery teams and the enablers of success.

Interviews were carried out with stakeholders who had a strategic remit (for example, involved in design and had oversight of SSF2 interventions) and an operational remit (for example, involved in day-to-day delivery of SSF2 interventions). Where possible, resident or business organisations involved in or impacted by the SSF2 interventions were also included. Interviews took place via telephone or on Zoom and lasted 45 to 60 minutes. Forty-six qualitative interviews were conducted between October 2021 and January 2022.

C1.5 Task 4: Post-implementation questionnaire

The final stage of the process evaluation was a second qualitative questionnaire following the core intervention implementation period and Fund delivery deadline. The post-implementation questionnaire focused on bid teams’ experiences of implementation (including barriers and enablers to effective delivery) broken down by intervention, early signs of impact (for example, through resident feedback and changes in police recorded crime (PRC)), experiences of business and community engagement, as well as delivery teams’ assessments of future programme sustainability. Verian delivered the questionnaire the same way as the pre-implementation questionnaire in Task 2.

Verian conducted the online questionnaire during May to June 2022. All 50 bid teams completed the questionnaire, reporting their experiences of SSF2 programme delivery for all successfully funded SSF2 bids. Verian received 46 responses submissions representing feedback from all 50 successfully funded SSF2 bid areas; 4 teams submitted combined responses for primary and secondary funded bids.

C2. Impact evaluation design and approach

C2.1 Analysis of police recorded crime data

Verian used the synthetic controls method to estimate the impact of the Fund on PRC during April to June 2022.

For each bid Lower Layer Super Output Area (LSOA), Verian considered either police recorded overall neighbourhood crime or specific crime types within neighbourhood crime (burglary, robbery, theft from the person and/or vehicle crime), depending on the focus of the bid. Wherever possible, the team used the Home Office’s DataHub data for this analysis, but in police force areas (PFAs) where the LSOA-level data were judged to be incomplete, the team used public data from police.uk data instead. The team used data as recorded in DataHub for all PFAs where the ‘completeness’ of the geotagged data (as determined by the Home Office) was at least 80% for each of the years ending March 2017 to 2021. For areas where the completeness fell below 80% at any point, Verian used the data downloaded from police.uk instead.

Greater Manchester Police has not returned granular data to DataHub or police.uk since July 2019. Verian therefore excluded LSOAs in the Greater Manchester PFAs from this analysis.

The synthetic controls method takes a ‘treatment area’ where an intervention was implemented and finds a weighted average from a set of similar ‘comparison areas’ such that the historic trends closely match those of the treatment area. The rationale is that if the historic trends are closely matched, it is more reasonable to think that this weighted average of the comparison areas can provide a good counterfactual estimate (that is, a good estimate of what would have happened in the treatment area without the intervention).

Verian designed synthetic controls to analyse aggregated data for any treated geographic units. This was the case for SSF2, where counts of PRC were aggregated at LSOA level and most bids covered a few LSOAs (bids covered between one and six LSOAs). In this case, synthetic controls provided a more robust framework for estimating the impacts of individual bids than alternative methods (such as difference-in-difference) which would be suited for averaging across bids.

Verian fitted a synthetic controls model for each bid LSOA to provide individual estimates of the SSF2 impact. To make sure the team was comparing bid LSOAs against similar LSOAs, Verian limited the pool of comparison in the following way:

• for bid LSOAs where data was drawn from DataHub, the team limited potential comparison areas to PFAs that had complete records; similarly for bid LSOAs with only police.uk data available
• to avoid potential spillover effects, the team excluded LSOAs within 3 miles of an LSOA involved in any round of SSF
• LSOAs were only eligible if they passed a ‘benchmark’ of level of crime recorded; the Home Office published these benchmarks as one of the application criteria for the Fund, calculated as the LSOA upper quartile crime rates (for more detail, see Home Office, 2022); the benchmarks were used for either all neighbourhood crime, burglary or vehicle crime, depending on the bid area’s focus
• the team excluded LSOAs with no PRC data available
• the team limited the pool of potential comparison areas to those above a similarity threshold to the bid LSOAs in terms of the following characteristics:
  • population size
  • proportion of the population aged 10 to 15
  • proportion of the population who are students
  • proportion of the population of an ethnic group other than white
  • size of the LSOA (in hectares)
  • housing density (number of properties per hectare)
  • workday to resident population ratio
  • proportion of homes which are flats
  • census data about employment in different industry sectors
  • number of recorded COVID-19 cases

These characteristics are consistent with those used in the SSF1 evaluation. Verian selected them to reflect the fact that patterns and trends for PRC will vary between different kinds of area. They are intended to cover a range of characteristics which are likely to be associated with local crime trends. For example, Langton et al. (2021) found the trend PRC over the COVID-19 pandemic varied substantially between different kinds of area; in particular, there were large declines in crimes such as theft and shoplifting in urban centres and commercial districts.

For each bid LSOA, Verian fitted the synthetic controls model using the comparison LSOAs that were above the similarity threshold. This approach ensured that the synthetic control generated for each treatment LSOA was a close match on quarterly historic crime data for the period quarter 1 year ending March 2017 (April to June 2016) to quarter 1 year ending March 2022 (April to June 2021).

C2.1.1 Limitations

One limitation of using PRC data is that the number of crimes recorded by police is not the same as the number of crimes occurring, hence changes in reporting or how police forces record incidents could affect the results. This also means that an increase in PRC could mean improvements in detection and reporting of crime rather than an increase in offences. However, this data still provides important information about levels of crime (at least for certain offences) and police activity.

It is possible that differences between the number of crimes recorded in bid areas and the counterfactual estimates are not caused by the SSF interventions but by other area-level factors. Verian cannot distinguish the effect of SSF interventions from the effect that local shocks and other interventions can have on the number of crimes recorded, which could add some bias to the analysis.

One key external factor is COVID-19. Since the pre-intervention trends used to build the models go up to quarter 1 year ending March 2022 (April to June 2021), most of the impact of COVID-19 will have occurred before this point and so is already incorporated into the impact estimates. However, the team cannot rule out the possibility that the ongoing effects of the pandemic will have had different impacts in different areas at different times, and that this could add some bias to the analysis^{[footnote 21]}.

A further limitation of a relatively short evaluation period is that it may not be possible to detect effects if they are small relative to the underlying variation in data trends. If an outcome usually varies a lot between time periods, it can be difficult to separate this natural fluctuation from impacts attributable to the Fund. A challenge of measuring impacts in this way relates to situations where many observed data points in the PRC data are very or extremely low; with some bids reporting less than 10 crimes in the 3-month reference period. This is called a statistical floor effect, meaning there is often very little variation across treatment and control areas and/or any detectable change over time. This may be hiding a possible effect of the SSF interventions; however, Verian’s initial impact analysis cannot confirm it at this stage.

Another challenge of measuring impact in situations where the outcomes are relatively rare events is that outcomes are hard to predict and, in turn, the results can be quite sensitive to small changes. This means that a few ‘extra’ crimes within an LSOA could seem like an overall increase in crime. The team was unable to control for any seasonal factors.

C2.1.2 Additional PRC data impact analysis for SSF1 intervention

To explore the possibility that it may have taken more time for impacts of SSF interventions to materialise in the PRC data, Verian analysed the impact on the 12 months after the introduction of SSF1. The team applied the same weights from the 3-month analysis to 12 months of PRC data (April 2021 to March 2022) to calculate a synthetic control estimate for each area, as well as the corresponding uncertainty intervals. This 12-month weighted average of the comparison areas provided a counterfactual to estimate the longer-term impact of the SSF1 interventions.

C2.2 Survey-based impact analysis

The impact of the Fund on residents’ perceptions was measured through a difference-in-difference (DiD) design. This analysis compared changes in treatment areas (SSF area residents) at 2 points in time (pre- and post-programme) with those found in the control group areas over the same period. The difference between these changes estimates the impact that may be attributed to the SSF interventions. This approach enabled control for unobservable and time-invariant area fixed effects (the characteristics of an area likely to influence the effectiveness of the interventions). It also allowed the team to distinguish the influence of various individual-level demographics and time-variant area-level characteristics (including an indicator of the COVID-19 pandemic prevalence) from the effect of the programme on changes in neighbourhood crime.

The DiD analysis rested on several assumptions requiring consideration when interpreting the impact estimates.

1. The team sampled and weighted the control areas to be as similar as possible to the treatment areas. However, it was not possible to remove all differences between areas, so that it is possible that some findings were due to remaining differences between areas rather than the effects of the interventions.
2. The DiD analysis relied on the assumption that if the SSF2 programme had not taken place, the change from baseline to follow-up in the treatment areas would have been the same as the equivalent change in the comparison areas.
3. For a few outcome measures, there was a noticeable difference between the treatment and comparison areas at the baseline, even after weighting (up to 8 percentage points for 2 outcomes). Strictly speaking, DiD does not require the baseline scores to be close. However, in these cases, there is a greater risk that the remaining differences between areas could also reflect differences in the expected trends in those areas.

Verian used the survey data for the overall analysis of the scheme, rather than for bid-level analysis, as the sample sizes would not allow for robust bid-level estimates.

C2.2.1 Control area selection

For the survey analysis, Verian employed the Coarsened Exact Matching (CEM) method (Iacus et al., 2009) to match each successful bid area to 2 control LSOAs. Selecting 2 comparison areas minimised local area effects. This approach involved identifying important matching variables and coarsening their values, so they were coerced in broader categories or bands. The variables used for categorisation were consistent with SSF1 analysis.

The team sorted the bid areas and potential comparison areas into strata – with all areas within a stratum sharing the same coarsened values for each of the matching variables. They then matched the bid areas to comparison areas within their strata. In some cases, they matched more than one bid area to a single stratum. Appendix F shows the matching bid areas, split by those focused on multiple neighbourhood crime (table 7), burglary (table 8), or vehicle crime (table 9). The team categorised successful bid areas based on the following information:

• trajectory in one of overall neighbourhood crime, burglary or vehicle crime rates over 5 years (police.uk data for (from year ending March 2016 to year ending March 2020) depending on the bid focus; data after March 2020 was not used due to the possible impact of COVID-19 on crime trajectories
• Index of Multiple Deprivation (IMD)
• percentage of homes that are flats (Valuation Office Agency (VOA) data)
• urban/rural classification (ONS data)
• Working to Residential Population WRP ratio (2011 Census) – to indicate level of commercialisation
• cumulative COVID-19 cases per 1,000 population (ONS, 2019) – data at lower-tier LA level
• geographical indicator – when selecting comparison areas for each stratum Verian prioritised:
  • looking within Police Force Area (PFA)
  • if no matches within PFA, looking within region
  • if no matches within region, matching on all the other criteria listed

The team selected a comparison sample from the list of LSOAs eligible for the Fund (based on their rates of neighbourhood crime over the years ending March 2018 to year ending 2020) but which were not awarded funding in either rounds 1 or 2. This would also help select comparison areas with similar levels of crime to the successful bid areas. To minimise the risk of potential spillover effects in these comparison areas, the team excluded any LSOA within 3 miles of a successful bid area.

The team segmented these eligible LSOAs in the same way as the successful bid areas, then randomly selected 2 LSOA per bid within its strata. This ensured that the comparison areas selected for the study reflected the area profile of the successful bid areas. The team drew a sample of addresses from this pool of comparison LSOAs to include in the survey.

Within each treatment and common control area, the team randomly sampled addresses (with equal probability) from Royal Mail’s database of delivery points (PAF).

C2.2.2 Survey fieldwork

In the selected treatment and control areas, Verian conducted surveys at pre- and post-intervals using its Address-Based Online Surveying (ABOS) methodology (William, 2016). They sent an advance letter to sampled addresses and directed residents to a website to take part in the survey. The letter included login details for up to 4 adults in each household. To maximise response rates and minimise the risk of non-response bias, Verian offered a conditional incentive of £5 and sent up to 2 reminders to addresses where any residents had not responded. To ensure the survey was inclusive (and covered the offline population), paper questionnaires were included in the final reminder. Residents could also request a paper questionnaire from Verian, by emailing or calling the project freephone number.

The baseline survey took place between 9 August and 12 September 2021. It achieved 3,327 individual responses, comprising 1,198 paper completes and 2,129 online completes. Of the completes, 74% (2,456) were from SSF bid areas. Based on in-person surveys (such as the Crime Survey for England and Wales), it can be assumed that approximately 8% of sampled addresses were not eligible residential addresses. Taking this into account, the household response rate for the survey was 21.9%, calculated as the number of addresses with at least one adult responding / (the number of sampled addresses * 0.92).

Based on Census data, it can be assumed that the average number of eligible individuals in sampled addresses was approximately 1.9. Taking this into account, the individual response rate for the survey was 18.9%, calculated as the number of responses / (the number of sampled addresses * 0.92 * 1.9).

The follow-up survey took place between 6 May and 19 June 2022. It achieved 3,311 individual responses, comprising 971 paper completes and 2,340 online completes. Of the completes, 74% (2,455) were from SSF bid areas. The household response rate was 19.1% and the individual response rate was 16.4%.

These response rates were in line with the response rates from the SSF1 evaluation. They are also broadly in line with other similar studies using similar methods. For example, the equivalent individual response rate for the Community Life Survey for year ending March 2021 was 22.8% (DCMS, 2021), and the equivalent for the BEIS Public Attitudes Tracker was 10.2% (BEIS, 2022). Differences between studies may be due to factors such as differences in fieldwork procedures (for example, respondents in the Community Life Survey received a larger incentive than respondents in the SSF2 survey), questionnaire topics and timing.

C2.2.3 Survey data weighting

Verian used weighting to ensure that the sample achieved by the survey for the successful bid areas represented the population at the overall level. The team weighted the survey data using reliable Office for National Statistics (ONS) LSOA-level population statistics for the following variables:

• age*gender (ONS LSOA population estimates mid-2020)
• ethnicity (ONS Census data 2011)
• household composition (ONS Census data 2011)
• disability status (ONS Census data 2011)
• housing tenure (ONS Census data 2011)

Verian then aggregated these LSOA benchmark population statistics to the whole scheme-level.

The team weighted each bid area equally (regardless of the population of each area). This is because they focused on the mean of SSF2 across the 50 successful bid areas (rather than giving more weight to schemes that covered a larger population).

Propensity Score Matching was used to weight the control area sample from each wave to match the profile of the successful bid area samples (in terms of demographic profile and the area characteristics used in the CEM). This helped increase the confidence that any differences observed between the bid areas and control areas could be attributed to the SSF2 funding rather than to differences in the local population profiles.

C2.2.4 Additional analysis covering SSF1 and SSF2 survey data

Verian pooled the data from SSF1 and SSF2 to generate a larger dataset that combined the responses from both baseline surveys and both follow-up surveys, for just the outcome variables that were included in both rounds. They carried out this additional analysis to exploit the potentially higher statistical power of a larger survey sample size. Each round had its own separate comparison group, although the method of selecting the comparison sample was consistent between the 2 rounds. The team achieved a pooled sample size of 14,576 responses (7,938 from SSF1 and 6,638 from SSF2); 52% corresponded to baseline and 48% to follow-up surveys. Before conducting the DiD analysis, the team scaled the SSF2 weighted sample so that each of the 4 analysis groups (baseline treatment, follow-up treatment, baseline control, follow-up control) had the same size of SSF1, to ensure that both rounds contributed equally to the analysis.

Appendix D: Impact analysis of police recorded crime data

At the bid-level, the analysis showed some differences in whether there are statistically significant indications of impact. Specifically, in 6 bid areas, the impact analysis detected a significant reduction in police recorded crime (PRC) per Lower Layer Super Output Areas (LSOA) (than would be expected without the Fund) over the 3-month period. It is not possible to determine what factors have enabled this improvement in crime levels through analysing the interventions and areas targeted. Also, in 3 bid areas, the number of crimes recorded was higher than would be expected without the Fund. The main body of the report discusses possible reasons for significant increases in crime. Table 5 below outlines the impact estimates for each of the bid areas. Bid area names have been removed for confidentiality reasons.

Table 5: Detailed bid-level impact estimates for each SSF2 bid (in the order of smallest to largest p-value)

Please see Appendix D: Table 5.

Appendix E: Difference-in-difference analysis

Table 6: Difference-in-difference analysis of residential survey data

Please see Appendix E: Table 6.

Appendix F: Strata characteristics

Verian used Coarsened Exact Matching (CEM) to select comparison areas. CEM involves sorting all areas available to be a comparison area into broader categories or bands, with the 50 bid areas and potential comparison areas sorted into strata. Verian selected 2 comparison LSOAs from the same stratum as each bid area.

The tables below show more detail about the variable groupings in each strata used to classify bid areas:

• trajectory in one of: all neighbourhood crime, burglary or vehicle crime rates over four years, depending on the bid focus
• Index of Multiple Deprivation (IMD)
• percentage of homes that are flats (VOA data)
• urban/rural classification (ONS)
• Working to Residential Population (WRP) ratio (2011 Census)
• cumulative COVID-19 cases per 1,000 population (ONS)

Table 7: Strata characteristics and matching bid areas targeting all neighbourhood crime or a combination of two or more types of neighbourhood crime

Table 8: Strata characteristics and matching bid areas targeting burglary

Table 9: Strata characteristics and matching bid areas targeting vehicle crime

Please see Appendix F: Tables 7 to 9.

References

BEIS (2022) BEIS Public Attitudes Tracker: Technical Report. Available from: https://www.gov.uk/government/publications/beis-public-attitudes-tracker-technical-report [accessed 10/05/2023]

College of Policing (2022a) ‘Neighbourhood crime: Situational crime prevention interventions – Evidence briefing’. College of Policing. Available from: https://assets.college.police.uk/s3fs-public/2022-01/Neighbourhood-crime-situational-crime-prevention-initiatives.pdf [accessed 10/05/2023]

College of Policing (2022b) What is situational crime prevention? Available from: https://www.college.police.uk/guidance/neighbourhood-crime/what-situational-crime-prevention [accessed 10/05/2023]

College of Policing (2022c) Safer Streets Fund toolkit updated. Available from: https://www.college.police.uk/article/safer-streets-fund-toolkit-updated [accessed 10/05/2023]

DCMS (2021) Community Life Survey: 2020/21 Technical Report. Available from: https://www.gov.uk/government/statistics/community-life-survey-202021 [accessed 10/05/2023]

Heeks, M., Reed, S., Tafsiri, M. and Prince, S. (2018) The economic and social costs of crime (second edition), Research Report 99, Home Office. Available from: https://www.gov.uk/government/publications/the-economic-and-social-costs-of-crime [accessed 10/05/2023]

Home Office (2021) Safer Streets Fund (2021-2022) Prospectus. Home Office, London. Available from: https://www.gov.uk/government/publications/safer-streets-fund-application-process/safer-streets-fund-2021-to-2022-prospectus-accessible-version [accessed 10/05/2023]

Home Office (2023) Evaluation of the Safer Streets Fund round 1, year ending March 2021. Home Office, London. Available from: https://www.gov.uk/government/publications/safer-streets-fund-evaluation-round-1-year-ending-march-2021/evaluation-of-the-safer-streets-fund-round-1-year-ending-march-2021 [accessed 10/05/2023]

Iacus, S.M., King, G. and Porro, G. (2009) ‘CEM: Software for Coarsened Exact Matching’, Journal of Statistical Software, vol. 30(9). Available from: https://gking.harvard.edu/files/gking/files/jss-paper.pdf [accessed 10/05/2023]

Langton S., Dixon., A, and Farrell, G. (2021) ‘Small area variation in crime effects of COVID-19 policies in England and Wales’. Journal of Criminal Justice, vol. 75. Available from: https://doi.org/10.1016/j.jcrimjus.2021.101830 [accessed 10/05/2023]

Manning, M., Wong, T.W.G., and Vorsina, M. (2016) Manning Cost-Benefit Tool. Centre for Social Research and Methods, Australian National University, Canberra. Available from: https://csrm.cass.anu.edu.au/research/projects/manning-cost-benefit-tool [accessed 10/05/2023]

ONS (2022) Crime in England and Wales: year ending September 2022. Office for National Statistics. Available from: Crime in England and Wales: Appendix tables - Office for National Statistics (ons.gov.uk)

Piza, E.L., Welsh, B.C., Farrington, D.P. and Thomas, A.L. (2019). ‘CCTV surveillance for crime prevention’. Criminology & Amp; Public Policy, 18(1), pp. 135-159. Available from: https://academicworks.cuny.edu/cgi/viewcontent.cgi?article=1275&context=jj_pubs [accessed 10/05/2023]

Welsh, B.C. and Farrington, D.P. (2008). ‘Effects of Improved Street Lighting on Crime’. Campbell Systematic Reviews, vol. 4(1), pp. 1-51. Available from: https://www.politieacademie.nl/kennisenonderzoek/kennis/mediatheek/PDF/76003.pdf [accessed 10/05/2023]

William, J. (Kantar Public) (2016) ‘An introduction to address-based online surveying’, Social Research Practice, Issue 3, Winter 2016/17. Available from: https://the-sra.org.uk/Common/Uploaded%20files/Social%20Research%20Practice%20Journal/social-research-practice-journal-issue-03-winter-2017.pdf [accessed 10/05/2023]

1. Statistical significance testing helped to determine which changes could be attributed to the programme, rather than any random error. The difference in differences (DiD) analysis focused on impact findings that were statistically significant at the 5% level. This means that if the survey were to be conducted 100 times, there would be a finding of the same nature on at least 95 occasions. ↩

2. The definition of ASB that was included in the survey is as follows: “Any behaviour that causes people nuisance, annoyance, alarm or distress, including those that are aggressive or intimidating or that impact upon a person’s quality of life”. ↩

3. Robbery is defined as stealing personal belongings from a person using force. In comparison, theft from person does not involve the use of force. ↩

4. Situational prevention is a crime deterrence strategy that focusses on locations where crime occurs to increase the risk and reduce the reward to offenders of committing crimes (College of Policing, 2022). ↩

5. Safer Streets Fund 2021 to 2022 prospectus (Home Office, 2021) ↩

6. The DiD method is a quasi-experimental approach that compares the changes in outcomes over time between a population enrolled in a program (the treatment group) and a population that is not (the comparison group). ↩

7. Regression analysis is a set of statistical processes for estimating the relationships between a dependent variable and one or more independent variables. ↩

8. This impact analysis does not cover one bid for which data was not available. ↩

9. A data set that contains geodemographic classification of households. ↩

10. Jaggaer is an online platform used by the Home Office to publish bid-relevant information to all successful bid teams and used as part of the application process. ↩

11. Business Improvement Districts - GOV.UK (www.gov.uk) ↩

12. PCCs can find the Process Implementation Checklist in the Knowledge Hub of the SSF Library, or contact saferstreets@homeoffice.gov.uk for further information. ↩ ↩² ↩³

13. The Cube is a portable wireless security device designed for filming on private property, for example, in sheds or farmyards. ↩

14. Both brands use property marking solutions that are invisible and permanent and require users to register the property they have marked. ↩

15. The definition of ASB that was included in the survey is as follows: “Any behaviour that causes people nuisance, annoyance, alarm or distress, including those that are aggressive or intimidating or that impact upon a person’s quality of life”. ↩

16. However, this does not necessarily or categorially allow us to conclude that a shorter impact measurement time frame (for example, 3 months) was not a factor that hampered the evaluation’s ability to detect impacts (for both SSF1 and SSF2). This is because it is not possible to confidently say exactly how long it would take for these interventions to have an impact on PRC incidents, alongside further contextual factors related to SSF1 and SSF2 differences in the interventions and locations targeted. ↩

17. CI is defined as the confidence interval. Statistically significant results (at the 5% level) are indicated by an asterisk (*). ↩

18. Statistical significance is defined as being within the 95% confidence interval. ↩

19. Verian understands that, since the drafting of this report, Police Crime Prevention Initiative (PCPI) launched a new website in February 2023, which is open to all SSF practitioners and supports local projects with a variety of resources to assist with effective delivery of SSF interventions. ↩

20. Logic models were developed as part of the SSF1 process evaluation case studies. At the time of publishing this report, the Safer Streets Fund round 4 Invitation to Tender indicated a request for the evaluators to develop a theory of change and logic model. ↩

21. For further information on the local impact of COVID-19, see Langton et al., (2021). ↩