Developing and implementing 20-mph speed limits in Edinburgh and Belfast: mixed-methods study

Edinburgh implementation

Edinburgh took a citywide approach to implementing the intervention, that is, streets that fell within the City of Edinburgh Council area. Prior to implementation, 50% of streets already had speed limits of 20 mph. Edinburgh implemented the 20-mph intervention in a further 30% of streets, which equated to an additional 1572 roads being reduced to 20 mph. This implementation resulted in 80% of streets in Edinburgh having speed limits of 20 mph (771 miles/1240.3 km), with a coherent network of 30-mph and 40-mph speed limits in the remaining 20% of streets. The 20-mph network was implemented under one citywide speed limit order. The city was split into seven implementation zones, and the intervention was implemented over three phases (Figure 1 and Table 2). Each implementation phase took approximately 16 weeks over a total period of 24 months (starting in July 2016 and finishing in March 2018).

FIGURE 1.

Map of the 20-mph implementation zones in Edinburgh. Copyright City of Edinburgh Council, contains Ordnance Survey data. © Crown copyright and database right (2021). All rights reserved. Ordnance Survey Licence Number 100023420.

Copyright City of Edinburgh Council, contains Ordnance Survey data. © Crown copyright and database right (2021). All rights reserved. Ordnance Survey Licence Number 100023420.

Belfast implementation

Belfast implemented the 20-mph speed limit as part of a wider pilot project. Five pilot 20-mph speed limit zones were implemented in Northern Ireland. Four were residential areas only; the fifth and largest site was in the central area of Belfast, encompassing a total of 76 streets. This is the part of the city centre with the highest levels of pedestrian movement, cycle activity and bus facilities. Twenty streets were subject to a Prohibition of Traffic Order (pedestrian zone) and seven were partially subject to a Prohibition of Traffic Order. Comparable with Edinburgh, the 20-mph streets in Belfast are surrounded by a coherent network of 30-mph and 40-mph streets in the city centre. In Belfast the intervention was implemented in a single phase in a similar geographical area (starting in February 2016) (Figure 2).

FIGURE 2.

Map of the 20-mph implementation zones in Belfast. Provided by the Department for Infrastructure, Northern Ireland. Reproduced from Land and Property Services data with the permission of the Controller of Her Majesty’s Stationery Office, © Crown copyright and database rights MOU203.

Reproduced from Land and Property Services data with the permission of the Controller of Her Majesty’s Stationery Office, © Crown copyright and database rights MOU203.

One of the main stages in the initial application to NIHR was to develop a programme theory to understand what change in outcomes were expected to occur as a result of the intervention, and the potential pathways. This programme theory was used to develop our research questions and frame the evaluation.

First iteration of the 20-mph programme theory

A very initial programme theory was developed (Figure 3) in discussion with the City of Edinburgh Council, but needed refinement. We therefore undertook a review of existing literature to refine the theory, and spoke to key stakeholders. This led to the programme theory that was used in the research bid to NIHR.

FIGURE 3.

First iteration of the programme theory.

Second iteration of the 20-mph programme theory (review of evidence)

The second iteration of the programme theory built on the first by drawing on the wider research evidence, complemented by discussions held with, and documents provided by, local authority officers delivering the 20-mph speed limit intervention in Edinburgh. There follows a summary of the evidence that we used to develop the theory.

Edinburgh pilot implementation

An area-wide 20-mph limit was implemented across south-central Edinburgh, with the scheme launched in March 2012. The speed limit was changed to 20 mph on 28 streets; these streets, along with 20 streets that retained a 30-mph limit, were monitored before (May and June 2011) and after implementation (May and June 2013) for traffic speeds. In the 20-mph streets, average speeds fell by 1.9 mph (from 22.8 to 20.9 mph); in 12 streets where average speeds exceeded 24 mph before implementation, a greater average reduction of 3.3 mph was found. Four streets saw slight increases in average speed after implementation, and four streets had average speeds of > 24 mph. Speeds also reduced on the roads that were maintained at 30 mph, by a smaller magnitude of 0.9 mph (from 26.3 to 25.4 mph). More than 1000 household surveys were carried out before and after implementation. Before implementation, 68% supported 20-mph speed limits; after implementation, this figure rose to 79%. Strong support increased from 14% to 37%. Opposition to the limits reduced from 6% to 4% after implementation. There was a net increase of 7% for journeys on foot, an increase of 5% for journeys by bike and a decrease of 3% for car journeys. This evaluation was small scale and did not utilise the robust methods that were used in the evaluation reported here.

The second iteration of the programme theory was based on the existing evidence base, and hypothesised theories of the effect of the intervention on the outcomes. Although we knew what had happened in other cities, we did not know what would happen in Edinburgh and Belfast.

Figure 4 shows the programme theory we used in the application for funding.

FIGURE 4.

Second iteration of the programme theory. Reproduced with permission from Turner et al. 4 © 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).

Reproduced with permission from Turner et al. 4 © 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).

Third iteration of the programme theory

The research team undertook a qualitative study that involved engagement with a range of stakeholders, including the general public. 3 The inclusion of stakeholders in theory development is considered good practice. 52 In the third iteration, the previous (second) iteration was further informed and refined through two streams of activity: semistructured interviews with key stakeholders involved with, or who had a vested interest in, Edinburgh’s scheme, and focus groups and interviews with members of the general public covering a range of demographics. Data collected were used to help identify mechanisms of change explaining how intervention activities were proposed to lead to purported changes in health outcomes (Figure 5).

FIGURE 5.

Third iteration of the programme theory. AT, active travel; CEC, City of Edinburgh Council; DRDNI, Department for Regional Development Northern Ireland.

The programme theory outlined in this version demonstrates the complexity of the pathways (i.e. perceptions, behaviours) through which the reduced peak and average speed is purported to lead to more objective health-related outcomes such as active travel. Consequently, this study was designed to address not only the question of the effects, impacts and costs of the intervention, but also the question of how and why the intervention and effects/impacts occurred or did not occur.

The fourth iteration (Figure 6) provided us with an outcomes framework, which we used to facilitate our appraisal of the various data sources to be used. We organised the outcomes into useful categories by theme and time. Once the data sources were appraised, the outcomes framework further helped us plan and conduct the analysis and reporting of some of the outcomes we observed.

FIGURE 6.

Fourth iteration of the programme theory.

The final project used a range of theoretical frameworks to assist in the interpretation of the results. These theories provide different ‘lenses’ through which to analyse the policy process, and provide a theoretically grounded underpinning to the analysis. A theory-based evaluation enables the use of a realist perspective to understand what worked, for whom, in what circumstances and why. 53 During the initial period of the grant development, we developed our programme theory further. However, as described in previous sections, this was a hypothesised model. We did not have any evidence to suggest that the intended inputs or activities would be available and/or implemented in the two cities, nor did we have evidence that the input, activities and implementation strategies would result in the desired change in outcomes. The process of developing this model was iterative, and involved important contributions from the Study Steering Committee (SSC) when we presented early drafts of the findings and our explanations. During the evaluation, other smaller side projects were undertaken to contribute to the evidence base, and to help answer some of the questions raised by the SSC; for example, a small-scale master’s dissertation looking at the impact on air pollution was undertaken, supervised by a member of the SSC (Dr Stefan Reis) (see Chapter 4, Air pollution).

At the end of the project, based on the findings from all work packages (WPs), our programme model was further refined to create our final model, as shown in Figure 7. A description of each box on the final model is included in Appendix 1.

FIGURE 7.

Final programme model.

Aim

The overall aim of this research was to evaluate and understand the processes and effects of citywide 20-mph legislation in Edinburgh and Belfast. The focus was on health-related outcomes (active travel and casualties) and the pathways and processes that cause this transport policy to have public health benefits. Alongside this was research into the political, historical and policy factors (conditions) that led to the decision to implement the new speed limit, with a view to understanding possible transferability and national impact.

Objectives and research questions

The objectives and research questions were developed to link in broadly with the programme theory.

Changes to the original protocol

Over the 3.5 years of the evaluation, we had to make some changes to the study protocol. The main change was the adjustment of our analyses of active travel. We had originally anticipated that the data that we received from the local jurisdictions and Sustrans would be suitable to be able to undertake longitudinal analyses and to measure the effects of the intervention on cycling and walking. However, this proved to be impossible for a number of reasons, linked to how, when and where the data were collected (see Chapter 4, Active travel). This affected not only the outcome evaluation, but also our ability to do an economic evaluation (see Chapter 6). This meant that the cost–utility analysis based on the relationship between changes in active travel and physical activity-related health improvements, with associated changes in length and quality of life, could not be estimated. In addition, emissions data were not available and the data gathered in the study suggested that perceptions of safety had not changed much following the introduction of 20-mph speed limits. In practice, the final analyses planned were not possible owing to changes in the role of the health economist co-leading the economic evaluation (NC). He is a government economist and was moved full time to COVID-19-related work.

Therefore, in the report, we present the findings from the analyses carried out to inform the progression decision, update these where we can with the data we have, reflect briefly on how these updated findings may affect the overall conclusions from the economic evaluation and give pointers to the further work that would be required to complete the planned analysis.

In our investigation of implementation (see Chapter 3), the original aim was to conduct interviews with 20 implementation agents: six in Belfast and 14 in Edinburgh. In Edinburgh, we originally intended to investigate implementation processes at both the citywide level (six interviews) and the local level, following implementation at each geographical phase (totalling eight interviews). Through the early interviews, it became apparent that there was less of a direct implementation role at the local level and the role of individuals/organisations previously identified as potential implementation agents was minimal. Therefore, no interviews took place at the local level; instead, we increased the number of interviews held with agents involved at the citywide level.

We had originally anticipated recruiting members of the general public to focus groups (see Chapter 5) via sampling based on responses to routinely collected quantitative surveys (e.g. the Edinburgh People Survey conducted by City of Edinburgh Council). However, only anonymised data sets were provided to the research team, meaning that this was not possible. We instead maximised the recruitment of participants via other means.

We had intended to evaluate the impact on health inequalities, but the data were not suitable. We had wanted to use rates by Index of Multiple Deprivation to investigate casualties by area of deprivation. We examined the multiple deprivation profiles of the implementation areas in Edinburgh and Belfast for the control site selection process. The difficulty was that these areas range in size from 11 to 170 Data zones or small areas, which are the geographic units at which the multiple deprivation measures are allocated. However, the distributions were mostly heavily skewed. Although we sampled the focus groups to reflect low-income areas, and people from different ethnicities, we found no data suggesting that there were differences in impact perceived by the different populations.

Finally, we had expected to conduct workshops to discuss the implementation and transferability of the intervention, but we were not able to because of the COVID-19 pandemic. We have, however, gained additional funding from the Wellcome Trust to undertake other forms of dissemination activities (see Chapter 7, Maximising the impact of the findings: dissemination).

The idea that complexity is paramount in public health and other policy arenas has gained widespread traction in recent years. 57–59 The MRC has published a series of guidelines over the years2,60 to help researchers manage the tricky issues that need to be grappled with when doing research on matters defined as complex. Various guidelines have been developed and the problems with doing evidence synthesis when complexity is involved are widely discussed (e.g. Noyes et al. 61). The current project sits firmly within the complexity arena: the actual intervention was multiplex, and variable things were done at different times, in different places, in different ways, in environments with varying histories, politics, organisations and personnel. The initial programme theory was the first opportunity to describe the complexities involved. Through subsequent iterations, we elaborated the theory to take account of local political and organisational dynamics and to identify some of the uncertainties involved. The interventions were not under the control of the researchers, and the fidelity of the individual components was pretty well impossible to assure. Although we framed our work in terms of being theory based, we have not overtheorised our work; instead, we have tried to disentangle the component parts of the activities in both cities and empirically get as much purchase as possible on what was going on. Our report reflects this approach, and, we hope, provides insights into the effectiveness of this type of intervention, as well as the difficulties in doing research in real-world settings and with the constraints described here.

Ethics arrangements and data management

This was a complex evaluation comprising several studies. Different types of data were being utilised, which required different levels of ethics approval. The study was conducted in line with MRC guidelines62 and the Economic and Social Research Council ethics framework. 63 Ethics approval was sought from the Moray House School of Education Ethics Committee, University of Edinburgh (see Chapters 3 and 4 for details). The committee has three levels of approval:

• level one – study that includes only anonymised data (e.g. large data sets)

• level two – study that includes primary data collection (qualitative and quantitative)

• level three – study that includes primary data collection from vulnerable groups (e.g. children, those who do not have capacity to consent).

The evaluation required level one and level two, but not level three, approval. When there were minor changes in research instruments, further ethics committee advice was sought.

The evaluation utilised a mixture of routinely collected and primary data:

• Routinely collected data – many of the data were collected by other parties (e.g. local councils and Sustrans) as part of routine monitoring and evaluation. We received data from these sources only after they had been fully anonymised.

• Primary data – the main primary data collected by the evaluation team were qualitative and survey data. For these types of data, we sought informed consent to participate. All data will be held for 5 years on university password-protected computers. Paper documents such as consent forms were stored in locked filing cabinets.

Patient and public involvement

For a project such as this (evaluating a natural experiment/an intervention that we were not delivering), our approach to patient and public involvement was about engaging the public and various public bodies. Considerable impetus for getting the 20-mph limit to the stage of implementation in Edinburgh and Belfast originally came from public opinion and public interest groups such as Sustrans and Living Streets. Both organisations worked closely with the team as the original research proposal was developed. They were particularly helpful concerning the collection of process and outcomes data. Sustrans also partly funded the intervention in Edinburgh and a senior member of Sustrans was a co-applicant on the grant. Sustrans has significant experience in collecting data on active travel and collected many of the data in Edinburgh and Belfast. Living Streets also advised on data collection methods for the liveability assessments. Both organisations were represented on the SSC and were involved in dissemination.

We conducted qualitative work (focus groups and interviews) with members of the public, which informed the development of the original research proposal and objectives. The data collected from this work allowed us to produce the original programme theory that framed what (and how) we evaluated in the project, an example being the development of questions for the perceptions survey. This qualitative work involved 37 members of the public from various professions and public groups: teachers, NHS workers, young professionals, older adults, school parent councils, students and residents. Stakeholders were also involved with the evaluation strategy and the planning and design of the research. We conducted 17 stakeholder interviews from a broad range of public service organisations: the council, charitable organisations (Sustrans, Living Streets), public transport companies (bus and taxi), driver groups (Institute for Advanced Motorists, Motorcycle Action Group), the local health board, Transport Scotland and the Scottish Fire and Rescue Service.

The SSC had representatives from public transport (Lothian Buses); driver groups (Institute for Advanced Motorists); charitable organisations (Living Streets, Sustrans, Belfast Healthy Cities); the local health board; Transport Scotland; the Department for Infrastructure, Northern Ireland; Police Scotland; and the City of Edinburgh Council.

During the project we engaged with 29 stakeholders from a broad range of public service organisations [the council, charitable organisations (Sustrans), the local health board, government agencies (Transport Scotland; Department for Infrastructure, Northern Ireland), police], 159 members of the public (across 24 focus groups in the two cities), the parents of young/school-aged children, older adults, pedestrians, cyclists, motorised transport users, young drivers, residents/community councillors, public transport companies (bus and taxi) and driver groups (Institute for Advanced Motorists, advanced motorcyclists).

Other groups, such as Belfast Healthy Cities; the Department for Infrastructure, Northern Ireland; the Police Service of Northern Ireland; Belfast City Council; and the City of Edinburgh Council 20-mph implementation team, also sat on the SSC and assisted with providing access to data and with dissemination.

Structure of the report

Rather than have a single methods section followed by results, each of the consequent chapters focuses on a different research objective (Table 3) that relates to our initial programme theory. We undertook the work in different WPs.

Each of these chapters focuses on explaining and/or evidencing the programme theory, which continued to evolve during the evaluation period. The final programme theory is described step by step in the next chapters. Each chapter focuses on specific objectives and sections of the programme theory, highlighting where effects were found or not found (Figure 8).

FIGURE 8.

Programme theory and relationship to chapters and objectives.

Documentary analysis

We conducted searches of key websites to identify documents relevant to 20-mph speed limits, including UK-wide developments, as well as national and local activity. We were interested in identifying any legislation, policy documents, research reports and official statistics, as well as other written records of events including political speeches, official announcements, committee reports and debates. The websites searched included those of the national governments (Scotland and Northern Ireland) and the City of Edinburgh Council. We did not look for local council documents in Belfast as the scheme was led and managed by the government of Northern Ireland (through the Department for Regional Development, which is now the Department for Infrastructure).

We compiled a timeline of the publication of relevant documents for each of the two cities. These were shared with a range of stakeholders in each city to confirm their comprehensiveness. Any additional documents that were identified by the stakeholders were added to the list. Each interviewee (see Stakeholder interviews for details) was also asked to verify the completeness of the list. In total we identified 19 documents for Edinburgh, published between 2000 and 2016, and 20 documents for Belfast, published between 2001 and 2016. These documents consisted of policy statements, reports of council decisions and reports of responses to public consultations. We obtained a copy of all documents and conducted thematic content analysis using NVivo 12 software (QSR International, Warrington, UK). An inductive approach was used, whereby no themes were proposed a priori. Documents relevant to Edinburgh and Belfast were analysed separately. Data were extracted independently by two members of the research team (KM and MPK), and the final coding framework for each city was agreed through discussion.

Stakeholder interviews

Interviews were conducted with 16 stakeholders: eight in Edinburgh and eight in Belfast. These were the same stakeholders as discussed in Chapter 3, Stakeholder interviews (more details of the methods can be found there). The Edinburgh-based participants consisted of council officers (n = 2), elected members of the council (n = 3), Sustrans officers (n = 2) and a civil servant. In Belfast, seven interviews took place with eight interviewees including representatives from government departments (n = 6), the Police Service of Northern Ireland (n = 2), a public transport organisation (n = 1) and a third-sector organisation (n = 1). Interviewees could represent more than one department. Interviews took place in November 2017: 25 months after the legislation went live, and 22 months after the signage was implemented. An additional interview (interview 8) took place with two representatives from the Department for Infrastructure (responsible for implementing the intervention) in March 2019 (38 months post implementation of signage) following completion of the initial ‘3-year pilot’ phase.

Briefly, a semistructured interview guide was used to discuss topics such as involvement in the scheme, the main drivers for getting it onto the policy agenda, barriers to and facilitators of moving it forward and key individuals or groups. The audio files were transcribed verbatim and uploaded to NVivo 12 for analysis. As with the document analysis, the analysis of the interviews was approached inductively using thematic content analysis and conducted separately for each city. Data were independently coded by two members of the research team (KM and MPK), and the final coding framework for each city was agreed through discussion. The findings from the documents and interviews for each city were triangulated to provide a more complete picture of the political processes and events in each city.

Results from the document analysis and interviews

The findings from the documents and interviews from the two cities are presented together under 11 key themes that emerged from the analysis (Table 4).

Print media methods

One of the ideas that we encountered when we were planning the project was the view that local print media had been hostile, especially in Edinburgh, to the whole idea of 20-mph restrictions. To explore the ways that the reporting was done, we conducted a thematic and narrative analysis of print media. This would allow us to understand the process of construction of the 20-mph policies, as portrayed by the print media.

Free-text searches were conducted on the LexisNexis^® (New York, NY, USA) database for the main print media relevant to Edinburgh and Belfast, including a mixture of local and national publications, and including articles from 1990 to October 2019. Search terms included ‘20mph’, ‘speed controls’ and ‘traffic calming’. Searches were also made on print media outlet websites, when not available on LexisNexis. Hits were downloaded and deduplicated. Two reviewers (CF and KM) screened articles for inclusion. Articles were excluded if they did not mention 20-mph schemes and were not related to the study areas and the environs. Excluded studies mainly related to criminal acts involving speed or collisions, or used the term ‘20 mph’ but not in relation to a speed reduction intervention, for example in car reviews. Full-text articles were assessed against three eligibility criteria: (1) published in a print outlet relevant to Belfast or Edinburgh; (2) mentioned 20 mph relating to road traffic, speed, congestion, collisions or safety; and (3) published before October 2019.

Eligible media articles were categorised into one of two types of coverage, defined as ‘news reports’ and ‘editorials or opinion pieces’, with details on key actors recorded within themes. Data were extracted from each article to determine key themes, and these themes were categorised according to the three ‘streams’ of Kingdon’s Multiple Streams Framework (problems, policies and politics). 73 The themes were also mapped in publication order to examine the shifts in media focus on news and opinions across the three streams over time. Articles could be classified across more than one stream.

Results from print media sources

The electronic searches of databases and media websites identified 242 hits (post deduplication). Post title screening, 154 full-text media articles were assessed for eligibility, leaving 84 articles meeting the inclusion criteria. Articles were excluded for not being relevant to 20 mph or for being related to criminal acts (Figure 9).

FIGURE 9.

Flow chart for identification of records (articles).

The majority of hits for Edinburgh were derived from six Scottish media sources, whereas there were only three key media sources for Belfast. This is probably because of the greater number of local and national newspapers in Scotland. The types of stories were more frequently classified as news rather than opinion based. Owing to the very low number of media stories related to 20 mph from Belfast (12 articles), thematic results are presented only for Edinburgh and the environs.

Each stream existed contemporaneously, rather than independently, as articles were coded across streams and codes appeared in each, especially pro- and anti-20-mph views and evidence. Examples of article texts for each of the streams can be found in Appendix 2.

Problem stream

In the problem stream, the impacts of traffic speeds, collisions and pedestrian fatalities were reported long before the adoption of the Edinburgh 20-mph scheme in the earlier articles published since 1993. The key themes identified placed traffic-related problems across three types of advocacy (professional, public and universal) against anti-20-mph reporting (Figure 10).

FIGURE 10.

Timeline (1993–2019) of Scottish media coverage by themes in the problem stream.

The personal impact and consequences of traffic-related problems were most profoundly expressed in the personal stories of child fatalities. The professional and public advocacy responses were universal in calling for a solution to the problem of speeding and accidents, and placed 20-mph schemes as an immediate solution for safer streets, especially in places where deaths had occurred. Evidence sources were also reported from outside Scotland to support the call for safer streets from professional advocacy groups. These combinations of both professional advocacy and public advocacy merged into a universal advocacy call for improvements to this problem:

Demonstrators lay in the street yesterday bearing 103 crosses to symbolise the projected death toll on Britain’s roads before the Government launches its safety review in 11 days’ time.

Protest over road deaths, The Herald (Glasgow), 21 February 2000.

Reproduced with permission from Newsquest Media Group

As 20-mph schemes were being more widely reported, opposition to such ideas were also reported. Nearly all of these anti-20-mph views were expressed as editorial or opinion articles stressing a political positioning of anti-speed and depersonalised positions or non-enforcement:

The Green Party have been advocating this, but do any of these people actually drive a car? Or could they be cyclists? And have they tried to drive at 20 mph?

Efforts to get drivers on go-slow must have been dreamed up by cyclists, Evening Express, 12 May 2018.

Reproduced with permission from Aberdeen Journals Archive

Policy stream

The policy stream saw a continued use of evidence presented in the problem stream for 20-mph schemes, the start of calls for using 20-mph speed limits as a solution to these problems and the 20-mph policy adoption by the Green Party. Anti-20-mph views were found in articles on issues of enforcement and cost. Key themes identified were the remessaging of evidence (i.e. developing positive statements about the evidence to support the views of the advocates of the speed limits) and the idea that policy solutions to these issues already exist. These were supported by political advocacy, but also received anti-20-mph pushback as policy options were reported (Figure 11).

FIGURE 11.

Timeline (1993–2019) of Scottish media coverage by themes in the policy stream.

The problem stream’s pro-20-mph speed limit evidence of impacts had focused on speed and child fatalities, pre 2000. Now new types of evidence appeared in the policy stream that had more breadth in terms of the range of issues reported, across health, economic, environmental and social benefits for individuals and communities. The problem of traffic was now expressed in policy language in terms of economic cost. This broadening of benefits was coded as a theme of remessaging of evidence as it was reported and as it was being trumpeted by advocates and, for the first time, political individuals and parties. Despite 20-mph schemes being sought as a solution to the problem, there were few calls for a political and policy solution pre 2000; however, articles post 2000 introduced examples of ‘successful’ implementation of 20-mph schemes in Scotland, reporting additional and potential benefits, proposing 20 mph as a solution, and calling for political support. The impact of 20-mph schemes in other parts of the UK and within Scotland were more frequently reported post 2000 and connected with solving similar problems elsewhere. This powerful connection was expressed in terms that indicated that a solution already exists and is being used by others like us. For example, redevelopment plans in Aviemore included 20 mph as part of their traffic safety measures:

Two roundabouts at either end of Grampian Road will take traffic at 20 mph up to a new system of streets organised around the green, designed as the heart of the village, its commercial as well as its community life.

Blueprint unveiled for new face of Aviemore, The Herald (Glasgow), 9 September 1997.

Reproduced with permission from Newsquest Media Group

The impact of the pilot of the 20-mph Edinburgh scheme on road safety was reported as successful and as the principal reason for expansion of the scheme across the whole city in 2014, with the endorsement of a local pressure group.

Editorial and opinion articles reported the problem of traffic and its impacts on Scottish communities, with the potential solution of 20-mph schemes, as early as 1994. Twenty-miles-per-hour schemes were presented as the local street solution to the local speeding and accident problem. The concept of universal justice (higher accident rates in poorer communities), environmental impacts and pollution were problems also connected to the potential policy solution of 20-mph schemes:

Creating 20 mph zones around schools, which the innocent among us might think could be enacted almost overnight and which has been proven to save lives, is apparently hindered by a classic Catch 22 . . .

Pollution and danger on the walk to school, Scotland on Sunday, 30 October 1994.

Reproduced with permission from Scotland on Sunday

Pushback against this proposal and anti-20-mph speed limit reporting included a broader range of reasons to oppose the scheme, including problems with cost, maintenance, enforcement and even 20-mph zones causing more speeding as a result of driver frustration.

Politics stream

The politics stream was identified across the majority of articles post 2000. Themes included the political adoption and remessaging of evidence, supporting political activation and the agreement that 20-mph speed limits have worked (and worked elsewhere), riposted by anti-20 mph public arguments. Although these themes appeared similar to those found in the problems and policy streams, what made them also political was the reporting of these issues with a local or national political figure (Figure 12).

FIGURE 12.

Timeline (1993–2019) of Scottish media coverage by themes in the politics stream.

Political parties began to connect the problems of traffic and 20-mph schemes as a policy option to champion:

Robin Harper, Green Party, Lothian, expressed sympathy with the petition, but added: ‘I hope that every council in Scotland will introduce home-zones and 20 mph areas.’

Speed freeze, The Scotsman, 24 May 2000

The effect of political support for 20-mph schemes was reported within the policy stream, and the critical efforts of one Green Member of Scottish Parliament (MSP) to champion 20-mph schemes. The reporting quoted ‘the postcode lottery’ and inequality that some areas had these schemes already whereas others did not. This political advocacy and offering of a policy solution shows the interaction of both the problem and policy streams preceding the political stream in which legislation was agreed.

The benefits of 20-mph schemes were similar, but tended to be reported in clear groups: children and vulnerable users, protecting schools and saving money by stopping speeding and ‘rat runs’. The evidence quoted in articles from politicians tended to be in sound bites or short summaries, whereas those opposed to 20-mph schemes used emotive short anti-20-mph slogans.

Convergence of the three streams: the policy window

The theme of political activation reflected news and editorial articles that brought together evidence, the 20-mph bill and political support, leading to the convergence point of the problem, policy and politics streams (Figure 13).

FIGURE 13.

Themes within the problem, policy and politics streams from Scottish news articles, 1993–2019.

The underlying principle of this theme was the clear assumption and undisputed fact that 20-mph schemes work. This editorial shows this convergence and undeniable need for a 20-mph scheme in Edinburgh and beyond, published in 2018:

Several cities have launched or experimented with 20 mph limits or zones in built-up areas, and pressure is growing for Scotland as a whole to follow suit . . . The case for 20 mph limits cannot be overstated, but the same surely applies to the importance of such provisions.

Safer-roads bill deserves much wider support, The Herald (Glasgow), 2 October 2018.

Reproduced with permission from Newsquest Media Group

Strengths and limitations of the media study

There are some limitations to our study. We were unable to search the Edinburgh Evening News as it was not online. Despite extensive searching, we located very few articles from Belfast, compared with Edinburgh. We used the same search methods for both locations, but identified fewer publications related to Belfast and the environs. We are aware that the media is in itself biased and we cannot be sure if the media presented the real story of what happened. The media may present a simplified public view of what happened, or perhaps from only one perspective, with other local voices being excluded; however, we were able to find examples of different views (e.g. anti-20 mph) being reported. We acknowledge the limitation that we did not examine social media (although we have done this since the main project concluded, see Semwal, et al. 74), which may have given us greater insight into new issues or broader views. However, some articles were views from the paper’s social media activity. We feel that these missing media may not be so critical to Edinburgh; our analysis predates the decision to approve the Edinburgh scheme in 2015, whereas the largest expansion of social media use for adults aged > 30 years was post 2016. 75 Finally, we are aware that we bring our own biases and reflexivity to this analysis at all stages, but, by checking key decisions at each stage, this may be reduced. 76 The strength of our study is the combination of systematic searching, screening and coding frames, and use of a theoretical and practical framework to understand processes through the lens of a key social mechanism that shapes opinion and reflects views. Our study also sits as part of broader analysis of local documents to understand the political processes to create the 20-mph policies in both cities.

Objective 2

Objective 2 aimed to describe and understand the what and how of implementation (i.e. the implementation processes) of two 20-mph speed limit interventions: Edinburgh (citywide) and Belfast (city centre). We aimed to explore the role of the agents responsible for implementing the multiple intervention activities and to identify emerging changes in implementation.

The objective 2 research questions were as follows:

• How is the 20-mph speed limit intervention being implemented in each city?

• To what extent is the intervention being delivered as intended in each city, and what adaptations were made to how the interventions were delivered?

• What are the barriers to and facilitators of successful implementation in the two cities?

This chapter focuses on reporting the findings on the implementation of the different components [(see Table 1 and ‘Chapter 3, objective 2’ section of the programme theory (see Figure 8)] from both stakeholders and those living in the two cities. Based on our programme theory and informed by guidance on conducting process evaluations,56 we aimed to examine key processes related to 20-mph speed limit interventions. This is an area with a limited evidence base, but it is important to elucidate it to provide insight into how such interventions may work. These processes included the ‘what’ and ‘how’ of implementation of 20-mph speed limit interventions. Context was considered throughout the process evaluation,79 acknowledging that there were likely to be multiple contextual factors influencing the following: implementation; how the interventions were experienced and perceived; the impact on the outcomes assessed; and the mechanisms through which change may, or may not, have occurred. 78

Sampling and recruitment

We developed a sampling framework to identify all implementation agents likely to have been involved with the interventions, considered to be after the final passing of the traffic regulation order (or citywide speed limit order) legislation in each city (for Belfast: October 2015; for Edinburgh: January 2016). We initially used purposive sampling, aiming for representation from organisations involved in implementing the main intervention activities: legislation (e.g. city council transport convenors), road signage (e.g. council design teams, roads services and outsourced contractors), education and awareness-raising (e.g. council programme delivery teams) and enforcement (e.g. police services), accounting for the contextual differences in terms of stakeholders involved in the two cities. We also utilised snowball sampling, asking each interviewed stakeholder to identify any other relevant individuals or organisations.

Sampling and recruitment for focus groups

We developed a pragmatic sampling framework based on the concept of maximum variation80 and accounting for what was possible within the time frame and resources of the study. An important concept that underpinned this sampling was the potential for the exposure of the public to the 20-mph speed limit interventions (through residing in, working in or travelling through the cities), given the scale of implementation across Belfast (city centre) and Edinburgh (citywide). Thus, we endeavoured to sample by several variables and sociodemographic groups where the existing literature provided an indication that there may be a differential impact of the interventions, or to include different travel behaviour/modes. Key groups included different pavement and road users (e.g. pedestrians, cyclists and motor vehicle drivers such as professional bus and taxi drivers); those of different age groups (e.g. parents of children, older adults); and those from lower socioeconomic status areas. We also sampled by the geographical phased implementation within Edinburgh. Within each site, a range of recruitment methods was used to invite the potential participants from the target groups to a focus group discussion. These involved the following:

• contacting key individuals (i.e. previously identified key contacts and gatekeepers) known of within the research team

• contacting individuals who took part in the perceptions survey (see Chapter 4, Analysis of public perceptions) and who had provided consent to be contacted (Edinburgh only)

• community councils (as residents, not community council members)

• online databases of certain groups (e.g. school parent groups)

• snowballing to new groups’ key contacts from groups already conducted.

Recruitment continued until data saturation occurred. The final sample size was 159: 99 participants took part in 15 focus groups in Edinburgh and 60 participants took part in nine focus groups in Belfast (see Appendix 3 for full details of focus group participants).

In Edinburgh, 99 participants took part in 15 focus groups {mean age 56.2 years [standard deviation (SD) 16.8 years]}. Fifty-nine participants were male (59%), and the majority of participants were white (n = 92, 92%). Nearly all participants identified as having no disability or medical conditions affecting their mobility (n = 97, 98%). Participants included advanced motorcyclists (n = 11), public transport users (n = 6), car/private motorised vehicle users (n = 7), people using active travel modes (n = 3) and cyclists of mixed experience (n = 13). One group consisted of taxi drivers (n = 4), and another of bus drivers and a bus company management representative (n = 3). Participants also included older adults (n = 19), parents of toddlers (n = 7) and parents of school-aged children (n = 6). Finally, participants also included those from a semirural location within the intervention area (n = 9), and community councillors and residents from South-Central Edinburgh (n = 6), and the west of Edinburgh (n = 5). It was ensured that participants covered a wide geographic spread of the intervention area, to ensure an understanding of any geographical differences in intervention experience among participants.

Stakeholder interviews

The interview schedule (see Appendices 4 and 5 for the Edinburgh and Belfast interview guides, respectively) was guided by the researchers’ understanding of the implementation of 20-mph speed limit interventions achieved through the development of the programme theory3 literature and policy documents, and informal contact with implementation agents in both sites. The intervention activities (signage, education and awareness-raising, enforcement) were explored in relation to what each consisted of, how each had been implemented, changes over time, and whether or not these activities had been delivered as intended (fidelity). We investigated barriers to and enablers of implementation and considered setting and contextual differences between Belfast and Edinburgh.

Focus groups

The focus group schedule (see Appendix 6) was guided primarily by the overall project programme theory. 3 In general, the focus groups sought to understand the public’s attitudes to, and perceptions of, 20-mph speed limits, and their experiences of, and behavioural responses to, the implemented 20-mph speed limit intervention activities, exploring the mechanisms through which any behaviour change occurred. The key health-related outcomes of interest included traffic speed, active transport (walking and/or cycling), and liveability. Specifically, the topics included in the guide were as follows:

• awareness of the 20-mph speed limit intervention and perceived rationale

• attitudes towards the 20-mph speed limit intervention

• views on, and experiences of, the 20-mph signage

• views on, and experiences of, enforcement of the 20-mph speed limit intervention

• behaviour change across a range of outcomes (see Chapter 4)

• views on the impact on liveability (see Chapter 5).

The topic guide was amended iteratively, initially after the first focus group following discussions with the research team, and then tailored for different groups focusing on specific behaviours of interest (e.g. driver behaviour among professional drivers, liveability among young parents’ groups). Questions from several iterations of the perceptions survey (see Chapter 4) were also incorporated into the topic guide.

Signage and road-marking activities

Education and awareness-raising activities

Enforcement activities

Partnership involvement and working: stakeholders

A key theme generated relating to implementation from the stakeholders’ perspective was partnership involvement and working. This theme was important in describing the role of the implementation agents involved, but was also a key barrier to or enabler of implementation, depending on the city and agents who discussed it.

In Edinburgh, there were multiple implementation agents engaged in delivering the intervention activities. The process started with the creation of a dedicated ’20-mph team’ (a programme manager and a professional officer) within the local authority with associated ring-fenced funding. This team was viewed as being essential, given the scale of the intervention:

. . . basically, it was a much bigger job to do it to the whole city, so it required a separate team, it was more than our team could handle, so there’s a separate team dealing with that . . . It’s just, literally, workload logistics, all sorts of things, you need somebody who is overseeing the whole thing and not doing other things.

Edinburgh, council

The ‘20-mph team’ worked closely with teams within the transport department, including media and communications, active travel, and road safety, and with external partners (i.e. outsourced contractors, Police Scotland, Sustrans and a range of local partners). In addition, a 20-mph steering group and a specific 20-mph communications steering group were created in the local authority to guide implementation and to share ideas, resources and networks. These agents worked closely on the different intervention activities, developing strong relationships and establishing clear lines of communication. This partnership working was consistently cited as an important enabler of effective implementation of intervention activities, ensuring that efforts were in tandem and supportive:

I think the relationship that we’ve had with schools has helped, actually. Partly because we have two officers, road safety officers, who liaise with the schools on a regular basis. So we’ve been able to tap in to that audience really well.

Edinburgh, council

I think so. It’s certainly been done in partnership really well. You know, it’s one o’ the main things I would think back of over the last few years, other than some significant events that have been run in Edinburgh where there’s been that real cohesive work between the local authority and the police to push a safety programme . . . Yeah, I think so. It’s been really first class, actually. And it’s been hand in glove. It’s not something the council have implemented on their own and we’ve come along later on to look at the bit that ‘Oh, we’ll be enforcing some of this’. Actually, it’s not – it’s been a sort o’ real partnership approach and I do think they’ve done a first-class job around that.

Edinburgh, police

In Belfast there was no direct involvement from the local council or external organisations in implementing the intervention, potentially owing (in part) to the intervention scale. Implementation was solely under the control of the roads team in the Department for Instructure. There was evidence that the creation of a dedicated and collaborative team would have acted as an important implementation enabler, providing additional resources and knowledge. This lack of stakeholder involvement was perceived as a barrier to implementation of some of the intervention activities (i.e. the education and awareness-raising campaign). The formation of a steering group was also viewed as being beneficial to guide implementation, create public ‘buy-in’ and counter any noted negative public perceptions:

If I was doing it again, I think what would be useful to have would be almost a small steering group committee which ensured a better buy-in. They wouldn’t have done any work but you could say ‘oh this was implemented by a steering group comprising of Police Service [of] Northern Ireland, Department for Infrastructure, Belfast City Council, Belfast Chamber of Trade and Commerce, Belfast City Centre’ . . . Yes, because it wouldn’t be big government . . . it would be this is not just the road service, this was all of those organisations saying this is a good idea.

Belfast, government department

Other influencing factors: stakeholders

There were several factors identified as having the potential to influence the 20-mph speed limit interventions. From a political context, the Belfast intervention was implemented when the Northern Ireland Assembly was not in session, an issue described by agents as a barrier to sustainable transport promotion:

Unfortunately, it would be the budget situation that would keep us back, because without a government it’s really difficult . . . with the budget being what it is, we have a difficulty around pace [of sustainable transport delivery].

Belfast, government department

In addition, other transport interventions (i.e. the Belfast Rapid Transit scheme, and the Belfast on the Move project) were being delivered during the implementation of 20-mph speed limit interventions, with many schemes involving several of the same agents as those implementing 20-mph speed limits, such as the local authority and the police:

. . . the impact of the Belfast on the Move project, which was we’re putting in more bus lanes and cycle lanes and all, it included measures to deter through traffic in Belfast, it constructed a new link road just to keep traffic away from the city centre and that has resulted in reduced traffic levels.

Belfast, government department

I think it’s [police involvement with cycling safety] been driven unintentionally probably by all the extra work we’ve done on the 20s, even doing the Close Pass, we’re doing a whole range of cycle safety things and the like, that actually it’s just a . . . we’ve got a sort of nice overview in the city.

Edinburgh, police

Strengths and limitations

One of the strengths of the qualitative work is the range of implementation agents recruited from both cities, which allowed us to examine in-depth how the interventions were implemented and who was involved, as well as how the interventions were experienced by those affected (i.e. residents and commuters within the cities). This is an important strength, as examination of the processes involved in implementation of 20-mph speed limit interventions has not been a focus of previous evaluations. 22,77 Our programme theory allowed for clear identification of certain processes thought to be integral to implementation to be examined, acknowledging that it may not be possible to investigate all plausible causal assumptions, or uncertainties, around a complex intervention. Thus, we adapted interview schedules and topic guides in an iterative manner, focusing on specific areas of the programme theory and allowing different implementer roles and population subgroups and contexts to be considered. These nuanced, rich data are complementary to the quantitative impact and outcome evaluation, as described in Chapter 4.

We acknowledge a number of limitations. The process and outcome teams were not fully separated, with some researchers involved in both, and thus aware of emerging outcomes. However, it was not possible to delay outcome analysis until the process evaluation was complete, and in fact this was not desirable as, originally, it was anticipated that focus group sampling would be informed by differential impacts of the 20-mph interventions (i.e. by geographical area of population subgroup). This sampling by differential impacts on key health outcomes was ultimately not possible. We also found it challenging to recruit groups from socioeconomically deprived areas; thus it was not possible to qualitatively examine the impact of the interventions on health inequalities as originally intended. A further limitation was that, as noted, funding for this evaluation was received post implementation in Belfast and during implementation in Edinburgh; therefore, this may have led to an element of recall bias.

Methods

The Speed Limits Perceptions Survey (SLiPS) developed for this study is provided in Appendix 7. Although the perception items were similar to those in the questionnaire developed by Tapp et al. ,87 additional sections on liveability, mode of transport and perceptions of 20-mph speed limits were developed using focus group data from Turner et al. 3 The survey was also adapted to investigate changes in relation to a specific intervention and to be more inclusive of non-drivers.

As the project was not funded until after the 20-mph limits had been implemented in Belfast and part way through the implementation in Edinburgh, it was possible to collect pre-intervention perceptions data in only three of the implementation zones in Edinburgh (implementation zones 4, 5 and 6). Two of these implementation zones were next to each other (zones 4 and 5) and the policy was implemented in them on the same date; therefore, we treated them as a single site for the perceptions survey. The study (which used the perceptions survey) was reviewed and approved by the Moray House School of Education Ethics Subcommittee, University of Edinburgh (reference number 1114).

In Edinburgh data collection was repeated 6 and 12 months after implementation in three areas to explore long-term changes. In Belfast, the data were collected 2 years after the speed limit change. At each data collection time point, we aimed for 500 participants. To efficiently collect this number of data within a defined geography, it was decided to recruit visitors to amenities within each implementation zone in Edinburgh, in particular large shopping centres, sports/leisure centres and a hospital. In Belfast, data collection took place in the city centre and public transport hub. Dates of data collection and type of location are provided in Appendix 8, with local university students trained to collect the data. Prior to completing the survey, written informed consent was gathered from each participant. Participants were required to be aged > 16 years, and to able to provide consent and complete the survey in English. Assistance was provided if participants had reading or writing difficulties. Entry into a competition to win a £100 shopping voucher was the incentive offered at each data collection.

Each data collection was cross-sectional in line with the objective to evaluate public opinion, not to assess changes of opinions within individuals. Although this was an efficient method for collecting the data, the systematic random sampling, ordered by urban–rural classification, Scottish and Northern Irish Index of Multiple Deprivation rank and postcode, as specified in the protocol, was not feasible. Subsequently, the plan was to weight the responses received to support the validity of the comparisons between survey waves. These weightings would account for differences between samples and would not attempt to reflect the demographics of each city.

Confirmatory factor analysis had been conducted during the development of the survey to test the assessment of support for the policy. An exploratory factor analysis of the baseline pre-intervention responses was conducted as part of the final analysis of the survey. The SLiPS comprised 53 questions, 39 of which related to perceptions; therefore, factor analysis helped to minimise the number of variables for analysis. The exploratory factor analysis also revealed the main public perceptions related to 20-mph limits in each city within the responses. The challenges faced in the evaluation of liveability and active travel outcomes led to the inclusion of additional items in the perceptions survey on these outcomes. However, as these were not included until after first data collection, the responses could not be included in the factor analysis. When testing whether or not the inclusion of these additional items altered the factors identified, it was found that the variables loading into each factor remained consistent, with the loadings altering by, at most, ± 0.15. As the survey responses were either binary or ordinal, the factor analysis used a polychoric correlation matrix and promax rotation.

Perception findings: Edinburgh

Across the two sites and three time points in Edinburgh, survey responses were collected from 3485 members of the public. Once linked through their home postcode to the Scottish Index of Multiple Deprivation of their data zone, complete data on the perceptions questions were available for 2253 (64.6%) of respondents. The demographic characteristics and travel habits of participants are summarised in Table 7. Full details regarding the responses to each of the perceptions questions are provided in Appendix 9.

The polychoric correlation matrix of the perceptions responses is given in Appendix 10. Five factors were identified with eigenvalues of > 1. The loadings for each question into each factor are provided in Appendix 10. The table in Appendix 12 lists the loading values for each question into the five factors; when a question made only a minor contribution to the factor (loading value of < 0.3), it was removed, to highlight the questions contributing most to each factor. Factor names were created based on the questions loading into each factor and the direction of the loading. Those perceptions assessed on a five-point Likert scale were scored from 1 ‘strongly agree’ to 5 ‘strongly disagree’, meaning that negative loadings reflect agreement with the perception statement.

The five names chosen for the factors were as follows:

• factor 1 – detraction and resistance

• factor 2 – support (for the implementation of 20-mph speed limits)

• factor 3 – rule-following

• factor 4 – child safety

• factor 5 – walking safety.

With the factors deriving from categorical survey responses, the five factors were not normally distributed; consequently, it was not possible to apply weightings to increase the validity of the comparisons between survey waves. Kruskal–Wallis tests were used to compare perception factor scores across the three time points (Figure 14). Detraction and resistance (factor 1) decreased, while support (factor 2) and rule-following (factor 3) increased. Factors 4 and 5, child and walking safety, did not change significantly.

FIGURE 14.

Box plots of each perception factor at baseline and at 6 and 12 months. p-values based on Kruskal–Wallis tests.

Perception findings: Belfast

Survey responses were collected from 490 members of the public. The mean age of the sample was 37.43 years (SD 17.22 years), with a median age of 32.00 years. The majority of participants were under the age of 40 years (n = 273, 55.7%), female (n = 212, 43.3%), with no disability or medical conditions (n = 381, 77.8%) and reported their ethnicity to be white (n = 402, 82.0%) (for complete details, see Appendix 11).

The majority of participants (62.1%, n = 304) reported that the 20-mph speed limits would make people drive more slowly (strongly, or tend to, agree), but 58.7% (n = 288) agreed that 20-mph speed limits were needed only where safety is an issue. The majority (60.8%, n = 298) also agreed that, although the speed limits will take time to get used to, they will eventually be accepted as the normal speed. There were mixed views as to whether or not a 20-mph speed limit would increase cycling and walking, but there was a slight tendency towards perceiving that it would (see Table 8 and Appendix 12).

Participant reports were generally positive in response to traffic-related outcomes, including agreeing that there would be a decrease in the number of collisions (46.8%, n = 229) and in the number of severe collisions (53.1%, n = 261), and better traffic flow (31.2%, n = 153). However, there was a tendency to think that there would be no difference in vehicle noise. For three of the (unintended) traffic-related outcomes, the majority of participants had no strong views on whether there would be more congestion (36.5%, n = 179), more air pollution (41.0%, n = 201) or a decrease in fuel efficiency (45.1%, n = 221). In contrast, 52.5% (n = 258) of participants agreed that journey times would be longer with 20-mph speed limits (Table 9).

The majority of respondents reported that the 20-mph speed limit would increase how pleasant the area is to live or work in (38.5%, n = 189) and make the streets safer (66.1%, n = 324), but were ambivalent as to whether or not it would provide more opportunities to socialise (43.9%, n = 215). In terms of active travel, approximately 10% said that they would cycle more, walk more or let children walk more if more roads had 20-mph speed limits

Comparing the findings of the perceptions surveys in Edinburgh and Belfast

Despite the absence of baseline data from Belfast, the collection of the data 2 years after the implementation in Belfast helped to gain some insights into the longer-term impact of 20-mph limits. Across the statements reported in Table 9, the responses from Belfast were less negative (strongly disagree) than those from Edinburgh (at any time point), usually with an increase in ‘strongly agree’ responses. However, there was also an increase in ‘neither agree nor disagree’ responses, particularly in terms of cycling, walking and collisions, for which the respondents in Belfast were more likely than those in Edinburgh to agree that there had been a positive change. For the statements related to traffic flow, noise, congestion, air pollution, journey times and fuel efficiency, the responses from Edinburgh showed more agreement with these statements, whereas Belfast respondents were more ambivalent. When the factor loadings from Edinburgh were applied to the Belfast data, detraction and resistance scores were lower in Belfast and support scores were higher, but the other factors did not differ markedly. These comparisons could reflect an ongoing positive shift in the norms relating to 20-mph speeds. However, they might also reflect sociocultural differences in the cities, or the distinct effect of the different scales of intervention. The limited area of the Belfast intervention meant that respondents had a lower dose of the intervention than people in Edinburgh.

Strengths and limitations of the perceptions survey

The SLiPS offers a robust method for evaluating relevant public perceptions of speed limit interventions. Repeat data collection before and after (in three implementation areas) the policy change has provided useful insights into the changes in public perceptions, although this was not possible in Belfast. The data collection methods and sample size limited the extent to which the data could be weighted or considered representative of the city population. Consequently, we cannot be highly confident that the changes (or lack of) in perception observed are not the result of differing samples at each time point. However, the use of repeat cross-sectional data collection is in line with the objective to assess public perceptions. Finally, prior to the perceptions survey data collection in Edinburgh, > 50% of roads in the city had historically been 20 mph, and three zones had been converted to 20 mph as part of the intervention being evaluated. Consequently, at baseline we were not assessing the perceptions of people who had not experienced the intervention. Therefore, different or larger changes in perceptions may be observed in locations where 20-mph limits were being implemented for the first time.

Methods

The overall statistical summaries employed in this trial included measures of central tendency, dispersion and basic t-tests. More statistical summaries are presented for Edinburgh because more streets were analysed, and the data set was richer.

Results for speed and volume: Edinburgh

Overall city-level changes

The mean and median speeds reduced by 1.34 mph and 0.47 mph, respectively, when comparing pre with post data at 12 months. There were also reductions across the distribution of speeds [interquartile range, and range (maximum to minimum)]. The greatest reductions were observed for the maximum speed observed (1.79 mph) and the third quartile (1.78 mph). A box plot summary of the speed distributions at a city level before and after implementation is shown in Figure 15. This illustrates the way the distribution of speeds shifted ‘down’ to reflect lower post-implementation speed.

FIGURE 15.

Box plots for before-and-after speed distributions using the combined records for all 66 monitored streets.

These distributions are further presented as frequency plots at 1-mph intervals in Figure 16. The speed distribution both before and after the 20-mph speed limits appears to be bimodal. Comparison of the distributions shows how the right-hand peak has shifted left in the post implementation distribution, indicating reductions in the frequency of higher traffic speeds being observed. It also suggests that the left-hand peak has shifted from 22 mph to 20 mph, indicating a reduction in speeds at these lower speeds.

FIGURE 16.

Histogram of average speeds on 20-mph streets in Edinburgh for each of the 66 monitored sites (n = 12,672 observations): (a) before 20-mph speed limits; and (b) after 20-mph speed limits.

In terms of traffic volume, the results show an average, per measurement site, of 3641, a reduction of 87 in the average number of vehicles per day at each site. Overall, the difference in vehicular volume after the speed limit implementation was minimal (reduction of 2.4%) and non-significant (Table 10).

TABLE 10 - Summary statistics of (7-day) average vehicle volume per site

Category	Mean	SD	Median	Interquartile range (first to third quartiles)	Minimum, maximum
Volume before	3641	2633	3738	743–5862	144, 9343
Volume after	3555	2592	3391	716–5860	154, 9788
Difference	–87	810	–347	First quartile: –27	Minimum: 10
				Third quartile: –2	Maximum: 446

Changes in speed by day of week and time of day

When the times were categorised into periods of the day (early morning, working day and night), the reductions in average speed were again found to be very similar in magnitude. Specifically, the average pre–post differences in speed for early morning, working day and night were found to be –1.29 mph [95% confidence interval (CI) –1.43 to –1.16 mph], –1.35 mph (95% CI –1.42 to –1.28 mph) and –1.36 mph (95% CI –1.45 to –1.26 mph), respectively. Figure 17 shows the change in average speed with time of day treated as a continuous variable. Here, it is apparent that relatively higher average speeds are observed between the times of 01:00 and 06:00. Figure 17 also depicts the trend in the percentage difference in the average speed before and after implementation, with the greatest variation occurring between 01:00 and 06:00.

FIGURE 17.

(a) Average speeds on 20-mph streets in Edinburgh, with time of day as a continuous variable (from 00:00 to 23:45 by 15-minute intervals); and (b) percentage difference in average speeds before and after implementation. Note that, for each plot, the data were aggregated over 96 time points by two time frames (n = 192).

Speed and volume findings: Belfast

Strengths and limitations of speed and volume analyses

For Edinburgh, it was not possible to include comparison or control streets in the analysis. If this were possible, it could have helped to identify the independent effects of the intervention. The analysis here cannot rule out that changes in speeds could have happened anyway (secular trends), or determine how much of the observed changes can be attributed to the intervention (rather than other events or interventions within the time frame under analysis).

However, the theory-based approach that we utilised in our evaluation, the pre-stated protocol52 and the logical proximal relationship between speed limits changing and traffic speed changing mean that the results presented here add substantial weight to the evidence base for 20-mph limits leading to reduced speeds at 12 months.

Methods

The methods employed for this study were as follows:

• monthly comparisons of the mean number of casualties and collisions before and after the 20-mph limit change in Edinburgh

• data visualisations of trends

• rate calculations

• modelling of road traffic collisions and casualties (Edinburgh).

The data visualisations focused on the time series of road traffic collisions and casualties, and on the trend in collisions by time of day. Rate calculations were performed to compare collision rates overall and at various aggregations, such as implementation zones and road categories. For casualties, rate calculations were performed by level of casualty severity and for vulnerable groups.

The models considered in this study are Elvik’s power and exponential models,89 generalised linear models, generalised additive models90 and hidden Markov models. 91 The Elvik models estimate the rate of collisions given the observed change in traffic speed, whereas the generalised additive models and the hidden Markov models focus on the features of the overall temporal trend of road traffic collisions. The generalised linear model was used to model the observed road traffic casualties with key explanatory variables being age of driver, accident severity and time of day. Overall, the approach taken for analysing road traffic collisions and casualties is multipronged, with analysis of trends, rates and counts, with data visualisation.

As collision and casualty data were collected across the whole of Scotland and Northern Ireland, it was possible to identify control zones. As described in the protocol, the control zones were matched using the separate domains for the Scottish Index of Multiple Deprivation (2016)92 and the Northern Ireland Multiple Deprivation Measure (2017),93 the urban–rural classification of the area and the population density. As each of the intervention areas include more than one data zone (Scotland), or output or small area (Northern Ireland), the mean, median or mode of each of the matching variables was used depending on the nature and distribution of the variable.

The 20-mph limits in Belfast were implemented across roads in 10 small areas and 11 output areas; matching across the variables listed above identified Ward Central as the most appropriate control zone. The implementation zones in Edinburgh varied in size from being equivalent to 21 data zones up to 173 data zones. Owing to their size, it was not possible to find a matched control for the largest implementation zones, South Central/East. But approximately suitable matches were found for the other implementation zones, allowing the evaluation of collisions and casualties to be compared against secular trends from matched areas (difference in differences).

Results: Edinburgh

Here we discuss the results relating to the observed road traffic collisions and casualties for the study period. For the citywide calculations, the number of years considered pre 20-mph speed limits is 3 years, and the number of years considered after 20-mph speed limits is 3 years and 10 months. The following results build a narrative to link the observed changes in traffic speed with observed changes in the number of road traffic collisions and casualties within a polyphonic evaluation process. The results (Table 16) found that there were statistically significant reductions in counts of casualties and collisions across road types and speed limits, with the largest reduction being observed for main streets that switched to 20 mph. In particular, we note a statistically significant average reduction of 14 collisions (and 16 casualties) on main roads that switched to 20 mph, four collisions (and five casualties) on local roads that switched to 20 mph and eight collisions (and nine casualties) on roads that remained at 30 mph.

Overall, a 40% reduction in road traffic collision rates was observed when considering the pre-20-mph period of 3 years (2013–15), and for the post-20-mph period of 3.72 years (2016–19). When the secular trend within the pre- and post-20-mph periods was incorporated, the resulting percentage change in road traffic collisions was 36%.

We employed the impact model (power and exponential models) proposed by Elvik89 to investigate the causal link between traffic speed reduction and road traffic collision rates. Both models, power and exponent, indicated a reduction in road traffic collisions. The predicted percentage rate reductions are 11% and 1% for the exponent and power models, respectively. For these models, the prediction of road traffic collision rates is lower than that observed; these predictions, however, provide a ‘finger on the pulse’ on the direction of impact of the observed changes in traffic speed. A larger reduction in collisions and casualties than would be expected based on the change in speed is supportive of an intervention effect.

When considering the individual road types and speed limits within the city of Edinburgh, we note that the roads that retained limits of 30 mph experienced the smallest (in magnitude) percentage reduction in rates. The largest percentage reduction in rates was observed for 20-mph main roads. The road types and speed limits in this comparison were 20-mph existing streets, 20-mph residential streets, 20-mph main streets and 30-mph streets. In terms of the time of day, the largest percentage reduction in road traffic collision rates was observed for the period 08:00–18:00 (45%).

Of the seven implementation zones considered, the largest percentage reduction in road traffic collision rates was observed in zones 5 and 6, with 47% and 45% reductions, respectively. Zones 5 and 6 represent Edinburgh West and Edinburgh South. Reductions in road traffic collisions were observed for all the implementation zones.

We employed two different approaches to model the trend and seasonality of the road traffic collisions as well as to investigate the effect of the introduction of the 20-mph policy. 94 In our analysis we detected a switch from a slow-declining to a fast-declining state shortly after the introduction of the 20-mph policy. In addition, we detected a faster decline in road traffic collisions post 20 mph with a significant ‘implementation effect’ observed when the intervention was modelled using the generalised additive models referred to previously.

The main findings were that, when controlling for other variables in the model, the log of the number of road traffic casualties (1) increased by 0.14 when considering males (vs. females), (2) decreased by 0.005 when age of casualty considered was increased by 1 year, (3) reduced by 0.723 when considering rural locations (vs. urban), (4) decreased by 0.738 after the speed limit implementation (vs. before the implementation) and (5) decreased by 0.344 when considering local streets that switched from a limit of 30 mph to one of 20 mph (vs. streets that remained at 30 mph before and after the 20-mph speed limit implementation). Further details can be found in Table 17.

After the 20-mph speed limit implementation, there was also a reduction in the number of road traffic casualties (Table 18). Reductions in casualty rates were observed citywide, at each casualty severity level, and for each vulnerable road user considered in this study. In particular, the overall percentage reduction in casualty rates was 39% (40% was observed for collision rates). The percentage reduction for each level of severity was 23%, 33% and 37% for casualty levels fatal, serious and slight, respectively.

Results: Belfast

Overall, there was a reduction of 2% in collision rates in Belfast when considering the study time frames (Table 19). Reductions in road traffic casualties were also observed for fatal and slight severities, and for pedestrians. Interestingly, an increase in road traffic casualties for motorcyclists and pedal cyclists was observed.

Strengths and limitations

The nature of this study, being a natural experiment evaluation, presents researchers with the need to acknowledge confounding variables and make adjustments in analyses when possible. The main confounder in this natural experiment is related to the citywide efforts to increase active travel during the study period. The attribution of a quantified impact of these efforts on road traffic collisions and casualties has been a challenge as representative data on active travel for the purposes of this study were not available.

Further work will involve a multiagency effort at designing efficient data collection of active travel data using emerging technology such as used in the ‘track and trace’ NHS applications. We are in the process of seeking funding with other collaborators on this venture. The data generated will facilitate the incorporation of a proxy for the attribution of the impact of active travel promotions on road traffic collision and casualties. In addition, the incorporation of prior knowledge such as estimates from Elvik’s models and from relevant systematic reviews within a Bayesian framework will allow for a broader modelling approach to the evaluation of the impact of 20-mph speed limits on road traffic collisions.

Methods

Liveability findings

For Edinburgh, GSV data were available for 91 of the 100 streets (91.0%) pre intervention, 80 of the 100 streets (80.0%) post intervention and a total of 76 streets (76.0%) at both time points (i.e. pre and post intervention).

Results showed that the mean liveability score in Edinburgh prior to the implementation of the 20-mph speed limit intervention was 56.80 (SD 8.13) (Table 20). This score was found to have increased significantly, to 58.93 (SD 8.92), when assessed post intervention. The mean change between time points was 2.38 (SD 3.31), which was found to be significant at the level of p = 0.000 (see Table 20).

When the analysis was performed for each of the nine constructs of liveability, three constructs were found to have statistically increased from pre to post intervention: (1) places (i.e. presence of amenities and facilities, and building maintenance and aesthetics) (p = 0.017), (2) traffic/transport (as above) (p = 0.000) and (3) pavements (i.e. street amenities, litter, obstructions to walking, street lights, crossings, pavement presence and quality, buffer presence and quality, and pavement obstructions) (p = 0.036) (see Table 20). Finally, when the proxy indicators of behaviour were investigated, no significant changes were found (p > 0.05) (Table 21).

For Belfast, GSV data were available for 68 of the 76 streets (89.5%) pre intervention, 73 of the 76 streets (96.1%) post intervention and 68 streets (89.5%) at both time points (i.e. pre and post intervention). The results showed that the mean liveability score in Belfast prior to the implementation of the 20-mph speed limit intervention was 65.93 (SD 18.31) (see Table 20). This score was found to increase significantly, to 66.22 (SD 18.52), when assessed post intervention. The mean change between time points was 1.38 (SD 3.08), which was significant at the level of p = 0.000.

When analysis was performed for each of the nine constructs of liveability, (1) traffic/transport (i.e. presence of public transport facilities, speed signage, speed-calming measures and traffic lanes) (p = 0.000) and (2) places (i.e. presence of amenities and facilities and building maintenance and aesthetics) (p = 0.041) were found to have increased significantly, pre to post intervention (see Table 20). Finally, when the proxy indicators of behaviour were investigated, a significant increase was found in relation to bicycle rack capacity (p = 0.028) and significant decreases were found for the number of parked cars (p = 0.037) and the total number of cars (i.e. moving and parked) (p = 0.035) (see Table 21).

Strengths and limitations

Strengths include the implementation of MAPS-Liveability, a reliable tool specifically designed to assess liveability at the level of the street and to include information regarding microlevel environmental features. In addition, as MAPS-Liveability was found to have good–excellent reliability for use with GSV, this enabled historic, pre-intervention implementation data to be collected, allowing pre- and post-intervention comparisons to be performed. Furthermore, each of the 76 20-mph streets in Belfast, and a stratified sample of 20-mph streets in Edinburgh, was assessed for liveability, using the nine liveability constructs and the proxy indicators of behaviour. Limitations of the study using GSV include the following: it is not possible to assess variables such as air or noise pollution, or perceived safety; temporality (i.e. an image is from one point in time and it is not possible to know time, day or date); and contemporality (i.e. time lapses between GSV images and data collection).

Driving behaviour

Overall, there were inconsistent views concerning the impact that the 20-mph intervention had on driving behaviour. In Edinburgh, there was some evidence of a positive impact on driver behaviour, with several participants reporting that they were adhering to the limits, expressed by some as simply sticking to the rules. A small number of participants discussed utilising features within the car such as ‘cruise control’ and satellite navigation systems to aid in monitoring speed and adhering to the speed limit.

In Belfast, there was no evidence that driving behaviour had changed, and in Edinburgh some participants also reported not adhering to the limits. There were several reasons proposed for this, including that 20 mph was perceived as being a difficult and unnatural speed to drive at, and road conditions, driver skill and knowledge were deemed to be more important factors influencing driver speed through the areas of new limits:

I think it is, especially if the roads are clear and it’s an off-peak time, to drive at 20 mph is very frustrating. You think, ‘why am I going so slow? There’s nobody here to go slow for’. That puts me off a wee bit.

Belfast, city-centre worker

Some participants also reported a social pressure to drive at speeds in excess of 20 mph, thus affecting their compliance with the limit:

At the moment I think it’s ’cause so many other people are doing it [driving at > 20 mph], you don’t feel bad, you know, going above the speed limit ’cause you feel like you have to. You have to go above speed limit for people behind you.

Edinburgh, FG11

There were also mixed reports from professional drivers in Edinburgh, with some stating that they had not adhered to the new limit, whereas others viewed driving to the local limits and conditions as part of their role: ‘[w]e have to slow down. It’s our job. We have to do that’ (Edinburgh, FG15).

There was evidence that the intervention had increased people’s awareness of both their own driving behaviour and the driving behaviour of others. In relation to perceptions of other drivers’ behaviour, there was a consistent, but not conclusive, view from participants that other drivers were adhering to the limits, particularly in certain areas such as residential streets. Again consistently, it was perceived that driving at precisely 20 mph was being done by only a minority, but what the intervention had succeeding in doing was reducing the overall traffic speed within the city by a smaller extent, often from a speed that had been in excess of the previous limit:

I didn’t note carefully what speed people drove at beforehand, but I have a feeling it was, you know, it was maybe, well often it would be 30, a wee bit over 30, you know. My . . . you know, having been a wee bit more observant of what speed people are actually driving at since it came in, I would say quite often it’s 23, 24, 25.

Edinburgh, FG3

As noted above, there was no evidence of 100% compliance with driving at a reduced speed (i.e. not all drivers were adhering to the new limits). Several conflicting views were expressed about this. For example, some professional drivers stated that they complied with the speed limit, but perceived the general public not to, whereas members of the public expressed an opposing view that professional drivers were among the worst offenders for not adhering to the reduced speed limit:

. . . and the buses are under huge pressure to stick to their timetables, and they don’t go at 20, and it would be safer for cyclists if the buses did go at 20.

So they go higher than 20?

Oh hell, aye, regularly.

In both cities, there was also the perception that driving behaviour had worsened as an unanticipated consequence of the 20-mph interventions. For example, participants discussed observing an increase in overtaking as a result of increased frustration with the slower-moving traffic, with the perception that this may have a negative impact on the number of collisions and casualties:

If everyone was doing 30, you wouldn’t get the overtaking that you do when people are doing 20 [. . .] if we did 20 mph, people would pass us.

Edinburgh, FG14

Yes, they’re pushing you on, beeping the horn or getting road rage or speeding past you.

Belfast, multimodal transport user

This issue involved both motor vehicle drivers and cyclists, with the reduced on-road speed differential causing potentially dangerous interactions and overtaking situations.

Walking and cycling

From the qualitative discussions, there was almost no evidence that the 20-mph interventions had affected levels of walking and cycling in either city. In both Edinburgh and Belfast groups, there was a small amount of discussion that the reduced speed limit had created a more pleasant cycling experience through the previously discussed reduced speed differential between cyclists and motorists, and an increased confidence to ride in the primary position on-road:

I would like to think I’m more confident, I never liked that really before, that route, I would like to think I’m a bit more, maybe I’m more aggressive now with the 20 mile, I feel I can sit on the road more, but no, I do use that more, I think perhaps, certainly much more than I would have done.

Edinburgh, FG3

As a cyclist, it reassures you a wee bit more, because you’ll generally be travelling 15–20, max[imum], I’d say, on a bike. It gives you a bit more assurance that you’re a bit safer.

Belfast, cyclists

However, there was no evidence from participants that the introduction of the 20-mph speed limits had led to any substantial increase in cycling, through either an increase in journeys in existing cyclists or new cyclists. Only one participant commented that a friend had taken up cycling as a result of the implementation of the 20-mph speed limit. Participants cited continuing concerns over safety, and the need for investment in cycling infrastructure (such as segregated cycle lines), as factors of greater influence around decisions on whether or not to cycle through the cities:

Going at 20 miles an hour, that is not going to make me a cyclist, I can tell you, never am I going to get on a bike, because I’ve seen the buses, I’ve seen how people drive, I’m not going to put myself at that risk.

Edinburgh, FG8

Participants in Edinburgh reported being aware of, or perceiving, an increase in the level of cyclists during the time period of the implementation of 20-mph speed limits. However, it was viewed that this was likely to be attributable to the range of other cycling promotion initiatives introduced in the city, rather than being solely attributable to the 20-mph intervention. Similarly, there was no evidence from either city that the 20-mph intervention had resulted in an increase in walking, either in terms of participants’ own behaviour or a wider perception of the behaviour of others. A minority of participants, mainly from parent groups in Edinburgh, reported that the intervention had led to a safer environment that was more conducive to walking:

I guess, feeling a bit better about everyone’s safety when we’re going somewhere, you know, around Parliament, or anywhere around there you know. You don’t have to handgrip everyone extra tight, you know, I do feel that everyone, the kids are a little bit more at ease, yeah. I think in general; I feel a bit better.

Edinburgh, FG12

However, this view was not a dominant one across the groups, with several participants stating that their experience as a pedestrian had not changed.

Overall, there was less discussion of walking and pedestrian behaviour than of cycling, with the purported positive impact of the reduced speed limit not immediately apparent to many. Indeed, in both Edinburgh and Belfast, several participants reported crossing the road at non-designated crossing points, or had observed other pedestrians doing so, because of a perception of a slower driving speed. Although the ability to cross the road in this manner was perceived positively by some pedestrians, this increased risk-taking was viewed negatively by some drivers, acting as a source of frustration and as a safety concern. Drivers expressed the view that pedestrians should have to follow road safety guidelines and laws, just as the driver has to, and they should be penalised for walking in front of cars.

Pre implementation and process of adopting the 20-mph limits

There was evidence that external factors, such as the political context and leadership, influenced decision-making and implementation in the two cities. Although there were similarities in the processes involved in the decision-making and implementation of the 20-mph speed limit interventions in Edinburgh and Belfast, there were also major differences between the two cities. The contrast between the two cities demonstrates that it is possible to implement 20-mph speed limits in varying ways, and that local context is critical for informing how 20-mph speed limits might best be operationalised in a given location. The media analysis shows that there is already relevant narrative in the media long before these speed reduction schemes exist. By understanding what are the arguments and positions in place, local jurisdictions can manage their paths to implementation with clear and coherent media and communication strategies. This would now also include social media.

Implementation

The intervention activities were viewed as being broadly implemented as intended in both cities; for example signage was implemented as intended, probably, in part, because of the rigid parameters afforded by the legislation, with only minor amendments being made. Enforcement activities, specific to the 20-mph limits, were limited by finite resources and competing priorities in both sites, and over time became ‘daily business’. Public experiences of these activities varied, but an important finding was the disjunction between agents (e.g. police services) and the public in terms of how the interventions should be enforced. The processes associated with rolling out such a large scheme in Edinburgh were identified as challenging; a dedicated ‘20-mph team’ was created within the local authority to address this. The creation of a dedicated, official and strong partnership and joined-up working were identified as key facilitators in broad implementation and in the delivery of a tailored education and awareness-raising campaign in Edinburgh. In Belfast, government reticence was seen as a potential barrier to there being no formal awareness-raising activities in Belfast. This latter point may help to explain the difference in levels of awareness of the 20-mph speed limits that was evident between participants from the two cities.

Impact

Comparison with the Department for Transport evaluation

The Department for Transport (DfT) commissioned a large research project to evaluate the effectiveness of 20-mph signed-only speed limits. This was published in 2018. 77 Sometimes known as the ‘Atkins Study’, this evaluation was based on 12 case study schemes in England, with various comparator areas with 30-mph limits in place. It is helpful to compare our findings with those of this key report:

• Although we observed an important and significant reduction in collision and casualty rates in Edinburgh, the DfT evaluation reported ‘no significant change in collisions and casualties in the short term’. 77 Although some individual case studies showed a reduction in the number of collisions/casualties, the results were based on very small sample sizes and the authors concluded that it was ‘not possible to attach any confidence to their significance’. 77 Therefore, our results from Edinburgh, based on city-level analysis, are an important step forward in the evidence base, as the findings were based on a large sample, thereby allowing greater confidence in the observed effect.

• Our findings of only minor changes to the numbers of collisions and casualties in Belfast are consistent with the DfT evaluation observation that effects may differ by case study area. In this way, our study adds weight to the notion that a 20-mph signed-only intervention may have differential effects by area.

• The DfT evaluation reported a median reduction in speed of 0.8 mph in residential case study areas, and of 0.6 mph in city-centre case study areas. Our findings for reductions in speed in Edinburgh (mean reduction of 1.34 mph, median reduction of 0.47 mph) are consistent with the direction and order of magnitude of these results. The DfT evaluation reported that faster drivers reduced their speeds more, and our Edinburgh results add city-level evidence to support this. This is a particularly important finding as these are the speeds associated with greater safety risk, so observing this at a city level advances the evidence base. As with collisions and casualties, the DfT report observed a range of effects across the case study areas and our results from Belfast city centre also suggest that effects can be variable by location.

• The DfT evaluation reported general support for 20-mph limits. It found that ‘local residents and other road users generally perceive the 20-mph limits as beneficial for local residents, pedestrians and cyclists’. 77 Our findings are consistent with this, while adding some further insight into how local perceptions evolve after implementation.

As an overall reflection, our findings are consistent with those of the DfT evaluation, and so add weight to the developing evidence base on 20-mph speed limits. Our study advances knowledge by (1) being based on a larger, citywide intervention, thereby giving greater confidence in a number of key findings; and (2) being able to conduct more in-depth informative analysis on the various outcomes, thereby developing our level of insight.

Cost-effectiveness

Interim analyses to inform the progression decision suggested that it was plausible that the benefits of the scheme in Edinburgh, associated with the reduction in the numbers of collisions and casualties, would exceed the costs. The observed increases in liveability strengthen this conclusion.

Sustainability and diffusion

One feature that the team did not fully recognise when putting the original proposal together was the ‘what next?’ question in the two cities, and its connectedness with our programme theory. Our programme theory took us as far as the outcomes, at the points at which we measured them, but the way in which the schemes did or did not become embedded in the future, we now recognise, ought to have been an explicit part of our programme theory. What had happened historically and during implementation was our focus, but clearly what happens next, and the extent to which the gains in terms of casualty reductions might be maintained, is another important question in its own right and one that is closely integrated in the processes and outcomes that had occurred up to the point when our investigation ended.

Since the project finished, three members of the team (RJ, KM and MPK) have leveraged further funds from the Wellcome Trust to carry out a range of diffusion activities. Videos and an interactive logic model are available online. 106

The breadth of the evaluation

We set out not only to evaluate a range of public health outcomes (including cost-effectiveness), but also to understand the systems (political, cultural, policy) in which the interventions were implemented. To our knowledge, this is one of the most comprehensive studies of a public health policy intervention that has ever been undertaken. In public health, there is a growing consensus that the systems and context in which an intervention is implemented can be as important as the intervention itself, and indeed can be the main factors influencing effectiveness. Through our extensive documentary, historical, stakeholder and media analysis, we were able to construct a plausible explanation as to the factors that enabled the 20-mph limit policy to move from an idea onto the policy agenda and, finally, into reality. Our qualitative work with stakeholders and the public enabled us to understand more about whether or not the implementation was implemented as intended. This is important because the processes involved in the implementation of 20-mph speed limits have not been a focus of previous evaluations. The outcomes evaluation enabled us to evaluate whether or not it was effective, and, finally, our economic evaluation enabled us to make some judgement as to the cost-effectiveness of the intervention. With all these pieces of the ‘jigsaw puzzle’, it is likely that other cities and areas will be able to move forward with developing their own policies. An outcomes evaluation alone would have told them only that it worked, not why and how. Although a standard outcomes evaluation with a tranche of numbers, p-values and SDs may be viewed by some as more ‘scientific’, it will undoubtedly be less useful to those who need to implement based on the findings.

The tale of two cities

By choosing two cities with different contexts and different implementation strategies, we were able to develop a much better understanding of the factors that influence decision-making. This enables us understand how the intervention could be transferred to other cities and jurisdictions.

The second key strength was the partnerships we had in both cities with local and national stakeholders. Through working closely with them while developing the evaluation proposal, we were able to gain a good understanding of the questions they wished to have answered and the data that were available. We continued to work closely with them throughout the intervention period, and shared data and results. They were instrumental in providing us with key information for the evaluation, from the documentation we needed to the people we should talk to. Without this strong partnership, we would not have been able to access the data and the key stakeholders. The impact of our work was also greatly enhanced by this partnership. Ruth Jepson (the principal investigator) was asked to present evidence to both local government and national government during the period of evaluation. It was also instrumental as evidence for extending the roll-out of 20-mph speed limits in Edinburgh to more streets in the city.

The Study Steering Committee

Another key strength was the composition of the SSC. As well as academics working in the area, the SSC included representatives from the local bus company, police, advanced car drivers and the local council, and those who had been involved with the development of the 20-mph schemes. Having a range of stakeholders meant that we were able to ask for advice and support, as well as get their feedback on what the results meant to them. They were also instrumental on helping us get access to data and specific participants (such as taxi drivers, police and bus drivers). Without their unique insights we may not have picked up on many of the complexities of the implementation and how it affected the police, the bus drivers and others.

The implementation of 20-mph limits in Edinburgh and Belfast provided a unique opportunity to observe a natural experiment and undertake a comprehensive evaluation. However, policy interventions such as 20-mph speed limits are often associated with a number of challenges and limitations that are outside the control of the researchers. There were two main challenges: (1) timings related to funding and implementation of the policy, and (2) the data that were available to the researchers for analysis.

Timings

The research team had no control over the timetabling of the introduction of the limits. We developed the bid hoping to undertake baseline data collection in all the implementation zones in Edinburgh and in Belfast. However, the funding timeline did not align well with the policy and implementation timeline. Funding was received only after implementation was complete in Belfast and after it had begun in Edinburgh (it been implemented in 50% of the zones). This prevented us from collecting baseline data. Of course, this is a potential issue in any study seeking to investigate real events in real time.

Data

Another challenge related to the data that we had access to on active travel (see Chapter 4, Active travel). When we initially put together the proposal, we were hopeful that we would have several different measures of active travel. We wanted to limit the number of new data that were collected by the research team to avoid significant additional costs. We were anticipating that the data would be suitable for analysis, but we encountered several problems. This meant that we were not able to undertake analysis of the active travel data as robustly as we had originally planned. This had implications not only for the research objectives relating to active travel, but also for our plans for the economic evaluation. We discussed these issues with the SSC, and provided an alternative plan for the economic evaluation, which was approved.

Media analysis

The strength of this analysis is the combination of systematic searching, screening and coding frames, and the use of a theoretical and practical framework to understand processes through the lens of a key social mechanism that shapes opinion and reflects views. Key limitations included the inability to include some important media as they were not available online to our study. Moreover, we are aware that the media are in themselves biased and we cannot be sure if the media presented the real story of what happened. The media may present a simplified public view of what happened, or perhaps from only one perspective, with other local voices being excluded; however, we were able to find examples of different views (e.g. anti-20 mph) being reported. We acknowledge the limitation that we did not examine social media, which may have given us greater insight into new issues or broader views.

Outcomes evaluation

The nature of this study, being a natural experiment evaluation, presents researchers with the need to acknowledge confounding variables and make adjustments in analyses when possible. The main confounder in this natural experiment is related to the citywide efforts to increase active travel during the study period. The attribution of a quantified impact of these efforts on the rates of road traffic collisions and casualties has been a challenge because representative data on active travel for the purposes of this study were not available.

Qualitative work

One of the strengths of the qualitative work is the range of stakeholder and members of the public recruited from both cities, which allowed us to examine, in-depth, how the interventions were implemented and who was involved in this, as well as how the interventions were experienced by those affected (i.e. residents and commuters within the cities). The main limitation was that we found it challenging to recruit groups from socioeconomically deprived areas; thus, it was not possible to qualitatively examine the impact of the interventions on health inequalities, as originally intended. A further limitation was that, as noted, funding for this evaluation was received post implementation in Belfast and during implementation in Edinburgh; therefore, this may have led to an element of recall bias.

Economic analysis

The key strength is the DTA that we adopted. It offers a useful framework for planning and interpreting the results of an economic evaluation where there is substantial uncertainty about the likelihood of finding significant effects (in public health terms) and to ensure a rounded assessment of impacts and cost-effectiveness when valuable interventions might otherwise be rejected on the basis of statistical insignificance. There are risks with the approach, however. In particular, transparency of the DTA process is vital to reduce the risk of publication and other biases that might otherwise lead to ad hoc and partial use of evidence to justify a prior position on dubious grounds. In addition, we were not able to undertake a full economic evaluation nor a full DTA for the reasons discussed in Chapter 6. The health economist was unable to complete the work within the project timelines as a result of being a member of the Scottish Government response team working on COVID-19.

International bodies and their policy statements and guidance

• A number of high-level policies on transport and health have been produced by international agencies such as the WHO and the European Union.

• These policies provide a general policy direction for national governments to follow.

Background legislation and policy

• The national government typically sets the policy framework for road safety.

• It is usually the responsibility of local authorities to implement measures locally.

• A key early step is to ensure that a 20-mph speed limit complies with the national framework.

• When a 20-mph speed limit does not fit within the national framework, lobbying for change at the national level may be required.

Local jurisdictions

• In Edinburgh, the scheme was initially led by the Lothian Regional Council, and then the City of Edinburgh Council (following a shift in governance structures). The formal legal mechanism for implementing speed restrictions in Scotland is via traffic regulation orders, which follow a statutory process including a robust consultation.

• Local jurisdictions did not play a role in Belfast, as the scheme there was led and managed by the government of Northern Ireland – the Department for Regional Development/Department for Infrastructure (the department changed its name in May 2016).

Local histories

• Local history is important for understanding traffic issues, previously tested interventions, and public perceptions and attitudes. We found a 20-year timeline to be helpful, although this is likely to vary from place to place.

• There may be particular historical factors that remain important, for example medieval and Victorian street design, and railway developments and closures, but these will be specific to localities.

• Local traffic data (e.g. on volume, speed and collisions) can be used to ‘make the case’ for speed reduction interventions around schools or in other ‘accident hotspots’.

• If small-scale interventions are shown to be effective, then this provides local data that a 20-mph speed limit is an appropriate solution to address road safety.

• Implementing small-scale interventions in different locations (e.g. around schools) can help to create a societal ‘norm’.

Politics

• Strong political leadership is required to push the agenda forward.

• Securing support across political parties helps to prevent the issue from becoming party political.

Government and local authority officers

Officers within the City of Edinburgh Council and civil servants in Belfast were central to making the schemes happen. These officers supported the politicians and political decisions, and implemented the relevant processes to take 20-mph speed limits from an idea to a reality.

Evidence and data

• The evidence discussed mostly related to vehicle speed and the related risks, as opposed to the effectiveness of 20-mph restrictions specifically.

• The limited available evidence on the effectiveness of 20-mph restrictions on slowing traffic and reducing the number of collisions and casualties was important in persuading different stakeholder groups that the speed limit policy may be worth pursuing. This evidence came mostly from evaluations of 20-mph ‘zones’ with traffic-calming infrastructure, as opposed to ‘limits’, which do not use physical traffic-calming measures.

• Without this evidence, it is likely that the idea would never have ‘taken off’.

Enforcement and the police services

• Speed limits of 20 mph have no physical traffic-calming measures.

• A concern around 20-mph sign-only schemes is the perception that enforcement is required to ensure compliance, which puts additional strain on the police service.

• Policy enforcement was implemented in both Edinburgh and Belfast, although some stakeholders felt that more visible enforcement would be beneficial.

• A combination of awareness, enforcement, education, driver training, and community and statutory involvement was recommended for future interventions.

The public

• Engaging the public in a democratic decision-making process can help to ensure buy-in and support.

• Public support can add weight to political debates.

• Running ‘pilot’ schemes can allow the public to experience a 20-mph speed limit on a small scale.

• Evidence from Edinburgh and Belfast indicates that some of those who are opposed to 20-mph limits can be ‘won over’ when they see the intervention in practice.

The political process including opposition

Interestingly, in both Edinburgh and Belfast, there were no clear political divides in opinion on 20-mph speed limits; instead, there were individuals who were for and individuals who were against the idea across all parties.

The use and non-use of evidence

• Once 20 mph made it onto the political agenda, evidence played only a small role in decision-making, although, from time to time, evidence was used to defend political arguments, for example when challenged in the public consultation.

• As the 20-mph speed limit moved closer to becoming a reality, little reference was made to the existence of evidence on whether or not the intervention ‘works’ (i.e. reduces speed and improves road safety). Other types of data became of interest to specific stakeholder groups, for example the impact that lower speed limits would have on bus timetabling, taxi drivers, carbon emissions and even pizza delivery times.

• A key concern in Belfast was that the speed limit would deter people driving to the shopping centre, and thus would have a detrimental impact on local businesses.

• Over time, the narrative shifted from road safety to a broader ideological vision of better communities and supporting wider economic and environmental agendas.

Professional advocacy

• Advocacy groups play an important role in bringing topics such as road safety and 20-mph speed limits to the attention of policy-makers.

• Such groups included 20’s Plenty for Us, Sustrans, Living Streets and cycling organisations.

• These groups can make a compelling case for the need to act, based on evidence and statistics, but also on case studies.

Process: what actually happened

• In Edinburgh and Belfast the approach was gradualist.

• Small schemes were implemented around schools as a first step: child safety was an ‘easy win’.

• There was a range of steps in the build-up to implementation, including consultations with the public, pilot schemes, and discussions around the design and implementation of the schemes.

• A large pilot scheme in the south of Edinburgh was particularly critical in gaining widespread public support for the wider roll-out in Edinburgh.

Legislation

Both cities had to issue legislative traffic regulation orders to implement 20-mph limits.

Signage and road markings

• Signage was implemented in both Edinburgh and Belfast, including advanced warning signs (i.e. new limit in force), terminal signs (on entering the 20-mph area) and repeater signs (throughout the area).

• Road markings were implemented in Edinburgh, and to a lesser extent in Belfast.

• In Edinburgh, ‘buffer zones’ were introduced between areas that had national speed limits and those that had 20-mph limits. These involved the installation of speed limits of 30 mph on sections of road to ease the transition from 60 mph to 20 mph, and vice versa.

• In Belfast there were some changes to signage once implementation had begun. Approximately 8 months into implementation, the original signs with a white background were replaced with signs with a yellow background to increase conspicuousness.

Awareness-raising and education

• In Edinburgh, an education and awareness-raising campaign was integral to the intervention and was implemented from the outset, with dedicated staffing and funding. This included physical adverts (e.g. bus backs/shelters, billboards) and press, media and social media releases delivered prior to signage being constructed, at the ‘go-live’ stage and throughout implementation. These activities were often tailored to a specific geographical area and informed by public feedback about preferences.

• In contrast, this component of implementation was delivered on a smaller scale with few delivery partners in Belfast. In retrospect, stakeholders and members of the public felt that a larger campaign in Belfast would have helped to raise awareness and build support for the intervention.

Enforcement

• A general view from police stakeholders was that 20-mph limits should be self-enforcing, requiring limited (if any) additional police resourcing.

• In Edinburgh, additional officers were trained in speed detection in advance of the speed limit, speed checks were carried out and community officers had ‘responding to 20 mph issues’ added to their remit. In addition, Police Scotland supported the city council’s media releases.

Self-enforcement of limits

• Some members of the public perceived that the 20-mph limit was not being enforced, and therefore there was no reason to stick to the new limit.

• The public expressed that a stronger police presence, with strict and visible enforcement (e.g. through the installation of speed cameras and the issuing of penalty points and fines), would improve compliance with the new speed limit, but acknowledged that greater enforcement may lead to negative public attitudes towards the police.

Awareness of 20-mph speed limits

• The scheme in Edinburgh was perceived by the general public as being highly visible, attributable in part to the education and awareness-raising activities that took place.

• In contrast, lower levels of awareness of the 20-mph speed limit initiative in Belfast are likely to stem from the relatively low-key public education and awareness-raising activities.

Changing attitudes to speed limit/speeding

• In Edinburgh, using the SLiPS, an increase in support and decrease in resistance to the 20-mph speed limit intervention was identified 6 and 12 months after the implementation. This was also supported by qualitative focus group data obtained from members of the general public.

Traffic flow (including volume)

• The volume of traffic on the roads decreased slightly in both cities.

• The average 7-day volume of traffic decreased by 87 vehicles (95% CI –112 to 286 vehicles) in Edinburgh.

• The average 7-day volume of traffic in Belfast city centre decreased by 133 vehicles (95% CI –252 to –15 vehicles).

Reduction in speed

• One year after the policy change, average speeds on monitored roads in Edinburgh had reduced by 1.34 mph (95% CI 0.95 to 1.72 mph).

• In Belfast, the average speed reduced by 0.91 mph (95% CI –3.01 to 1.18 mph).

• The observed speed reductions in Edinburgh and Belfast are consistent with other signage-only schemes, which have also shown an approximate 1-mph reduction in speed.

Changes in speed differential (between motorised and non-motorised)

• In both the Edinburgh and Belfast focus groups, there was a small amount of discussion that the reduced speed limit had created a more pleasant cycling experience through a reduced speed differential between cyclists and motorists.

• However, some members of the general public also discussed how this reduced speed differential could lead to potentially dangerous interactions and overtaking situations.

Speed limit of others imposed by self-enforcers

Feelings of frustration when caught behind a slow driver did not change markedly on the perceptions survey in Edinburgh, although in the focus groups it was reported that drivers were becoming frustrated with slower drivers, leading to increased overtaking.

Drivers copy self-enforcers

• In Edinburgh, using the SLiPS, members of the public reported an increase in following the behaviour of others on the road (rule-following) 6 and 12 months after implementation.

• In Belfast, drivers reported not changing their driving speed and not self-enforcing a 20-mph driving speed, which was attributed to a lack of awareness of the 20-mph speed limits, the perception that there was no need for the 20-mph speed limits within the city centre and a lack of enforcement of the 20-mph speed limits.

Perceptions of journey time and effort

• In Edinburgh, using the SLiPS, members of the public reported a reduction in perceptions that 20-mph speed limits would lengthen their journey times, worsen traffic flow and increase congestion. These findings were supported by the qualitative data gathered via focus groups with the general public.

• In Belfast, those who participated in the focus groups perceived that traffic congestion had increased as a result of a reduction in driving speed; however, they also reported that congestion was an ongoing issue in the city centre prior to the implementation of the 20-mph speed limits.

Perceptions of safety and traffic-related dangers

• The SLiPS in Edinburgh did not identify marked changes in the perceptions of safety on the city’s roads. In Edinburgh, the general public did not report any enhanced feelings of safety as a driver of a motor vehicle, as a cyclist or as a pedestrian.

• Stakeholders in Belfast perceived that the 20-mph speed limits would improve cyclist safety; however, if the limits were not being enforced or drivers were not adhering to the new limits, they felt that the 20-mph speed limit would be ineffective and that there would be no improvements in road safety.

• In Belfast, it was perceived that the reduced driving speeds could result in drivers becoming irate/frustrated, which may cause an increase in the rate of collisions and casualties, although a reduced driving speed could reduce the severity of collisions and casualties.

Changes in judgement of speeds by pedestrians

• The SLiPS in Edinburgh did not identify marked changes in the perceptions of walking safety near the city’s roads or when crossing the road.

• In Belfast and Edinburgh, the general public perceived that reduced driving speeds caused pedestrians to walk in front of moving cars. It was discussed that pedestrians walking in front of moving cars could cause drivers to become irate/frustrated, potentially leading to an increase in the rate of collisions and casualties. Consequently, this was viewed as an unintended consequence of the 20-mph speed limits.

Changes in driver attention

The public perceived that 20-mph limits may lead to reductions in driver attention, as a result of concentrating on the speedometer and drivers checking their phone when travelling at slower speeds. This was perceived to have the potential to result in an increased rate of collisions and casualties. This was not substantiated by the collision and casualty data, highlighting a discrepancy between the effectiveness of the intervention and its perceived effectiveness among the public.

Changes in driving style (e.g. braking and acceleration)

• The general public in both cities perceived that driver behaviour had worsened to some extent as a result of the 20-mph speed limits. For example, they perceived that there had been an increase in overtaking due to driver frustration with slower moving traffic.

• In Belfast it was also perceived that an increase in breaking/accelerating could lead to increased noise and air pollution.

Liveability and pleasantness of streets

• Using the MAPS-Liveability tool, liveability significantly increased in both Edinburgh and Belfast.

• In Edinburgh, the liveability constructs of traffic/transport, places and pavements were found to have significantly increased.

• In Belfast, the liveability constructs of traffic/transport and places were found to have significantly increased.

Increase in journeys made by public transport

• Insufficient data were available to determine the impact of the intervention on the number of journeys made by public transport. There was very little change in the volume of vehicles on the roads, which may suggest that the 20-mph speed limits did not lead to a change in travel mode from the car to public transport.

• The general public in Belfast and Edinburgh did not report changes in their selected mode of transport as a result of the 20-mph speed limit intervention.

Increase in journeys made by bike

• Insufficient data were available to determine the impact of the intervention on the number of journeys made by bike.

• The general public in Belfast and Edinburgh did not report a change in the number of journeys they made by bike; however, they did perceive that the 20-mph speed limits may improve cyclist safety.

Increase in journeys made on foot

• Insufficient data were available to determine the impact of the intervention on the number of journeys made by foot.

• The general public in Belfast and Edinburgh did not report a change in the number of journeys they made on foot as a result of the 20-mph speed limit intervention.

Reduced emissions

• Insufficient data were available to determine the impact of the intervention on emissions.

• In Belfast, the general public perceived that the reduced speed limits could lead to an increase in air pollution through reduced car efficiency and increased breaking/accelerating.

Reductions in car use

• No data were available to determine the impact of the intervention on overall car use.

• Vehicle volume did not decrease as a result of the new limit, which may suggest that people continued to use their cars in their usual way.

• The general public in Belfast and Edinburgh did not report a change in car use as a result of the 20-mph speed limit intervention.

Increase in ease of crossing roads

Data from focus groups with the general public in Belfast and Edinburgh suggested that the intervention was perceived to lead to pedestrians more freely crossing the road, and less frequently utilising formal crossing points. It was suggested that this could have the unintended consequence of more pedestrians being hit by cars.

Reduction in number and severity of collisions involving cyclists

No data were available on the number and severity of collisions involving cyclists.

Reduction in number and severity of collisions involving pedestrians

No data were available on the number of collisions involving pedestrians; however, the number of pedestrian casualties decreased by 53% (–591 casualties).

Reduction in number and severity of collisions involving motorists and passengers

• In Edinburgh, the number of collisions in a year reduced by 367 (40% reduction), with 409 fewer casualties (39% reduction). The number of road traffic fatalities in Edinburgh decreased by 11 (23% reduction).

• There was a 2% reduction in collisions and a 6% reduction in casualties in Belfast city centre up to 3 years following the introduction of the 20-mph speed limits. Road traffic fatality rates decreased by 44.3%.

Increase in population levels of physical activity

No data were available to determine the impact of the intervention on overall physical activity levels, which we proposed would change as a result of increases in walking and cycling.

Reduction in the number of air pollution-related respiratory conditions

• No data were available to determine the impact of the intervention on the number of air pollution-related respiratory conditions.

• The general public in Belfast perceived that the 20-mph speed limits could lead to an increase in air pollution through reduced car efficiency and increased breaking/accelerating.

Problem stream

Evidence sources from professional advocacy groups outside Scotland were also cited, to support the call for safer streets:

The first, full independent survey into the effect of school traffic on road safety and traffic congestion is to be carried out by the AA Foundation for Road Safety Research over the next nine months.

Dangers of congestion highlighted, The Scotsman, 7 May 1999

These combinations of both professional advocacy and public advocacy merged into a universal advocacy call for improvements to the problem:

Demonstrators lay in the street yesterday bearing 103 crosses to symbolise the projected death toll on Britain’s roads before the Government launches its safety review in 11 days’ time. Around 200 families affected by road accidents gathered at Whitehall, in London, to draw the Government’s attention to the fact that half the 3400 people killed or seriously injured on Britain’s roads every year would be saved by 20 mph speed limits in built-up areas.

Protest over road deaths, The Herald (Glasgow), 21 February 2000.

Reproduced with permission from Newsquest Media Group

As 20-mph schemes were being more wildly reported, opposition to such ideas was also reported. Nearly all of these anti-20-mph views were expressed as editorial or opinion articles stressing a political positioning of anti-speed and depersonalised positions or non-enforcement:

The Green Party have been advocating this, but do any of these people actually drive a car? Or could they be cyclists? And have they tried to drive at 20 mph?

Efforts to get drivers on go-slow must have been dreamed up by cyclists, Evening Express, 12 May 2018.

Reproduced with permission from Aberdeen Journals Archive

Policy stream

The problem stream pro-20-mph speed limit evidence of impacts had focused on speed and child fatalities, pre 2000. Now new types of evidence appeared in the policy stream that had more breadth in terms of the range of issues reported, across health, economic, environmental and social benefits for individuals and communities. The problem of traffic was now expressed in policy language, in terms of economic cost:

The GCPH [Glasgow Centre for Population Health] found that imposing a similar limit in Scotland could cut road casualties by more than 10%, and up to £39.9m per year in costs to emergency services . . . Under the higher estimate, there could be 755 fewer casualties, including saving save [sic] five lives per year and £39.9m.

Cutting the speed limit to 20 mph ‘would save lives and £40m a year’, The Herald (Glasgow), 20 September 2018.

Reproduced with permission from Newsquest Media Group

This broadening of benefits was coded as a theme of remessaging of evidence, as it was reported as being trumpeted by advocacy groups and, for the first time, political individuals and parties:

Mark Lazarowicz [Scottish Labour] said . . . ‘Improving the pedestrian environment will be good for business, good for local people and good for visitors and tourists. What we are bringing forward will create a city centre to rival those of the great cities in Europe’.

Drivers face total Princes St ban, The Scotsman, 15 February 2000

Despite 20-mph schemes being sought as a solution to the problem, there were few calls for a political and policy solution pre 2000; however, articles post 2000 introduced examples of ‘successful’ implementation of 20-mph schemes in Scotland, reporting additional and potential benefits, proposing 20-mph speed limits as a solution and calling for political support:

Edinburgh could become one of the first cities . . . to introduce a 20 mph limit across its city centre to reduce accidents and improve the pedestrian environment . . .

The main reason behind the scheme is to reduce accident figures. Over the last three years there have been 550 accidents in the city centre.

20 mph – and that’s the limit, The Scotsman, 16 November 2000

The impact of 20-mph schemes from other parts of the UK and within Scotland become more frequently reported post 2000, and connected with solving similar problems elsewhere. This powerful connection was expressed in terms that a solution already exists and is being used by others like us. For example, redevelopment plans in Aviemore included 20-mph speed limits as part of the traffic safety measures:

Two roundabouts at either end of Grampian Road will take traffic at 20 mph up to a new system of streets organised around the green, designed as the heart of the village, its commercial as well as its community life.

Blueprint unveiled for new face of Aviemore, The Herald (Glasgow), 9 September 1997.

Reproduced with permission from Newsquest Media Group

The pilot of the 20-mph Edinburgh scheme was reported as successful and as the principal reason for the expansion of the scheme across the whole city in 2014, with the endorsement of a local pressure group:

Edinburgh is to become the first city in Scotland to impose a 20 mph speed limit across its streets . . . The plan follows a successful pilot in the Marchmont, Grange and Prestonfield areas of the city last year.

Edinburgh to have 20 mph speed limit, The Herald (Glasgow), 14 January 2014.

Reproduced with permission from Newsquest Media Group

This is not a particularly poor or problem area, nor is Edinburgh a city where traffic problems are as ghastly as in many . . . Poorer areas in general have worse roads: the poorest children in Britain are four times likelier to be killed in a traffic accident than children from social class one . . ..

Pollution and danger on the walk to school, Scotland on Sunday, 30 October 1994.

Reproduced with permission from Newsquest Media Group

[Person 4] said he faced problems trying to stick [to] the limit, writing: ‘Driving through Portobello the other day, doing 20 mph with the driver behind practically in the boot! I agree 20 zones should only be near schools, residential areas etc but not on main roads’.

Facebook responses – Evening News readers have their say on 20 mph limits, The Scotsman, 4 July 2017