WebTAG Unit 3.10.6 – an alternative draft

Where this document comes from

The Department for Transport (DfT) first released this new WebTAG Unit for consultation in November 2011.

At the instigation of Campaign for Better Transport, in January 2012 DfT officials met with a group of experts in the field of Smarter Choices who were concerned that the Unit downplayed the potential of Smarter Choices, particularly when applied as packages of measures.

Following that meeting, these experts worked together with us to submit an annotated and referenced alternative draft Unit, which was submitted to the DfT consultation in March 2012 along with references to a range of other evidence sources. This alternative draft was, of course, quite different from the document that the experts would have produced if working on a new Unit from first principles, but they felt that by accepting some aspects of the DfT’s approach, they would be more likely to influence the draft constructively.

When the Unit was transferred to IN DRAFT status and published in May 2012, we were very disappointed to see that only a minority of the suggestions we had submitted had been taken on board by the DfT.

We have decided, therefore, to open up the process by publishing an updated version of the amended WebTAG Unit text that was compiled by the expert group for the DfT’s consultation, and asking for more suggestions and additional evidence from practitioners and academics in the field of Smarter Choices, aiming to develop it more fully as an alternative draft.

Please submit comments and further evidence to Campaign for Better Transport who will be co-ordinating this process, at the latest by 11 October2012. The contact address is: .

We have included some specific questions in the text, but would be grateful for comments on any aspect of the guidance, not simply the issues that are raised.

Please also send any new evidence material to the DfT itself, as the WebTAG revision process remains ongoing for their official draft. The DfT contact address is:

The contributors to this text include:

Keith Buchan, Director, MTRU

Sally Cairns, Transport Research Laboratory and University College London

Andy Cope, Director of Research & Monitoring, Sustrans

Phil Goodwin, Emeritus Professor of Transport Policy, University College London and University of the West of England.

Lynn Sloman, Transport for Quality of Life

All the contributors write in their personal capacity, and the suggestions and new content in this draft do not necessarily represent the views of the institutions mentioned.

This draft was based on:

TAG UNIT 3.10.6

Modelling Smarter Choices

November 2011

Department for Transport

Transport Analysis Guidance (TAG)

Technical queries and comments on the DfT version of this TAG Unit should be referred to:

Transport Appraisal and Strategic Modelling (TASM) Division

Department for Transport

Zone 2/25 Great Minster House

33 Horseferry Road

London

SW1P 4DR

Tel 020 7944 6176

Fax 020 7944 2198

Contents

1 Modelling Smarter Choices

/

4

1.1 Introduction

1.2 Classification of Smarter Choice Initiatives

1.3 General Approach to Modelling Smarter Choice Initiatives

1.4 Benchmarking Expected Impacts of Smarter Choice Initiatives

1.5 Modelling the ‘Hard’ Components of Smarter Choice Packages

1.6 Modelling ‘Soft’ Components of Smarter Choice Packages

1.7 Modelling Workplace Travel Plans

1.8 Modelling School Travel Plans

1.9 Modelling Targeted Marketing Initiatives

1.10 Reporting Requirements

2 Further Information

/

21

3 References

/

21

4 Document Provenance

/

23

TAG Unit 3.10.6

Modelling Smarter Choices – AN ALTERNATIVE DRAFT

1 Modelling Smarter Choices

1.1 Introduction

1.1.1

This Unit provides advice on incorporating the impacts of Smarter Choice initiatives and packages in models which are generally compliant with the advice in Variable Demand Modelling (TAG Units 3.10.1 to 3.10.4).

1.1.2

The structure of the Unit is as follows:

  • The type and nature of the Smarter Choice initiatives considered in this guidance are discussed;
  • The recommended general approach to modelling Smarter Choice initiatives is explained;
  • The evidence of the effects of Smarter Choice initiatives is reviewed and guidance is provided on how impacts may be ‘benchmarked’;
  • Smarter Choice packages often include ‘hard’ components and guidance is provided on the ways in which these may be modelled;
  • Guidance is provided on the general ways in which the ‘soft’ components of Smarter Choice packages may be modelled; and
  • The modelling of workplace travel plans, school travel plans and targeted marketing initiatives is considered in more detail.

1.2 Classification of Smarter Choice Initiatives

1.2.1

The report Smarter Choices – Changing the Way We Travel (Department for Transport, July 2004) identified ten Smarter Choice measures: workplace and school travel interventions, personalised travel planning, travel awareness campaigns, public transport information and marketing, car clubs, car sharing schemes, teleworking, teleconferencing, home shopping, and support, training and incentives for active travel, including cycling. Since that time, the term ‘smarter choices’ has expanded to encompass a range of additional initiatives, including, for example, residential travel plans, destination based travel plans, station travel plans, walking and cycling initiatives (that include both hard and soft components); new forms of car and bike rental schemes, etc. However, at the time of writing, workplace travel interventions, school travel interventions, personalised travel plans and various sorts of marketing activities are perhaps the longest established, and the most thoroughly evaluated, of the various initiatives, and therefore form the primary focus for this guidance.

1.2.2

Personalised travel planning, travel awareness campaigns, and public transport information and marketing share common features of targeted marketing and there are no sharp dividing lines between them. These initiatives are therefore treated together in this guidance under the heading of ‘targeted marketing initiatives’.

1.2.3

In summary, then, this guidance focuses on three packages of Smarter Choice initiatives: workplace travel plans, school travel plans, and targeted marketing. The first two packages potentially involve both ‘hard’ and ‘soft’ measures, while targeted marketing initiatives are, by definition, all ‘soft’.

Some elements of other Smarter Choice measures are considered where possible. For example, in the discussion of workplace travel plans, there is some consideration of car sharing, and of telework and teleconferencing. However, it is beyond the scope of this Unit to consider all forms of Smarter Choice measures, not least given that this is a very dynamic area, where the evidence base is growing rapidly. Appraisers are advised to check on the latest available information for individual schemes and refer to the literature given in the list of references

1.2.4

For the purposes of this guidance, ‘hard’ measures bear directly on the time and money components of generalised cost while ‘soft’ measures change travellers’ response to differences or changes in generalised cost.

1.2.5

It is important that the effects of Smarter Choice components should not be double counted by modelling a single measure as both a ‘hard’ and ‘soft’ choice, but also should not be ‘half-counted’ by only considering one or other aspect of them. There is some evidence that a combination of hard and soft components can achieve greater change than either in isolation (see for example Cairns, 2007).

1.3 General Approach to Modelling Smarter Choice Initiatives

1.3.1

The evidence available from monitoring studies about the effects of Smarter Choice initiatives has grown substantially in recent years, especially due to comprehensive evaluation of the three Department-funded Sustainable Travel Towns (Darlington, Peterborough and Worcester) (Sloman et al., 2010) and similar initiatives elsewhere (such as in Richmond and Sutton, see for example, MVA 2011). This research provides considerable evidence of the effects of packages of measures in aggregate, as well as a number of specific interventions in the ‘targeted marketing’ category. There is also evidence from a number of individual project evaluations, relating to personalised travel plans (e.g. Sustrans and Socialdata, 2008a, 2009, 2010a, 2010b, 2010c, Parker et al., 2007), workplace travel plans (e.g. Cairns, Newson & Davis, 2010) and school travel plans (Cairns & Newson, 2006), as well as the Department-funded Cycling Demonstration Towns programme (Sloman et al., 2009). This evidence can inform the specification of how these measures may be modelled. The ‘hard’ measures often included in Smarter Choice packages can often be represented in models of the kind specified in Variable Demand Modelling (TAG Units 3.10.1 to 3.10.4) in generally conventional ways.

1.3.2

Bearing in mind the current state of knowledge, the following general approach to modelling Smarter Choice initiatives is suggested:

  • Benchmark the expected impacts of the Smarter Choice package, based on the available evidence, taking account of the proposed intensity of application compared with the intensity of application to which the evidence relates. This is discussed further in Section 1.4;
  • Model the ‘hard’ components of the Smarter Choice packages explicitly (where possible); check that the impacts are less than the benchmark for the package and adjust the model if necessary. This is discussed further in Section 1.5;
  • Model the ‘soft’ components of the Smarter Choice package by means of assumed adjustments to the model parameters; check that the impacts are plausible in comparison with the benchmark and the impacts of the ‘hard’ measures and that the combined impacts of the hard and soft measures are consistent with the benchmark. This is discussed further in Section 1.6; and
  • Assess traffic impacts on the road network, especially in terms of congestion, and also assess impacts on public transport patronage and revenues.

1.3.3

Whatever changes are made to the model, for the final forecasts, it should be run to convergence in the normal manner so that the induced traffic effects are properly accounted for. However, if the benchmark excludes induced traffic effects, it may be appropriate to adjust the model parameters so that the first iteration results match the benchmark, before running the model to convergence.

1.3.4

Given the uncertain nature of the assumptions necessary for the third step, consideration should be given to conducting sensitivity tests using different assumptions.

1.4 Benchmarking Expected Impacts of Smarter Choice Initiatives

1.4.1

The first of the steps listed above requires information about measured or observed impacts of Smarter Choices packages of measures so that an appropriate benchmark or target car trip reduction can be established for the Smarter Choice package to be modelled.

1.4.2

In order to establish a benchmark, two steps are required:

  • First, the substantial evidence about the impacts of Smarter Choice measures needs to be understood;
  • Second, the scale of the impacts indicated by the evidence needs to be related to the intensity of the proposed application so that a benchmark of an appropriate scale can be derived. The evidence is considered first, followed by advice on derivation of appropriate benchmarks.

Evidence

1.4.3

Evidence about the effects of Smarter Choice initiatives is growing and results from a number of studies have been published:

  • Smarter Choices: Changing the Way We Travel (Cairns et al., 2004 and 2008)
  • The report on “The Effects of Smarter Choice Programmes in the Sustainable Travel Towns”: Full Report (Sloman et al., 2010);
  • The Smarter Travel Richmond and Smarter Travel Sutton Programmes:
  • Several evaluations of Sustrans’ personalised travel planning projects (also known as individualised travel marketing) (Sustrans and Socialdata, 2008a, 2009, 2010a, 2010b, 2010c, Parker et al 2007).
  • Evaluation of workplace travel plans (Cairns, Newson & Davis, 2010)
  • Evaluation of school travel plans (Cairns & Newson, 2006)
  • Evaluation of the six Cycling Demonstration Towns (Sloman et al., 2009).[1]

1.4.4

Many of these reports provide results specific to personalised travel planning (PTP) or other single-intervention projects, and some provide evidence of the aggregate effect of packages of Smarter Choices measures. The headline indicator in PTP evaluations is often change in mode share (proportion of trips by mode), although effects on distance travelled by private car are also estimated. The Sustainable Travel Towns evaluation (Sloman et al., 2010) provides considerable detail on the effect of Smarter Choices measures on trip distances by different modes. This information can be used in determining a benchmark or target traffic reduction.

1.4.5

Möser and Bamberg (2008) carried out a meta-analysis of Smarter Choices project evaluations which synthesised data from 141 project evaluations from around the world (93 in the UK). This was understood to be critical of some of the earlier work it used. However, as Wall et al. (2011) have pointed out, there are serious problems with this analysis in terms of its categorisation of measures, and inaccuracies in its reports and interpretations of findings and methods from other studies. In any case, many of the conclusions of the Möser and Bamberg study were within a percentage point or two of the other studies.

1.4.6

As such it may be more useful to refer to the UK evaluations listed in paragraph 1.4.3 when setting benchmark trip or traffic reductions. Given the range of interventions delivered and the depth and breadth of research associated with the programme, the Sustainable Travel Towns evaluation is perhaps the most useful single source. However, users of this guidance are strongly encouraged to seek out local examples of Smarter Choices evaluation and to consider examples which exhibit similarities with the particular context being modelled.

1.4.7

Nonetheless, for illustration of the magnitude of effect that can be expected from an integrated package of Smarter Choices measures across a town, and for guidance in the absence of other suitable information, a number of headline results from the Sustainable Travel Towns are provided below (Sloman et al. 2010).

  • Car driver trips per resident of the three towns taken together fell by 9% between 2004 and 2008.
  • Car driver distance per resident fell by 5%7% (for trips of 50km or less). (Car use per head also fell nationally in comparable (medium-sized) urban areas during this period, but by a much smaller amount: a change of -1.2% for car driver trips and -0.9% for car driver distance.)
  • Overall reductions in car traffic (based on counts) of the order of 2%, and more substantial reductions in inner areas, of the order of 78% overall.
  • Bus use grew substantially in Peterborough and Worcester during the period of the Sustainable Travel Town work, whereas it declined in Darlington. Bus trips per resident of the three towns taken together increased by 10%22% (for trips of 50km or over) whereas there was a national decline of bus trips in medium-sized towns of 0.5% over the same period.
  • There were positive results for cycling in all three towns, with particularly substantial growth in Darlington. Cycle trips per resident of the three towns taken together increased by 2630%, whereas, according to the National Travel Survey, there was a national decline of cycle trips in medium-sized towns over an approximately similar period.
  • Walking trips by residents grew in all three towns during the period of the Sustainable Travel Town work. Walk trips per resident of the three towns taken together increased by 10%13%, whereas, according to the National Travel Survey, there was a national decline in walk trips in medium-sized towns of at least 9% over an approximately similar period.
  • The growth in bus use, cycling and walking cannot be explained by trip generation. In fact, at the aggregate level, the total number of trips per capita by all modes, as recorded in household surveys, fell by 1.1%.
  • Although the largest behaviour changes were seen in short car driver trips, the largest reductions in distance travelled as a car driver came from medium and longer distance trips. Of the reduction in distance travelled for trips of <50km, about 45% of the reduction in car driver kilometres came from trips of 1050km; about 40% from trips of 310km; and about 15% from trips of less than 3km.

1.4.8

It is notable that despite problems with interpretation of data from other studies (Wall et al, 2011), Möser and Bamberg (2008) reach broadly similar conclusions about the effect of Smarter Choices measures to the Sustainable Travel Towns evaluation (Sloman et al., 2010), and in both cases, the conclusions are in line with the anticipated ranges of effects that were estimated in the original smarter choices work, and which are given in Table 1. These findings provide a starting point for modelling the effect of a particular Smarter Choice intervention, although the need for appreciation of local context cannot be overstated. It is recommended that benchmarks be set using judgement based, as far as possible, on reasoned arguments from the available evidence and its applicability to a particular situation. These reasoned arguments and the evidence supporting them should be fully documented.

1.4.9

Potential effects from smarter choices initiatives

Table 1: Summary of Smarter Choice Impacts
Smarter Choice Measure / Reduction in Car Trips*
Workplace Travel Plan / 1030%
School Travel Plan / 815%
Targeted Marketing / 715% (urban areas)~

*In all cases, it should be noted that these figures were averages across a range of schemes – with examples of good practice, or higher intensity application, often achieving considerably greater reductions than this.

~The evidence base for targeted marketing in rural areas is considerably less than for the other results given here, such that informed judgement may provide a better indicator of impacts than previous evaluation results.

1.4.10

In some cases it may be appropriate to use lower levels of impact for the benchmark for reasons other than intensity of application (for example, where only limited measures are introduced). (See Cairns et al., 2004, for detailed discussion of intensity of application.) These reasons need to be documented, along with reasoned arguments and evidence supporting the assumptions actually used. Otherwise, it is reasonable to assume that the impact will be broadly proportional to the scale of effort.