Counterfactual analysis
Purpose
As outlined within HMT Magenta book (HM Treasury guidance on what to consider when designing an evaluation), it is often necessary to infer the impact of an intervention through statistical comparison to a group unaffected by the intervention. This group acts as a proxy for what would have happened to the impacted group in the absence of the scheme, this is called the counterfactual.
For this scheme, the counterfactual is a similar area that has not had a new station and acts as a proxy for what would have happened in East Linton if the new station has not been built. Findings at East Linton will be compared against the impacts at this location.
A limitation of this analysis is that no baseline data was collected for a counterfactual location prior to scheme delivery. For this reason, it is difficult to interpret whether the travel patterns or behaviours have worsened or improved over time.
Methodology
Counterfactual Selection
When measuring the counterfactual, it is essential that it is genuinely comparable to the intervention group. To ensure this, metrics including population, distance to rail and public transport connectivity were analysed to determine the most suitable location.
Population
Settlements of a similar demographic make up to East Linton were selected to determine the most appropriate counterfactual location. Settlements without a station but with a comparable distance (5-10 miles) of an existing station were identified. The populations of these settlements were compared with East Linton to see which matched most closely, as shown in Figure 49.
| Settlement | Population | Comparable to East Linton? |
|---|---|---|
| East Linton | 1,881 | - |
| Lauder | 1,878 | Yes |
| Gullane | 2,910 | No |
| Gifford | 783 | No |
| Duns | 3,127 | No |
| Greenlaw | 605 | No |
| Earlston | 1,765 | Yes |
| Selkirk | 5,521 | No |
This data reveals that as of the last census, Lauder and Earlston had the most comparable populations to East Linton and as such were retained whilst other possible locations were sifted out.
Demographic Comparisons
Additional lines of analysis including age, population density and area were used to determine whether these locations fit closely with East Linton. This is outlined in Table 50 below. A high-level assessment of average house prices was also undertaken to give a better picture of the local economy at each location.
| Metric | East Linton | Lauder | Earlston |
|---|---|---|---|
| Population | 1,881 | 1,878 | 1,765 |
| Area | 0.90 | 0.87 | 0.86 |
| Population Density | 2,092 | 2,165 | 2,041 |
| Annual Population Change | 0.76 | 0.29 | -0.07 |
| % Male | 47.3 | 48.7 | 49.2 |
| % Female | 52.7 | 51.3 | 50.8 |
| %0-17 | 18% | 25% | 17% |
| %18-64 | 52% | 53% | 57% |
| %65+ | 30% | 22% | 26% |
| Average Property Price (2026) | £408,575 | £300,265 | £190,402 |
| Average Property Price (2022/23) | £419,945 | £283,564 | £210,401 |
The data revealed many similarities between East Linton and these possible counterfactual locations, namely very similar geographical areas, population densities and age profiles. For this reason, transport related statistics including distance to rail and public transport journey times were also considered to help refine this selection process.
Distance to Rail
Distance to rail is important to consider as without the rail station in place, residents of East Linton were forced to either travel by alternative modes or travel longer distances to access the rail network. Counterfactual locations should therefore be a similar distance away from existing stations as East Linton was prior to scheme delivery.
| Location | Existing Stations | Distance (miles) |
|---|---|---|
| East Linton | Dunbar | 5.9 |
| East Linton | Drem | 6.1 |
| Lauder | Stow | 5.5 |
| Lauder | Galashiels | 11.0 |
| Lauder | Tweekbank | 11.3 |
| Earlston | Galashiels | 9.4 |
| Earlston | Tweedbank | 7.0 |
As shown in Table 51, this assessment revealed that Lauder has an existing station within 6 miles of the settlement (Stow) which is comparable to East Linton which has Dunbar and Drem stations within 6.1 miles.
Access to Rail
Public transport connectivity to the rail network was also assessed as, prior to the delivery of the station, bus services were available for East Linton residents connecting them to Dunbar and North Berwick where they were able to access the rail network.
| Location | Existing Stations | Journey Time (mins) |
|---|---|---|
| East Linton | Dunbar | 19 |
| East Linton | North Berwick | 25 |
| Lauder | Stow | 89 |
| Lauder | Galashiels | 50 |
| Lauder | Tweekbank | 63 |
| Earlston | Galashiels | 30 |
| Earlston | Tweedbank | 26 |
As shown in Table 52, Earlston has public transport connections to Galashiels and Tweedbank with journey times of approximately 30 minutes. This is most comparable to the c.20-25 minute journey by bus between East Linton and Dunbar/North Berwick. Lauder has notably higher public transport journey times to nearby stations as direct bus services are not available.
Counterfactual Selection
Based on this assessment it was decided that Earlston was the most suitable counterfactual location as it was most comparable to East Linton in terms of population as well as transport accessibility/connectivity. For this reason, Earlston was considered a suitable location for establishing what might have happened at East Linton without a station being delivered.
Earlston is located in the Scottish Borders on the A68 road. It has an estimated population of c.1,800 people. As with East Linton prior to the station opening, residents must travel to an alternative settlement in order to access rail services, with the closest station located approximately 8 miles away at Tweedbank.
Data Collection
Survey Material
A counterfactual survey was undertaken within the settlement of Earlston. The counterfactual survey was developed collaboratively between IBP, Jacobs and Transport Scotland. It addressed the following issues:
- Your travel – questions designed to explore the modes of transport being used, journey purposes, frequency of travel and key destinations that are visited.
- Use of the Rail Network – questions designed to identify residents’ use of the rail network at this stage in terms of stations used, modes used for access, frequency of travel and satisfaction with the East Coast Main Line.
- Views on public transport - current views and how these have changed.
- About You – questions designed to identify the demographic profile of survey respondents, including gender, age, income, health and ethnicity.
Where possible, questions were constructed in a way to maximise comparability with the East Linton Residents Survey.
The counterfactual survey was similar to the resident survey but had less focus on rail travel and more focus on current travel patterns. In the absence of any baseline data collection, the survey also incorporated retrospective questions designed to understand how travel patterns may had changed over time. This was compared with East Linton residents to understand whether any changes in terms of travel choices can be attributed directly to the scheme or are as a result of wider market changes.
A copy of the counterfactual survey questionnaire delivered in Earlston is included as Appendix A.
Target Area
As with the East Linton Residents Survey, the counterfactual survey was focused specifically on the settlement of Earlston itself.
The location of the settlement is shown in Figure 81 below:
Data Gathering Method
The survey was conducted as a face-to-face household survey, with IBP interviewers undertaking doorstep interviews across the settlement. A target of 100 interviews was set, with an achieved sample of 101 interviews being achieved. Interviewing was undertaken over the weekend of 10th and 11th April.
Research Limitations
The counterfactual survey was undertaken to provide a comparator settlement against which findings from East Linton could be considered. While Earlston was selected on the basis of broad similarities in size, location and transport context, no two settlements are identical and differences in demographic profile, local economy, service provision and travel patterns may influence survey findings. In addition, the survey captures stated behaviours and perceptions at a single point in time rather than directly observing how residents would have responded under alternative circumstances. The counterfactual survey should therefore be viewed as one strand of evidence within the wider evaluation framework, providing contextual comparison and indicative insight rather than definitive proof of causation or attribution.
Data & Analysis
Data collected via the counterfactual survey has been compared with the residential survey outputs for East Linton to determine the impact of the scheme. In the absence of a baseline survey for East Linton, it is difficult to attribute these differences directly to the introduction of the station; but it does help to demonstrate differing travel behaviours between East Linton and Earlston.
Demographics
Demographic information has been collected via both the counterfactual and resident surveys and has been used to determine similarities between the two areas in greater detail. Particular emphasis has been placed on the age profile, average income and employment/unemployment of each area. These metrics will be further analysed during future monitoring as the scheme becomes more established.
Figure 82 demonstrates a similar age profile between East Linton and Earlston with 35-54 and 65-74 age groups making up the greatest proportions of respondents.
Similarities are also apparent between East Linton and Earlston when considering employment across these areas. Figure 83 shows that in both cases, residents who are either in full time employment or retired make up the greatest proportion of respondents.
Figure 84 presents the annual household income for respondents in East Linton and Earlston. In both cases many of the respondents preferred not to disclose their income; however, when discounting these responses many of the respondents outlined income from £20,000 to £60,000.
Access to a Vehicle
Similar to East Linton, a significant proportion of the respondents have full access to a vehicle as a driver (53%) with just 21% having no access to a car at all. This is shown in Figure 85 below.
Commuting Travel
Respondents were asked a series of questions relating to their commuting patterns in terms of frequency, mode share and journey times. The survey findings are presented below.
Commuting Frequency
As shown in Figure 86, within Earlston, 64% of commuters travel at least five days per week. This is markedly more than at East Linton where more hybrid working is adopted with residents commuting 2-5 days per week.
Commuting Mode Share
Table 53 presents the main mode of transport used for commuting in Earlston and East Linton. In Earlston car comprises the main mode of transport for the majority (84%) of commuter journeys. There are no users of rail for access to work and education.
In contrast, despite still demonstrating a high proportion of commuter journeys by car (49%), 32% of East Linton residents use rail for commuting journeys. It is thought that the introduction of the station had a significant influence on this, contributing to a reduction in private car use and increase in rail use.
| Mode of Transport | East Linton Residents | Earlston Residents |
|---|---|---|
| Walk / Wheel | 4% | 11% |
| Cycle | 2% | 2% |
| Bus | 2% | 0% |
| Rail | 32% | 0% |
| Bus & Rail | 4% | 2% |
| Park & Ride (Bus) | 1% | 0% |
| Park & Ride (Rail) | 2% | 0% |
| Car / Van (dropped off) | 5% | 7% |
| Car / Van (drive myself) | 49% | 77% |
| Other | 1% | 0% |
Commuting Journey Times
Figure 87 presents the typical commuting journey times for Earlston residents. 9% of respondents have travel times greater than one hour with a third (33%) having a journey time of between 31 minutes and an hour.
Figure 88 demonstrates the places of work identified by the respondents. Lengthy journey times are most likely for those travelling long distances to locations such as Edinburgh and Dunbar for work. Without a direct rail connection, these respondents either rely on travel by private car or interchanging between various modes of public transport.
Leisure Travel
Leisure Destinations
Table 54 shows that the majority of leisure-based trips are to Galashiels (48%) with journeys to Edinburgh (29%) also taking place. Very few respondents undertake cross border journeys to Carlise (1%) and Newcastle (4%) respectively.
| Location | 2026 |
|---|---|
| Berwick Upon Tweed | 7% |
| Carlisle | 1% |
| Edinburgh | 29% |
| Galashiels | 48% |
| Hawick | 3% |
| Jedburgh | 5% |
| Kelso | 9% |
| Melrose | 13% |
| Newcastle | 4% |
| Peebles | 7% |
| St Boswells | 2% |
| Tweedbank | 18% |
| Other (please specify below) | 2% |
| I have not travelled for personal or leisure purposes in the past 30 days | 21% |
| Base | 101 |
Leisure Mode Share
Table 55 below presents the modal share for leisure-based journeys to various destinations. This data demonstrates car dependency within Earlston, with rail only being used for journeys to Edinburgh and longer distance journeys to Newcastle. This varies quite significantly when compared to the East Linton resident survey where there is evidence of rail being used to access various long and short distance destinations for leisure-based journeys.
| Location | Walk / Wheel | Cycle | Bus | Bus & Rail | Rail | Park & Ride (Bus) | Park & Ride (Rail) | Car / Van (non-driver) | Car / Van (driver) |
|---|---|---|---|---|---|---|---|---|---|
| Berwick Upon Tweed | 0% | 0% | 29% | 0% | 0% | 0% | 0% | 0% | 71% |
| Carlisle | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 100% |
| Edinburgh | 0% | 0% | 4% | 11% | 21% | 0% | 11% | 7% | 46% |
| Galashiels | 0% | 2% | 19% | 0% | 0% | 0% | 0% | 15% | 65% |
| Hawick | 0% | 0% | 33% | 0% | 0% | 0% | 0% | 33% | 33% |
| Jedburgh | 0% | 0% | 20% | 0% | 0% | 0% | 0% | 40% | 40% |
| Kelso | 0% | 0% | 11% | 0% | 0% | 0% | 0% | 0% | 89% |
| Melrose | 0% | 0% | 31% | 0% | 0% | 0% | 0% | 15% | 54% |
| Newcastle | 0% | 0% | 0% | 0% | 25% | 0% | 0% | 25% | 50% |
| Peebles | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 100% |
| St Boswells | 0% | 0% | 100% | 0% | 0% | 0% | 0% | 0% | 0% |
| Tweedbank | 6% | 0% | 11% | 0% | 0% | 0% | 0% | 11% | 72% |
Leisure Frequency of Travel
Respondents were asked to identify how frequently they have visited various destinations over the course of the last 30 days for leisure-based journeys. Results in Table 56 below show that travel to many destinations is mostly infrequent, with the majority of destinations being visited up to five times over a 30-day period. More frequent journeys are evident to destinations such as Galashiels, Tweedbank and Kelso which are in close proximity to Earlston and therefore more accessible by existing public transport or car travel.
| Location | Up to 5 times | 6 to 10 times | 11 to 15 times | 16 to 20 times | 21 to 30 times | More than 30 times | Base |
|---|---|---|---|---|---|---|---|
| Berwick Upon Tweed | 71% | 29% | 0% | 0% | 0% | 0% | 7 |
| Carlisle | 100% | 0% | 0% | 0% | 0% | 0% | 1 |
| Edinburgh | 69% | 21% | 3% | 0% | 7% | 0% | 29 |
| Galashiels | 17% | 31% | 29% | 21% | 2% | 0% | 48 |
| Hawick | 67% | 0% | 33% | 0% | 0% | 0% | 3 |
| Jedburgh | 80% | 0% | 20% | 0% | 0% | 0% | 5 |
| Kelso | 33% | 44% | 11% | 11% | 0% | 0% | 9 |
| Melrose | 23% | 46% | 23% | 8% | 0% | 0% | 13 |
| Newcastle | 100% | 0% | 0% | 0% | 0% | 0% | 4 |
| Peebles | 43% | 14% | 14% | 29% | 0% | 0% | 7 |
| St Boswells | 0% | 100% | 0% | 0% | 0% | 0% | 2 |
| Tweedbank | 6% | 22% | 61% | 11% | 0% | 0% | 18 |
Journey Times
Respondents were asked to identify their usual journey times to various destinations. Table 57 below shows that it takes in excess of 30 minutes to reach many destinations from Earlston and in excess of one hour to reach destinations including Carlisle, Newcastle, Berwick-upon-Tweed and Edinburgh.
| Location | Up to 10 minutes | 11 to 15 minutes | 16 to 30 minutes | 31 minutes to 1 hour | More than 1 hour | Base |
|---|---|---|---|---|---|---|
| Berwick Upon Tweed | 0% | 0% | 0% | 86% | 14% | 7 |
| Carlisle | 0% | 0% | 0% | 0% | 100% | 1 |
| Edinburgh | 0% | 0% | 0% | 72% | 28% | 29 |
| Galashiels | 2% | 36% | 53% | 9% | 0% | 47 |
| Hawick | 0% | 0% | 0% | 100% | 0% | 3 |
| Jedburgh | 0% | 0% | 80% | 20% | 0% | 5 |
| Kelso | 0% | 0% | 100% | 0% | 0% | 9 |
| Melrose | 8% | 46% | 46% | 0% | 0% | 13 |
| Newcastle | 0% | 0% | 0% | 0% | 100% | 4 |
| Peebles | 0% | 0% | 57% | 43% | 0% | 7 |
| St Boswells | 50% | 0% | 50% | 0% | 0% | 2 |
| Tweedbank | 33% | 67% | 0% | 0% | 0% | 18 |
Travel by Rail
Respondents were asked to identify whether they have travelled by rail over the last twelve months. Within the sample for Earlston, 30% had used rail services over the last year. Conversely, at East Linton, 84% of respondents had used rail services over the last twelve months. This demonstrates a greater rail use among residents local to East Linton compared to Earlston residents where rail use is markedly lower. In the absence of baseline data, it is difficult to attribute this uplift directly to the station; however, it is considered a contributing factor to greater rail patronage in East Linton.
Most Frequently used stations
Respondents who identified as rail users were subsequently asked to identify the rail station they use most and the frequency of their use. Results shown in Figures 89 and 90 below reveal that residents within Earlston mostly travel to Tweedbank (87%) with rail journeys being undertaken typically less often than once per week.
The frequency of rail use aligns broadly with the journey purposes for rail trips in Earlston, which is dominated by leisure-based travel.
Modes used to Access Rail
Respondents were asked to identify the main modes used to access the rail network at their chosen station. As shown in Table 91 below, journeys to stations were dominated by car travel (representing 84% of journeys) and owing to the distance that would need to be travelled, no respondents access stations using active modes.
There is a large disparity between Earlston and East Linton where car journeys represent 22% of journeys to East Linton station and 76% of journeys are undertaken by active modes.
Rail journey purpose
Figure 92 below shows that at Earlston and East Linton there is an observed tendency for passengers to travel by rail for leisure journey purposes. At East Linton 95% of rail users do so for leisure purposes.
At East Linton there is a greater number of respondents who use rail for travel to employment/work and/or education, representing 35% of the sample. This is likely due to the convenience of having a local station available making this a viable commuting travel choice.
Main reasons people don’t travel by rail
Residents within Earlston outlined numerous reasons for not travelling by rail, with the most common being the convenience of travelling by car. Without a local rail station Earlston residents can only access the rail network either by private car or bus which makes it less favourable.
- 29% said the convenience of travelling by car is a reason for not travelling by rail.
- 11% said they need their car for business purposes.
- 8% said travel by rail would take longer.
- 8% said rail is too expensive.
- 10% said accessing the train is challenging due to physical limitations/disabilities.
Satisfaction with public transport services
Figure 93 shows that 46% of residents at Earlston expressed satisfaction with the available public transport services with respect to their travel needs, with 7% expressing that they were dissatisfied.
Figure 94 shows that this was mostly influenced by the availability of public transport (33%), overall journey time (24%) and service frequency (21%).
Public Transport Perceptions
As part of the counterfactual survey, respondents were encouraged to consider improvements and limitations associated with public transport provision. This question has been included in the absence of any baseline data to understand local trends. Residents were instructed to consider any changes over a three-year period to help interpret whether perceptions regarding public transport are improving, declining or are unchanged.
Table 58 presents perceptions in Earlston regarding various public transport metrics over the last three years. This data reveals that over a third of respondents responded ‘Don’t know / Not relevant’ to all categories which limits the sample size for this data.
Notwithstanding this, those who responded to this question demonstrate broadly a more positive outlook on public transport:
- Availability of public transport services: 38% Improved, 21% no change, 9% worsened.
- Frequency of service: 30% Improved, 23% no change, 12% worsened.
- Public transport Fares: 12% Improved, 22% no change, 27% worsened.
- Length of operating day: 22% Improved, 31% no change, 9% worsened.
- Journey time by public transport: 19% Improved, 38% no change, 9% worsened.
- Reliability of public transport: 25% Improved, 30% no change, 10% worsened.
- Public transport meeting travel needs: 30% Improved, 27% no change, 7% worsened.
| Location | Improved a lot | Improved a little | No change | Worsened a little | Worsened a lot | Don't know / Not relevant | Base |
|---|---|---|---|---|---|---|---|
| Availability of public transport services | 16% | 22% | 21% | 8% | 1% | 33% | 101 |
| Frequency | 4% | 26% | 23% | 12% | 0% | 36% | 101 |
| Fares | 1% | 11% | 22% | 18% | 9% | 40% | 101 |
| Length of operating day | 4% | 18% | 31% | 7% | 2% | 38% | 100 |
| Overall journey time | 5% | 14% | 38% | 7% | 2% | 35% | 101 |
| Reliability | 4% | 21% | 30% | 8% | 2% | 34% | 99 |
| Current public transport overall with respect to your household's travel needs | 10% | 20% | 27% | 7% | 0% | 35% | 99 |
Summary of Findings
The Earlston counterfactual survey provides a benchmark for how travel behaviour may have evolved in East Linton without a local rail station. Overall, travel patterns in Earlston remain strongly car dependent, with limited rail use and longer journey times, highlighting the station’s impact upon travel patterns in East Linton.
Commuting in Earlston is dominated by private vehicle use, with most residents travelling five or more days per week and around 84% commuting by car, with no rail use. By contrast, East Linton shows lower commuting frequency and a higher rail mode share, reflecting more flexible and multimodal travel patterns.
Leisure travel in Earlston is largely local, focused on nearby towns such as Galashiels and Kelso, with car use dominant. Rail is used only occasionally for longer trips, whereas in East Linton it plays a much greater role in both regional and longer distance journeys.
Rail usage overall is much lower in Earlston, with around 30% of residents using rail in the past year. Accessibility is also weaker, with longer journey times and indirect access to the rail network compared to East Linton’s faster, more direct connectivity.
Public transport perceptions are mixed, with around half satisfied with the public transport service provision in Earlston with respect to their travel needs, but cost, convenience and car preference remain key barriers.
In summary, the counterfactual survey demonstrates that, in the absence of a local station, travel behaviour remains dominated by car use, is less flexible and more locally focused, with limited rail uptake and weaker connectivity to regional centres. The comparison therefore supports the conclusion that East Linton Railway Station has enabled a meaningful increase in rail use and improved accessibility even though car use continues to play a major role in overall travel behaviour.