Official Statistics

NTS mid-year estimates, year ending June 2024: technical note

Published 16 April 2025

Applies to England

Chapter 1: Background

The National Travel Survey (NTS) provides up-to-date and regular information about personal travel within Great Britain and monitors trends in travel behaviour. First commissioned in 1965/1966, since July 1988 the NTS has been a continuous survey (that is, fieldwork conducted on a monthly basis starting in January and ending in December).

Since January 2002, the Department for Transport (DfT) has commissioned the National Centre for Social Research (NatCen), an independent social research institute, as the contractor for the NTS. NatCen is responsible for questionnaire development, sample selection, data collection and editing, data file production and building the database. The DfT is responsible for data analysis, publication, and archiving.

Reporting of NTS results has been annual, based on a calendar year of data collection. However, starting from NTS 2023, an additional mid-year data delivery has been commissioned for 12 months’ worth of data collected from July of the previous survey year to June of the most recent survey year. As outlined in this Technical Note, this mid-year data delivery therefore combines the first 6 months of NTS 2024 with the last 6 months of NTS 2023, to give a more recent 12-month data delivery than the 2023 annual data.

This note briefly describes the differences as to sample design, survey content, and data preparation. Please see the 2023 Technical Report for full information of the overall methodology and approaches used on NTS. The detailed 2024 Technical Report will be provided with the annual data in the summer of 2025.

Chapter 2: Fieldwork

Historically, the NTS has used two data collection methods: face to face (F2F) interviewing using computer assisted personal interviewing (CAPI) and respondent self-completion of a 7-day travel record (issued as a paper diary). However, in response to the pandemic, for respondents with COVID-19 or shielding or otherwise unable to take part face-to-face due to COVID-19, a phone back-up (PB) alternative has also been available. This phone back-up began in April 2022 and continued throughout 2023 and 2024, although gradually being used less over time. The month with the highest rate of PB during the period July 2023 to June 2024 was September 2023 (8.2% of fully productive cases) and the month with the lowest rate was May 2024 (1.1% of fully productive cases). The average PB rate in July to December 2023 was 6.4%, and in January to June 2024 it was 1.9%.

All approaches used the same CAPI questionnaire, but in cases where F2F was not possible the interviews were administered by telephone and the travel diaries were populated by interviewers collecting travel data from participants over the phone.

A small number of changes were made to the CAPI questionnaire for 2024, of which the main ones were:

• new questions about electric and plug-in hybrid vehicle charging
• new questions about how vehicles are owned and acquired
• the amendment of the Sex question so that it is directly asked of respondents (previously this was coded by the interviewer from observation), and reversal of the response order so that ‘Female’ now coded first
• the question about the type of goods ordered by phone, post or the internet (OrderA) now being asked every year, rather than every other year
• the follow-up question, now also asked every year, about the frequency of goods deliveries (Deliv) being split in two to separate takeaways and food shopping deliveries (now asked in Deliv1) from non-food shopping deliveries (now asked in Deliv2)
• the reinstatement of a question in the self-completion (CASI) module, to be asked every year and with updated response options, about things that are important to the respondent when buying a car or van (CarStyle)

For full details on these changes please contact the National Travel Survey team.

The overall sequence of tasks, from the initial sample selection to data analysis and reporting remained the same across face-to-face and telephone interviewing:

1. Sample selection from the Postcode Address File (PAF).

2. Questionnaire development and fieldwork preparation.

3. Fieldwork.

4. Data transmission and documents returned by interviewers.

5. In-house data input and editing.

6. Quality checking.

7. Data checking using SPSS.

8. Data file production, variable derivation and imputation and delivery to DfT.

9. Data analysis and reporting by DfT.

All aspects of the fieldwork sequence are as described in the 2023 Technical Report. This fieldwork sequence continued into NTS 2024.

For the beginning of 2024, administrative fieldwork information was managed using an electronic address record form (eARF), having previously been recorded on a paper ARF and then transferred to the CAPI questionnaire Admin block. This includes: eligibility and observations of the sampled address; refusal responses; a record of the number of dwelling units and households found at the address and (in cases with more than one dwelling unit or household) an account of the random selection process to select just one of those for interview; as well as a record of the number of contacts and contact attempts made at the address. Data recorded in the new eARF was compared to data from the paper ARF in the previous year to explore whether this new mode of recording the information was having an impact on the data, and no notable deviations were detected between the two survey years.

Chapter 3: Sample selection and size

The combined 2023 and 2024 mid-year data is based on a random sample of 25,608 private households in England, drawn from the Postcode Address File (PAF), which is split evenly between the 2023 and the 2024 portions of the sample (12,804 households in each 6-month period).

The 2023 and 2024 samples were drawn separately, but follow the same random sampling principles (including most of the stratification and the clustering), which are outlined in the 2023 Technical Report. For 2024 there was one minor change to the sampling approach compared to 2023. Census 2021 data had been released between the NTS 2023 and 2024 sampling, therefore the 2024 sampling used updated stratification variables. However, the Census 2021 data was gathered during COVID-19 lockdown restrictions, and this had an impact on the fourth stratification variable: the percentage of people working from home. After comparing Census 2011 and 2021 profiles for this variable, it was judged to be inappropriate to use the 2021 version as it had been skewed by temporary restrictions.

Accordingly, only the former three stratification variables were used to draw new NTS 2024 PSUs: International Territorial Level 2 (ITL2) region, urban-rural status, and the percentage of households with no car. The third of these variables was used in continuous form rather than in three categories (tertiles), as was the case for NTS 2023. This interim solution was required as the sampling stratification review had not yet evaluated the appropriateness of all NTS sampling stratification variables using new 2021 census data. This review of the 2023 sampling stratification has since been published. The review’s recommendations were implemented in the NTS 2025 sampling, replacing the percentage working from home with the percentage travelling to work by car or van as the fourth and final stratification variable.

For both 2023 and 2024, the fieldwork points (also known as assignments) included 22 addresses. Each of the 22 addresses in an interviewer’s point quota for 2023 and 2024 continued to have a 7-day period (Travel Week) allocated to it, during which participants’ travel details were recorded. Interviewers were given a unique Travel Week Allocation Card with 22 Travel Weeks for each point, the start dates of which were spread evenly across a period of 31 days. Interviewers followed the traditional, methodical process to allocate Travel Weeks to individual addresses from this card.

Chapter 4: Response

Tables 1.1 to 1.3 show the response rates for the combined 2023 and 2024 mid-year data, as well as the response rates for the two individual six-monthly periods. The “set sample” figure represents the total number of addresses in the sample invited to take part in the study, and all figures shown are at the household level.

Note: The response rate tables below show figures for the fieldwork outcomes at the household level. All fully productive households are given a weight, as are partially productive households where all household members have completed a full interview (but not necessarily completed a full travel diary). A small number of partially productive households are not given a weight where not all household members have completed a full interview. Therefore, base sizes in the final data may not always exactly replicate the figures in these response rate tables.

These tables demonstrate the similarity in response between the two periods. The standard response rate was 31% (3,649 households) for the first six months of 2024 compared to 30% (3,510 households) for the last six months of 2023. Based on this, a response rate of 31% (7,158 households) was achieved for the combined mid-year period overall (July 2023 to June 2024).

The tables also show that for eligible households rates of partially co-operating households (6%), refusals to co-operate and other unproductives (50%), and non-contacts (13%) were the same across both the 2023 and 2024 portions of the sample.

Table 1.1: NTS response rates, mid-year data (July 2023 to June 2024)

Category	Achieved Sample	Achieved Sample Response Rate	Standard Response Rate
Set sample	25,608	100%	not applicable
Ineligible or deadwood	1,915	7%	not applicable
Unknown eligibility	3,201	13%	not applicable
Eligible households	23,420	91%	100%
Fully co-operating	7,158	28%	31%
Partially co-operating	1,460	6%	6%
Refusal to co-operate and other unproductive	9,414	37%	50%
Non-contact	2,460	10%	13%

Table 1.2: NTS response rates, July to December 2023

Category	Achieved Sample	Achieved Sample Response Rate	Standard Response Rate
Set sample	12,804	100%	not applicable
Ineligible or deadwood	940	7%	not applicable
Unknown eligibility	1,841	14%	not applicable
Eligible households	11,706	91%	100%
Fully co-operating	3,510	27%	30%
Partially co-operating	740	6%	6%
Refusal to co-operate and other unproductive	4,569	36%	50%
Non-contact	1,204	9%	13%

Table 1.3: NTS response rates, January to June 2024

Category	Achieved Sample	Achieved Sample Response Rate	Standard Response Rate
Set sample	12,804	100%	not applicable
Ineligible or deadwood	975	8%	not applicable
Unknown eligibility	1,360	11%	not applicable
Eligible households	11,713	91%	100%
Fully co-operating	3,648	28%	31%
Partially co-operating	720	6%	6%
Refusal to co-operate and other unproductive	4,845	38%	50%
Non-contact	1,256	10%	13%

Note: The estimated number of eligible households is calculated by firstly deducting the ineligible or deadwood cases from the total sample, and then secondly deducting a proportion of the cases with unknown eligibility. The second deduction is made on the assumption that at least some of the addresses with unknown eligibility would have been ineligible or deadwood. The proportion used in this calculation is simply the inverse of the known eligibility rate.

Chapter 5: Data Processing protocols

As usual, the edited survey data was prepared for analysis and reporting before being delivered to DfT. However, as the data crosses two calendar years, some of the protocols that were followed during post-processing were unique to the combined data.

Please refer to Chapter 4 of the 2023 Technical Report for full information on how the data was processed for NTS 2023.

The protocols specific to the combined 2023 and 2024 mid-year data processing are described below.

5.1 Creating household income semi-deciles and quintiles

The protocol was applied as per NTS 2023, with 2022 income mid-points applied to the NTS 2024 months. Income mid-points for 2023 or 2024 were not available at the time of producing the data.

5.2 Adding Holidays data

The holidays database was extended in the usual manner to incorporate dates up to the end of September 2024, so that the NTS 2024 portion of the mid-year data can apply holiday status coding in the same way as the NTS 2023 portion of the data. This approach enables comparable analysis of trip data by travel day type for all months of the mid-year data delivery.

5.3 Adding concessionary travel data

The concessionary data does not change significantly year-on-year, so the 2023 concessions were applied to the NTS 2024 months for the mid-year data delivery.

Chapter 6: Weighting

The mid-year data combines the first 6 months from 2024 and the last 6 months from 2023. In practice, this includes a full year of survey data capturing travel behaviour and opinions across seasons. It was therefore possible to adapt the weighting scheme previously used for NTS calendar year data and apply it to the combined mid-year data. Both NTS 2023 and 2024 used only face-to-face (F2F) fieldwork, so there were no mixed mode impacts to adjust for.

The adjustments required for the mid-year data weighting (compared to the NTS 2023 weighting) are described in more detail below. Two recommendations from the Weighting Review published in April 2025 were tested in the mid-year weighting, specifically changing the socio-economic covariate in the household participation model and the removal of the full response model. No alterations were needed to the diary weighting method.

6.1 Household participation weights

The aim of the household participation weights is to attempt to reduce bias caused by systematic differences between the households that participated in the NTS (that is, for which a household interview was obtained) and those that did not. To generate the non-response weights, a logistic regression model was fitted with whether or not an eligible household participated as the outcome measure.

As two halves of independent NTS survey years were combined, the model was split by year whilst keeping the predictors consistent. This approach was also used in the combined 2022 and 2023 mid-year weighting. Splitting the model in this way makes it possible to account for any year-specific non-response bias. Although the data collection mode and response rates for the two halves of the data were similar, the NTS 2023 cases had a higher percentage of unworked addresses to adjust for. As ACORN codes were not available for NTS 2024 cases, this socio-economic measure was replaced for both 2023 and 2024 cases with Output Area Classification supergroups based on the 2021 Census (OAC21). This change is recommended by the Weighting Review. OAC21 was significant (that is, with a p-value less than 0.05) in both NTS 2023 and 2024 participation models, confirming that it is appropriate for inclusion.

6.2 Fully responding weights

Prior to NTS 2024, weights were generated to reduce the bias from the removal of households which did not fully respond (W4). The Weighting Review evaluated whether the model which generated the W4 weights was necessary for future survey years. The review found that the profiles of interview and fully responding samples were very similar and that W4 had minimal impact. Alternative fully responding weights without W4 were generated for NTS 2019 and 2023 data and were found to have extremely similar efficiency, bias, and impact on key NTS measures compared to the original weights. Consequently, the review recommended the removal of W4 in order to streamline the weighting method, and this recommendation has been implemented for the combined 2023 and 2024 mid-year data weighting. The fully responding weights were therefore generated using the same three steps as the interview weights (W1 × W2 × W3) and then calibrated to the same population estimates.

6.3 Calibration Weighting

For the combined 2023 and 2024 mid-year weights, the composite (household-level) weights from the previous stages were adjusted so that the distribution for groups defined by age and sex and region matched ONS 2023 mid-year population estimates of household residents. This calibration adjustment did not require trimming to improve efficiency, unlike in the NTS 2023 weighting.

The NTS 2023 full year weights were calibrated to age-sex categories and region (GOR), as well as 50% of the weights per half-year to account for more addresses being issued in the latter six months. This last adjustment was not necessary for the mid-year weights, as the same number of addresses were issued in each of the six-month periods. The mid-year weights were therefore calibrated to age-sex and region (GOR) only. The rescaling to adjust for region-specific bias within halves of the year used in NTS 2023 weighting was likewise no longer needed, so was not used for the mid-year weights.