Annual Population Survey - Frequency of speaking Welsh by local authority and year

Contents

Standard shorthand is used in this table: [a] = average, [t] = total, [u] = low reliability, [x] = not available.

View options

Filters

Data description ( of 2 selected2 selectable options)

Select all Data description|Select None of Data description

Frequency of speaking Welsh ( of 4 selected4 selectable options)

Select all Frequency of speaking Welsh|Select None of Frequency of speaking Welsh

Local Authority ( of 23 selected23 selectable options)

Select all Local Authority|Select None of Local Authority
Select all Wales|Select None of Wales

Year ( of 85 selected85 selectable options)

Select all Year|Select None of Year

Data valuesData descriptionFrequency of speaking WelshLocal AuthorityYear
1.5PercentageNeverNewportYear ending 31 Dec 2014
1.1PercentageNeverNewportYear ending 30 Sep 2014
1.2PercentageNeverNewportYear ending 30 Jun 2014
1.1PercentageNeverNewportYear ending 31 Mar 2014
1.1PercentageNeverNewportYear ending 31 Dec 2013
1.3PercentageNeverNewportYear ending 30 Sep 2013
1.4PercentageNeverNewportYear ending 30 Jun 2013
1.2PercentageNeverNewportYear ending 31 Mar 2013
1.2PercentageNeverNewportYear ending 31 Dec 2012
1.3PercentageNeverNewportYear ending 30 Sep 2012
1.1PercentageNeverNewportYear ending 30 Jun 2012
1.1PercentageNeverNewportYear ending 31 Mar 2012
1.3PercentageNeverNewportYear ending 31 Dec 2011
1.2PercentageNeverNewportYear ending 30 Sep 2011
1.0PercentageNeverNewportYear ending 30 Jun 2011
0.9PercentageNeverNewportYear ending 31 Mar 2011
0.8PercentageNeverNewportYear ending 31 Dec 2010
1.0PercentageNeverNewportYear ending 30 Sep 2010
1.3PercentageNeverNewportYear ending 30 Jun 2010
1.4PercentageNeverNewportYear ending 31 Mar 2010
1.5PercentageNeverNewportYear ending 31 Dec 2009
1.6PercentageNeverNewportYear ending 30 Sep 2009
2.2 [u]PercentageNeverNewportYear ending 30 Jun 2009
2.1 [u]PercentageNeverNewportYear ending 31 Mar 2009
2.3 [u]PercentageNeverNewportYear ending 31 Dec 2008
2.3 [u]PercentageNeverNewportYear ending 30 Sep 2008
2.4 [u]PercentageNeverNewportYear ending 30 Jun 2008
2.6 [u]PercentageNeverNewportYear ending 31 Mar 2008
2.5 [u]PercentageNeverNewportYear ending 31 Dec 2007
2.1 [u]PercentageNeverNewportYear ending 30 Sep 2007
1.9PercentageNeverNewportYear ending 30 Jun 2007
2.5PercentageNeverNewportYear ending 31 Mar 2007
Showing 13,401 to 13,432 of 13,432 rows
Page 135 of 135

Main information

Most recent update
25 September 2025
Next update expected
9 October 2025
Designation
Official statistics
Data provider
Office for National Statistics (ONS)
Data source
Annual Population Survey (APS)

Data notes

Revisions
  • 25 September 2025
  • 25 September 2025
Rounding applied

Figures are rounded to the nearest 100 and so there may be some apparent slight discrepancies between the sum of constituent items and the totals as shown.

Overview

Summary of dataset and variables

This dataset shows how often people aged 3 or older say they speak Welsh, by local authority and year ending date.

The information is broken down by:

  • Local authority – the area where people live.
  • Year ending – the period the estimate relates to (for example, year ending 31 March 2023).
  • Frequency category – how often people say they speak Welsh (e.g. daily, weekly, less often, never).
  • Measure – the estimated number or percentage of people in each frequency category.

You can use this dataset to compare how often Welsh is spoken in different parts of Wales and how this changes over time.

Data collection or calculation

The data come from the Annual Population Survey (APS), which is an annual dataset derived from the Labour Force Survey (LFS) carried out by the Office for National Statistics (ONS). Local authority level data have been available since 2001, when an enhanced sample (around 350% larger than the standard survey) was introduced.

Since 2004, the APS has been produced on a rolling annual basis, with each rolling annual average covering a 12-month period and updated every three months. This provides more timely data while overlapping periods by nine months. The rolling dataset is boosted with additional sampled respondents to improve the robustness of estimates, particularly for local-level geographies.

For the years 2001–2003, Welsh-specific weights are applied to better reflect population figures for local authorities.

In mid-March 2020, APS interviews switched from face-to-face to telephone because of the COVID-19 pandemic. The ONS monitored the impact and re-weighted the data accordingly. Face-to-face interviews were reintroduced in autumn 2023. Comparing pre-2020 face-to-face interviews with telephone interviews suggested a slight increase in respondents reporting Welsh-speaking ability, so caution is advised when interpreting trends across this period.

The survey collects information on the population’s ability to speak Welsh. Respondents aged 3 or older are asked whether they can speak Welsh, and estimates are presented as counts or percentages of Welsh speakers.

Standard analysis and weighting methods are applied to ensure the estimates are representative of the population.

Statistical quality

These estimates were previously classified as accredited official statistics. Due to a fall in sample sizes in the Annual Population Survey (APS) in recent years, and because the survey has not been reweighted to the latest population estimates, this accreditation has been temporarily suspended. The estimates are now designated as official statistics.

The statistics remain suitable for use, but users should note that there is increased uncertainty, particularly for smaller geographic areas or population sub-groups, where estimates are less reliable. Trends should be considered alongside other sources of information on Welsh speakers, such as the National Survey for Wales. The Welsh Government considers the census of population to be the key source of information to measure the number of Welsh speakers in Wales.

All APS data are sample-based, meaning results are estimates rather than exact counts. The true value of any measure lies within a range around the estimate, with variability increasing as data become more detailed. For example, local authority-level estimates are subject to greater uncertainty than national-level estimates.

APS responses are weighted to official population projections. Initially, 2020 data were based on 2018 population projections, which did not account for changes from the COVID-19 pandemic. In September 2021, APS data were reweighted using updated population estimates derived from HM Revenue and Customs Real Time Information (RTI), improving accuracy for overall rates and levels.

These changes reduce potential bias in higher-level aggregates, but some smaller breakdowns may still be affected due to reduced sample sizes during the pandemic, leading to more extreme variation in estimates.

Related reports

Published by

Organisation
Welsh Government
Contact email
welshlanguagedata@gov.wales

Select the filtered or whole dataset to download

Select download format

Contains only the dataset
Contains only the dataset
Contains only the dataset

Select number formatting

This includes rounding to decimal places and commas to separate thousands
This includes rounding to decimal places and commas to separate thousands

Select language

Metadata

The metadata for this dataset is not included when you download the data. You can download this separately. This contains all the information from the 'About this dataset' tab.

Download metadata