Quality

Data sources

1.7 The Wellbeing of Wales report relies on statistics and data from a wide range of sources. The dashboard covers 50 national indicators, and these data along with others are used as the basis of the narrative in the report. The data were selected for their accuracy and robustness as varied measures of the seven well-being goals. Wherever possible, existing data sources are used, and the statistics team has taken a pragmatic approach to using existing official statistics where possible, supplementing these with data from other sources where this will add value to the output.

1.8 The statistics team is proactive in improving coverage and quality of existing indicators and adding new indicators. This reflects a strong culture of improvement and innovation through the continuous process of indicator selection and refinement. For example, Welsh Government has:

1. added four new indicators following extensive consultation and engagement to ensure that the indicator set remained relevant in the light of societal changes following the pandemic. These include the percentage of people who have confidence in the justice system, and an indicator about housing affordability. The team is in the process of developing questions for the National Survey for Wales to measure the former, and using data from an existing source – the Family Resources Survey – for the latter indicator, which had widespread interest from users.
2. Improved some indicators. For example, the team has worked with experts in the sector to develop a new method to fill a gap in the indicator set on biodiversity. More generally the team has improved coverage of protected characteristics, or signposted to where this information can be found, across a range of indicators.

Documentation about quality

1.9 The narrative within the report allows the statisticians to incorporate information about quality and uncertainty of the statistics on a case-by-case basis to help users interpret the statistics appropriately. For example, the 2021 report includes statements such as “Most labour market data is survey-based and volatile, and short run changes should not be over-interpreted” which gives a really clear sense of the uncertainties around the data. In other cases, the report notes where differences are statistically significant. However, statements such as “The percentage of pensioners in relative income poverty fell by 1 percentage point for the second consecutive period following a gradual rise since around 2013” from the 2021 report would be better if they provided greater context around whether such a small numerical value actually represents a real-world change or not. We welcome the statistics team’s commitment to review how it presents uncertainty within the report.

1.10 The report also outlines where data are from official statistics (including those designated as National Statistics) or other sources to help support the understanding of data quality. The report is accompanied by an extensive range of information about the quality of the data used. Separate quality reports for indicator data and for other contextual information used in the report provide information about quality and links to further information, to help users interpret the statistics. The technical descriptions and source of each indicator are comprehensively documented along with a range of links, all of which helps users to understand the quality of the sources and the rationale for inclusion. For example, the inclusion of “Percentage of live single births with a birth weight of under 2,500g” is explained with reference to the fact that low birth weights are associated with health risks in an infant’s first year of life, alongside the source and link to the data.