Data Sources
To understand the spatio-temporal variations in air quality and happiness across England on both a national and a regional scale, we must select our data sources carefully. Our investigation will include data on different indicators of air pollution (modelled PM2.5 & PM10 background pollution levels) from the Department for Environment, Food & Rural Affairs (DEFRA) and mean rating responses to questions about happiness, contained in a well-being survey published by the Office for National Statistics (ONS).
Department for Environment Food & Rural Affairs (DEFRA)
Office for National Statistics (ONS)
Happiness Data: Happy Means
Source: We will use ‘happy means’ scores- the average ranking of the response to: ‘Overall, how happy did you feel yesterday?’ on a scale from 0-10 or ‘not at all’ to ‘completely’-- reported to the ONS in Life Satisfaction Surveys (LSS) in the years 2011-2021 at the Local Authority level-- an ONS boundary concept, each assigned an area code. The data is available here (it is April 2020 to March 2021 Local Authority Update Edition).
​
Analysis: While ONS is an independent agency whose published statistics are generally trusted-- a UK Statistics Authority survey reported an 85% public approval rating in 2018 (Morgan & Cant, 2018, p. 3)-- the outputs must be scrutinised as the data may be manipulated for political gain. The ONS now forms the ‘single organisation for the collection, primary analysis and publication of statistics in government’ and holds data transparency as one of the guiding principles for statistics to be published (Economic Secretary to the Treasury, 1998, p. 11), yet it has a troubling past. In 1980 Sir Derek Rayner suggested that statistical information should be published to ‘satisfy the needs of the Government first’ (Economic Secretary to the Treasury, 1998, p. 8).
Air Pollution Data: PM2.5 & PM10 Modelled Background Pollution Data
Source: We selected modelled data on PM2.5 and PM10 concentrations-- measured in µg/m3 at a 1x1km resolution across sites in England from 2011-2021: each point was assigned a UK Grid Code and labelled using coordinates. The data for PM2.5 is available here, and the data for PM10 is available here. While pollution data is available for the years 2001-2021, "happy means" is only available from 2011-2021: therefore we only selected the sheets covering this timeframe.
​
Analysis: DEFRA (the Department for Environment, Food & Rural Affairs) is a ministerial department led by appointed Members of Parliament. While it is an 'advisory body, with no executive powers or responsibilities' (Gov.uk, n.d.), the pollution data could be subject to scrutiny as the department has a history of poor operational procedures. DEFRA has the broken the Code of Practice for Official Statistics for various matters, including a pre-release access breach for data on UK emissions of air pollutants in 2009 (Department for Environment, Food & Rural Affairs (a), 2022). This is problematic, considering that one of DEFRA’s priorities is to ‘deliver new approaches to tackling all sources of air pollution’ (DEFRA, 2018, p. 5). Additionally, certain datasets are only available for DEFRA contractor use (Department for Environment Food & Rural Affairs (b), 2022), which limits the scope of public investigation on important environmental matters.
A Note on Bias
While bias cannot be completely eliminated from an investigation, our approach will attempt to minimise the effect that response bias may have (from the survey data from the Office of National Statistics), by evaluating the differences between areas on a much smaller scale (across sites in the the North and South of England). While this does not increase the reliability of this method, the smaller scale analysis could produce more valid results as it does not make sweeping generalisations about the relationship between happiness and air pollution on a national scale. By assuming air pollution can be represented as monitored PM2.5 and PM10 levels, this could procure more precise results. However, it may also reduce the validity of the results as it assumes that these indicators best represent how pollution exists. This may have real social consequences if direct policy action is taken to alleviate the problem through this lens only.