Comparability of LFS data over time
There is a broadly consistent time series, with the exception of 2020/21, for:
- workplace injury from 2000/01;
- work-related illness from 2001/02;
- working days lost due to workplace injury from 2000/01 and due to work-related illness from 2001/02.
Whilst the small change in survey period from December-February for years prior to 2006/07 to January-March from 2006/07 onwards may have potentially introduced a discontinuity to these data series, investigations suggested that this change in survey design has not affected the top level injury, ill health and working days lost data.
The coronavirus (COVID-19) pandemic and the government’s response to its impact has had a significant effect on the labour market in 2020/21. This combined with both methodological challenges and issues with data collection mean it is difficult to interpret trends in health and safety outcomes across the latest years. The approach adopted for putting the 2020/21 data in context has been to comment on the level of workplace injury and work-related ill health in 2020/21 compared to the pre-coronavirus level and to note the pre-coronavirus trend. The effects of the pandemic mark 2020/21 out as an anomalous year and as such it is not directly comparable with earlier years. No working days lost data is available for 2020/21 as reliable estimates consistent with earlier years could not be produced. See more details on the impact of the coronavirus pandemic on health and safety statistics.
The above surveys had the same design and coverage (with the exception of the small change in survey time period in 2006/07) and used the same methods to adjust for non-response to the screening questions in each survey.
It is only possible to make broad comparisons with the earlier surveys, and even these need to be treated with caution. This is due to the differences in:
- level of information collected;
- methods used to adjust for non-response to the screening question in each survey;
- use of pre-2001 census population weights (the 2001 Census showed that the UK population was around 1 million lower than previously estimated).
To achieve even these broad based comparisons for illness data, a number of adjustments have to be made to the data to ensure comparability.