There are four areas of change for this release, summarised below. More information about the methods used in this release is available in the accompanying methodology.
Changes in demand indicators
Two indicators have been removed from the analysis:
- Skills shortage vacancy density has been removed because the latest release of this data (Employer Skills Survey 2024) does not include the occupational breakdown used in this analysis (SOC 2020 unit groups).
- Annual change in contract or temporary workers has been removed following a review of the quality and relevance of this indicator, which concluded the data was not robust enough to provide a reliable indication of demand.
Three indicators have changed:
- Visa application density has been replaced with visa grant density, following consultation with experts on the most appropriate measure to capture demand.
- Annual change in hourly wage is now based on a percentage change instead of an absolute change, to account for differences in average wage across occupations.
- Online job advert density now uses quarterly snapshot job advert data rather than monthly new job adverts; these snapshots are more stable so are more suitable for analysing long term trends.
Assessment relative to historical trends
In the previous methodology, data for all indicators were converted to the same scale to allow them to be compared. This scale was based on the distribution of data within the year of analysis. In the new methodology, the scale is based on the distribution of data within a historical comparison period. This means that occupations can be assessed as substantially higher or above average compared to historical trends, and comparisons can be made between years.
The visa density indicator is an exception to the above and will continue to be scaled relative to the year of analysis. This is because changes in visa policy make it more difficult to define an appropriate historical reference period.
Accounting for uncertainty
Many of the inputs into the Occupations in Demand analysis contain uncertainty, for example because they use data from sample surveys which means the survey estimate might differ from the ‘true value’ the survey is trying to measure. In the new methodology, information about this uncertainty (confidence intervals) has been used to calculate a lower estimate for each indicator, which reflects how much lower the ‘true value’ might be. It is this lower value that is used for demand categorisation, to reduce the chance of random variation in the data moving an occupation into higher demand levels.
Where there are other forms of uncertainty, for example trends unrelated to demand have been identified which may be influencing the data, this is indicated by uncertainty variables in the accompanying datasets.
When data at a 4-digit SOC 2020 level falls below suppression thresholds it is imputed, when possible, using 3-digit SOC 2020 data. If imputed data is used then demand for that indicator is capped at elevated demand, to reduce the chance of moving an occupation into the highest demand level based on data influenced by other occupations in the 3-digit group. Previously, if a large number of demand indicators were imputed then the occupation was excluded from the analysis.
Changes to overall demand classification
In the previous methodology, overall demand categories were produced by creating a weighted average of all individual indicators called the ‘demand index’. Occupations which were substantially higher (above the mean plus one standard deviation) or above average on this index were categorised as critical and elevated demand respectively.
In the new methodology, occupations are first categorised as critical or elevated for each individual indicator using the thresholds specific to that indicator. Occupations are then categorised as in critical or elevated demand overall based on the number of indicators that are in critical or elevated demand. The criteria are set out in full in the methodology.
This change provides greater transparency over the contribution of individual indicators to the final categorisation. It also ensures that there must be signs of demand from at least two indicators for an occupation to be in one of the high demand categories, and occupations are not penalised where an occupation shows low demand on an indicator for technical reasons, e.g. ineligibility for visas.
A demand index has been produced for this release; however, it is provided for information only and is not used in the categorisation process.