106 . In practice , this means that in the analysis that follows , the group of learners being compared with those from Uni Connect areas has the same mix of these listed characteristics . The key difference is that one group was living in Uni Connect areas in Key Stage 4 , while the other was not . Although there will of course be other unobserved differences and even differences within the categories of matched characteristics , such as the exact GCSE grades achieved by each learner beyond the number of ‘ standard passes ’.
107 . The group of learners from non-Uni Connect areas is known as the ‘ matched counterfactual ’ group , because it represents a hypothetical situation where learners from Uni Connect areas had instead come from non-Uni Connect areas . We repeated this hypothetical situation 1,000 times to ensure the findings we obtained were not simply by chance . This gives us 1,000 different matched counterfactual groups whose application outcomes we can then compare against the same group of learners from Uni Connect areas each time .
108 . The datafile associated with this report contains the estimates for all 1,000 matched counterfactual groups , in addition to the sampling rates for each of these groups .
Findings
Application rates
109 . Figure 13 below shows the difference in application rates between learners from Uni Connect areas and the average application rate of all 1,000 matched counterfactual groups . This gap is clearly much smaller than the observed gap for the whole population , as was shown in Figure 8 . In fact , in the 2021 application cycle , the observed gap in application rates for the whole population stood at 17.1 percentage points , compared with just 6.1 percentage points after underlying differences in characteristics between the two groups were taken into account through matching .
110 . This reduction in the gap after matching suggests that , as previously stated , at least some of the characteristics used for matching are : ( a ) associated with application rates ; and ( b ) unequally distributed across the two groups of learners . In other words , matching ensures that the remaining gap in application rates can no longer be accounted for by differences in matched characteristics . Instead , the only known difference between the two groups is that one was living in a Uni Connect target area and the other was not ( although there will remain other unobserved differences ).
111 . Therefore , if this gap is found to be narrowing over time , this would suggest that the Uni Connect programme is associated with a relative improvement in application rates in targeted areas , after controlling for differences in characteristics between the two groups of learners in terms of matched characteristics . This would suggest the primary aim of the Uni Connect programme , to raise participation in underrepresented areas , is being met .
112 . However , Figure 13 appears to show that the average gap across all 1,000 matched counterfactual groups was wider in the 2021 application cycle ( which was affected by the COVID-19 pandemic ), at 6.1 percentage points , compared with the 5.0 percentage points gap in 2016 , before the Uni Connect programme began . This appears to have reversed a narrowing of the gap in the 2020 cycle , which then stood at 4.6 percentage points .
40