Sub-analysis of Strugglers identified as APC attendees or course terminations vs controls [posted as supplied by author]

The Strugglers were identified in one of four ways, but could not be strictly defined as the groups sometimes overlapped, eg an APC attendee might also have a course suspension. However, those identified primarily as APC attendees or course terminations might be expected to be experiencing the most severe difficulties, so a multivariate analysis was carried out in these two sub-groups combined against the controls. The results are shown below; as in the main analyses, we have looked at all students, excluding ethnicity as a variable, and UK students only, including ethnicity. Independent predictors are similar to those shown before with some variation in likelihood ratios and confidence intervals.

Predictor	LR (95% CI)	p
All APC & terminated vs control students, ethnicity excluded, n = 502

Negative comments in reference	3.31 (1.80 to 6.08)	<0.001
Male	2.50 (1.36 to 4.61)	0.003
Late offer of a place	2.31 (1.18 to 4.53)	0.015
Lower A level MEG	2.22 (1.16 to 4.23)	0.016
Lower GCSE science MEG	2.51 (1.09 to 5.81)	0.031


All APC & terminated vs control students, UK only, ethnicity included, n = 436

Non-white ethnicity	6.84 (3.13 to 14.92)	<0.001
Lower A level MEG	3.16 (1.43 to 6.99)	0.004
Male	2.91 (1.38 to 6.13)	0.005
Negative comments in reference	2.23 (1.06 to 4.71)	0.035
Late offer of a place	3.16 (1.43 to 6.99)	0.004