Table 2

Testing changes in psychological distress over time among young adults aged 16–24 living in England, by different subgroups

Subgroups	Between 2009–2010 and 2018–2019 (n=5873)			Between 2018–2019 and April–November 2020 (n=4302)
Subgroups	AME of time on GHQ	95% CI	Interaction p value	AME of time on GHQ	95% CI	Interaction p value
All	1.66	1.25 to 2.07		1.96	1.27 to 2.66
Sex			0.024			0.364
Male	1.25	0.73 to 1.76		1.65	0.64 to 2.66
Female	2.09	1.51 to 2.66		2.26	1.26 to 3.16
Age			<0.001			0.695
16–18	2.64	2.04 to 3.23		1.28	−0.71 to 3.26
19–21	1.15	0.53 to 1.78	0.001	2.02	1.10 to 2.94	0.474
22–24	0.86	−0.01 to 1.72	0.001	2.23	1.21 to 3.25	0.396
Economic activity			0.103			0.899
FT employed	0.84	0.11 to 1.58		1.89	0.93 to 2.85
PT employed	2.16	1.05 to 3.27	0.049	1.88	0.71 to 3.04	0.991
Unemployed	2.15	0.68 to 3.62	0.114	1.95	−0.57 to 4.47	0.962
FT education	1.61	1.11 to 2.11	0.083	2.03	0.76 to 3.30	0.851
Out of labour force	3.59	0.99 to 6.20	0.045	0.41	−2.50 to 3.31	0.343
Living arrangements			0.861			0.447
With parent(s)	1.68	1.26 to 2.09		2.04	1.29 to 2.78
Not with parent(s)	1.54	0.08 to 3.01		1.41	−0.06 to 2.88
Parental education			0.117			0.916
Degree	1.87	1.36 to 2.39		1.93	1.04 to 2.81
No degree	1.24	0.62 to 1.86		1.99	1.10 to 2.87
Area deprivation			0.642			0.021
Most deprived	1.29	0.46 to 2.12		4.11	2.41 to 5.81
Second most deprived	1.92	1.11 to 2.74	0.525	1.36	0.22 to 2.51	0.454
Second least deprived	1.53	0.77 to 2.29	0.902	1.80	0.44 to 3.15	0.823
Least deprived	1.86	1.16 to 2.55	0.301	1.21	0.30 to 2.11	0.002
Ethnicity			0.006			0.394
White UK	1.96	1.50 to 2.42		1.78	0.98 to 2.59
White other	2.08	−0.68 to 4.84	0.933	2.55	−1.61 to 6.71	0.722
Mixed	−0.24	−1.56 to 1.10	0.002	4.43	2.04 to 6.81	0.037
Indian	1.54	−0.38 to 3.46	0.673	2.28	−0.69 to 5.24	0.755
Pakistani and Bangladeshi	0.35	−1.03 to 1.74	0.029	2.25	1.05 to 3.46	0.514
Black	0.04	−1.68 to 1.77	0.033	0.87	−1.77 to 3.51	0.510
Other ethnic groups	−0.98	−2.75 to 0.79	0.002	3.60	0.69 to 6.51	0.234

Understanding Society, from 2009–2010 to 2018–2019 and from 2018–2019 to 2020.
Estimates come from fully adjusted linear models in pooled samples of observations using the wave-specific cross-sectional weights, clustering on individuals. Interactions were then tested for each variable in separate models. AMEs are based on complete-case models adjusted for all other predictors (age, sex, economic activity, living arrangements, parental education, area deprivation and ethnicity).
P values reported next to variable names refer to global tests of differences in AMEs across categories. P values reported next to variable categories refer to test of differences in AMEs with respect to the reference group.
Significant estimates at p<0.05 are in bold.
AME, average marginal effect; FT, full time; GHQ, General Health Questionnaire; PT, part time.