Young People’s Subjective Wellbeing in the Wake of the COVID-19 Pandemic: Evidence from a Representative Cohort Study in England

Jake Anders; Erica Holt-White

doi:10.1007/s11205-025-03776-7

. 2026 Jan 23;181(2):62. doi: 10.1007/s11205-025-03776-7

Young People’s Subjective Wellbeing in the Wake of the COVID-19 Pandemic: Evidence from a Representative Cohort Study in England

Jake Anders ^1,^✉, Erica Holt-White ²

PMCID: PMC12830484 PMID: 41582941

Abstract

The COVID-19 pandemic and the disruption it has caused had substantial short-term effects on young people. These effects have been found to be highly unequal, exacerbating existing inequalities in society, including those associated with socio-economic status, gender and ethnicity. But, just as importantly, it is believed that they continue to cast a long shadow over some young people’s lives. In this paper we use data from the COVID Social Mobility & Opportunities study (COSMO) — a representative cohort study of over 13,000 young people in England aged 14–15 at pandemic onset whose education and post-16 transitions were acutely affected by the pandemic’s disruption through their remaining education and subsequent transitions — to highlight inequalities in young people’s subjective wellbeing and mental health in the wake of the pandemic. We document the substantial differences in subjective wellbeing — especially highlighting differences by gender — after adjusting for other demographic characteristics, self-reported levels of social support, and experience of adverse life events. We estimate how wellbeing differs by young people’s own perceptions of the ongoing impact of the pandemic: those who indicate an ongoing negative impact in their lives have substantially lower subjective wellbeing scores. Finally, we find a link between adverse life experiences during the pandemic and lower post-pandemic wellbeing, but do not find evidence that this is mediated by demographic characteristics or social support.

Keywords: COVID-19, Young people, Subjective wellbeing, Inequalities, Adverse life events, Social support

Introduction

The COVID-19 pandemic and the disruption it caused had substantial short-term effects on young people’s lives around the world, with evidence of significant impacts on young people’s wellbeing and mental health (De France et al., 2022; Wolf & Schmitz, 2024). Young people in England, the focus of this paper, were no exception: extended periods in which in-person schooling was suspended (Anders, 2024) interrupted pupils’ learning (Jakubowski et al., 2025) and social lives (Kalenkoski & Pabilonia, 2024), with consequent rises in loneliness a clear symptom of this (Kung et al., 2023). This widespread disruption had widely documented short-term effects on young people’s wellbeing (e.g. Attwood & Jarrold, 2023; Banks & Xu, 2020; Neugebauer et al., 2024; Newlove-Delgado et al., 2021; Quintana-Domeque & Zeng, 2023), the magnitude of which was found to be linked with the intensity of lockdown restrictions (Owens et al., 2022), and the immediacy of which is reflected in wellbeing increasing and decreasing as restrictions tightened and eased (Creswell et al., 2021). A review by Kauhanen et al. (2023) summarised the international picture as “a longitudinal deterioration in symptoms for different mental health outcomes especially for adolescents and young people”.1

Existing analyses suggest that effects of the disruption were unequal, often exacerbating existing demographic inequalities in society, including those associated with socioeconomic status (e.g., Ravens-Sieberer et al., 2022), gender (e.g., Anders et al., 2023; Davillas & Jones, 2021), and ethnicity (e.g., Proto & Quintana-Domeque, 2021). Wolf and Schmitz (2024) finds that older adolescents were particularly affected, perhaps as these are such formative years for social relationships and critical years for education and subsequent transitions.

Variation in experiences and support during the pandemic has also been found to be important for young people’s wellbeing. Restrictions on social activities and the closure of schools reduced physical activity for some, which has been linked to worse mental health outcomes (Samji et al., 2022); other aspects of the pandemic are likely to have exacerbated the prevalence of adverse life events that previous studies have shown affect wellbeing (Cleland et al., 2016). Conversely, social support has been identified as a potential buffer to negative impacts (Demaray et al., 2005; Racine et al., 2021; Siedlecki et al., 2014) of such negative stressors. These highlight the potential importance of experiences and social support during the pandemic for young people’s wellbeing and, hence, the need to consider these in understanding differences in wellbeing.

While short-term impacts are important, we should be especially concerned if the impacts of the pandemic have continued to affect young people’s lives, including their subjective wellbeing, once restrictions ended. Concern was expressed from early in the pandemic that its negative effects on wellbeing would persist (Sonuga-Barke & Fearon, 2021), something that has been identified in some (Quintana-Domeque & Proto, 2022) but not all (Henseke & Schoon, 2025) studies of the general population. We explore the extent to which young people’s own perceptions of the ongoing impact of the pandemic on their wellbeing are associated with their post-pandemic subjective wellbeing, following in the spirit of research that seeks to understand the informational value of individuals’ own assessments of their situation (e.g. Fernandez-Urbano & Samuel, 2024 on perceived coping with the context of the pandemic).

Moreover, we should care about inequalities in young people’s wellbeing whether or not these are (entirely) due to the pandemic. Indeed, we may be more concerned about the persistence of inequalities that pre-date the pandemic — including in gender (Yoon et al., 2023) and socioeconomic status (Verhulst & Tiemeier, 2020) — for which there is no particular reason to expect them to subside. As such, taking stock of the current situation is important in its own right given correlation between subjective wellbeing and later economic and wider outcomes (Deaton, 2008; Lyubomirsky et al., 2005).

In our analyses of these issues, we are guided by Social Production Function (SPF) theory (Ormel et al., 1999), which enumerates five components contributing to subjective well-being: stimulation, comfort, status, behavioural confirmation, and affection. While this study does not engage individually with all five factors, SPF nevertheless provides a helpful framework, including in distinguishing between long-term factors such as status, linked with socioeconomic and demographic characteristics, and more acute potential impacts of changes to stimulation, comfort and affection presented by the disruption of the pandemic and specific events during its course. Furthermore, the potential buffering role of social support can be seen as integral to the SPF components of status and affection. Focussing particularly on the context of the COVID-19 pandemic, Chesters (2025) posit that its restrictions may have negatively affected young people’s access to affection, when not able to spend time with friends and extended family; stimulation, due to restrictions on activities; comfort, both material through potential financial distress, and emotional through adverse life events; and behavioural confirmation, through the disruption to routines and societal expectations.

Using data from the COVID Social Mobility & Opportunities study (COSMO) — a representative cohort study of over 13,000 young people in England aged 14–15 at pandemic onset whose education and post-16 transitions were acutely affected by the pandemic’s disruption through their remaining education and subsequent transitions — we provide new evidence regarding these issues, specifically through the following research aims:

to estimate differences in post-pandemic wellbeing among this cohort by demographic characteristics;
to validate and quantify the informational value of young people’s own perceptions of the impact of the pandemic on their wellbeing, along with the extent to which this may be explained by differences in social support and;
to explore the role of adverse experiences during the pandemic in explaining differences in post-pandemic wellbeing, again accounting for the potential importance of differences in social support.

The paper proceeds as follows. In Sect. 2, we describe the data that we use, the steps taken to prepare it for analysis, and conduct descriptive analyses to provide initial evidence on our research aims. In Sect. 3, we describe our use of regression modelling, presenting results in Sect. 4. Finally, we discuss our findings and conclude in Sect. 5, noting implications for policy and practice.

Data and Descriptive Analyses

We use data from the COVID Social Mobility & Opportunities study (COSMO), a longitudinal cohort study following a representative sample of young people (and their parents) who were in Year 10 (i.e., aged 14–15) at pandemic onset (March 2020), who participated at both waves 1 (Anders et al., 2024a), carried out October 2021–March 2022 (ages 16–17), and 2 (Anders et al., 2024b), carried out October 2022–March 2023 (ages 17–18). In both cases the majority of interviews were carried out within the first two months of fieldwork; we also control for month of interview in our regression models (further details below).

COSMO has a clustered and stratified design with oversampling of those from smaller (e.g., ethnic minorities), more disadvantaged and harder to reach demographic groups to improve statistical power when exploring inequalities between such groups. Furthermore, there was initial non-response and attrition between the first two waves. As such, it is important to account for the deliberate and modelled sample disproportionalities, as well as implications of clustering and stratification for statistical inference. We take these features into account in analyses using study-provided clustering and stratification variables, and design and non-response weights (Adali et al., 2022, 2023).

To ensure consistency across analyses, we restrict our sample to those with valid data on the key variables for our analyses. This includes the primary outcome of self-reported wellbeing score, along with key predictors and demographic variables. However, we are mindful of the potential implications of sample selection caused by complete case analysis. We robustness check our findings in Sect. 9, re-running our core analyses having only restricted the sample based on the primary outcome (wellbeing score) and the main predictors (impact of pandemic on mental health and adverse life events reporting) and multiply imputing across 10 datasets all other predictors using a highly flexible classification and regression tree approach (Lumley, 2019; van Buuren & Groothuis-Oudshoorn, 2025).

Subjective Wellbeing

To measure self-reported wellbeing, we use the UK Office for National Statistics’ official measure of life satisfaction (Office for National Statistics, 2018), which is widely recognised as an important dimension of subjective wellbeing (Petersen et al., 2022). This asks participants to respond to the prompt “Overall, how satisfied are you with your life nowadays?” on a scale ranging from 0 “Not at all satisfied” to 10 “Completely satisfied”. This measure has been used in national UK surveys since 2011 and increasing numbers of academic studies, hence providing a useful benchmark for this concept in UK-based surveys. This measure is found to be a reliable measure of subjective wellbeing in young people (Levin & Currie, 2014), performing as well as the more in-depth Satisfaction with Life Scale (Jovanović, 2016), for example, although we do recognise that it will not capture all dimensions of wellbeing (Ruggeri et al., 2020). It is also worth noting that, while they are distinct constructs, a clear correlation between lower wellbeing and increased risk of poor mental health (Lombardo et al., 2018).

As COSMO was established in response to the pandemic, there are no pre-pandemic baseline measures. As such, we emphasise that our estimates of differences are between individuals all of whom have experienced the pandemic, but experienced it differently, rather than between their current situation and a counterfactual in which the pandemic did not happen. Others have used survey experiment methods to attempt to get closer to such a counterfactual (Andreoli et al., 2024), or pre-existing longitudinal studies to explore change in mental health across the pandemic period (Henseke & Schoon, 2025).

We have measures of wellbeing from two post-pandemic waves and use these to explore evidence of change in wellbeing between the two waves both overall, and between sub-groups of the data where this might be expected. We plot the overall distribution of reported wellbeing in both Waves 1 (age 16/17) and 2 (age 17/18) in Fig. 1.

Fig. 1 — Histogram of distribution of subjective wellbeing in Wave 1 and 2. Notes: Histogram weighted for survey design and non-response

Young people report a mean wellbeing score of 6.41 in Wave 1 and 6.43 at Wave 2, with the standard deviation declining slightly from 2.05 to 1.97. These are not particularly substantial changes, providing little evidence of change between these two post-pandemic time points. However, in interpreting this (lack of) aggregate change, we must be mindful of this cohort’s wider context.

One interpretation would be that, as we know there was a decline in mental health and wellbeing among young people at the onset of the pandemic and its restrictions (Newlove-Delgado et al., 2021), we would hope to see an upward trajectory in wellbeing in subsequent years to be confident of a ‘bounce back’, with this lack of change suggesting a plateau at a lower level than before the pandemic. That could be the case. A finding of minimal change is consistent with the findings of Henseke et al. (2022) (albeit for a wider age range of young people aged 16–29). Similarly, the UK Office for National Statistics’ annual population survey also suggests that life satisfaction has not returned to pre-pandemic levels in the general population (Office for National Statistics, 2023).

Fundamentally, using these data alone we are unable to adjudicate between multiple potential plausible scenarios. Others, using a wider range of datasets are better placed to do so. For example, Henseke and Schoon (2025) suggest that young people’s wellbeing may have already returned to pre-pandemic levels, thus explaining a lack of trend for this reason. These findings would also be consistent with an upward post-pandemic trend being cancelled out by a countervailing age effect (for example) implied by the wider literature on life course wellbeing (Blanchflower, 2021).

However, this is not our paper’s focus. Aggregate stability does not mean that there are not individual-level differences or differential change in wellbeing. The correlation between the reported measures in Waves 1 and 2 is 0.54. While some of this likely reflects natural fluctuation in young people’s wellbeing due to daily idiosyncratic shocks, it provides a basis to explore evidence of systematic difference in change between the two waves, along with the differences in levels at each wave.

Social Support

Social support is concerned with the extent to which an individual is, or perceives they are, “cared for, esteemed, and valued by people in [their] social network” (p. 691, Demaray et al., 2005). As such, it directly relates to the SPF framework, specifically status and affection. As such, we anticipate that individuals with greater social support will have higher levels of wellbeing (Li et al., 2021; Magson et al., 2021; Siedlecki et al., 2014). Furthermore, because of the potential for substitution between components of the SPF in the production of wellbeing (Ormel et al., 1999), we also anticipate social support buffering shocks to other aspects; this has been observed empirically with social support buffering shocks to wellbeing in the face of adversity (Aksoy et al., 2024; Kearns et al., 2015; McMahon et al., 2023).

To capture this factor, we use the social provisions scale (Cutrona & Russell, 2018), specifically a shortened three-item variant available in COSMO in which young people are asked to respond (using the categories “Not true”, “Partly true” or “Very true”) to the statements:

I have family and friends who help me feel safe, secure and happy.
There is someone I trust whom I would turn to for advice if I were having problems.
There is no one I feel close to [Negatively coded].

Following standard practice, we sum over the values of the three items and standardise the resulting variable (mean zero; standard deviation one) for the purposes of interpretation. We plot the distribution of the social provisions scale in Fig. 2. There is some evidence of a ceiling effect — most respondents score the maximum of 6 — but with a decent spread below this. We use this as a continuous measure in our analyses.

Fig. 2 — Distribution of social provisions scale. Notes: Distribution of social provisions scale. The scale is standardised to have mean 0 and standard deviation 1 in the analysis sample. Weighted for survey design and non-response

Demographic Characteristics

The impact of the pandemic on young people’s wellbeing differed depending upon their demographic characteristics (e.g., Anders et al., 2023). Both to estimate differences between young people based on these characteristics, and to control for these measures in other analyses, we make use of the rich set of demographic measures collected in COSMO. Specifically, we construct the following measures of demographic characteristics.

Gender: There are longstanding concerns about differences in wellbeing by gender (Yoon et al., 2023), which have only been exacerbated by the pandemic (Davillas & Jones, 2021). We use young people’s reported gender at either wave (giving precedence to the subsequent reports if they differ) and group cohort members into ‘female’, ‘male’ and ‘non-binary+’, where the final category is a combination of those who explicitly report being non-binary or choose to identify in any other way (since these other groups are too small for analysis).
Ethnicity: As with gender, our measure is based on self-reports at either wave (where a subsequent report is given precedence if they differ), young people are grouped into ‘White’, ‘Mixed’, ‘Black’, ‘Asian’ and ‘Other’. While these categories are broad, they are chosen for consistency with the UK’s major ethnic group classifications while avoiding groups that are too small for analysis purposes.
Parental education: Generally viewed as a core component of socioeconomic status, which may affect wellbeing through the status component of the SPF (Ormel et al., 1999), we construct our measure using the highest level of education reported by either parent at either wave, grouping parents into ‘Graduate’, ‘Below Graduate’ and ‘No Quals’.
Housing tenure: Housing tenure is another component of socioeconomic status, hence with potential implications for young people’s wellbeing. Our measure is based on parental reports at either wave (giving precedence to the subsequent if they differ), grouping families into those who own their home (either with a mortgage or outright; ‘Own House’) and all others (predominantly social and private renting; ‘Other’).
Area deprivation: We also include an area-based measure of deprivation of participants’ homes, both as a correlate of socioeconomic stats due to residential sorting and given more direct implications this can have for potentially wellbeing-enhancing amenities. COSMO provides decile groups of the UK’s Income Deprivation Affecting Children Index (IDACI), constructed at the ‘lower-layer super-output area’ (the smallest geographical areas in UK statistical geography, containing an average population of 1,500).

To allow exploration of differences in wellbeing by socioeconomic status (SES) in a simple way, we create a combined index of SES (with mean 0 and standard deviation 0 in our analysis sample) across our measures of parental education, housing tenure and home neighbourhood deprivation. We describe how we do this and demonstrate that it captures the underlying SES measures on which it is based in Sect. 6.1.

Having constructed this set of measures, we report the prevalence of demographics in our cohort along with mean levels of self-reported wellbeing by these categories at Wave 1, Wave 2, and mean difference between the two in Table 1.

Table 1.

Mean subjective wellbeing score by demographic characteristics

Characteristic	N	Prevalence (%)	Wave 1	Wave 2	Difference
Overall	7,723		6.41	6.43	0.017
Gender
Male	3,475	50	6.76	6.76	0.007
Female	4,030	48	6.13	6.15	0.021
Non-Binary+	218	2.6	4.90	5.04	0.136
Ethnicity
White	4,877	77	6.43	6.44	0.014
Mixed	477	5.7	6.09	6.09	−0.008
Black	1,503	10	6.51	6.48	−0.030
Asian	684	5.0	6.34	6.43	0.094
Other	182	2.2	6.44	6.64	0.201
Parental Education
Graduate	3,807	55	6.48	6.45	−0.024
Below Graduate	2,962	36	6.35	6.37	0.024
No Quals	871	7.6	6.25	6.54	0.286
Unknown	83	0.8	6.42	6.37	−0.049
Housing Tenure
Own House	4,224	65	6.50	6.54	0.037
Other	3,499	35	6.24	6.22	−0.020
Unknown	0	0
IDACI Quintile Group
1 (High Deprivation)	2,306	22	6.24	6.23	−0.007
2	1,678	19	6.39	6.42	0.032
3	1,351	19	6.34	6.46	0.118
4	1,231	20	6.54	6.56	0.023
5 (Low Deprivation)	1,157	20	6.56	6.49	−0.072
SES Quintile Groups
1 (Low SES)	2,257	20	6.26	6.26	0.005
2	1,770	20	6.28	6.37	0.088
3	1,405	20	6.42	6.40	−0.019
4	1,266	21	6.53	6.53	0.006
5 (High SES)	1,025	19	6.58	6.58	0.002

Variable, N = 7723	No (64%)¹	Negative (32%)¹	Don’t know (2%)¹	Positive (2%)¹	Overall (100%)¹	p-value²
Wave 1	6.81	5.62	6.37	6.40	6.41	< 0.001
Wave 2	6.91	5.46	6.29	6.48	6.43	< 0.001
Difference	0.11	−0.16	−0.08	0.08	0.02	< 0.001

Variable, N = 7723	Fewer (36%)¹	Average (30%)¹	More (33%)¹	Overall (100%)¹	p-value²
Wave 1	7.04	6.46	5.67	6.41	< 0.001
Wave 2	7.02	6.47	5.75	6.43	< 0.001
Difference	−0.03	0.00	0.08	0.02	0.2

Variable	L1	L2	L3	L4	L5	L6	L7	L8
Gender	Included			Included	Interacted w/Ethnicity and SES	Included	Interacted w/Social Support	Included
Ethnicity		Included		Included	Interacted w/Gender and SES	Included	Interacted w/Social Support	Included
SES			Included	Included	Interacted w/Gender and Ethnicity	Included	Interacted w/Social Support	Included
Social Support						Included	Interacted w/Gender, Ethnicity and SES	Included
Adverse Events								Included

Variable	P1	P2	P3	P4
Perceived Impact	Included	Included	Included	Interacted with Demographics, SES and Social Support
Demographics		Included	Included	Interacted with Perceived Impact
SES		Included	Included	Interacted with Perceived Impact
Social Support			Included	Interacted with Perceived Impact

	L1		L2		L3		L4		L5		L6		L7		L8
Characteristic	Beta ¹	SE	Beta ¹	SE	Beta ¹	SE	Beta ¹	SE	Beta ¹	SE	Beta ¹	SE	Beta ¹	SE	Beta ¹	SE
(Intercept)	6.7***	0.062	6.4***	0.057	6.2***	0.076	6.6***	0.087	6.5***	0.117	6.6***	0.081	6.6***	0.081	6.6***	0.079
Gender
Male	—	—					—	—	—	—	—	—	—	—	—	—
Female	−0.63***	0.057					−0.63***	0.056	−0.53***	0.123	−0.60***	0.051	−0.60***	0.051	−0.50***	0.051
Non-Binary+	−1.9***	0.216					−1.9***	0.218	−1.7***	0.387	−1.4***	0.208	−1.4***	0.230	−1.2***	0.202
Ethnicity
White			—	—			—	—	—	—	—	—	—	—	—	—
Mixed			−0.33*	0.133			−0.27*	0.127	−0.34	0.278	−0.12	0.128	−0.14	0.124	−0.12	0.126
Black			0.08	0.086			0.09	0.087	0.20	0.191	0.20**	0.078	0.20*	0.077	0.16*	0.075
Asian			−0.10	0.103			0.02	0.103	0.13	0.222	0.23*	0.092	0.22*	0.091	0.18*	0.089
Other			0.03	0.229			0.06	0.224	0.16	0.513	0.22	0.189	0.22	0.184	0.22	0.193
SES Quintile Groups
1 (Low SES)					—	—	—	—	—	—	—	—	—	—	—	—
2					0.03	0.094	0.00	0.093	0.05	0.146	−0.04	0.083	−0.03	0.081	−0.03	0.080
3					0.15	0.099	0.15	0.097	0.23	0.163	0.09	0.087	0.09	0.086	0.04	0.085
4					0.26**	0.090	0.27**	0.089	0.33*	0.140	0.16	0.082	0.16	0.082	0.09	0.081
5 (High SES)					0.30**	0.094	0.30**	0.094	0.36*	0.142	0.21*	0.086	0.21*	0.085	0.12	0.084
SES Quintile Groups * Gender
2 * Female									−0.09	0.175
3 * Female									−0.24	0.185
4 * Female									−0.05	0.170
5 (High SES) * Female									−0.01	0.172
2 * Non-Binary+									−0.23	0.553
3 * Non-Binary+									0.46	0.697
4 * Non-Binary+									0.23	0.588
5 (High SES) * Non-Binary+									−1.3*	0.587
SES Quintile Groups * Ethnicity
2 * Mixed									0.29	0.387
3 * Mixed									0.26	0.363
4 * Mixed									−0.25	0.333
5 (High SES) * Mixed									0.11	0.332
2 * Black									−0.02	0.246
3 * Black									−0.05	0.239
4 * Black									−0.23	0.287
5 (High SES) * Black									−0.35	0.303
2 * Asian									−0.11	0.220
3 * Asian									0.07	0.258
4 * Asian									0.04	0.387
5 (High SES) * Asian									−0.66	0.935
2 * Other									−0.06	0.551
3 * Other									0.78	0.672
4 * Other									−0.15	0.664
5 (High SES) * Other									−0.08	0.949
Gender * Ethnicity
Female * Mixed									0.01	0.249
Non-Binary+ * Mixed									−0.22	0.617
Female * Black									−0.09	0.161
Non-Binary+ * Black									0.43	0.678
Female * Asian									−0.09	0.220
Non-Binary+ * Asian									−0.92	1.03
Female * Other									−0.36	0.454
Non-Binary+ * Other									1.4*	0.613
Social Provisions Scale											0.90***	0.028	0.89***	0.075	0.83***	0.029
Gender * Social Provisions Scale
Female * Social Provisions Scale													0.02	0.054
Non-Binary+ * Social Provisions Scale													−0.07	0.157
Ethnicity * Social Provisions Scale
Mixed * Social Provisions Scale													−0.17	0.109
Black * Social Provisions Scale													−0.12	0.074
Asian * Social Provisions Scale													−0.09	0.083
Other * Social Provisions Scale													0.03	0.194
SES Quintile Groups * Social Provisions Scale
2 * Social Provisions Scale													0.08	0.081
3 * Social Provisions Scale													0.06	0.083
4 * Social Provisions Scale													−0.02	0.089
5 (High SES) * Social Provisions Scale													0.03	0.086
Adverse Event Index															−0.35***	0.028
W1 Month of Interview
Sep 2021	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—
Oct 2021	0.10	0.067	0.14*	0.067	0.12	0.067	0.09	0.068	0.10	0.068	0.03	0.061	0.03	0.061	0.03	0.060
Nov 2021	0.41*	0.193	0.42*	0.185	0.37*	0.184	0.36	0.188	0.37*	0.183	0.38*	0.169	0.39*	0.169	0.38*	0.165
Dec 2021	0.32*	0.135	0.30*	0.135	0.30*	0.135	0.29*	0.132	0.31*	0.135	0.14	0.116	0.14	0.116	0.13	0.113
Jan 2022	0.49	0.250	0.50	0.260	0.51*	0.257	0.48	0.246	0.47*	0.241	0.47	0.260	0.47	0.262	0.43	0.256
Feb 2022	−0.33	0.234	−0.21	0.251	−0.23	0.246	−0.37	0.233	−0.36	0.229	−0.50*	0.243	−0.49*	0.243	−0.49*	0.246
Mar 2022	−0.12	0.093	−0.10	0.095	−0.10	0.096	−0.12	0.093	−0.12	0.093	−0.14	0.084	−0.14	0.084	−0.13	0.084
Apr 2022	−0.04	0.102	0.00	0.105	−0.01	0.106	−0.06	0.103	−0.05	0.102	−0.08	0.096	−0.08	0.096	−0.10	0.093
N	7,723		7,723		7,723		7,723		7,723		7,723		7,723		7,723
Residual DoF	757		755		755		749		717		748		738		747

Variable, N = 7723	Fewer (36%)¹	Average (30%)¹	More (33%)¹	Overall (100%)¹
Parent lost job	0	13	23	12
Couldn’t afford food	0	2.8	23	8.4
Couldn’t afford bills	0	4.6	28	11
Seriously ill	0	2.6	7.0	3.1
Close family member seriously ill	0	45	54	32
Close family member died	19	28	51	33
More arguments with parents	0	28	72	32
More arguments between parents	0	7.8	60	22
Moved home	0	6.6	16	7.3
Parents separated	0	1.5	10	3.8
Number of events (grouped)
0	81	0	0	29
1	19	60	0	25
2	0	40	23	20
3+	0	0	77	26
Number of events (mean)	0.19	1.40	3.44	1.64

Characteristic	Beta¹	SE
(Intercept)	3.3***	0.133
Gender
Male	—	—
Female	-0.48***	0.051
Non-Binary+	-1.2***	0.179
Ethnicity
White	—	—
Mixed	-0.10	0.113
Black	0.15*	0.072
Asian	0.17*	0.079
Other	0.18	0.175
SES Quintile Group
Q1 (Low)	—	—
Q2	-0.10	0.082
Q3	0.04	0.080
Q4	-0.01	0.076
Q5 (High)	0.07	0.080
Social Provisions Scale	0.64***	0.021
Adverse Event Index	-0.50***	0.036
Wave 1 Survey Month
Sep 2021	—	—
Oct 2021	0.03	0.057
Nov 2021	0.38*	0.156
Dec 2021	0.10	0.107
Jan 2022	0.33	0.246
Feb 2022	-0.45*	0.209
Mar 2022	-0.12	0.080
Apr 2022	-0.08	0.088
N	9,307

Characteristic	Beta¹	SE
(Intercept)	4.8***	0.128
Negative continuing impact of pandemic on mental wellbeing
No	—	—
Yes	-1.0***	0.055
Gender
Male	—	—
Female	-0.32***	0.049
Non-Binary+	-0.96***	0.165
Ethnicity
White	—	—
Mixed	-0.17	0.108
Black	0.08	0.066
Asian	0.05	0.087
Other	0.30	0.161
SES Quintile Group
Q1 (Low)	—	—
Q2	0.09	0.076
Q3	0.07	0.075
Q4	0.15	0.077
Q5 (High)	0.26**	0.085
Social Provisions Scale	0.39***	0.021
Adverse Event Index	-0.39***	0.035
Wave 2 Survey Month
October 2022	—	—
November 2022	-0.11*	0.052
December 2022	0.09	0.103
January 2023	0.24	0.288
February 2023	0.35	0.207
March 2023	0.41*	0.183
April 2023	0.02	0.176
N	9,307

Characteristic	Beta¹	SE
(Intercept)	3.5***	0.169
Negative continuing impact of pandemic on mental wellbeing
No	—	—
Yes	-0.84***	0.052
Gender
Male	—	—
Female	-0.45***	0.050
Non-Binary+	-1.1***	0.184
Ethnicity
White	—	—
Mixed	-0.11	0.113
Black	0.14	0.072
Asian	0.18*	0.080
Other	0.18	0.171
Parental Education	-0.02	0.042
Housing Tenure	-0.05	0.057
IDACI Quintile Group
1 (High Deprivation)	—	—
2	0.07	0.077
3	0.03	0.084
4	0.17*	0.082
5 (Low Deprivation)	0.20*	0.088
Social Provisions Scale	0.67***	0.020
Wave 1 Survey Month
Sep 2021	—	—
Oct 2021	0.02	0.057
Nov 2021	0.33*	0.160
Dec 2021	0.10	0.110
Jan 2022	0.	0.249
Feb 2022	-0.54**	0.199
Mar 2022	-0.12	0.080
Apr 2022	-0.06	0.087
N	9,307

Characteristic	Beta¹	SE
(Intercept)	4.0***	0.179
Adverse Event Tertile Groups
Low	—	—
Medium	-0.34***	0.060
High	-0.75***	0.067
Negative continuing impact of pandemic on mental wellbeing
No	—	—
Yes	-0.70***	0.052
Gender
Male	—	—
Female	-0.39***	0.050
Non-Binary+	-1.0***	0.178
Ethnicity
White	—	—
Mixed	-0.10	0.113
Black	0.11	0.071
Asian	0.14	0.078
Other	0.17	0.172
Parental Education	-0.02	0.042
Housing Tenure	-0.03	0.056
IDACI Quintile Group
1 (High Deprivation)	—
2	0.05	0.076
3	0.01	0.084
4	0.17*	0.081
5 (Low Deprivation)	0.15	0.086
Social Provisions Scale	0.63***	0.021
Wave 1 Survey Month
Sep 2021	—	—
Oct 2021	0.02	0.056
Nov 2021	0.36*	0.162
Dec 2021	0.08	0.107
Jan 2022	0.33	0.248
Feb 2022	-0.53**	0.202
Mar 2022	-0.12	0.080
Apr 2022	-0.09	0.086
N	9,307

Characteristic	Beta¹	SE
(Intercept)	5.3***	0.176
Adverse Event Tertile Groups
Low	—	—
Medium	-0.34***	0.060
High	-0.67***	0.068
Negative continuing impact of pandemic on mental wellbeing
No	—	—
Yes	-1.0***	0.054
Gender
Male	—	—
Female	-0.33***	0.050
Non-Binary+	-0.99***	0.163
Ethnicity
White	—	—
Mixed	-0.16	0.110
Black	0.05	0.067
Asian	0.07	0.089
Other	0.30	0.163
Parental Education	0.04	0.046
Housing Tenure	-0.15**	0.057
IDACI Quintile Group
1 (High Deprivation)	—	—
2	0.13	0.074
3	0.12	0.080
4	0.19*	0.080
5 (Low Deprivation)	0.17*	0.085
Social Provisions Scale	0.40***	0.022
Wave 2 Survey Month
October 2022	—	—
November 2022	-0.11*	0.052
December 2022	0.08	0.105
January 2023	0.25	0.292
February 2023	0.35	0.204
March 2023	0.42*	0.184
April 2023	0.01	0.175
N	9,307

Characteristic	Beta¹	SE
(Intercept)	3.6***	0.179
Wave 1 Wellbeing	0.41***	0.016
Adverse Event Tertile Groups
Low	—	—
Medium	-0.20***	0.055
High	-0.36***	0.060
Negative continuing impact of pandemic on mental wellbeing
No	—	—
Yes	-0.75***	0.051
Gender
Male	—	—
Female	-0.17***	0.043
Non-Binary+	-0.59***	0.160
Ethnicity
White	—	—
Mixed	-0.11	0.090
Black	0.00	0.058
Asian	-0.01	0.081
Other	0.24	0.150
Parental Education	0.05	0.039
Housing Tenure	-0.14**	0.052
IDACI Quintile Group
1 (High Deprivation)	—	—
2	0.10	0.066
3	0.11	0.075
4	0.11	0.072
5 (Low Deprivation)	0.09	0.074
Social Provisions Scale	0.14***	0.021
Wave 1 Survey Month
Sep 2021	—	—
Oct 2021	0.01	0.052
Nov 2021	0.26*	0.125
Dec 2021	0.36**	0.113
Jan 2022	0.20	0.173
Feb 2022	0.22	0.287
Mar 2022	-0.04	0.069
Apr 2022	-0.04	0.075
Wave 2 Survey Month
October 2022	—	—
November 2022	-0.08	0.047
December 2022	0.12	0.095
January 2023	0.40	0.266
February 2023	0.19	0.198
March 2023	0.44**	0.152
April 2023	0.00	0.174
N	9,307

PERMALINK

Young People’s Subjective Wellbeing in the Wake of the COVID-19 Pandemic: Evidence from a Representative Cohort Study in England

Jake Anders

Erica Holt-White

Abstract

Introduction

Data and Descriptive Analyses

Subjective Wellbeing

Fig. 1.

Social Support

Fig. 2.

Demographic Characteristics

Table 1.

Perceived Ongoing Impact

Table 2.

Adverse Life Events

Table 3.

Analytical Approach

Demographic Differences in Subjective Wellbeing

Table 4.

Importance of Perceived Impact of the Pandemic on Wellbeing

Table 5.

Importance of Adverse Life Events During the Pandemic

Table 6.

Results

Demographic Differences in Subjective Wellbeing

Fig. 3.

Fig. 9.

Fig. 10.

Table 9.

Table 10.

Table 11.

Perceived Continuing Impact of the Pandemic on Wellbeing

Fig. 4.

Table 12.

Table 13.

Table 14.

Fig. 12.

Adverse Life Events

Table 15.

Table 16.

Table 17.

Fig. 5.

Fig. 6.

Discussion, Conclusions and Limitations

Acknowledgements

Appendix: Construction of measures

Construction of SES measure

Fig. 7.

Table 7.

Construction of Adverse Life Events Measure

Table 8.

Fig. 8.

Appendix: Full Regression Tables

Demographic Differences in Wellbeing

Perceived Continuing Impact of the Pandemic On Wellbeing

Adverse Life Events

Appendix: Supplementary Results

Demographic Differences in Wellbeing

Fig. 11.

Perceived Continuing Impact of the Pandemic On Wellbeing

Appendix: Multiple Imputation

Table 18.

Table 19.

Table 20.

Table 21.

Table 22.

Table 23.

Table 24.

Table 25.

Table 26.

Declarations

Competing interests

Footnotes

References

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles