Relative educational mobility

Published: 29 October 2026
see all updates

Data on relative educational mobility – the strength of the link between people's educational level as adults and their parents' educational level.

Summary main findings

Longitudinal data (1971 to 2021)

According to the ONS Longitudinal Study (ONS-LS), relative educational mobility has increased since 1991. People are now less likely to end up with the same educational status as their parents.

This link between parents’ and children’s education was high in 1991. It fell sharply by 2001 and saw another clear drop by 2021. Today, opportunities to earn a degree between young people from graduate and non-graduate families are more equal than they used to be.

Other data (2020)

According to findings from the UK Household Longitudinal Survey (UKHLS), relative educational mobility was similar:

  • across all ethnic groups in the UK
  • for people with or without a long-term illness or disability

Relative educational mobility was greater in Scotland, but did not vary significantly across other parts of the UK.


By year and sex

Visualisation

Relative education mobility – degree: no-degree log odds ratios for people aged 28 to 37, by sex and Census year (England and Wales, 1991 to 2021)

Click or tap on legend items to toggle visibility

Data

Relative education mobility – degree: no-degree log odds ratios for people aged 28 to 37, by sex and Census year (England and Wales, 1991 to 2021)
Year Men Women
2021 3.94 4.04
2011 7.40 7.37
2001 7.80 7.65
1991 9.39 10.41
  • Download the data (CSV, 80B)
  • Download

    For the full download file, see Download the data.


    By ethnicity

    Visualisation

    Parent:child odds ratios relating to university degrees, respondents aged 25 to 64 years, by ethnicity (UK, 2020)

    Data

    Parent:child odds ratios relating to university degrees, respondents aged 25 to 64 years, by ethnicity (UK, 2020)
    Ethnicity Parent:child odds ratio
    Bangladeshi 4.58
    Chinese No data
    Indian 4.17
    Pakistani 4.29
    Black African 4.83
    Black Caribbean 2.90
    Mixed 4.72
    White British 4.18
    White Other 3.60
    Other 4.85
  • Download the data (CSV, 197B)
  • Download

    For the full download file, see Download the data.


    By disability status

    Visualisation

    Parent:child odds ratios relating to university degrees, respondents aged 25 to 64 years, by disability status (UK, 2020)

    Data

    Parent:child odds ratios relating to university degrees, respondents aged 25 to 64 years, by disability status (UK, 2020)
    Disabled Parent:child odds ratio
    No 4.07
    Yes 4.30
    All 4.33
  • Download the data (CSV, 65B)
  • Download

    For the full download file, see Download the data.


    By area

    Visualisation

    Parent:child odds ratios relating to university degrees, respondents aged 25 to 64 years, by area (UK, 2020)

    Data

    Parent:child odds ratios relating to university degrees, respondents aged 25 to 64 years, by area (UK, 2020)
    Region Parent:child odds ratio
    Yorkshire & Humberside 5.59
    East of England 5.51
    West Midlands 4.46
    North East 4.44
    South West 4.41
    London 4.10
    North West 3.95
    South East 3.85
    Wales 3.76
    East Midlands 3.18
    Scotland 3.13
    Northern Ireland 3.08
    Average for all regions 4.35
  • Download the data (CSV, 286B)
  • Download

    For the full download file, see Download the data.


    About the data

    Data source

    • ONS Longitudinal Study
    • UK Household Longitudinal Survey (UKHLS)

    Time period

    • Longitudinal Study (1971 to 2021)
    • UKHLS (1991 to 2020)

    Geographic area

    • Longitudinal Study: England and Wales
    • UKHLS: United Kingdom

    What the data measures

    Longitudinal data (1971 to 2021)

    Education is treated as a binary measure: holding an undergraduate degree qualification versus not holding a degree. Families where one or both parents obtained a degree themselves are treated as a graduate family, and those where neither parent gained a degree are treated as non-graduate families.

    The data shows the strength of the relationship between people and their parents with respect to gaining a university degree. The data shows trends over time among young people aged 28 to 37 in the relevant census.

    Relative mobility compares the mobility chances of 2 (or more) different groups. Log odds ratios are used to compare how likely alternative outcomes are in one group (in this case people coming from graduate families) versus another group (in this case people coming from families where neither parent had a university degree). We measure the relative odds of people from graduate and non-graduate families gaining (or not) a university degree themselves.

    Results are presented as log odds ratios, where 0 indicates equal opportunity between groups.

    UKHLS

    The data measures relative educational mobility among 25 to 64 year olds in the UK.

    It shows the link between parents’ and children’s education level by sex, ethnicity, disability status and area.

    For this data, education is used as a binary measure of attainment of an undergraduate degree qualification versus not.

    Results are presented as odds ratios, where an odds ratio of 1 indicates equal opportunity between groups. A higher odds ratio means a greater link between parents’ and children’s education levels (‘intergenerational persistence’), while a lower odds ratio (closer to 1) means greater relative mobility.

    Things you need to know

    Because these measures use different mathematical scales, do not directly compare numerical values between the ONS-LS and UKHLS datasets.

    The error bars show 95% confidence intervals. Read more about confidence intervals.

    Longitudinal data (1971 to 2021)

    This analysis uses log odds ratios. A value of 0 represents complete relative mobility (no link between parent and child degree attainment). Positive values indicate higher intergenerational persistence.

    The census does not directly measure people’s social backgrounds. It links censuses to identify people’s families in an earlier census when they were growing up and still living at home. This is done for specific birth cohorts and a sample of birth dates. In the case of educational mobility we focus on young adults aged 28 to 37 at the time of the destination census and link their data with that for their parents from the census 20 years earlier.

    This method means that recent migrants who were not living in England or Wales in the relevant earlier census cannot be linked. Recent migrants from countries outside England and Wales are therefore excluded.

    The questions asked in the census about educational qualifications, and their coding, differ across the relevant censuses but the census has consistently asked about degree-level qualifications. It is possible to measure trends with respect to higher education consistently over time but not trends in school qualifications.

    See section 3.4 of our technical annex for more information.

    Future analysis for State of the Nation will use the ONS-LS to analyse results by ethnic group and disability status.

    UKHLS

    This analysis uses odds ratios. A value of 1 represents complete relative mobility (no link between parent and child degree attainment). Values greater than 1 indicate higher intergenerational persistence.

    Formal modelling of the data with logistic regression shows that:

    • the odds ratios for men and women are significantly different from each other
    • the odds ratios for the different ethnic groups are not significantly different from each other
    • the odds ratio for Scotland is significantly lower

    Odds ratios for the Chinese ethnic group could not be calculated due to the small sample size.

    For data by disability status, respondents were asked if they had a physical or mental impairment, illness or disability for at least 12 months. This is different from Labour Force Survey data, which specifies that the condition should be both long-term and limiting. We do not know how long respondents had the illness or disability, and it is possible that they completed their education before the illness or disability occurred. Because of this, the data is likely to underestimate the effects of disability.

    Type of data

    • Longitudinal Study: Admin data
    • UKHLS: Survey data, weighted using UKHLS population weights.

    Full report

    Read more in State of the Nation 2026 on GOV.UK.


    Download the data

    Download full dataset (CSV, 4KB)

    This file contains the following variables:

    • Indicator code
    • Indicator name
    • Area type
    • Area name
    • Time period
    • Age
    • Sex
    • Category type
    • Category
    • Value
    • Lower confidence interval
    • Upper confidence interval
    • Unit

    Page history

    Publication release date:

    29 October 2026

    Editions of this page: