Living alone and mental health: a longitudinal study

1
Living alone and mental health a longitudinal
study

• Zhiqiang Feng, Peteke Feijten, Paul Boyle
• Longitudinal Studies Centre for Scotland
• School of Geography and Geosciences
• University of St Andrews
• Scotland, UK

2
Introduction

• Living alone has become a common phase in the
  lifecourse with one person households now making
  up a third of all households in Britain
• This is a rising trend in western societies
• Proportion of 16-59 year olds living alone in
  Britain has trebled from 5 in 1971 to 16 in
  2002

3
Proportion of households and people in one
person households
Source Social Trends No 36, 2006, ONS
4
Introduction

• Increase of people who live alone has many
  social, economic and policy implications
• The trends are regarded as symbolic of excessive
  individualism (Bauman 1995).
• Also optimistically as a democratisation of
  personal life

5
Introduction

• Increase of living alone leads to more housing
  demand, more poverty, social inequality, care
  demand, decrease of fertility
• And health problems which are not widely
  researched.

6
Does living alone matter?

• Living alone is a type of social isolation
• Social isolation may be detrimental to
  individuals mental health
• Faris, (1934)
• Any form of isolation that cuts the person off
  from intimate social relations for an extended
  period of time may lead to this form of mental
  disorder.

7
Does living alone matter?

• People living alone are less likely to develop
  close personal relationship and the same level of
  commitment to other people than people who living
  together
• Persons living alone much less likely to be
  engaged in mutual obligations and mutual
  reinforcement of a primary type, particularly on
  a daily basis

8
Does living alone matter?

• Persons living alone less likely involve
  emotional gratification and personal
  satisfaction.
• Although individually people living alone may be
  better off they are worse off because of pooling
  effects for people living as a family
• Living alone more likely to have pathological
  behaviour, smoking, drinking

9
Does living alone matter?

• However, social integration involves not only
  benefits but costs
• Social relations can be a source of emotional,
  physical, and financial tension (House, et al
  1988)
• At an extreme, of all social ties, family
  relations are most likely to be negative and
  damaging to individual well-being.

10
Does living alone matter?

• Marriage has been shown to have protective
  effects on peoples health
• Married people enjoy better mental health than
  other demographic groups (Lillard Panis, 1996)

11
Previous studies

• Supportive
• Smith et al 2005 Rogers et al 2000 Denton
  Walters, 1999 Hughes Waite 2002 Joutsenniemi
  et al 2006
• Unsupportive
• Hughes Gove 1981 Denton Walters 1999

12
What will this study do?

• This research focuses on the effect of living
  alone on mental health
• Longitudinal data

13
Research questions

• Does living alone affect mental health?
• Does living alone affect mental health
  independent of marital status?
• Does the transition into living alone affect
  mental health?
• Are there gender differences in the effect of
  living alone?

14
Data

• British Household Panel Survey (BHPS)
• A national representative sample
• Prospective samples-interviewees are followed
  annually from 1991
• All household members over 16 are interviewed
• 1991-2003, 5000 households and 10000 people

15
Measurement of mental health

• General Health Questionnaire (GHQ)
• 12 questions like
• Have you recently
• Lost much sleep over worry?
• Felt constantly under strain?
• Felt you could not overcome your difficulties?
• Been feeling unhappy and depressed?
• Been losing confidence in yourself?
• Been thinking of yourself as a worthless person?

16
Measurement of mental health

• We use the caseness method, each response is
  coded 0, and 1
• Total 12 items end up with 0 to 12
• A dichotomous variable is constructed with a
  threshold of 4
• Individuals aged 16 to 64 are included in our
  study

17
Proportion of poor mental health by gender and
living arrangements
Source BHPS 1991-2003
18
Proportion of living alone by gender and age
Source BHPS 1991-2003
19
Modelling issues

• Longitudinal data allow us to
• Control unobserved heterogeneity
• Ability to cope with stress
• Childhood experience
• Study transitions between different living
  arrangements

20
Modelling issues

• Fixed effect logistic models for models of being
  alone
• Logistic models for effects of transition into
  living alone (excluding never married or never
  cohabiting)
• Control for age, household income, urban rural
  areas of residence, education, employment status,
  smoking, number of friends, and region of
  residence

21
Model 1 - Odds ratios by marital status
Men Women Married 1
1 Widowed 2.052 (1.142-3.688) 2.407
(1.682-3.446) Divorced 1.193
(0.912-1.561) 1.009 (0.834-1.222) Separated 3.28
1 (2.470-4.358) 1.913 (1.540-2.378) Never
Married 1.124 (0.934-1.353) 0.991 (0.835-1.175
) N 2460 3098 Observation 22307
27181 Log-likelihood -8239.5
-10798.2 plt0.05 plt0.01 95 confidence
intervals in bracket
22
Model 2 Odds ratio by living arrangements

• Men Women
• Living together 1 1
• Living alone 1.229 (1.056-1.431) 1.053
  (0.900-1.231)
• N 2460 3098
• Observation 22307 27186
• Log-likelihood -8274.9 -10829.2
• plt0.05 plt0.01
• 95 confidence intervals in bracket

23
Model 3 Odds ratios by marital status and living
arrangements

• Men Women
• Married 1 1
• Widowed 2.438 (1.297-4.582) 2.497 (1.727-3.
  609)
• Divorced 1.148 (0.839-1.569) 1.027 (0.845-1.
  249)
• Separated 3.320 (2.397-4.600) 1.941 (1.558-
  2.418)
• Never Married 1.172 (0.945-1.454) 1.006 (0.84
  5-1.199)
• Living together 1 1
• Living alone 0.924 (0.758-1.125) 0.930 (0.786
  -1.100)
• N 2294 3098
• Observation 20240 27181
• Log-likelihood -7486.1 -10797.0
• plt0.05 plt0.01
• 95 confidence intervals in bracket

24
Model 4 Odds ratios by marital status and living
arrangements

• Men Women
• Married together 1 1
• Married alone 1.434 (0.608-3.384) 0.610
  (0.270-1.379)
• Widowed together 2.975 (1.092-8.100)
  2.645 (1.686-4.148)
• Widowed alone 2.068 (1.026-4.168) 2.193
  (1.415-3.400)
• Divorced together 1.444 (0.929-2.242) 1.063
  (0.862-1.311)
• Divorced alone 0.978 (0.709-1.350) 0.858
  (0.623-1.181)
• Separated together 4.281 (2.542-7.208)
  1.917 (1.505-2.443)
• Separated alone 2.843 (2.024-3.993)
  1.847 (1.190-2.866)
• Never married together 1.130 (0.905-1.412)
  0.981 (0.821-1.173)
• Never married alone 1.142 (0.894-1.459)
  1.004 (0.789-1.279)
• N 2295 3098
• Observation 20244 27185
• Log-likelihood -7485.7 -10797
• plt0.05 plt0.01
• 95 confidence intervals in bracket

25
Model 5 Odds ratios by transitions between living
arrangements

• Men Women
• Married together 1 1
• Married to wid/div/sep 5.812 (3.596-9.392)
  4.525 (3.425-5.977)
• together
• Married to wid/div/sep 6.034 (4.155-8.763)
  7.881 (4.724-13.146)
• alone
• N 3474 3568
• Observations 19268 18630
• Log likelihood -6630.9 -8168.2
• plt0.05 plt0.01
• 95 confidence intervals in bracket

26
Proportion of people with poor mental health by
transition of living arrangements
Source BHPS 1991-2003
27
Conclusions

• Being living alone tends to negatively affects
  mental status for men but not women
• Being living alone does not have independent
  effects on mental health from marital status
• Transition into living alone has strong effects
  on mental health and increases risks of mental
  disorders in short-term

28
Future studies

• Better methodologies in tackling endogeneity
  using SEM modelling or 2SLS
• Dynamic panel models taking account of state
  dependency
• Increase sample size by incorporating more data
  from the latest waves
• More analysis of duration effects

29
Acknowledgements

• The BHPS data are made available through the UK
  Data Archive and were collected by the ESRC
  Research Centre on Micro-social Change at the
  University of Essex, now incorporated within
  Institute for Social Economic Research.