Quality of life in the third age : key predictors
of the CASP-19 measure
RICHARD D. WIGGINS*, PAUL F. D. HIGGS#, MARTIN HYDE#
and DAVID B. BLANE**
ABSTRACT
This article aims to identify and analyse the lifecourse and contextual factors that
influence the quality of life in early old age. We conceptualise quality of life as
distinct from the factors which influence it, and employ a model of the quality of
life that is derived from an explicit theory of human need. The operational
measure (CASP-19) consists of 19 Likert-scaled items which cover four theoretical
domains: control, autonomy, self-realisation and pleasure. A postal questionnaire
was sent to 286 British people aged 65–75 years who were members of the sample
for the 1930s Boyd-Orr study of health and diet and who had been followed up
through retrospective interviews during the late 1990s. The 286 were broadly
representative of their age group. The survey’s response rate was 92 per cent.
Respondents provided information on a number of contextual influences on their
quality of life in early old age, notably social support and participation, the quality
and quantity of social contact, feelings of trust and reciprocity about the local
neighbourhood, health and financial security. In the analyses reported here, a
series of conceptual and operational influences on quality of life in early old age is
identified using block regression models. Finally, the relative impact of each
predictor on CASP-19 is examined. The findings suggest that the legacy of the
past tends to be best captured by people’s feelings about the adequacy of their
pensions and their status as owner-occupiers as well as a feeling that the area in
which they live is deprived. The quality of the social contact people describe and
how close they feel to those around them will ameliorate the negative impacts of
the past and the immediate environment. In addition, we must recognise that
people remain vulnerable to the impact of loss : recent bereavement, and major
illnesses can impact on a person’s quality of life.
KEY WORDS – quality of life, Casp-19, third age.
* City University, London.
# University College, London.
** Imperial College of Science, Technology and Medicine, London.
694
Richard D. Wiggins et al.
Background
There is widespread agreement that as society changes, so also does
people’s experience of ageing and later life (Phillipson 1998; Gilleard
and Higgs 2000 ; Vincent 2003). During the last century, the population
structure and vital statistics of the United Kingdom changed markedly,
with clear consequences for life expectancy and health (Office for National
Statistics 1998; Blaikie 1999). Alongside the increased health and longevity
of the older generation, there has been an increase in early retirement
(Gruber and Wise 1999). Furthermore, the increase in the availability of
occupational and private-pension provision to supplement the basic state
pension means that a section of those retiring have an income that is (often
considerably) more than subsistence level (Department for Work and
Pensions 2000). Some authors describe these phenomena as indicators of
the move into ‘disorganised capitalism ’ (Lash and Urry 1997) or late
modernity (Giddens 1991). Whatever the conceptual status of such changes,
they pose serious challenges for the understanding of the lifecourse.
In a Fresh Map of Life, Peter Laslett (1996a) portrayed four distinct ‘ ages’
of a person’s lifecourse. He represented the ‘ First Age’ as the period of
childhood dependency, the ‘ Second Age’ as the time of independence,
employment and maturity, the ‘Third Age’ as the period during which
people are freed from work and family constraints and have time to pursue
a good quality of life. Finally, he saw the ‘Fourth Age’ as characterised by
dependence and declines in health. The focus of this paper is on people in
early old age (aged around 65–74 years), who might best be described as
potential members of the ‘Third Age’. For many authors, the ‘ Third Age’
is in historical terms a new lifecourse stage (Young and Schuller 1991;
Midwinter 1992 a; Laslett 1996b). As Gilleard and Higgs (2002) have
`
argued, the lives of those in retirement in fin de siecle Britain are defined by
circumstances that have greatly changed from those experienced by their
predecessors. The Third Age seems to be based upon and to offer a later
life that is agentic and reflexive. It is based on health and leisure rather
than infirmity and poverty. At the same time, older people are neither
socially nor economically homogeneous (Baltes et al. 1993). This paper uses
the outcome measure CASP-19 to examine the extent to which being in
the Third Age implies freedoms from family constraints and to pursue
leisure activities, a time for self-realisation, and opportunities for fun
(Hyde et al. 2003). Specifically, we explore the contributions of gender,
ageing, lifecourse events and variations as well as contemporary contex-
tual factors on the quality of life of a unique sample of people in early
old age drawn from the 1930s Boyd-Orr study of health and diet (Gunnell
et al. 1996).
Predictors of the quality of life in the third age
695
Conceptual framework
Our conceptual map was developed from previous research and publica-
tions about several key potential influences on a person’s quality of life in
the Third Age. It represents a framework for the analyses that follow. First,
we describe the theoretical foundations of our outcome measure; and
continue with a brief description of each of the factors that influence or
shape a person’s quality of life. We argue that quality of life may be shaped
by ageing, gender, the past as captured by lifetime exposure to poor ac-
commodation and environment, as well as by markers of passage into
early old age, pension provision, current housing tenure, health status and
access to a car. We have collectively referred to these items as ‘ legacy of
the lifecourse ’. To capture the present context we present people’s re-
flection of the quality of their neighbourhood and feelings of safety under
the heading ‘social capital’. A person’s current social network is captured
not only in terms of the frequency with which they communicate with
family and friends but the quality of these exchanges and the closeness
they derive from them. Finally, there is a domain to capture the impact of
recent life events.
The CASP-19 measure of the quality of life
We have argued elsewhere that quality of life remains under-theorised and
ill-defined (Hyde et al. 2003 ; Higgs et al. 2003). The evidence that signifi-
cant numbers of older people are living longer, healthier and more active
lives requires us to rethink the way in which we conceptualise and measure
quality of life in this age group. The lack of a theoretically-informed
measure of quality of life in early old age has meant that for a long time it
has been measured by various proxies, such as health (Bowling 1997).
These proxy measures rest on an implicit set of normative assumptions
about each person’s quality of life and often neglect the range of experi-
ences of people in older age. Such scales should be seen as measures of the
influences on quality of life rather than measures of quality of life itself.
Partly as a reaction to this criticism, recent attempts to measure quality of
life have sought to emphasise the importance of subject-led accounts in the
construction of measures of quality of life in early old age (Farquhar 1995;
Bowling 1995). Researchers seek to understand what quality of life means
for each person by asking respondents to rate the most important things in
their lives. This approach suffers from some of the same problems as
‘ expert-defined’ quality of life measures, in that both tend to generate
measures of the influences on the quality of life rather than of quality of life
itself.
696
Richard D. Wiggins et al.
In developing the CASP-19 measure, the initial premise was that any
quality of life measure should be distinct from the factors that influence it.
This conceptualisation was derived from a theory of human need that
recognises the social and biological components as equal (Doyal and
Gough 1991). This ‘ needs satisfaction’ approach assumes that quality of
life should be assessed as the degree to which human needs are satisfied.
Grounding the model of quality of life in this ontology enables meaningful
comparisons of different quality of life scores. The model implies that
although there are common basic human needs, such as food and shelter,
it is equally important to recognise that being human is principally an
active and reflexive process (Giddens 1990 ; Turner 1995). Based on the
range of human needs, four domains of the quality of life are con-
ceptualised : ‘control ’, ‘ autonomy’, ‘ self realisation ’ and ‘ pleasure ’. The
first letter of each domain and its 19 items create the acronym CASP-19
that names the measure. ‘Control ’ and ‘ autonomy’ are prerequisites of
free participation in society. Much gerontological research has however
failed to consider the more active and reflexive dimensions of being old in
an increasingly complex society in which ‘ post-materialist ’ values prevail
(Inglehart 1997). The narrowness of its focus has also been criticised for
ignoring the agency of older people (Gilleard and Higgs 1998). It is the
more active aspects of old age that we aim to map with the other two
dimensions of our model, ‘self realisation’ and ‘pleasure ’.
The operationalisation of the theoretical model into the self-enumer-
ated CASP-19 measure of the quality of life has been described in more
detail elsewhere Hyde et al. (2003). Each domain consists of four or five
Likert-scaled agreement items, e.g. ‘I look forward to each day ’, and in total
there are 19 items, which are summed to form the overall score. The range
of the scale is from zero, which represents a complete absence of quality of
life, to 57, which represents total satisfaction across the four domains.
Internal-consistency analysis reveals a reliable overall measure (Cronbach’s
alpha=0.7), which has demonstrable concurrent validity. Second-order
factor analysis also confirms the structure of four domains (control, auton-
omy, self-realisation and pleasure), for all load on to a single underlying
latent variable that represents an intrinsic measure of the quality of life.
CASP-19 has been adopted by several important studies, notably the
English Longitudinal Study of Ageing (Marmot et al. 2003) and the retirement
module in wave 11 of the British Household Panel Survey (BHPS) (Taylor 2003).
Age, gender and lifecourse hazards
Many writers have expressed a concern that the gendered dimension of
ageing and older age is neglected (Bury 1995; Ginn and Arber 1999). This
Predictors of the quality of life in the third age
697
study therefore set out to discover whether a person’s age and gender had
any impact on their quality of life. For convenience, age was dichotomised
above and below 70 years. Blane and colleagues (1998) have shown that
exposure to social disadvantage, as indicated by exposure to environ-
mental health hazards, accumulates across the lifecourse and is related in
complex ways to health in early old age (Holland et al. 2000 ; Montgomery
et al. 2000; Berney et al. 2000 a, b). In the light of these findings, we were
interested to see if the same accumulated hazard exposures directly affect
the quality of life. The hazard variables used in the present analyses refer
to the domestic sphere of life (‘ household hazards’) : the number of years
of exposure to residential dampness, atmospheric pollution in the resi-
dential locality, and inadequate nutrition.
These hazards or factors are often represented by cross-sectional vari-
ables but we regard them as particular stages of processes or pathways into
early old age. For example, the choice whether to retire or to continue
working, current health status (poor or otherwise), car ownership or access
to a car, an individual’s subjective assessment of whether or not they feel
that they have enough income to meet their future needs, and their cur-
rent housing status (owner occupier or not), collectively mark the legacy of
the past. These markers represent past decisions and the consequences, or
life chances, as people negotiated their passage through life. They are
sensitive to the heterogeneity of individuals in the Third Age, many of
whom will have retired early, be in good health and relatively affluent
(Hardy and Hazelrigg 1999 ; Gruber and Wise 1999; Scase and Scales
2000). Following Arber’s (1991) argument that housing tenure and car
access are more sensitive to women’s positions when compared with social
status indicators based on male occupations, they were adopted as the
indicators of social and material advantage.
Social capital
One of the fastest-growing areas of study in sociology and health in-
equalities has been social capital (Putnam 1995; Kaplan et al. 1996;
Wilkinson 1996; Kwachi and Kennedy 1997 ; Hawe and Shiell 2000). Its
meaning and measurements are still highly contested (Muntaner et al.
2000; Cattell 2001). The measure of social capital used in the present study
has been informed mainly by the ideas of Wilkinson, Putnam and Kwachi,
who argue that feelings of ‘ trust’ and ‘ reciprocity’ at the local or com-
munity level are important for understanding health inequalities. A
scale was devised to test whether these two concepts had an equally im-
portant effect on the quality of life among people in aged between 65 and
75 years.
698
Richard D. Wiggins et al.
Social networks and recent life events
There is growing evidence that social networks and social support have an
impact on a person’s physical and mental health and their risk of in-
stitutionalisation (Bosworth et al. 2000; Hinkikka et al. 2000). However, the
relationship between social support and measures of subjective wellbeing
is inconsistent (Bowling, Farquhar and Browne 1991). Bowling (1995) re-
ported that when asked to rate the most important things in their lives,
many older people mentioned friends and family. We were interested to
find out the extent to which social support influenced the quality of life in
the Third Age. Following Ward, Sherman and LaGory (1984), it was be-
lieved important to measure both the quality and quantity of social con-
tacts, and for this purpose questions were adapted from the local area
study carried out by the National Centre for Social Research (Veenstra 2000). A
single question to measure the extent to which our respondent’s had close
or confiding relationships outside those in the home was added. Finally,
we felt that it was important to contextualise a person’s account of their
quality of life in terms of the impact of any stressful recent life events.
The study sample
Respondents were drawn from a unique sample. As children they were
surveyed between 1937 and 1939 by a team of medical and nutritional
scientists under the direction of Sir John Boyd Orr (Gunnell et al. 1996).
This sample was drawn from 16 locations in Great Britain and comprised
1,352 families. A range of social, dietary, health and anthropometric data
was collected. During 1996, 99 per cent of the household records and 95
per cent of the children’s medical records were retrieved from the Rowlett
Research Institute and entered into an electronic database at the De-
partment of Social Medicine at the University of Bristol. Using the National
Health Service Central Register, the Office for National Statistics successfully traced
85 per cent of the children who participated in the 1937–39 study.
A stratified random sample using three childhood socio-economic cat-
egories was drawn from the traced survivors who were aged 5–14 years at
the time of the original study and for whom complete records, including
physical measurements, were available. Recruitment continued until there
were 100 people in each of the three strata. The 300 people represented 43
per cent of those who were invited to participate in the study : 296 of them
were interviewed in their own homes during 1997–98. Retrospective
information was assembled onto a life-grid (Berney and Blane 1997) with
items on life since childhood (family and household, residential and oc-
cupational histories) and current social circumstances and self-reported
and objective health. The sample is broadly representative of their age
Predictors of the quality of life in the third age
699
group within the British population. As children, their fathers had a social
class distribution which was similar to all men aged 25–44 years in the
1931 population census. In early old age, the study sample was similar
socio-demographically to those aged 65–74 years in the 1991 census; and
objective measures of their current health were similar to those of
people aged 65–74 years in the 1995 Health Survey for England (Blane et al.
1999).
In 2000 these individuals were contacted again as part of the Economic
and Social Research Council’s Growing Older programme. Of the 296
people interviewed in 1997, 12 had either died or were untraceable. The
remaining 284 were mailed a self-completion questionnaire about their
quality of life. After two reminder mailings and telephone calls, a 92 per
cent response rate was achieved. All of our analyses are based on an
achieved sample of 263 people aged between 65–74 years.
Operationalising the measures
The CASP-19 measure
The distribution of the CASP-19 scores is remarkably symmetrical
(mean=median=42) with a slight negative skewness (range 13–57). The
higher the score the higher a person’s self-reported quality of life. Item
non-response for any of the 19 items forming the scale was handled using
multiple imputation (Rubin 1996; Hill 1997). Just over 25 per cent (65
cases) had at least one missing item. The imputation procedure for this
degree of missing information required us to complete the data matrix for
three replicates. The analyses are then repeated for each replicate and the
estimates are adjusted (Schafer 1997; Schafer and Graham 2002).1
Age and gender
Women slightly outnumbered men in the sample (140 or 53%) and they
tended to be older, for 61 per cent were aged 70 or more years, compared
with 51 per cent of the men. Overall, there was no statistical association
between age and gender (x2=2.19, p<0.14). The mean CASP-19 scores
for men and women were not significantly different (p<0.62), but there
was a significant difference between the younger (43) and older (39) age
groups (p<0.01).
Household hazards
Hazards were seen as cumulative, and the scores for each person were the
summation of the number of years of exposure to each. Thus a person
700
Richard D. Wiggins et al.
who spent one year in a damp house (score=1), which was situated close
to a source of industrial atmospheric pollution (score=1), would get an
exposure score of two hazard-years. The measure is age-adjusted and the
length of exposure was calculated from the life-grid interviews conducted
during 1997 (Blane et al. 1999).
Legacy of the lifecourse
Altogether five dichotomous items made up this domain : whether or not a
person felt that they had made an adequate provision for their pension ;
whether or not they felt that they had a choice about when to retire ; their
owner occupancy status ; whether or not they had access to a car; and
whether or not they felt that they were in good health (denoted by the
absence of a reported long-standing illness).
Social capital
The postal questionnaire items were derived from the British Household
Panel Survey. Its interviewers asked each respondent to name five ‘ good
things’ and five ‘ bad things’ about their local area. From the responses, a
list was compiled of the most frequent statements among the respondents
aged 65–74 years. From this list we chose a mixture of 14 positive and
negative statements that we believed captured aspects of trust and reci-
procity (see Table 1). Respondents were asked to endorse which of the 14
items applied to their local area.
To understand the nature of this domain and to reduce the number of
items for the regression analysis, an exploratory factor analysis was un-
dertaken. The varimax rotation revealed four distinct dimensions or
components, which explained over 53 per cent of the total variance. The
dimensions were labelled ‘ misery’, ‘ sense of community ’, ‘ deprivation’
and ‘affluence’. Misery is a general dimension that taps into feelings of
wanting to leave the area, high levels of crime and graffiti, no sense of
community and deprivation. The ‘sense of community’ dimension is
positive and indicated by: friendly people who enjoy living in the area,
people looking out for each other, and a good community spirit. The
‘ deprived’ dimension paints a bleak assessment, and includes deprivation,
graffiti, lack of safety and unfriendly [people]. Finally, ‘ affluent’ marks out
a sense of comfort : it associates with privacy, a safe place, a sense of
affluence, and ‘ few neighbours but acceptable’. Table 1 presents the
component loadings. Summary scores for each of the dimensions were
calculated by summing the endorsements for each domain (adjusting for
the sign of each loading).
of the CASP-19 measure
RICHARD D. WIGGINS*, PAUL F. D. HIGGS#, MARTIN HYDE#
and DAVID B. BLANE**
ABSTRACT
This article aims to identify and analyse the lifecourse and contextual factors that
influence the quality of life in early old age. We conceptualise quality of life as
distinct from the factors which influence it, and employ a model of the quality of
life that is derived from an explicit theory of human need. The operational
measure (CASP-19) consists of 19 Likert-scaled items which cover four theoretical
domains: control, autonomy, self-realisation and pleasure. A postal questionnaire
was sent to 286 British people aged 65–75 years who were members of the sample
for the 1930s Boyd-Orr study of health and diet and who had been followed up
through retrospective interviews during the late 1990s. The 286 were broadly
representative of their age group. The survey’s response rate was 92 per cent.
Respondents provided information on a number of contextual influences on their
quality of life in early old age, notably social support and participation, the quality
and quantity of social contact, feelings of trust and reciprocity about the local
neighbourhood, health and financial security. In the analyses reported here, a
series of conceptual and operational influences on quality of life in early old age is
identified using block regression models. Finally, the relative impact of each
predictor on CASP-19 is examined. The findings suggest that the legacy of the
past tends to be best captured by people’s feelings about the adequacy of their
pensions and their status as owner-occupiers as well as a feeling that the area in
which they live is deprived. The quality of the social contact people describe and
how close they feel to those around them will ameliorate the negative impacts of
the past and the immediate environment. In addition, we must recognise that
people remain vulnerable to the impact of loss : recent bereavement, and major
illnesses can impact on a person’s quality of life.
KEY WORDS – quality of life, Casp-19, third age.
* City University, London.
# University College, London.
** Imperial College of Science, Technology and Medicine, London.
694
Richard D. Wiggins et al.
Background
There is widespread agreement that as society changes, so also does
people’s experience of ageing and later life (Phillipson 1998; Gilleard
and Higgs 2000 ; Vincent 2003). During the last century, the population
structure and vital statistics of the United Kingdom changed markedly,
with clear consequences for life expectancy and health (Office for National
Statistics 1998; Blaikie 1999). Alongside the increased health and longevity
of the older generation, there has been an increase in early retirement
(Gruber and Wise 1999). Furthermore, the increase in the availability of
occupational and private-pension provision to supplement the basic state
pension means that a section of those retiring have an income that is (often
considerably) more than subsistence level (Department for Work and
Pensions 2000). Some authors describe these phenomena as indicators of
the move into ‘disorganised capitalism ’ (Lash and Urry 1997) or late
modernity (Giddens 1991). Whatever the conceptual status of such changes,
they pose serious challenges for the understanding of the lifecourse.
In a Fresh Map of Life, Peter Laslett (1996a) portrayed four distinct ‘ ages’
of a person’s lifecourse. He represented the ‘ First Age’ as the period of
childhood dependency, the ‘ Second Age’ as the time of independence,
employment and maturity, the ‘Third Age’ as the period during which
people are freed from work and family constraints and have time to pursue
a good quality of life. Finally, he saw the ‘Fourth Age’ as characterised by
dependence and declines in health. The focus of this paper is on people in
early old age (aged around 65–74 years), who might best be described as
potential members of the ‘Third Age’. For many authors, the ‘ Third Age’
is in historical terms a new lifecourse stage (Young and Schuller 1991;
Midwinter 1992 a; Laslett 1996b). As Gilleard and Higgs (2002) have
`
argued, the lives of those in retirement in fin de siecle Britain are defined by
circumstances that have greatly changed from those experienced by their
predecessors. The Third Age seems to be based upon and to offer a later
life that is agentic and reflexive. It is based on health and leisure rather
than infirmity and poverty. At the same time, older people are neither
socially nor economically homogeneous (Baltes et al. 1993). This paper uses
the outcome measure CASP-19 to examine the extent to which being in
the Third Age implies freedoms from family constraints and to pursue
leisure activities, a time for self-realisation, and opportunities for fun
(Hyde et al. 2003). Specifically, we explore the contributions of gender,
ageing, lifecourse events and variations as well as contemporary contex-
tual factors on the quality of life of a unique sample of people in early
old age drawn from the 1930s Boyd-Orr study of health and diet (Gunnell
et al. 1996).
Predictors of the quality of life in the third age
695
Conceptual framework
Our conceptual map was developed from previous research and publica-
tions about several key potential influences on a person’s quality of life in
the Third Age. It represents a framework for the analyses that follow. First,
we describe the theoretical foundations of our outcome measure; and
continue with a brief description of each of the factors that influence or
shape a person’s quality of life. We argue that quality of life may be shaped
by ageing, gender, the past as captured by lifetime exposure to poor ac-
commodation and environment, as well as by markers of passage into
early old age, pension provision, current housing tenure, health status and
access to a car. We have collectively referred to these items as ‘ legacy of
the lifecourse ’. To capture the present context we present people’s re-
flection of the quality of their neighbourhood and feelings of safety under
the heading ‘social capital’. A person’s current social network is captured
not only in terms of the frequency with which they communicate with
family and friends but the quality of these exchanges and the closeness
they derive from them. Finally, there is a domain to capture the impact of
recent life events.
The CASP-19 measure of the quality of life
We have argued elsewhere that quality of life remains under-theorised and
ill-defined (Hyde et al. 2003 ; Higgs et al. 2003). The evidence that signifi-
cant numbers of older people are living longer, healthier and more active
lives requires us to rethink the way in which we conceptualise and measure
quality of life in this age group. The lack of a theoretically-informed
measure of quality of life in early old age has meant that for a long time it
has been measured by various proxies, such as health (Bowling 1997).
These proxy measures rest on an implicit set of normative assumptions
about each person’s quality of life and often neglect the range of experi-
ences of people in older age. Such scales should be seen as measures of the
influences on quality of life rather than measures of quality of life itself.
Partly as a reaction to this criticism, recent attempts to measure quality of
life have sought to emphasise the importance of subject-led accounts in the
construction of measures of quality of life in early old age (Farquhar 1995;
Bowling 1995). Researchers seek to understand what quality of life means
for each person by asking respondents to rate the most important things in
their lives. This approach suffers from some of the same problems as
‘ expert-defined’ quality of life measures, in that both tend to generate
measures of the influences on the quality of life rather than of quality of life
itself.
696
Richard D. Wiggins et al.
In developing the CASP-19 measure, the initial premise was that any
quality of life measure should be distinct from the factors that influence it.
This conceptualisation was derived from a theory of human need that
recognises the social and biological components as equal (Doyal and
Gough 1991). This ‘ needs satisfaction’ approach assumes that quality of
life should be assessed as the degree to which human needs are satisfied.
Grounding the model of quality of life in this ontology enables meaningful
comparisons of different quality of life scores. The model implies that
although there are common basic human needs, such as food and shelter,
it is equally important to recognise that being human is principally an
active and reflexive process (Giddens 1990 ; Turner 1995). Based on the
range of human needs, four domains of the quality of life are con-
ceptualised : ‘control ’, ‘ autonomy’, ‘ self realisation ’ and ‘ pleasure ’. The
first letter of each domain and its 19 items create the acronym CASP-19
that names the measure. ‘Control ’ and ‘ autonomy’ are prerequisites of
free participation in society. Much gerontological research has however
failed to consider the more active and reflexive dimensions of being old in
an increasingly complex society in which ‘ post-materialist ’ values prevail
(Inglehart 1997). The narrowness of its focus has also been criticised for
ignoring the agency of older people (Gilleard and Higgs 1998). It is the
more active aspects of old age that we aim to map with the other two
dimensions of our model, ‘self realisation’ and ‘pleasure ’.
The operationalisation of the theoretical model into the self-enumer-
ated CASP-19 measure of the quality of life has been described in more
detail elsewhere Hyde et al. (2003). Each domain consists of four or five
Likert-scaled agreement items, e.g. ‘I look forward to each day ’, and in total
there are 19 items, which are summed to form the overall score. The range
of the scale is from zero, which represents a complete absence of quality of
life, to 57, which represents total satisfaction across the four domains.
Internal-consistency analysis reveals a reliable overall measure (Cronbach’s
alpha=0.7), which has demonstrable concurrent validity. Second-order
factor analysis also confirms the structure of four domains (control, auton-
omy, self-realisation and pleasure), for all load on to a single underlying
latent variable that represents an intrinsic measure of the quality of life.
CASP-19 has been adopted by several important studies, notably the
English Longitudinal Study of Ageing (Marmot et al. 2003) and the retirement
module in wave 11 of the British Household Panel Survey (BHPS) (Taylor 2003).
Age, gender and lifecourse hazards
Many writers have expressed a concern that the gendered dimension of
ageing and older age is neglected (Bury 1995; Ginn and Arber 1999). This
Predictors of the quality of life in the third age
697
study therefore set out to discover whether a person’s age and gender had
any impact on their quality of life. For convenience, age was dichotomised
above and below 70 years. Blane and colleagues (1998) have shown that
exposure to social disadvantage, as indicated by exposure to environ-
mental health hazards, accumulates across the lifecourse and is related in
complex ways to health in early old age (Holland et al. 2000 ; Montgomery
et al. 2000; Berney et al. 2000 a, b). In the light of these findings, we were
interested to see if the same accumulated hazard exposures directly affect
the quality of life. The hazard variables used in the present analyses refer
to the domestic sphere of life (‘ household hazards’) : the number of years
of exposure to residential dampness, atmospheric pollution in the resi-
dential locality, and inadequate nutrition.
These hazards or factors are often represented by cross-sectional vari-
ables but we regard them as particular stages of processes or pathways into
early old age. For example, the choice whether to retire or to continue
working, current health status (poor or otherwise), car ownership or access
to a car, an individual’s subjective assessment of whether or not they feel
that they have enough income to meet their future needs, and their cur-
rent housing status (owner occupier or not), collectively mark the legacy of
the past. These markers represent past decisions and the consequences, or
life chances, as people negotiated their passage through life. They are
sensitive to the heterogeneity of individuals in the Third Age, many of
whom will have retired early, be in good health and relatively affluent
(Hardy and Hazelrigg 1999 ; Gruber and Wise 1999; Scase and Scales
2000). Following Arber’s (1991) argument that housing tenure and car
access are more sensitive to women’s positions when compared with social
status indicators based on male occupations, they were adopted as the
indicators of social and material advantage.
Social capital
One of the fastest-growing areas of study in sociology and health in-
equalities has been social capital (Putnam 1995; Kaplan et al. 1996;
Wilkinson 1996; Kwachi and Kennedy 1997 ; Hawe and Shiell 2000). Its
meaning and measurements are still highly contested (Muntaner et al.
2000; Cattell 2001). The measure of social capital used in the present study
has been informed mainly by the ideas of Wilkinson, Putnam and Kwachi,
who argue that feelings of ‘ trust’ and ‘ reciprocity’ at the local or com-
munity level are important for understanding health inequalities. A
scale was devised to test whether these two concepts had an equally im-
portant effect on the quality of life among people in aged between 65 and
75 years.
698
Richard D. Wiggins et al.
Social networks and recent life events
There is growing evidence that social networks and social support have an
impact on a person’s physical and mental health and their risk of in-
stitutionalisation (Bosworth et al. 2000; Hinkikka et al. 2000). However, the
relationship between social support and measures of subjective wellbeing
is inconsistent (Bowling, Farquhar and Browne 1991). Bowling (1995) re-
ported that when asked to rate the most important things in their lives,
many older people mentioned friends and family. We were interested to
find out the extent to which social support influenced the quality of life in
the Third Age. Following Ward, Sherman and LaGory (1984), it was be-
lieved important to measure both the quality and quantity of social con-
tacts, and for this purpose questions were adapted from the local area
study carried out by the National Centre for Social Research (Veenstra 2000). A
single question to measure the extent to which our respondent’s had close
or confiding relationships outside those in the home was added. Finally,
we felt that it was important to contextualise a person’s account of their
quality of life in terms of the impact of any stressful recent life events.
The study sample
Respondents were drawn from a unique sample. As children they were
surveyed between 1937 and 1939 by a team of medical and nutritional
scientists under the direction of Sir John Boyd Orr (Gunnell et al. 1996).
This sample was drawn from 16 locations in Great Britain and comprised
1,352 families. A range of social, dietary, health and anthropometric data
was collected. During 1996, 99 per cent of the household records and 95
per cent of the children’s medical records were retrieved from the Rowlett
Research Institute and entered into an electronic database at the De-
partment of Social Medicine at the University of Bristol. Using the National
Health Service Central Register, the Office for National Statistics successfully traced
85 per cent of the children who participated in the 1937–39 study.
A stratified random sample using three childhood socio-economic cat-
egories was drawn from the traced survivors who were aged 5–14 years at
the time of the original study and for whom complete records, including
physical measurements, were available. Recruitment continued until there
were 100 people in each of the three strata. The 300 people represented 43
per cent of those who were invited to participate in the study : 296 of them
were interviewed in their own homes during 1997–98. Retrospective
information was assembled onto a life-grid (Berney and Blane 1997) with
items on life since childhood (family and household, residential and oc-
cupational histories) and current social circumstances and self-reported
and objective health. The sample is broadly representative of their age
Predictors of the quality of life in the third age
699
group within the British population. As children, their fathers had a social
class distribution which was similar to all men aged 25–44 years in the
1931 population census. In early old age, the study sample was similar
socio-demographically to those aged 65–74 years in the 1991 census; and
objective measures of their current health were similar to those of
people aged 65–74 years in the 1995 Health Survey for England (Blane et al.
1999).
In 2000 these individuals were contacted again as part of the Economic
and Social Research Council’s Growing Older programme. Of the 296
people interviewed in 1997, 12 had either died or were untraceable. The
remaining 284 were mailed a self-completion questionnaire about their
quality of life. After two reminder mailings and telephone calls, a 92 per
cent response rate was achieved. All of our analyses are based on an
achieved sample of 263 people aged between 65–74 years.
Operationalising the measures
The CASP-19 measure
The distribution of the CASP-19 scores is remarkably symmetrical
(mean=median=42) with a slight negative skewness (range 13–57). The
higher the score the higher a person’s self-reported quality of life. Item
non-response for any of the 19 items forming the scale was handled using
multiple imputation (Rubin 1996; Hill 1997). Just over 25 per cent (65
cases) had at least one missing item. The imputation procedure for this
degree of missing information required us to complete the data matrix for
three replicates. The analyses are then repeated for each replicate and the
estimates are adjusted (Schafer 1997; Schafer and Graham 2002).1
Age and gender
Women slightly outnumbered men in the sample (140 or 53%) and they
tended to be older, for 61 per cent were aged 70 or more years, compared
with 51 per cent of the men. Overall, there was no statistical association
between age and gender (x2=2.19, p<0.14). The mean CASP-19 scores
for men and women were not significantly different (p<0.62), but there
was a significant difference between the younger (43) and older (39) age
groups (p<0.01).
Household hazards
Hazards were seen as cumulative, and the scores for each person were the
summation of the number of years of exposure to each. Thus a person
700
Richard D. Wiggins et al.
who spent one year in a damp house (score=1), which was situated close
to a source of industrial atmospheric pollution (score=1), would get an
exposure score of two hazard-years. The measure is age-adjusted and the
length of exposure was calculated from the life-grid interviews conducted
during 1997 (Blane et al. 1999).
Legacy of the lifecourse
Altogether five dichotomous items made up this domain : whether or not a
person felt that they had made an adequate provision for their pension ;
whether or not they felt that they had a choice about when to retire ; their
owner occupancy status ; whether or not they had access to a car; and
whether or not they felt that they were in good health (denoted by the
absence of a reported long-standing illness).
Social capital
The postal questionnaire items were derived from the British Household
Panel Survey. Its interviewers asked each respondent to name five ‘ good
things’ and five ‘ bad things’ about their local area. From the responses, a
list was compiled of the most frequent statements among the respondents
aged 65–74 years. From this list we chose a mixture of 14 positive and
negative statements that we believed captured aspects of trust and reci-
procity (see Table 1). Respondents were asked to endorse which of the 14
items applied to their local area.
To understand the nature of this domain and to reduce the number of
items for the regression analysis, an exploratory factor analysis was un-
dertaken. The varimax rotation revealed four distinct dimensions or
components, which explained over 53 per cent of the total variance. The
dimensions were labelled ‘ misery’, ‘ sense of community ’, ‘ deprivation’
and ‘affluence’. Misery is a general dimension that taps into feelings of
wanting to leave the area, high levels of crime and graffiti, no sense of
community and deprivation. The ‘sense of community’ dimension is
positive and indicated by: friendly people who enjoy living in the area,
people looking out for each other, and a good community spirit. The
‘ deprived’ dimension paints a bleak assessment, and includes deprivation,
graffiti, lack of safety and unfriendly [people]. Finally, ‘ affluent’ marks out
a sense of comfort : it associates with privacy, a safe place, a sense of
affluence, and ‘ few neighbours but acceptable’. Table 1 presents the
component loadings. Summary scores for each of the dimensions were
calculated by summing the endorsements for each domain (adjusting for
the sign of each loading).
Comentários
Postar um comentário