Stepparenting and Mental Health

Abstract

Demographic changes in the Western world over the last few decades, such as later marriages, lower fertility, increasing divorce
rates and rising rates of cohabitation, have brought about significant changes in household formation and composition. One
outcome is a growing number of stepfamilies, where a parent, whether never married, separated, widowed or divorced, forms
a new marriage or partnership. Nowadays, most stepfamilies result from divorce, while in the past they were more likely to
result from widowhood.

Full text

Chapter 11
STEPPARENTING AND MENTAL HEALTH
Paul Boyle1,2, Peteke Feijten1,2, Zhiqiang Feng,1,2, Vernon Gayle3 and Elspeth Graham1
1School of Geography and Geosciences, University of St Andrews, United Kingdom; 2
L
ongitudinal Studies Centre
– Scotland, University of St Andrews, United Kingdom; 3Department of Applied Social Science, University o
f
Stirling, United Kingdom.
Abstract: This chapter describes a study of the effects of being a stepparent or a partner of a stepparent
on mental health. Using longitudinal cohort data from the National Child Development Study
(NCDS), it was found that adults aged 33 living in stepfamilies have a higher risk of having
poor mental health than otherwise comparable adults in ‘first families’. It was also shown that
this was partly due to selection of respondents with prior mental health problems into
stepfamilies. Among those who had no prior mental health problems, only adults in
stepfamilies where both partners are stepparents to each others’ children had an increased risk
of having poor mental health. For those with prior mental health problems, being in any type
of stepfamily increased the risk of poor mental health compared to first families, suggesting
that stepfamily life poses an extra burden on the already frail mental health of this group.
Keywords: birth cohort; mental health; National Child Development Study (NCDS); parenthood;
longitudinal analysis; selection effects; stepfamily.
1. INTRODUCTION
Demographic changes in the Western world over the last few decades, such as later
marriages, lower fertility, increasing divorce rates and rising rates of cohabitation,
have brought about significant changes in household formation and composition. One
outcome is a growing number of stepfamilies, where a parent, whether never married,
separated, widowed or divorced, forms a new marriage or partnership. Nowadays,
most stepfamilies result from divorce, while in the past they were more likely to result
from widowhood.
Despite the rising incidence of stepfamilies and the demographic and social
differences between stepfamilies and traditional families with two biological parents,
researchers concerned with family life and parenting were relatively slow in
acknowledging the importance of such non-traditional families (Ferri and Smith 1998,
Utting 1995). While there has been an increase in social science research on
stepfamilies in recent years much remains to be done (Coleman et al. 2000). For
example, numerous researchers have explored the effect of living in a stepfamily on
children, including studies of their psychological well-being but, perhaps surprisingly,
little research has considered the potential psychological impacts on the adults
(stepparents and their partners) of living in such a household arrangement. This is the
focus of this study.
1

2. BACKGROUND
In the UK most people live in traditional households, but non-traditional households
are becoming more common. Of those marrying during the 1990s, nearly 50% will
end up divorced if current trends continue (Allan 1999), and a growing number of
divorcees are starting new relationships. Haskey (1994) estimated that 12% of British
children will live in a stepfamily before their sixteenth birthday. Recent estimates for
Britain suggest that about 40% of mothers will experience being a lone parent and
about 75% of lone mothers will go on to form a stepfamily (Ermisch and Francesconi
2000). Currently nearly 90% of stepfamilies involve children living with their mother
and a new male partner (Finch 2002). Importantly, stepfamilies differ from traditional
families demographically because they tend to include more and older children than
do first families (Haskey 1994). The greater complexity of intra-household
relationships in stepfamilies provides scope for tensions to arise and, perhaps,
increases the potential for negative health impacts for both the children and the
parents.
Numerous studies highlight the strains that every-day life in stepfamilies may
entail and the effects this may have on stepchildren’s health and well-being (Brown
and Booth 1996; Pryor and Rodgers 2001). Not only do stepchildren experience the
breakdown of their parent’s relationship but they often feel relatively neglected by the
biological parent. This may be combined with the potentially disruptive effects of
having to divide their time between two homes. Most studies of the effects of
remarriage on children fail to show a benefit, despite the financial advantages that
usually result (Fergusson et al. 1994; Pagani et al. 1998; Walper 1995; Duncan and
Hoffman 1985; Zill 1988). Some findings point to negative effects, with stepchildren
performing worse at school (Pong 1997; Teachman et al. 1996), five-year old children
in stepfamilies being significantly more at risk of behavioural and developmental
problems than children in traditional families (Wadsworth et al. 1985), and a higher
risk of drinking alcohol, drug abuse and problem behaviour among schoolchildren
living with a stepfather (del Carmen et al. 2002; Mekos et al. 1996). While Joshi et al.
(1999) found that maternal educational attainment and, to a lesser extent, family
economic circumstances eliminated the relationship between family structure and
children’s cognitive and behavioural outcomes, the majority of quantitative studies
suggest that stepchildren are at greater risk of a range of problems (Coleman et al.
2000; Ram and Hou 2003).
There is also a considerable literature on the effects of family arrangements
and marital status on adult health. Higher mortality rates among the unmarried, those
who live alone and the divorced, compared to those who are married, are well
established (Seeman et al. 1987; Trovato and Lauris 1989; Gardner & Oswald, 2004).
More depressive symptoms are apparent among both the recently separated (Neff and
Schluter 1993) and those who have been separated for longer (Richards et al. 1997),
even when mental health status prior to separation/divorce is taken into account
(Wade and Pevalin 2004). In some studies, the beneficial effects of marriage are
found for men but not women (Berkman and Syme 1979; Avlund et al. 1998). Lone
parenthood has been studied in some detail. Hope et al. (1999) found that lone parents
suffer higher levels of mental distress than other parents, although this may be related
to the significantly higher poverty levels they experience (Keirnan and Mueller 1998;
Shouls et al. 1999). Notably, within this large literature, there are virtually no studies
examining explicitly the effects of living in a stepfamily on stepparents’ and their
partner’s health.
2

Another strand of the literature which is relevant to this study considers
‘marital quality’ and how people cope with the redefinition of kinship that follows
divorce and remarriage (overviews are provided by Coleman & Ganong, 1990 and
Pasley et al., 1993). Remarriage has been found to lead to lower marital quality and
satisfaction than first marriage (Clingempeel, 1981; White and Booth, 1985), and this
may have an impact upon mental health. In a qualitative study by Simpson (1994), it
was argued that the roles of remarried persons compared to married persons, and
stepparents compared to biological parents, are less well defined and thus harder to
fulfil with confidence and satisfaction. This is sometimes labelled as the ‘incomplete
institution hypothesis’ (Cherlin, 1978). We found one study specifically focusing on
the experience of stepparenting and how this affected marital quality and the
stepparent-stepchild relationship. Ambert (1986) showed that an intimate relationship
between stepparents and stepchildren was harder to establish when stepchildren were
not resident and that the birth of a common child into the stepfamily made the
stepparent-stepchild relationship better for men, but not for women.
Thus, stepfamily arrangements are potentially stressful since they involve the
negotiation of different intra-household relationships that may introduce new sources
of tension. For example, Hetherington and Jodl (1994) found that stepparents remain
less engaged and more authoritarian in parenting stepchildren than in parenting their
biological children, illustrating the scope for friction between biological parents and
stepparents. Yet a recent wide-ranging literature review on remarriage and
stepparenting cited only a small number of studies on the psychological health of
adults, the majority of which examined the general effects of remarriage (Coleman et
al. 2000). A study by Ferri and Smith (1998) suggested that adults in stepfamilies
were more likely to express ‘negative feelings’ and suffer from depression than those
in first families. Also, a recent study considered the relationship between depression
and being a parent in the US (Evenson and Simon 2005). Parents with young children
living at home were shown to have significantly higher rates of depression than non-
parents and those with adult children who have left home. However, no increased risk
of depression for stepparents with minor aged stepchildren was found and this study
only considered the mental health of stepparents, not their partners.
In addition to the lack of previous work on the topic, another challenge to any
quantitative study of stepparenting comes from relatively recent, largely qualitative
studies that have begun to undermine the idée fixe that divorce is always harmful, at
least to children (Smart 2003). These new approaches benefit from life course
perspectives and deploy different conceptual categories. In developing our
quantitative research design, we have tried to incorporate a greater subtlety into our
analysis than some previous studies, by recognising that there are many types of
stepfamily and that health impacts may change over time.
Coleman et al. (2000) state that more longitudinal quantitative studies of the
effects of stepparenting are required. One particularly important reason for this is that
cross-sectional studies cannot control adequately for selection effects. Thus, while
stepparenting may result in poorer mental health, the opposite effect is also possible;
those prone to poorer mental health may be more likely to end up living in
stepfamilies (Amato 2000). Only with longitudinal data could this be explored. The
National Child Development Study (NCDS) collects data for a large birth cohort
based on all children born in a single week in 1958 and provides information from
birth throughout childhood and adolescence into young and middle age adulthood.
These data made it possible to control for adolescent characteristics which may
influence subsequent mental health. This is a simple but effective way of providing
3

improved statistical control for the potential of increased representation of mentally ill
people into stepparenthood.
In sum, there is a dearth of research on the mental health of adults living in
stepfamilies, a need for a large scale quantitative longitudinal study in the UK and a
demand that such a study gives due consideration to new and challenging
conceptualisations of family arrangements. Our study, a longitudinal analysis using
secondary data from a British birth cohort study, is designed to respond to all three of
these points.
3. RESEARCH AIMS AND HYPOTHESES
The study aimed to fill the gap in knowledge about the relationship between
stepparenting and mental health in the UK. Our main research question was, ‘how
does the mental health of adults in stepfamilies differ from the mental health of adults
in otherwise comparable first families?’ We were also interested in several
dimensions of stepfamilies that are identified in the literature as affecting marital
quality and personal well-being. These relate to the characteristics of the stepparent,
such as age, income and attitude, and the characteristics of the children in the
stepfamily, such as age and whether they are resident or non-resident. Thus, we
derived the following hypotheses:
1. Stepparents have worse mental health outcomes than parents in first families.
2. The partners of stepparents have worse mental health outcomes than parents in
first families.
These first two hypotheses are at the heart of our study, as described above.
3. The poorer the stepfamily, the greater the negative effect on mental health for both
partners.
This has been identified as an understudied area in stepfamily research, but we base
our hypothesis on clinicians’ reports that financial issues are one of the primary
sources of stress in stepfamilies (Coleman & Ganong, 1990).
4. The younger the stepparent, the greater the negative effect on mental health for
both partners.
Firstly, young age indicates a lack of experience with children. Palisi et al. (1991)
found that stepparents who had previous parenting experience performed better at
stepparenting than those with no experience. Secondly, young age can be interpreted
as a proxy for time spent in the stepfamily. There is some evidence that the family
situation improves when the stepparent is longer in the family (Pasley et al. 1993).
5. The presence of a child born to the two partners in a stepfamily reduces the
likelihood of poor mental health outcomes for both partners.
Studies in the past (Berman, 1980; Ambert, 1986) have found that a common child
increases the quality of the marriage of partners in a stepfamily, and this leads us to
expect that it also positively affects the mental health of the partners. However, the
quality of the relationship between stepparents and stepchildren has not been found to
improve as a result of the birth of a common child (White and Booth, 1985).
6. The presence of adolescent stepchildren increases the likelihood of poor mental
health outcomes for both partners in stepfamilies.
In a study on the well-being of stepchildren, Hetherington and Clingempeel (1992)
showed that stepparents found it particularly hard to cope with adolescent
stepchildren who were often coercive and hostile towards the stepparent, whereas
4

with younger children it was often easier for the stepparent to find a parenting style
that worked. The relationship between stepfathers and adolescent stepdaughters was
found to be especially problematic.
7. The likelihood of poor mental health outcomes for both parents in stepfamilies
increases where stepparents hold more traditional views.
Those with traditional views often have an authorative parenting style, and the
literature shows that authorative parenting styles do not work well with stepchildren
(especially older stepchildren) (Hetherington et al. 1988; Hetherington 1991).
Therefore, we expect traditional views to lead to an increased risk of poor mental
health for such parents.
8. The presence of non-resident children in the household increases the likelihood of
poor mental health outcomes for both partners in stepfamilies.
Relationships between stepparents and resident children have been shown to be closer
than between stepparents and non-resident stepchildren. Stepparents with non-resident
stepchildren may have to manage a relationship with the ex-spouse and it may also be
difficult to develop close bonds with stepchildren who rarely visit (Ambert, 1986).
These issues may add stress to the relationship and subsequently affect the
stepparent’s mental health.
4. METHODS
The data used for this study is the National Child Development Study (NCDS), which
is a sample of all babies born in the UK in one week in the spring of 1958. The
sample originally included 17,416 respondents and they have been returned to seven
times to collect a variety of data on issues including: mental health, partnership
histories, and other time-invariant and time-varying demographic, health and socio-
economic variables (University of London, 2007). We focused in this study on
characteristics as recorded in 1991 when the sample members were aged 33, but we
also included some variables from earlier in the life course which we expected to be
related to subsequent mental health status (University of London, 2000). The sample
(N=6,121) includes those with children, with a valid measures of behavioural/mental
problems at age 16 and age 33, and family status information. Descriptive statistics
for all the variables are given in Appendix 1.
Our outcome variable of interest is mental health. The measurement
instrument used to measure mental health in the NCDS is the Malaise Inventory Scale
(MIS) developed by Rutter et al. (1970). It consists of 24 questions which are
designed to capture depression and anxieties, obsessions and phobias. It has
commonly been used as a mental health screening instrument and several studies have
tested and confirmed the alpha reliability and internal consistency of this scale (e.g.
Cherlin et al. 1998; Hirst and Bradshaw 1983). The distribution of the MIS scale is
highly skewed, and a common solution to deal with this is to divide the scale into a
binary variable. As in other studies, such as Flouri and Buchanan (2002) and Power et
al. (1988), we defined a score of 7 or higher out of 24 as an indicator of poor mental
health. As Chase-Landsdale et al. (1995: 1619) point out, the MIS is merely a
screening instrument, and a score of 7 or higher must be interpreted as “a clinical cut
off score, indicating a high likelihood of the presence of mental illness and the need
for psychiatric help”. Twelve percent of adults aged 33 have a score over 6 on the
MIS.
5

We also wanted to control for proneness to poor mental health at the onset of
adulthood, because we hypothesised that this may influence whether people are more
or less likely to end up living in a stepfamily – a potential sample selection effect.
This was operationalised through the Home Behaviour Scale (HBS), also developed
by Rutter et al. (1970). This measure was collected in 1974 when respondents were
aged 16. The instrument consists of a 22-item scale that is meant to signal both
externalizing (where the child shows under-controlled behaviour such as aggression
or disobedience) and internalizing (where the child shows over-controlled behaviour
such as anxiety or depression) behavioural disorders in children. Again, a score of 7
or higher was chosen as a threshold for defining behavioural problems. Twenty one
per cent of children at age 16 have a score over 6 on the HBS.
In order to distinguish different types of (step)families, we created a
categorical variable with five categories: (1) respondents in first families, (2)
respondents in lone parent families, (3) respondents who are stepparents, (4)
respondents who are the partners of stepparents and (5) respondents who are
stepparents and partners of stepparents at the same time (i.e., both partners have
children from a previous relationship). We label this group ‘dual stepfamilies’.
Finally, we extracted additional explanatory variables expected to influence
mental health. We used variables from different points in the life course. Some reflect
the status of the sample respondents in the year of analysis, 1991, when respondents
were 33 years old. These included the sex of the respondent, their economic status,
highest educational qualification and social class. We also included the number of
children in the family, as stepfamilies are larger than first families on average. Then,
we included variables about respondents’ characteristics at ages 7 and 16, to control
for previous circumstances expected to be associated with mental health outcomes in
later life (see Cherlin et al. 1998 and Flouri and Buchanan 2003). These included
domestic tension at age 7, financial hardship in the child’s household at age 7,
whether the child lived with his/her natural mother at age 16, whether the father was
interested in the child’s education at age 16, and school abilities at ages 7 and 16.
The method used is logistic regression, with mental health status at age 33
(1991) as the binary dependent variable (0 = non-poor mental health; 1 = poor mental
health). All models include the control variables described above, and the results are
expressed as parameter estimates in the tables. We first estimated a model without
behavioural problems in adolescence, and then a model that included this variable,
where we were particularly interested in the effects of the interaction between
behavioural problems at age 16 and family status at age 33 on mental health status at
age 33. We also graphed the coefficients by family status. In order to better facilitate
appropriate visual comparisons between categories of the explanatory variable, we
have plotted comparison intervals as suggested by Firth (2003) and further illustrated
by Gayle and Lambert (2007).
5. RESULTS
5.1 Description of stepfamilies
We distinguish between stepparents, partners of stepparents, and families where both
partners are stepparents to each other’s children in our study (‘dual stepfamilies’). The
distribution of respondents over these family types is shown in Table 1 (column 1).
Table 1 also shows that adults living in stepfamilies are different from adults living in
6

first families and lone parent families, in terms of age, number of children in the
household and the age range of these children (see also Haskey, 1994). The second
and third columns show the average number of children per family type (excluding
and including non-resident children). Columns four and five show the average age
range of the children in the household, and the average age of (step)parents per family
type. It can be seen that adults in stepfamilies are on average older and have more
children, who are of a wider age range, than people in first families and lone parent
families.
Table 1. Demographic characteristics by family type
Distribution
of family
type (%)
Average
number of
children
(resident
children
only)
Average
number of
children
(incl. non-
resident
children)
Average age
range
children
(resident
children
only)
Average age
(step)parent
*
First family 75.6 2.0 2.0 2.8 33.3
Lone parent family 7.2 1.7 1.9 3.0 ---
All stepfamilies 17.3 1.8 2.8 4.5 34.7
stepparent 6.4 1.7 2.7 4.1 37.5
partner of stepparent 8.1 1.8 2.5 4.9 32.0
both stepparents 2.8 2.3 4.2 4.6 36.1
Source: NCDS sweep 1991
* This statistic is based on partners of NCDS cohort members, because cohort members themselves were all aged
33 at the time of the survey.
5.2 Association between being a (partner of a) stepparent and mental
health
Table 2 shows the numbers and percentages of people with poor and non-poor mental
health, by family type, at age 33. Adults in stepfamilies (both stepparents and their
partners) were more likely to suffer poor mental health than adults in first families,
but less likely than lone parents.
Table 2. Poor mental health by family type
Non-poor mental health Poor mental health
N % N %
First family 5631 90.7 575 9.3
Lone parent family 434 73.9 153 26.1
Stepfamily 176 83.1 240 16.9
Source: NCDS sweep 1991
5.3 Health selection into stepfamilies?
As argued above, it is possible that part of this association is due to health selection
into stepfamilies. Table 3 shows the numbers and percentages of people with and
without behavioural problems at age 16 who ended up in a stepfamily at age 33.
Those with behavioural problems at age 16 were significantly more likely to live in a
stepfamily as an adult (22.9%) than those who had no behavioural problems at age 16
(15.9%). This supports the idea that at least part of the worse mental health of
(partners of) stepparents is due to the selection of people prone to poorer mental
health into stepfamilies. This is therefore accounted for in the models below.
7

Table 3. Behavioural problems at age 16 by stepfamily status at age 33
Behavioural problems In a stepfamily at age 33?
at age 16? No Yes Total
N % N % %
No 4057 84.1 766 15.9 100
Yes 1001 77.1 297 22.9 100
Pearson chi2(1) = 34.9; p < 0.000
Source: NCDS sweeps 1974 and 1991
5.4 Modelling results
All hypotheses were tested in models with and without control variables, which
included: sex, household size, employment status, highest completed education, social
class, domestic tension at age 7, financial hardship at age 7, school abilities at ages 7
and 16, living with the natural mother at age 16, and father interested in education at
age 16. Table 4 shows the results from two multivariate analyses. Model 1 does not
take behavioural problems at age 16 into account while Model 2 does.
We see, first of all, that the number of children in the household does not
affect the risk of having poor mental health (Model 1). Gender does have an
influence: our result replicates the well-known finding that women have a higher risk
of poor mental health than men. The employment status variable shows that people
who do not work are at increased risk of having poor mental health, especially those
who are unemployed. Also, those with lower levels of education and from lower
social classes have an increased risk of poor mental health.
Childhood circumstances at age 7 also affect mental health at age 33. Living in
a household with domestic tension or financial hardship, or performing below average
at school increases the risk of being in poor mental health. Those who lived with their
natural mother or had a father who was interested in their education at age 16 have
better mental health at age 33. Also school abilities at age 16 are positively associated
with mental health at age 33. Our main variable of interest at age 16 is whether the
respondent had behavioural problems. Model 2 shows that those with behavioural
problems at age 16 had a significantly higher risk of poor mental health at age 33. To
summarise, most control variables, both at age 33 and in childhood, are associated
with mental health in the hypothesised directions.
Table 4. Modelling results for mental health status at age 33 (1991) (N=6121)
Model 1 Model 2
Variable Coef. p-value Coef. p-value
Constant -2.378 0.000 -2.470 0.000
Behavioural problems age 16
HBS score < 7 0.000
HBS score >= 7 0.554 0.000
Family type age 33
First family 0.000 0.000
Lone parent family 0.692 0.000 0.572 0.001
Stepparent 0.401 0.017 0.282 0.187
Partner of stepparent 0.329 0.017 0.211 0.240
Dual stepfamily 0.869 0.000 0.966 0.000
Behavioural problems age 16 * Family type age 33
8

First family
0.000
Lone parent family 0.194 0.468
Stepparent
0.257 0.452
Partner of stepparent 0.206 0.463
Dual stepfamily
-0.320 0.429
Number of children in the household (incl. non-resident ) -0.011 0.797 -0.021 0.630
Sex
Male 0.000 0.000
Female 0.658 0.000 0.624 0.000
Employment status age 33
Fulltime working 0.000 0.000
Part-time working -0.080 0.556 -0.076 0.576
Unemployed 0.628 0.001 0.572 0.003
Other not working 0.228 0.079 0.224 0.086
Highest completed educational level age 33
No education completed 0.000 0.000
CSE levels 2-5 -0.410 0.001 -0.373 0.002
O levels -0.523 0.000 -0.481 0.000
A levels -0.734 0.000 -0.663 0.001
Degree or subdegree -1.064 0.000 -0.985 0.000
Unknown -0.617 0.119 -0.528 0.184
Social class age 33
Unskilled 0.000 0.000
Partly skilled -0.157 0.371 -0.139 0.434
Skilled manual -0.235 0.212 -0.219 0.250
Skilled non-manual -0.431 0.018 -0.401 0.029
Managerial / technical -0.510 0.008 -0.469 0.016
Professional -0.802 0.063 -0.725 0.094
Unknown 0.195 0.202 0.221 0.278
Domestic tension age 7
No 0.000 0.000
Yes 0.497 0.004 0.428 0.015
Unknown -1.449 0.152 -1.454 0.158
Financial hardship age 7
No 0.000 0.000
Yes 0.370 0.012 0.360 0.016
Unknown 1.474 0.141 1.471 0.149
School abilities age 7
Most abilities on or above average 0.000 0.000
Most abilities below average 0.231 0.023 0.215 0.037
Unknown 0.070 0.736 0.080 0.702
Living with natural mother age 16
No 0.000 0.000
Yes -0.440 0.013 -0.433 0.015
Father interested in child's education age 16
No 0.000 0.000
Yes -0.326 0.008 -0.332 0.008
Unknown -0.279 0.018 -0.270 0.023
School abilities at age 16
Half or more of subjects above CSE 1/O levels 0.000 0.000
More than half of subjects below CSE 1/O levels 0.375 0.023 0.339 0.041
Unknown 0.425 0.007 0.389 0.014
9

5.5 Family type
Of particular interest were the effects of family type. Figure 1 graphs the regression
coefficients for the five family types (as in Model 1, Table 4). Compared to those in
first families, the risk of poor mental health is significantly higher for lone parents and
all three types of stepfamily. For lone parents, the odds of poor mental health are
twice as high (exp(0.692) = 2.0) as for those in first families, and for dual stepfamilies
2.4 times as high (exp (0.869) = 2.38). The difference between people in first families
and people in families with only one stepparent is smaller, yet significant at the 95%
confidence interval level. These initial results are consistent with the hypothesis that
the various strains involved in stepparenting may result in poorer mental health for
both stepparents and their partners compared to those in first families, and that these
strains are even stronger in more complex stepfamilies where both partners have
stepchildren.
Figure 1. Poor mental health at age 33 (1991) by family status (derived from Model 1, Table 4)
Poor mental health
Logistic regression coefficients (with 95% comparison intervals)
-0.2
0
0.2
0.4
0.6
0.8
1
1.2
1.4
first family lone parent stepparent partner of
stepparent dual stepfamily
However, as we argued earlier, it is also possible that those prone to poorer
mental health may be selected into stepfamilies, and we showed in section 5.3 that
those with a high HBS score at age 16 were more likely to end up in a stepfamily at
age 33 than those with a low HBS score at age 16. Therefore, we controlled for
behavioural problems in adolescence in Model 2, by including an interaction between
family type at age 33 and behavioural problem status at age 16.
Figure 2 shows the coefficients calculated from the main effects and
interaction effects of family type (at age 33) and behavioural problem status (at age
16) from Model 2 (Table 4). Of those who had few behavioural problems at age 16
(HBS<7), only lone parents and those couples where both partners were stepparents
had significantly worse health than those in first families at age 33. This fits in with
the finding from earlier studies that stepparent-stepchild relationships were more
distant if both parents had their own children living in the household (Coleman &
10

Ganong, 1990). In the more common stepfamilies, where only one of the adults was a
stepparent, neither the stepparents nor their partners had significantly worse health
than those in first families, if they had few behavioural problems at age 16.
On the other hand, the risk for those who had a high HBS score at age 16 was
considerably higher in every family status category. Comparing across the categories
for those with high HBS scores at age 16, we find that lone parents, stepparents and
partners of stepparents and partners who were both stepparents, all have a
significantly higher risk of poor mental health at age 33 than their counterparts in first
families (although only significant at the 90% confidence interval level for partners of
stepparents and dual stepfamilies; note that the larger confidence interval for dual
stepfamilies will partly be caused by the small number of observations in this
category (N=55)). Hence our results suggest that those who had behavioural problems
in adolescence are much more likely to suffer poor mental health in later life, but that
those who end up living in stepfamilies (and lone parent families) suffer a
significantly heavier burden on their mental health.
Finally, we compare those with and without behavioural problems at age 16
within each family type. For those in first families who had high HBS scores at age 16
the risk of having poor mental health at age 33 was significantly higher than for those
in first families with low HBS scores at age 16. Also, lone parents, stepparents and
partners of stepparents have a significantly higher risk of having poor mental health if
they had high HBS scores at age 16 than if they had low scores. Only for dual
stepfamilies, there is no difference in poor mental health risk for those with and
without behavioural problems in adolescence. For both groups the risk is high,
indicating that living in a stepfamily with children from both partners’ previous
relationships leads to stress and worries, no matter what the prior mental health status.
Figure 2. Poor mental health at age 33 (1991) by family status (derived from Model 2, Table 4)
Poor mental health
Logistic regression coefficients (with 95% comparison intervals)
-0.5
0.0
0.5
1.0
1.5
2.0
first family lone parent stepparent partner of
stepparent dual stepfamily
HBS<7
HBS>=7
The next five hypotheses, concerning the characteristics of the adults and the children
in stepfamilies, were tested in multivariate models with the same set of control
11

variables as shown in Table 4. The effects of the control variables in all models were
very comparable to those in Table 4 and are not discussed further.
5.6 Characteristics of the adults in stepfamilies
We hypothesised that being poor affects the mental health of those in stepfamilies,
because financial problems are a primary source of stress. We measured socio-
economic background in three ways: highest completed level of education, social
class and economic activity status (income information in the NCDS is not suitable
for this present analysis, so we did not use it). Each dimension shows the same
picture, but the results for labour market activity are the most marked, and remain
largely significant after controlling for background variables. The results show that
people who do not work have a higher risk of poor mental health than people who
work. Among those who work, those working part-time have a slightly elevated risk
of poor mental health compared to fulltime workers, but only in the model without
control variables. Overall, our findings confirm that adults in stepfamilies with poorer
socio-economic circumstances have a higher risk of having poor mental health.
However, we also find this effect for adults in first families and lone parent families,
so it is not exclusively true for stepfamilies. When we control for background
variables, the effects of level of education and social class remain in the same
direction, but their magnitude decreases and they become largely insignificant.
Our next hypothesis was that the younger the stepparent, the greater the
negative impact of stepparenting on mental health. In the NCDS, it is not possible to
test hypotheses about age differentiation directly, because NCDS cohort members are
all the same age. However, we explored this by comparing results at age 33 (NCDS
sweep 1991) and age 41/42 (NCDS sweep 2000). Comparison of the results shows
that differences in mental health between those in first families and those in
stepfamilies were larger at age 33 than at age 41/42, confirming our hypothesis. Of
particular interest is the finding that at age 33, adults in stepfamilies have a higher risk
of poor mental health than those in first families, and this difference has disappeared
at age 41/42 (although, only for those with behavioural problems at age 16 in the
model with control variables). When we compare those who did and did not have
behavioural problems in adolescence within each family type, we also see that
differences were bigger at age 33 than at age 41/42.
Last, we find some confirmation for our hypothesised effect of traditional
attitudes on the likelihood of poor mental health for adults in stepfamilies. We did this
analysis on the data at age 41/42, with attitudes as measured at age 33 (prior attitudes
should provide a better explanation of current behaviour than current attitudes). A
traditional attitude was measured as a cumulative score on several attitudinal items
such as “wives who don't have to work should not do so”. The results show that adults
in lone parent families and dual stepfamilies who held more traditional views at age
33 have a slightly higher risk of poor mental health at age 41/42 than their
counterparts who held less traditional views (although, only significant for those who
had no behavioural problems in adolescence). For people in first families and ‘single’
stepfamilies, attitude does not affect mental health. For people who had behavioural
problems in adolescence, attitude has no effect for stepfamilies and lone parent
families, but it does for first families (the more traditional, the higher the risk of poor
mental health).
12

5.7 Characteristics of the children in stepfamilies
The presence of a child born to both parents in a stepfamily was expected to reduce
the risk of poor mental health, because it can cement the bond between family
members and provide more role clarity to the stepparent. We found marginal support
for this, as having common children did reduce rates of mental illness, but none of the
differences between adults in stepfamilies with and without common children were
statistically significant, even in the models without control variables.
The presence of adolescent children (age 13-17) in the household does
increase the likelihood of poor mental health, but this is true for all family types, not
just stepfamilies. In fact, the effect is greatest for adults in first families. When we
control for background variables, the differences between families with and without
adolescent children disappears, except for those in first families and those in double
stepfamilies who had no behavioural problems in adolescence.
Lastly, we studied the effect of having non-resident children (that is, either or
both adults in the stepfamily have a child or children who live(s) somewhere else,
usually with the other biological parent) on mental health. We expected that having
non-resident children would increase the risk of mental health problems. Firstly, for
the parent him/herself because (s)he is always in the shadow of the parent with whom
the children live, and because (s)he may miss the child(ren) and suffer from feelings
of guilt that (s)he does not spend more time with them. The stepparents of non-
resident children may also find it difficult to cope with stepchildren who they only see
infrequently, and to support their partner. Our results show a consistent effect of
having non-resident children in both the models with and without control variables.
Lone parents and adults in stepfamilies with non-resident children are more likely to
have mental health problems than their counterparts who do not have non-resident
children. The differences between those with and without non-resident children are
particularly large among people who had behavioural problems in adolescence.
6. SUMMARY AND CONCLUSION
This study has shown that there are significant differences in mental health between
adults in first families and different types of stepfamilies. Stepparents themselves, and
their partners, are more at risk of having poor mental health than their counterparts in
first families. When both adults in a stepfamily are stepparents to each other’s
children, their mental health is even more at risk.
However, as we also showed, part of this association is due to the fact that
adults who live in stepfamilies are more prone to poor mental health in the first place.
This ‘proneness to poor mental health’ was captured using a variable that measured
behavioural problems at age 16, which is a good predictor of mental health problems
in later life. Those who had behavioural problems at age 16 were more likely to end
up in stepfamilies, and therefore make up a larger proportion of the stepfamily
population than they do of the first family population. The temporal ordering of these
events means that cross-sectional data would not allow this issue to be explored.
Evenson and Simon (2005), for example, show a higher likelihood of depression
among some types of stepparents compared to parents in first families, but they
acknowledge in their discussion that this may partly be due to selection into and out
of stepfamilies of people with different prior mental health status (interestingly, they
13

speculate that it may be people with better initial mental health who select themselves
into stepfamilies).
Once we took prior mental health into account, we found that stepfamily life is
particularly harmful to the mental health of people who had behavioural problems in
adolescence. They appear to find it hard to cope with the role of being a stepparent, or
the partner of a stepparent, presumably because of the problems and tensions in
stepfamilies that are largely absent in first families. Adults who did not have
behavioural problems as an adolescent only have an increased risk of poor mental
health if they live in dual stepfamilies or lone parent families.
Several other elements of stepfamilies were also shown to account for the
adverse effect of stepfamily life on mental health. When stepfamilies have non-
resident children, that is, one (or both) of the adults has children who live in another
household, this increases the risk of poor mental health. The age of the stepparent also
mattered: the younger, the higher the likelihood of poor mental health. Other factors
increasing the likelihood of poor mental health in stepfamilies are: not working, or
working part-time (compared to working full-time), and having adolescent
(step)children in the household. However, these factors also increase the risk of poor
mental health for adults in the other family types (first families and lone parent
families), and are thus not only true for stepfamilies.
This study is one of the first of its type to compare the risk of mental health
problems for stepparents, or partners of stepparents, compared to those in first
families. Although we could not directly compare our outcomes with any previous
studies, the results do conform to studies on related topics. Studies arguing that
stepfamilies are more stressful environments than first families (such as Brown and
Booth 1996; Pryor and Rodgers 2001) seem to be supported by our findings. Also, the
worse mental health of people in stepfamilies may affect marital quality and marital
stability of couples in stepfamilies. This would provide an additional explanation for
the markedly higher divorce rates found in remarriages that involve stepchildren.
Such marriages have been found to suffer from worse marital quality (e.g. Coleman &
Ganong, 1990; Pasley, 1993), but the explicit link with mental health has not been
made in such studies.
It is also possible that our results underestimate the adverse effect of
stepfamily life on mental health. If those who find it hard to cope with being a
(partner of a) stepparent are more likely to end the relationship, they will be ignored
in our analysis. While their mental health may have been most seriously affected by
stepparenthood, leaving the relationship may have caused improvements making it
difficult to estimate the effect of stepparenting. Only long-running panel data (with
frequent waves) would allow us to examine such a case, as we would be able to
observe people entering and exiting different types of households over their life
course, and relate their (changing) family type to their mental health status.
Our results demonstrate the value of longitudinal analysis, which has allowed
the subtlety of the relationship between stepparenting and mental health to be
explored and we support Coleman et al.’s (2000) plea for more longitudinal analyses
of complex family circumstances. Clearly, the health outcomes for stepparents and
their partners would appear to deserve more attention and the results of such analyses
may be of interest to family researchers, as well as clinicians and counsellors working
with adults who are struggling with stepparenting issues.
14

REFERENCES
Allan, G. (1999) The Sociology of the Family. Oxford, Blackwell.
Amato, P.R. (2000) The consequences of divorce for adults and children, Journal of Marriage and the Family, 62:
1269-1287.
Ambert, A.M. (1986) Being a stepparent: live-in and visiting stepchildren, Journal of Marriage and the Family,
48: 795-804.
Avlund, K., Damsgaard, M.T. and Holstein, B.E. (1998) Social relations and mortality. An eleven year follow-up
study of 70-year-old men and women in Denmark, Social Science and Medicine, 47: 635-643.
Berkman, L.F. and Syme, S.L. (1979) Social networks, host resistance, and mortality: A nine-year follow-up study
of Alameda County residents, American Journal of Epidemiology, 109: 186-204.
Berman, C. (1980) Making it as a stepparent: New roles/new rules. Garden City (NY), Doubleday.
Brown, S.L. and Booth, A. (1996) Cohabitation versus marriage: a comparison of relationship quality, Journal of
Marriage and the Family, 58: 668-678.
Chase-Lansdale, P.L., Cherlin, A.J. and Kiernan, K.E. (1995) The long-term effects of parental divorce on the
mental health of young adults: a developmental perspective, Child Development 66: 1614-1634.
Cherlin. A.J. (1978) Remarriage as an incomplete institution, American Journal of Sociology, 86: 634-650.
Cherlin, A.J., Chase-Lansdale, P.L. and McRae, C. (1998) Effects of Parental Divorce on Mental Health
throughout the Life Course, American Sociological Review, 63(2): 239-249.
Clingempeel, G. (1981) Quasi-kin relationships and marital quality, Journal of Personality and Social Psychology,
41: 890-901.
Coleman, M. and Ganong, L.H. (1990) Remarriage and stepfamily research in the 1980s: increased interest in an
old family form, Journal of Marriage and the Family, 52: 925-940.
Coleman, M., Ganong, L. and Fine, M. (2000) Reinvestigating remarriage: another decade of progress, Journal of
Marriage and the Family, 62: 1288-1307.
Del Carmen, M., Alcon, G., Pedersen, J.M., Maria, A., and Gonzalez, C. (2002) Greenlandic family structure and
communication with parents: influence on schoolchildren’s drinking behaviour, International Journal of
Circumpolar Health, 61: 319-331.
Duncan, G.J. and Hoffman, S.D. (1985) A reconsideration of the economic consequences of marital dissolution,
Demography, 22: 485-497.
Ermisch, J. and Francesconi, M. (2000) The increasing complexity of family relationships: Lifetime experience of
lone motherhood and stepfamilies in Great Britain, European Journal of Population, 16: 235–249.
Evenson, R.J. and Simon, R.W. (2005) Clarifying the relationship between parenthood and depression, Journal of
Health and Social Behaviour, 46: 341-358.
Fergusson, D.M., Lynskey, M.T. and Horwood, L.J. (1994) The effects of parental separation, the timing of
separation and gender on children’s performance on cognitive tests, Journal of Child Psychology and
Psychiatry and Allied Disciplines, 35: 1077-1092.
Ferri, E. and Smith, K. (1998) Step-parenting in the 1990s, Family Policy Studies Centre, London.
Finch, N. (2002) Demographic Trends in the UK, Social Policy Research Unit, University of York.
Firth, D. (2003) Overcoming the Reference Category Problem in the Presentation of Statistical Models,
Sociological Methodology, 33(1): 1-18.
Flouri, E. and Buchanan, A. (2002) The role of father involvement in children’s later mental health, Journal of
Adolescence, 26: 63-78.
Gardner, J. and Oswald, A. (2004) How is mortality affected by money, marriage, and stress?, Journal of Health
Economics, 23: 1181-1207.
Gayle, V. and Lambert, P. (2007) Using Quasi-variance to Communicate Sociological Results from Statistical
Models, Sociology, 41: 1191-1207.
Haskey, J. (1994) Stepfamilies and stepchildren in Great Britain, Population Trends, 76: 17-28.
Hetherington, E.M., Arnett, J.D. and Hollier, E.A. (1988) Adjustment of parents and children in remarriage. In
Wolchick, S.A. and Karoly, P. (eds.), Children of Divorce, University of Chicago Press, Chicago, pp.
67-107.
Hetherington, E.M. (1991) Presidential address: Families, lies and videotape, Journal of Research on Adolescence,
1: 323-348.
Hetherington, E.M. and Clingempeel, W.G. (1992) Coping with marital transitions: A family systems perspective,
Monographs of the Society for research in Child Development, 57: 2-3 (serial no. 227).
Hetherington, E.M. and Jodl, K.M. (1994) Stepfamilies as settings for child development. In Booth, A. and Dunn,
J. (eds.) Stepfamilies: Who benefits? Who does not?, Lawrence Erlbaum Associates, Hillsdale (New
Jersey), pp. 55-80.
Hirst, M.A. and Bradshaw, J.R. (1983) Evaluating the Malaise Inventory: A Comparison of Measures of Stress,
Journal of Psychosomatic Research, 27(3): 193-199.
Hope, S., Power, C. and Rodgers, B. (1999) Does financial hardship account for elevated psychological distress in
lone mothers?, Social Science and Medicine, 49: 1637-1649.
Joshi, H., Cooksey, E., Wiggins, R., McCulloch, A., Verropoulou, G. and Clarke, L. (1999) Diverse family living
situations and child development: a multi-level analysis comparing longitudinal evidence from Britain
and the United States, International Journal of Law, Policy and the Family, 13: 292-314.
15

Keirnan, K. and Mueller, G. (1998) Who are the divorced, and who divorces? CASEbrief 6, Centre for Analysis of
Social Exclusion, London.
Mekos, D., Hetherington, E.M. and Reiss, D. (1996) Sibling differences in problem behaviour and parental
treatment nondivorced and remarried families, Child Development, 67: 2148-2165.
Murphy, M., Glaser, K., and Grundy, E. (1997) Marital status and long-term illness in Great Britain, Journal of
Marriage and the Family, 59: 156-164.
Neff, J.A. and Schluter, T.D. (1993) Marital status and depressive symptoms: The role of race/ethnicity and sex,
Journal of Divorce and Remarriage, 20: 137-160.
Pagani, L., Tremblay, R.E., Vitaro, F., Kerr, M. and McDuff, P. (1998) The impact of family transition on the
development of delinquency in adolescent boys: A 9-year longitudinal study, Journal of Child
Psychology and Psychiatry and Allied Disciplines, 39: 489-499.
Palisi, B.J., Orleans, M., Caddell, D. and Korn, B. (1991) Adjustment to stepfatherhood: The effects of marital
history and relations with children, Journal of Divorce and Remarriage, 14 (3/4): 89-106.
Pasley, K., Dollahite, D.C. and Ihinger-Tallman, M. (1993) Clinical applications of research findings on the spouse
and stepparent roles in remarriage, Family Relations, 42: 315-322.
Pong, S.L. (1997) Family structure, school context, and eighth grade math and reading achievement, Journal of
Marriage and the Family, 59: 734-746.
Power, C., Manor, O., Fox, J. and Fogelman, K. (1988) Health Selection: An Explanation of Social Inequalities in
Health in Young Adults? NCDS Working Paper 28, City University, London.
Pryor, J. and Rodgers, B. (2001) Children in Changing Families: Life After Parental Separation, Blackwell,
London.
Ram, B.L. and Hou, F. (2003) Changes in family structure and child outcomes: Roles of economic and familial
resources, Policy Studies Journal, 31: 309-330.
Richards, M., Hardy, R., and Wadsworth, M. (1997) The effects of divorce and separation on mental health in a
national UK birth cohort, Psychological Medicine, 27: 1121-1128.
Rutter, M., Tizard, J. and Whitmore, K. (1970) Education, Health and Behaviour, Longman, London.
Seeman, T.E., Kaplan, G.A., Knudsen, L., Cohen, R. and Guralnik, J. (1987) Social network ties and mortality
among the elderly in the Alameda County Study, American Journal of Epidemiology, 126: 714-723.
Shapiro, A.D. (1996) Explaining psychological distress in a sample of remarried and divorced persons: The
influence of economic distress, Journal of Family Issues, 17: 186-203.
Shouls, S., Whitehead, M., Burstrom, B. and Diderichsen, F. (1999) The health and socio-economic circumstances
of British lone mothers over the last two decades, Population Trends, 95: 41-46.
Simpson, B. (1994). Bringing the `unclear' family into focus: divorce and re-marriage in contemporary Britain,
Man, New Series, 29: 831-851.
Smart, C. (2003) Introduction: new perspectives on childhood and divorce, Childhood, 10: 123-129.
Teachman, J.D., Paasch, K. and Carver, K. (1996) Social capital and dropping out of school early, Journal of
Marriage and the Family, 58: 773-783.
Trovato, F. and Lauris, G. (1989) Marital status and mortality in Canada: 1951-1981, Journal of Marriage and the
Family, 51: 907-922.
University of London (2000). Institute of Education. Centre for Longitudinal Studies, National Child Development
Study: Sweeps 1 (1965), 2 (1969) and 3 (1974) [computer file]. National Birthday Trust Fund, National
Children's Bureau and City University Social Statistics Research Unit [original data producer(s)]. UK
Data Archive, University of Essex, Colchester [distributor], 25 July 2000 (2nd Edition). SN: 3148.
University of London (2007). Institute of Education. Centre for Longitudinal Studies, National Child Development
Study: Sweep 5, 1991 [computer file]. City University. Social Statistics Research Unit, [original data
producer(s)]. Colchester, Essex: UK Data Archive [distributor], January 2007. SN: 5567.
Utting, D. (1995) Family and Parenthood: Supporting Families, Preventing Breakdown, Joseph Rowntree
Foundation, York.
Wade, T.J. and Pevalin, D.J. (2004) Marital transitions and mental health, Journal of Health and Social Behavior,
45: 155-170.
Wadsworth, J., Burnell, I., Taylor, B. and Butler, N. (1985) The influence of family type on children’s behaviour
and development at five years, Journal of Child Psychology and Psychiatry, 26: 245-254.
Walper, S. (1995) Family relationships and adolescents social development in nuclear, single mother, and
stepfather families, Zeitschrift für Entwicklungspsychologie und Pädagogische Psychologie, 27: 93-121.
White, L.K. and Booth, A. (1985) The quality and stability of remarriages: The role of stepchildren, American
Sociological Review, 50: 689-698.
Zill, N. (1988) Behavior, achievement, and health problems among children in stepfamilies: Findings from a
national survey of children health. In Hetherington, E.M. and Arasteh, J. (eds.) The Impact of Divorce,
Single Parenting and Stepparenting on Children, Lawrence Erlbaum Associates, Hillsdale (New Jersey),
pp. 325-368.
16

Appendix 1. Descriptive statistics (N = 6121)
Variable Number Percentage
Mental health age 33
Good health 5395 88.1%
Poor health 726 11.9%
Behavioural problems at age 16
HBS score < 7 4823 78.8%
HBS score >=7 1298 21.2%
Family type age 33
First family 4638 75.8%
Lone parent 420 6.9%
Stepparent 400 6.5%
Partner of stepparent 494 8.1%
Both stepparents 169 2.8%
Mean number of children by family type Mean S.d.
Overall mean 2.1 1.0
First family 2.0 0.8
Lone parent 1.9 1.0
Stepparent 2.7 1.2
Partner of stepparent 2.5 1.1
Both stepparents 4.2 1.5
Sex
Male 2714 44.3%
Female 3407 55.7%
Employment status
Fulltime working 3264 53.3%
Parttime working 1356 22.2%
Unemployed 195 3.2%
Other not working 1306 21.3%
Highest completed educational level
No education completed 1076 17.6%
CSE 2-5 1201 19.6%
O levels 2482 40.5%
A levels 604 9.9%
Degree 684 11.2%
Unknown 74 1.2%
Social class
Unskilled 280 4.6%
Partly skilled 1046 17.1%
Skilled manual 1137 18.6%
Skilled non-manual 1388 22.7%
Managerial / technical 1642 26.8%
Professional 212 3.5%
Unknown 416 6.8%
Domestic tension age 7
No 5194 84.9%
Yes 255 4.2%
Unknown 672 11.0%
School abilities age 7
Most abilities below average 4181 68.3%
Most abilities on or above average 1331 21.7%
Unknown 609 9.9%
Financial hardship age 7
17

No 5071 82.8%
Yes 378 6.2%
Unknown 672 11.0%
Living with natural mother age 16
No 245 4.0%
Yes 5876 96.0%
Father interested in child's education age 16
No 781 12.8%
Yes 3022 49.4%
Unknown 2318 37.9%
School abilities at age 16
More than half below CSE 1/O levels 1189 19.4%
Half or more above CSE 1/O levels 1989 32.5%
Unknown 2943 48.1%
18