SIBLING CONSTELLATION OF THE FAMILY OF ORIGIN AND

SIBLING CONSTELLATION OF THE FAMILY OF ORIGIN AND DIVORCE RISK

IN ADULTHOOD

KIERON BARCLAY AND SEYMOUR SPILERMAN

ABSTRACT. This study examines whether the sibling constellation of the family of origin of

individuals is related to the risk of divorce later in life. The duplication theorem has argued

that individuals develop role preferences based upon their childhood experiences in the sibling

group of origin, and particularly in response to their rank in the sibling group, and the gender

composition of the sibling group. In turn, the relative compatibility of pairings of these role

preferences between a husband and wife should inﬂuence the stability of their relationship. We

use Swedish administrative register data and use event-history analysis examine divorce in ﬁrst

marriages over the period 1990 to 2012 amongst individuals born 1950 to 1994. The unit of

analysis is the couple, and we adjust for a range of factors that previous research has shown to

be associated with divorce risk. We ﬁnd that pairings of an only born husband and an only born

wife have the highest risk of divorce. Marriages where one of the partners was an only child

have the second highest risk of divorce. Different combinations of pairings where both husband

and wife come from multi-child sibling groups do not vary in terms of the risk of divorce. These

results provide partial, but not full, support for the predictions of the duplication theorem.

2 KIERON BARCLAY AND SEYMOUR SPILERMAN

INTRODUCTION

The characteristics of the family of origin have been found to extend their inﬂuence far into

adulthood. Studies have shown that family size has an effect on educational attainment (Black

et al., 2005), intelligence (Blake, 1989), as well as mortality (Smith et al., 2009). Other re-

search has found that, independent of family size, birth order has a causal effect on educational

attainment (Black et al., 2005), intelligence (Bjerkedal et al., 2007), and mortality (Barclay and

Kolk, 2013). A great deal of research has also speculated about the extent to which the family

of origin, and in particular birth order, can affect personality (Toman, 1961; Sulloway, 1996).

However, only a small handful of studies have examined whether speciﬁc pairings of these fam-

ily of origin characteristics in marriage is associated with the risk of marital dissolution (Toman,

1962; Weller et al., 1974; Vos and Hayden, 1985). This study uses Swedish population register

data to examine how birth order pairings, family size pairings, and gender composition pairings

inﬂuence divorce risk in ﬁrst marriages between 1990 and 2012.

One theory that has been proposed to explain why birth order pairings in a romantic re-

lationship should matter is the duplication theorem (Toman, 1961). The duplication theorem

describes how early life experiences, in terms of an individual’s own birth order and the gender

composition of the sibling group of origin, help to determine role preferences (Toman, 1971). In

childhood the experience of birth order accustoms ﬁrst born individuals to feelings of seniority,

and later borns to expect leadership and protection from their older peer (Toman, 1971). Gender

composition, in turn, helps to shape the degree to which an individual is accustomed to spend-

ing time with an individual of the opposite sex, and the degree to which they are accustomed to

leading, or being led by, an individual of the opposite sex (Toman, 1971). The dual importance

of relative rank in the sibling group of origin, and the opportunity to interact with members

of the opposite sex, determines the likelihood of relationship stability in Toman’s duplication

theorem.

If we take a heterosexual marriage between two individuals, both from two-child sibling

groups, there are eight different sibling positions. There can be the older brother of a brother

– b(b); the older brother of a sister – b(s); the younger brother of a brother – (b)b; the younger

brother of a sister – (s)b; the older sister of a sister – s(s); the older sister of a brother – s(b); the

younger sister of a sister – (s)s; and ﬁnally, the younger sister of a brother – (b)s. The duplication

theorem states that complementarity between the experiences in the sibling group of origin and

the future romantic relationship will imply the greatest chance of success in that relationship

(Toman, 1971). So, for example, a man who was the older brother of a younger sister [b(s)],

would be most compatible with the younger sister of an older brother [(b)s]. The duplication

theorem does not argue that a man from an all-male sibling group cannot form or maintain a

relationship with a woman, but does argue that it will be less stable than these complementary

conﬁgurations.

Toman (1993) developed very speciﬁc predictions about how marital stability would vary ac-

cording to what experience from the sibling group of origin each member of the couple brought

to the marriage. These can be seen in Table 1, which is reproduced directly from Toman (1993,

pages 88-89). These explicit ranked predictions make it easy to test whether there is empiri-

cal support for the duplication theorem. Toman argued that individuals from one child sibling

groups, ‘only borns’, should have the highest risk of relationship instability, primarily due to a

SIBLING CONSTELLATION OF THE FAMILY OF ORIGIN AND DIVORCE RISK IN ADULTHOOD 3

TABLE 1. Predictions of Duplication Theorem, reproduced from Toman (1993,

pages 88-89)

Degrees of Complementarity Basic Types Other Examples

1a. Neither partner has a rank or sex b(s)/(b)s b(ss)/(b)s

conﬂict with the other (s)b/s(b) (ss)b/s(bbb)

1b. Each partner has a least one sibling b(sb)/(b)s

relationship that is neither in a rank (s)b(s)/s(bbs)

nor in a sex conﬂict with at least one (bs)b(s)/(b)s(bs)

sibling relationship of the partner

2a. Neither partner has a rank conﬂict (or b(s)/(s)s b(ss)/(s)s

only a partial one) and only one of the (s)b/s(s) (b)b/s(bbb)

partnes has a sex conﬂict b(b)/(b)s b(s)/(s)s(s)

(b)b/s(b) b(b)/(b)s(b)

2b. Both partners have either a complete b(s)/s(b) (ss)b/(bbb)s

(or partial) rank conﬂict, but no sex (s)b/(b)s (b)b(s)/s(bb)

conﬂict, or both have a sex conﬂict, b(b)/(s)s (b)b(b)/s(ss)

but no rank conﬂict (or only a partial (b)b/s(s) (b)b(b)/(s)(ss)

one)

2c. Both partners have a complete rank b(s)/s(s) (ss)b/(ss)s

conﬂict and one of the partners also has (s)b/(s)s b(bb)/s(bbb)

a sex conﬂict b(b)/s(b)

(b)b/(b)s

2d. Both partners have a complete rank b(b)/s(s) b(bb)/s(s)

and sex conﬂict. (b)b/(s)s (bb)b/(sss)s

3a. One partner is an only child, the other b(s)/s (s)b(sb)/s

has at least one sibling of the opposite b/s(b) b/s(bbb)

sex (s)b/s (s)b(s)/s

b/(b)s b/(b)s(s)

3b. One partner is an only child, the other has b(b)/s b(bb)/s

only (one or more) siblings of the same b/s(s) b/(s)s(s)

sex as himself (b)b/s

b/(s)s

3c. Both partners are only children b/s

general lack of peer interaction in the home environment (Toman, 1993). There is some sup-

port for this theory amongst young children, with research ﬁnding that teachers rate only borns

as having poorer social skills than their peers (Downey and Condron, 2004). Nevertheless, a

meta-analysis of studies examining the relationship between only-born status and personality in

adolescence indicates that there are no genuine personality differences between only borns and

their peers from larger sibling groups (Polit and Falbo, 1987).

Despite this evidence, more recent research using the one-child policy in China as a natural

experiment suggests that growing up as an only-born has a causal effect on trustworthiness,

with only-borns being less trusting, and less trustworthy (Cameron et al., 2013). Furthermore,

4 KIERON BARCLAY AND SEYMOUR SPILERMAN

the same study found only-borns to be less optimistic, less conscientious, and to have higher

scores on neuroticism (Cameron et al., 2013). While the Swedish context is clearly substantially

different from the Chinese context, this rare opportunity to study an exogenous inﬂuence on a

decision made by parents that is usually endogenous, suggests that growing up as an only-

born can have a substantial impact upon behaviour, and perhaps also personality. It can be

imagined that trust and trustworthiness may be important factors for marital stability. If these

results do indeed demonstrate the causal impact of growing up as an only-born, and if they are

generalisable, then this would predict that only-borns would have higher divorce risks.

Aside from role preferences, previous research has also suggested that birth order may in-

ﬂuence personality. Sulloway (1996) argued that competition for parental investment within

the family caused siblings to adapt their behaviour so as to occupy particular niches within the

family. First borns and only borns were the most likely to be aligned with their parents as they

experienced a period of time where they were the only child within the home, and where they

were the sole focus of parental care. This led them, on average, to identify with power and

authority, and thereby to become more conservative and socially dominant than their later born

siblings. Later born children, ﬁnding that they were disadvantaged from the very beginning

in terms of size and strength, were naturally more inclined to develop a personality that was

questioning of authority. Furthermore, in the scramble for parental investment, later borns were

forced to become more creative, original and follow risks so as to achieve that goal.

Only a handful of studies have evaluated the relationship between birth order pairings in mar-

riage and marital satisfaction. A study by Kemper (1966) also found that men with a younger

sister married to women with an older brother reported higher marital satisfaction, consistent

with Toman’s duplication theorem. Weller et al. (1974) examined how birth order pairings

in marriage was associated with happiness in the marriage amongst several hundred Israeli

women. They found that marriages where either one or both of the spouses were only borns

reported the highest rates of marital dissatisfaction. Weller et al. (1974) also found that the

respondents who reported the highest level of marital satisfaction were rank and gender compo-

sition complementary, consistent with the predictions of Toman’s duplication theorem. Finally,

Toman (1971) reported that there was a higher proportion of low complementary pairings than

high complementary pairings amongst his sample of divorced couples. Other studies, however,

have found no relationship between birth order pairings and marital adjustment (Levinger and

Sonnheim, 1965; Birtchnell and Mayhew, 1977; Vos and Hayden, 1985).

DATA AND METHODS

The data used for this study is Swedish administrative register data, and we focus ﬁrst mar-

riages among individuals born between 1950 and 1994. The unit of analysis is the couple. The

period of time over which we observe marriage and divorce is 1990 to 2012. Although the mar-

riage and divorce register is available from 1968-2012, we chose to focus on the period 1990

to 2012 because this period is where the data available for the control variables is at its highest

quality. This means that we study all ﬁrst marriages that occur from 1990 to 2012. The rationale

behind choosing cohorts born 1950 to 1994 is that we wanted to be sure that the marriage that

we study is the ﬁrst marriage for both individuals in the relationship. Since the marriage register

begins in 1968, those born in 1950 will have been 18 at that point in time, and so we can be

SIBLING CONSTELLATION OF THE FAMILY OF ORIGIN AND DIVORCE RISK IN ADULTHOOD 5

sure that the registers contain full and accurate information on their civil status over the entire

life course.

200

400

600

Frequency

0 100 200 300

Time to Divorce (Months)

FIGURE 1. Time to Divorce, in Months.

We study time to divorce using piece-wise exponential event-history analysis. In our analy-

ses we adjust for a number of covariates that previous reviews of the literature (Lyngstad and

Jalovaara, 2010; H

ark

onen, 2014) have shown to be important factors in terms of divorce risk.

These include: year of marriage; husband’s age at the time of marriage; spousal age differ-

ence; whether the husband had children with a woman other than his future spouse before the

marriage; whether the wife had children with a man other than her future spouse before the

marriage; whether the couple had any children before the marriage; the number of children the

couple have (time-varying); the husband’s educational level (time-varying); and, the difference

in the educational level between the husband and wife (time-varying).

RESULTS

Table 2 shows the results for divorce risk in ﬁrst marriages in the period 1990 to 2012 amongst

individuals born 1950 to 1994. The categorization used for birth order here is ﬁrst, middle, or

last, and we distinguish only borns from individuals from sibling groups with more than one

birth. This analysis adjust for the control variables listed in the data section. Relative risks are

shown in bold if the difference is statistically signiﬁcant at the 95% level. Table 2 shows that

relative to a partnership with two only-born individuals (the reference category is italicized, all

other couple pairings have a lower risk of divorce. It is also clear that in couples where one

of the partners was an only-born, the relative risk is higher than that found in couples where

neither child was an only-born.

6 KIERON BARCLAY AND SEYMOUR SPILERMAN

TABLE 2. Birth Order Pairings in Marriage and Relative Risk of Divorce

in Sweden: First Marriages 1990-2012, Cohorts Born 1950-1994. First

Borns, Middle Borns, and Last Borns come from Multi-child Sibling Groups.

N=470,527, Divorces=75,807

Wife

Only Born First Born Middle Born Last Born

Husband Only Born 1.00 0.81 0.79 0.77

First Born 0.88 0.45 0.45 0.43

Middle Born 0.86 0.42 0.42 0.41

Last Born 0.83 0.42 0.41 0.40

Table 3 shows the results where we test the duplication theorem (Toman, 1993) in marriages

where both couples came from a one- or two-child sibling group. Again, pairings where both

individuals were only-borns are the reference category. The clearest pattern evident is similar to

that seen in table 2, with all marriage pairings having a lower relative risk of divorce in compar-

ison to a partnership between two only-born individuals. Table 3 also shows that couples where

at least one of the partners was an only-born have a higher risk than those with another sib-

ling. This is consistent with part of the predictions of the duplication theorem shown in table 1,

which stated that couples with two only-born individuals should have the highest relationship

instability, followed by couples where at least one of the pair was an only-born. However, there

is no support in the results in table 3 for Toman’s distinction between high and medium com-

plementary types of marital pairings. That is, to give an example, the older brother of a younger

sister married to the younger sister of an older brother [b(s)/(b)s], have a divorce risk similar to

that of other couples made up of individuals who are both from sibling groups with more than

one child.

Further indication that there is no signiﬁcant difference between high and medium comple-

mentary pairings comes from the results shown in table 4. These results are based upon models

where both individuals in the partnership came from a two-child sibling group. In this case the

reference category is the older brother of a younger sister married to the younger sister of an

older brother [b(s)/(b)s]. As can be seen, there are few statistically signiﬁcant differences, and

the pattern of relative risks do not consistently point in the direction predicted by that of the du-

plication theorem. For example, there is an elevated risk of divorce when a man with a younger

sister [b(s)] is married to a sister with a younger brother [s(b)] or sister [s(s)], which would be

consistent with a lack of rank complementarity, but the difference is not statistically signiﬁcant

for the same pairings if the man has a younger brother rather than a younger sister [b(b)]. More

generally though, it can be seen that pairings between two ﬁrst borns have an elevated relative

risk of divorce, and pairings between two second-borns have a lower relative risk of divorce.

Table 5 shows the results for couples where both partners came from a sibling group with

between 1 and 3 children. Broadly speaking, the clearest result is one that was already evident

in tables 2 and 3, which is that marriages where both partners were only-borns have the highest

risk of divorce. The group with the second highest risk of divorce are those couples where at

least one of the partners was an only-born. It also appears that the risk of divorce is higher

when both partners were ﬁrst born in comparison to when either partner was last born, or both

SIBLING CONSTELLATION OF THE FAMILY OF ORIGIN AND DIVORCE RISK IN ADULTHOOD 7

TABLE 3. Birth Order and Gender Composition Pairings in Marriage and Rela-

tive Risk of Divorce in Sweden: First Marriages 1990-2012, Cohorts Born 1950-

1994. Set Sizes 1 and 2. N=171,243, Divorces=28,123

Wife

s s(b) s(s) (b)s (s)s

Husband b 1.00 0.83 0.81 0.80 0.79

b(b) 0.85 0.69 0.69 0.65 0.65

b(s) 0.92 0.76 0.72 0.65 0.67

(b)b 0.83 0.68 0.64 0.62 0.60

(s)b 0.80 0.68 0.62 0.63 0.65

TABLE 4. Birth Order and Gender Composition Pairings in Marriage and Rela-

tive Risk of Divorce in Sweden: First Marriages 1990-2012, Cohorts Born 1950-

1994. Set Sizes 2. N=100,601, Divorces=14,597

Wife

s(b) s(s) (b)s (s)s

Husband b(b) 1.06 1.06 0.99 0.99

b(s) 1.16 1.09 1.00 1.03

(b)b 1.05 0.98 0.94 0.92

(s)b 1.04 0.94 0.96 0.99

partners werer middle born. Table 5 presents many different coefﬁcients, and so it may be

useful to present the results from table 5 alongside Toman’s predictions. These are shown in

table 6 for the basic types of pairings that Toman makes predictions for. These results show that

the most important result is that two only-borns have the highest risk of divorce, followed by

partnerships where at least one of the partners was an only-born. There are no real differences,

however, between couples where neither individual was an only-born.

TABLE 5. Birth Order and Gender Composition Pairings in Marriage and Relative Risk of Divorce in Sweden: First

Marriages 1990-2012, Cohorts Born 1950-1994. Set Sizes 1 to 3. N=359,100, Divorces=56,763

Wife

s s(b) s(s) s(ss) s(sb) s(b[b/s]) (b)s(b) (s)s(s) (s)s(b) (b)s(s) (b)s (s)s (ss)s (bs)s ([b/s]b)s

Husband b 1.00 0.82 0.81 0.72 0.80 0.80 0.76 0.74 0.78 0.79 0.80 0.78 0.69 0.73 0.77

b(b) 0.85 0.69 0.69 0.61 0.67 0.69 0.71 0.65 0.67 0.69 0.65 0.65 0.57 0.67 0.66

b(s) 0.92 0.76 0.72 0.65 0.66 0.68 0.68 0.57 0.67 0.64 0.65 0.67 0.57 0.73 0.64

b(bb) 0.81 0.73 0.60 0.71 0.63 0.53 0.63 0.72 0.61 0.54 0.66 0.70 0.47 0.59 0.60

b(bs) 0.81 0.67 0.74 0.48 0.39 0.67 0.73 0.68 0.60 0.44 0.55 0.69 0.70 0.44 0.57

b(s[b/s]) 0.79 0.69 0.61 0.66 0.66 0.65 0.68 0.60 0.73 0.73 0.67 0.65 0.70 0.83 0.68

(b)b(b) 0.78 0.76 0.67 0.43 0.77 0.65 0.57 0.63 0.62 0.66 0.57 0.52 0.44 0.67 0.64

(s)b(s) 0.80 0.64 0.65 0.62 0.71 0.55 0.67 0.59 0.90 0.79 0.55 0.71 0.64 0.75 0.57

(s)b(b) 0.76 0.64 0.63 0.68 0.61 0.52 0.62 0.55 0.42 0.61 0.61 0.68 0.63 0.64 0.56

(b)b(s) 0.86 0.66 0.75 0.51 0.69 0.66 0.57 0.55 0.63 0.59 0.66 0.62 0.43 0.56 0.55

(b)b 0.83 0.69 0.64 0.59 0.61 0.62 0.58 0.61 0.59 0.62 0.61 0.60 0.55 0.58 0.60

(s)b 0.80 0.68 0.62 0.60 0.67 0.64 0.58 0.64 0.45 0.62 0.63 0.65 0.61 0.64 0.61

(bb)b 0.82 0.66 0.64 0.40 0.66 0.66 0.48 0.63 0.46 0.76 0.61 0.73 0.46 0.54 0.50

(sb)b 0.77 0.67 0.61 0.48 0.58 0.62 0.58 0.57 0.72 0.64 0.63 0.76 0.50 0.58 0.54

([b/s]s)b 0.78 0.66 0.65 0.50 0.59 0.49 0.70 0.59 0.58 0.46 0.57 0.63 0.65 0.49 0.56

Source: Swedish administrative register data, compiled by the author.

SIBLING CONSTELLATION OF THE FAMILY OF ORIGIN AND DIVORCE RISK IN ADULTHOOD 9

TABLE 6. Duplication Theorem: Basic Types and Results

Degrees of Complementarity Basic Types RR SE 95% CI

1a. Neither partner has a rank or sex b(s)/(b)s 0.76 0.03 0.70-0.81

conﬂict with the other (s)b/s(b) 0.68 0.03 0.63-0.73

1b. Each partner has a least one sibling

relationship that is neither in a rank

nor in a sex conﬂict with at least one

sibling relationship of the partner

2a. Neither partner has a rank conﬂict (or b(s)/(s)s 0.67 0.03 0.62-0.73

only a partial one) and only one of the (s)b/s(s) 0.62 0.03 0.57-0.67

partnes has a sex conﬂict b(b)/(b)s 0.65 0.03 0.60-0.70

(b)b/s(b) 0.69 0.03 0.64-0.74

2b. Bother partners have either a complete b(s)/s(b) 0.76 0.03 0.70-0.81

(or partial) rank conﬂict, but no sex (s)b/(b)s 0.63 0.02 0.58-0.68

conﬂict, or both have a sex conﬂict, b(b)/(s)s 0.65 0.03 0.60-0.70

but no rank conﬂict (or only a partial (b)b/s(s) 0.64 0.03 0.59-0.69

one)

2c. Both partners have a complete rank b(s)/s(s) 0.72 0.03 0.66-0.77

conﬂict and one of the partners also has (s)b/(s)s 0.65 0.03 0.60-0.70

a sex conﬂict b(b)/s(b) 0.69 0.03 0.64-0.74

(b)b/(b)s 0.61 0.02 0.57-0.66

2d. Both partners have a complete rank b(b)/s(s) 0.69 0.03 0.64-0.75

and sex conﬂict. (b)b/(s)s 0.60 0.02 0.55-0.65

3a. One partner is an only child, the other b(s)/s 0.92 0.03 0.87-0.98

has at least one sibling of the opposite b/s(b) 0.82 0.03 0.77-0.88

sex (s)b/s 0.80 0.03 0.75-0.86

b/(b)s 0.80 0.03 0.75-0.86

3b. One partner is an only child, the other has b(b)/s 0.85 0.03 0.80-0.91

only (one or more) siblings of the same b/s(s) 0.81 0.03 0.75-0.86

sex as himself (b)b/s 0.83 0.03 0.78-0.89

b/(s)s 0.78 0.03 0.73-0.84

3c. Both partners are only children b/s 1.00

10 KIERON BARCLAY AND SEYMOUR SPILERMAN

DISCUSSION

This study has found that the sibling constellation in the family of origin has an association

with divorce risk in adulthood. The clearest result is that a partnership between two only-born

individuals has the highest risk of divorce. Furthermore, marriages where at least one of the

partners was an only-born have a lower risk that partnerships with two only-borns, but a higher

risk than partnerships without any only-born individuals. These results provide partial support

for the duplication theorem (Toman, 1993). Toman (1993) had predicted that partnerships with

two only-born individuals would have the highest risk of marital instability, and partnerships

with at least one only-born individual the next highest risk. However, his predictions about

the relative importance of rank within, and the gender composition of, the sibling group of

origin are not borne out by our results. Although we did not ﬁnd complete support for the

duplication theorem, Toman’s prediction are highly detailed, and it should be noted that a theory

is implicitly easier to disprove when it makes speciﬁc predictions (Western, 2001).

The vast majority of birth order research has adopted the approach used in this study, com-

paring individuals of different birth order across different families. However, this approach has

proved controversial, with a number of researchers pointing out that almost all theories that

predict a relationship between birth order and later life outcomes describe within-family pro-

cesses (Rodgers, 2001). Furthermore, because families differ in many ways that are difﬁcult

to measure, studies of birth order that compare individuals across different families suffer from

an unknown degree of residual confounding that are likely to bias the results. We have tried

to address this issue by adjusting for a number of factors that previous research has shown are

correlated with divorce risk. Nevertheless, there are important potential sources of residual

confounding that remain. For example, studies show that women are more likely to divorce if

they have only one child (Andersson, 1997). This would mean that only-borns would be more

likely to experience the divorce of their parents. Research on the inter-generational transmission

of divorce suggests that this would increase the probability that they would themselves would

divorce (Kiernan and Cherlin, 1999). We will attempt to address this issue by adjusting for

whether parents had any further children outside of the ﬁrst relationship.

In addition to the intergenerational transmission of divorce, it should be noted that there is a

strong two-child norm in Sweden (Andersson, 1999). It is possible that the parents who choose

to have only one child are selected in some way that might be associated with an individual

growing up as an only-born having a less happy childhood. This might results in an only-born

individual being more likely to divorce in adulthood. In additional analyses it will also be

possible to examine only partnerships where neither partner had parents who had children with

more than one partner. This would be suggestive of union stability, and it would be possible to

examine whether the pattern of results observed in our analyses so far would be consistent for

individuals who grew up in more stable home environments. While this is not a perfect measure

of union stability, cohabitation is Sweden is very common, and just measuring whether the

parents of our index couples were divorced would not be able to fully capture parental union

stability; indeed, many of them would never have been married in the ﬁrst place (Trost, 1975).

As between-family comparisons require careful adjustment to account for confounding, some

researchers have turned to sibling ﬁxed effects, or within-family comparisons, to deal with

this issue. We have also estimated within-family comparisons, not presented here, but these

can only be done for sibling groups with at least two same-sex siblings. This means that we

SIBLING CONSTELLATION OF THE FAMILY OF ORIGIN AND DIVORCE RISK IN ADULTHOOD 11

cannot analyse the clearest pattern in our data, regarding the association of divorce with the

presence of an only-born in the relationship, using a within-family comparison. The sample

size is reduced dramatically in these within-family comparisons. The reason for this is that

a within-family comparison requires variance in the outcome within the sibling group. This

means that there must be at least two brothers or sisters in a sibling group, at least two of them

should have married, and at least one of them must have divorced. Furthermore, if we take

a sibling group with two brothers, they both must be married to an individual of the opposite

sex with the same birth order (e.g. they must both be married to a ﬁrst-born woman). The

use of a within-family comparison clearly creates very speciﬁc selection criteria. Nevertheless,

the results for the within-family comparisons of multi-child sibling groups indicate that there is

no clear association between birth order and divorce risk, consistent with our between-family

results for marriages between individuals from sibling groups with at least two children

An additional factor to consider is that we are studying divorce in Sweden. While there has

been an increasing revival in the popularity of marriage in Sweden in recent years (Ohlsson-

Wijk, 2011), it remains true that cohabitation rates in Sweden are very high in a comparative

sense (Kiernan, 2002). Indeed, Sweden is the only country where the majority of ﬁrst births

occur in a cohabiting rather than a marital union (Thomson and Eriksson, 2013). While study-

ing partnership dissolution in cohabiting unions would make an important contribution to the

current study, at present it is only possible to know if a cohabiting but non-married couple are

living together if they have a child together (Thomson and Eriksson, 2013). This means that we

would have a very particular sample of cohabiting unions, presumably pre-selected on a degree

of union stability as they will have chosen to have a child together. Nevertheless, it will also be

possible to examine these cohabiting couples with children in the coming months.

12 KIERON BARCLAY AND SEYMOUR SPILERMAN

REFERENCES

Andersson, G. (1997), ‘The impact of children on divorce risks of swedish women’, European

Journal of Population/Revue europ

eenne de D

emographie 13(2), 109–145.

Andersson, G. (1999), ‘Childbearing trends in sweden, 1961-1997’, European Journal of Pop-

ulation 15(1), 1–24.

Barclay, K. J. and Kolk, M. (2013), ‘Birth order and mortality: A population-based cohort

study’, Stockholm Research Reports in Demography 14, 1–30.

Birtchnell, J. and Mayhew, J. (1977), ‘Toman’s theory: Tested for mate selection and friendship

formation.’, Journal of Individual psychology 33(1), 18–36.

Bjerkedal, T., Kristensen, P., Skjeret, G. A. and Brevik, J. I. (2007), ‘Intelligence test scores and

birth order among young norwegian men (conscripts) analyzed within and between families’,

Intelligence 35, 503–514.

Black, S. E., Devereux, P. J. and Salvanes, K. G. (2005), ‘The more the merrier? the effect

of family size and birth order on children’s education’, Quarterly Journal of Economics

120(2), 669–700.

Blake, J. (1989), Family Size & Achievement, Vol. 3, University of California Press, Berkeley.

Cameron, L., Erkal, N., Gangadharan, L. and Meng, X. (2013), ‘Little emperors: Behavioral

impacts of china’s one-child policy’, Science 339(6122), 953–957.

Downey, D. B. and Condron, D. J. (2004), ‘Playing well with others in kindergarten: The beneﬁt

of siblings at home’, Journal of Marriage and Family 66(2), 333–350.

ark

onen, J. (2014), Divorce: Trends, patterns, causes, consequences., in J. K. Treas, J. Scott

and M. Richards, eds, ‘The Wiley-Blackwell Companion to the Sociology of Families’, John

Wiley & Sons, Chichester, pp. 303–322.

Kemper, T. D. (1966), ‘Mate selection and marital satisfaction according to sibling type of

husband and wife’, Journal of Marriage and the Family 28(3), 346–349.

Kiernan, K. (2002), Cohabitation in western europe: Trends, issues, and implications., in ‘Just

Living Together: Implications of Cohabitation for Children, Families, and Social Policy, Oct,

2000, Pennsylvania State U, PA, US; Based on the presentations and discussions from the

aforementioned national symposium.’, Lawrence Erlbaum Associates Publishers.

Kiernan, K. E. and Cherlin, A. J. (1999), ‘Parental divorce and partnership dissolution in adult-

hood: Evidence from a british cohort study’, Population Studies 53(1), 39–48.

Levinger, G. and Sonnheim, M. (1965), ‘Complementarity in marital adjustment: Reconsider-

ing toman’s family constellation hypothesis.’, Journal of Individual Psychology 21(2), 137–

145.

Lyngstad, T. H. and Jalovaara, M. (2010), ‘A review of the antecedents of union dissolution’,

Demographic Research 23(10), 257–292.

Ohlsson-Wijk, S. (2011), ‘Sweden’s marriage revival: An analysis of the new-millennium

switch from long-term decline to increasing popularity’, Population Studies 65(2), 183–200.

Polit, D. F. and Falbo, T. (1987), ‘Only children and personality development: A quantitative

review’, Journal of Marriage and the Family 49(2), 309–325.

Rodgers, J. L. (2001), ‘What causes birth order-intelligence patterns? the admixture hypothesis,

revived’, American Psychologist 56, 505–510.

Smith, K. R., Mineau, G. P., Garibotti, G. and Kerber, R. (2009), ‘Effects of childhood and

middle-adulthood family conditions on later-life mortality: evidence from the utah population

SIBLING CONSTELLATION OF THE FAMILY OF ORIGIN AND DIVORCE RISK IN ADULTHOOD 13

database, 1850-2002’, Social Science and Medicine 68, 1649–1658.

Sulloway, F. J. (1996), Born to Rebel: Birth Order, Family Dynamics, and Creative Lives, Little,

Brown and Company, London.

Thomson, E. and Eriksson, H. (2013), ‘Register-based estimates of parents’ coresidence in

sweden, 1969-2007’, Demographic Research 29(42), 1153–1186.

Toman, W. (1961), Family Constellation: Its Effects on Personality and Social Behavior,

Springer, New York.

Toman, W. (1962), ‘Family constellations of the partners in divorced and married couples’,

Journal of Individual Psychology 18(1), 48–51.

Toman, W. (1971), ‘The duplication theorem of social relationships as tested in the general

population.’, Psychological Review 78(5), 380.

Toman, W. (1993), Family Constellation: Its Effects on Personality and Social Behavior, 4th

edn, Springer, New York.

Trost, J. (1975), ‘Married and unmarried cohabitation: The case of sweden, with some compar-

isons’, Journal of Marriage and the Family 37(3), 677–682.

Vos, C. J. V. and Hayden, D. J. (1985), ‘Birth-order complementarity and marital adjustment’,

Journal of Marriage and the Family 47(3), 723–727.

Weller, L., Natan, O. and Hazi, O. (1974), ‘Birth order and marital bliss in israel’, Journal of

Marriage and the Family 36(4), 794–797.

Western, B. (2001), ‘Bayesian thinking about macrosociology’, American Journal of Sociology

107(2), 353–378.