Shorter birth intervals between siblings are associated with increased risk of parental divorce.

Birth intervals are a crucial component of fertility behaviour and family planning. Short birth intervals are associated-although not necessarily causally-with negative health-related outcomes, but less is known about their associations with family functioning. Here, the associations between birth intervals and marital stability were investigated by Cox regression using a nationally representative, register-based sample of individuals with two (N = 42,481) or three (N = 22,514) children from contemporary Finland (observation period 1972-2009). Shorter interbirth intervals were associated with an increased risk of parental divorce over a ten-year follow-up. Individuals with birth intervals of up to 1.5 years had 24-49 per cent higher divorce risk compared to individuals whose children were born more than 4 years apart. The pattern was similar in all socioeconomic groups and among individuals with earlier and later entry to parenthood. Our results add to the growing body of research showing associations between short birth intervals and negative outcomes in health and family functioning.

Introduction

The spacing between sibling births in a family is a crucial component of fertility behaviour and family planning. In Western high-income populations, birth intervals seem to have shortened during the late 20th and early 21st century [1-3]. A large proportion of birth intervals in these populations are now shorter than 2.5 years, which is shorter than what is considered to be the human species-typical birth interval of three to four years, found in hunter–gatherer societies [4]. Women’s increased educational levels and labour participation and increased age at first birth in Western countries motivate closer spacing of children [2,5,6]. Further, some family benefits have shown to encourage tight sibling spacing in welfare states [7-10]. Birth intervals have been extensively studied from an epidemiological viewpoint, but less is known about their effects for family functioning. In the current paper we investigate whether sibling spacing is associated with the risk of parental divorce.

Very short (less than 18–27 months) and very long (typically over 54–59 months) birth intervals are associated with perinatal health problems for both the mother and the child [11,12]. Long-term associations between shorter intervals and negative health outcomes have also been reported. For example, a few studies have found an association between short birth intervals and reduced longevity in the parents [13-15]. Further, closely spaced siblings have been found to have more behavioural [16] and mental health problems [17-20]. The main mechanism behind the association between short birth intervals and poorer maternal and infant health is hypothesized to be that the mother’s body has not had time to fully recover between pregnancies [21]. The risks are most evident in developing countries with limited maternal health care [11,12,21], while contradicting findings have been reported in developed Western countries with advanced health care systems and good nutritional status [22-27]. Importantly, many studies in high-income countries have found that while shorter birth intervals are associated with poorer child health, the association may not be causal [23,24,26,but see 27 for opposite findings].

Much less is known about the associations between birth intervals and sibling relations, parenting, and couple relations. Allegedly, age difference between siblings may affect the psychosocial functioning of family members and family relations, including the marital relationship. Being a parent, especially of young children, is demanding and may affect the wellbeing of parents and the quality of the marital relationship negatively [28,29]. On one hand, closely spaced children have been hypothesized to relief parenting burden because children of same age are assumed to provide each other companionship and because the demands of raising small children are concentrated into a shorter period [see 30 for an early contemplation on the matter]. On the other hand, having closely spaced children may be associated with an increased parenting burden because of the elevated need to provide intensive care simultaneously for many young children [30].

If shorter birth intervals between children were associated with higher parenting stress, we would expect to see an association between shorter intervals and higher likelihood of divorce. Higher parental stress in general [31] and increased parenting stress among parents of children with special needs [32,33] are associated with a higher likelihood of divorce. Giving birth to twins has also been associated with an increased risk of divorce compared to mothers with no twins [34], suggesting that bringing up children close in age may influence divorce risk (although this association may also result from an unexpected increase in the number of children rather than a short birth interval between two children). In addition to infants’ need for intensive parental investment, increased parenting stress may be induced by siblings with small age differences having more behavioural [16] and mental health problems [17-20]. Closely spaced siblings also seem to experience an increased risk of serious injury [35] and parental maltreatment [16,36] in childhood, and tend to have more conflictual relationships [37-39].

The evolutionary life history theory provides an additional explanation for the potential association between short birth intervals and higher risk of divorce. According to the theory, individuals vary in their timing, pace, and amount of reproduction in accordance with environmental conditions [40-42]. In resource scarce and dangerous environments with a lower life expectancy, it is more optimal to invest in current reproduction rather than future reproduction because the future is uncertain, and in quantity rather than quality of offspring because of lack of resources [40-44]. These features of faster reproductive strategy lead to shorter birth intervals between children [42]. Faster life history strategy also entails behaviours that destabilize romantic relationships, including decreased parental investment and increased mating effort, i.e., seeking opportunities with new partners [40,43,44]. Empirically, fast life history has been shown to associated with, for example, more negative parenting attitudes, risky sexual behaviour and higher likelihood of union dissolution [45-47]. Thus, based on life history theory, short birth intervals and increased risk of divorce may correlate even in the absence of a causal, e.g., stress-related, pathway between the two.

Even though absolute poverty is rare in today’s high-income societies, relative socioeconomic inequality has been shown to be associated with earlier reproduction, i.e., faster life history, in such societies in many studies. For example, living in poorer neighbourhoods with higher extrinsic mortality, unpredictability, and higher incidence of violent crime is associated with lower age at menarche and lower age at first birth, and more sexual and risky behaviour at young adulthood [e.g., 48–54]. These studies imply that even in prosperous environments, some psychological mechanisms may be sensitive to environmental cues of relative disadvantage and subtly steer our behaviour.

To our knowledge, no previous studies have examined if the length of birth intervals between singleton siblings is associated with parental divorce. Here, we use a nationally representative, register-based dataset of over 60,000 individuals from the late 20th century Finland, and hypothesize that shorter birth intervals of singletons are associated with an increased risk of parental divorce. To take into account potential differences in life history strategies, we examine whether the association is moderated by differences in individuals’ adulthood socioeconomic status (SES) and by individuals’ age at first reproduction (AFR). We consider lower adulthood SES (at age 35) to be indicative of a faster life history, because lower SES is strongly related to shorter life expectancy and shorter healthy life expectancy in particular [55-57], younger age at first birth [5], and a higher risk of divorce [58]. Earlier entry to parenthood also is, by definition, an indication of faster life history strategy [40]. Thus, among individuals with low SES and younger AFR, short birth intervals and higher divorce risk could correlate even in the absence of any causality between the two. In contrast, we assume people with a high adulthood SES and later entry to parenthood to be on a slower life history path, and an association between short birth intervals and higher divorce risk among them could be indicative of other mechanisms, such as parental stress due to closely spaced children.

Methods

Participants

We use register data from Finland, a secularized Nordic welfare state, between years 1972 and 2009. During the study period, the total fertility rate in Finland was between 1.7 and 1.8, and the median number of children among mothers two. The average interbirth interval between first and second births decreased by almost eight months from 3 years 10 months in women born in 1955 to 3 years 2 months in women born in 1975 [2]. In 1987, no-fault divorce legislation was introduced, after which divorce could be initiated without a named reason (e.g., adultery or violence). Consequently, the divorce rate increased considerably and stabilized on a higher level than before 1987. Currently almost 40 per cent of marriages end in divorce in Finland [59]. Further, cohabiting steeply increased during the study period [60], and nowadays it is common to have children out of wedlock in Finland [61]. To a large extent, many cohabiting unions can be seen as socially comparable to marriages. Thus, although this study concentrates on divorces, i.e., dissolving marriages, we decided to include also individuals who got married after the birth of the first child in order to keep the sample as representative as possible.

We use the FINNUNION data set, which is an 11-per cent sample of individuals from cohorts born in 1930–1990 residing in Finland between 1970 and 2010, drawn from the population register by Statistics Finland. The Finnish legislation does not require ethical approval for doing secondary data analysis on existing and anonymized register data [62].

The longitudinal dataset covers complete marital and childbearing histories of the sampled individuals until the end of 2009 (i.e., the data do not contain marital pairs). Finnish register data on divorces is available from January 1972 onwards. We restricted the analyses to cohorts born in 1955 to 1979, whose second children were born on January 1972 earliest. Two separate sets of analyses were performed, one on individuals with two, and only two, children with their first married partner (N = 42,481) and another on individuals with three, and only three, children with their first married partner (N = 22,514). Thus, two-child individuals who continued to have a third child were excluded from the analyses regarding individuals with two children, and three-child individuals who continued to have a fourth child were excluded from the analyses regarding individuals with three children. By thus keeping the study samples as homogenous as possible, we aim to control for some of the confounding exogenous variation, such as the parents of larger families having more traditional values, shorter birth intervals, and lower likelihood of divorce [see, e.g., 63]. Families with twins were excluded from the present study.

First marriages were chosen for the analyses to make the analysis sample as homogenous as possible, because divorce risk is higher on later marriages [58]. Participants who were included in the analyses had to be married before, and still married at, the start of the follow-up (i.e., at the birth of the second or third child, depending on the analysis). The descriptive statistics of the present sample are shown in Table 1.

Table 1

Descriptive statistics.

	Women		Men
	2 children	3 children	2 children	3 children
N	22,708	11,257	19,773	9,551
Divorcesa; n (%)	3889 (17)	1441 (13)	3430 (17)	1197 (13)
Censored personsa; n (%)	447 (2)	223 (2)	386 (2)	172 (2)
Average follow-up time in years; mean (SD)	7.68 (3.14)	7.77 (3.08)	7.37 (3.26)	7.48 (3.21)
Birth year;%
1955–1959	30	29	31	32
1960–1964	25	27	26	29
1965–1969	20	22	20	21
1970–1974	14	14	14	13
1975–1979	10	8	9	6
1st birth interval in years; mean (SD)	3.29 (1.96)	2.70 (1.48)	3.20 (1.85)	2.62 (1.38)
≤ 18 months; n (%)	1641 (7)	1267 (11)	1453 (7)	1057 (11)
18.01–24	3688 (16)	2702 (24)	3372 (17)	2337 (24)
24.01–30	3985 (18)	2326 (21)	3593 (18)	2051 (21)
30.01–36	3388 (15)	1715 (15)	3019 (15)	1518 (16)
36.01–42	2500(11)	1061 (9)	2207 (11)	899 (9)
42.01–48	1832 (8)	668(6)	1556 (8)	577 (6)
48.01–54	1403 (6)	452 (4)	1156 (6)	352 (4)
54.01–60	998 (4)	278 (2)	808 (4)	217 (2)
60.01–66	725 (3)	196 (2)	604 (3)	150 (2)
66.01–72	464 (2)	141 (1)	386 (2)	86 (1)
> 72	2084 (9)	451 (4)	1619 (8)	307 (3)
2nd birth interval in years; mean (SD)		4.26 (2.48)		4.15 (2.37)
≤ 18 months; n (%)		534 (5)		467 (5)
18.01–24		1213 (11)		1035 (11)
24.01–30		1167 (10)		1039 (11)
30.01–36		1234 (11)		1083 (11)
36.01–42		1023 (9)		875 (9)
42.01–48		983 (9)		836 (9)
48.01–54		870 (8)		714 (7)
54.01–60		772 (7)		677 (7)
60.01–66		656 (6)		591 (6)
66.01–72		471 (4)		429 (4)
> 72		2334 (21)		1805 (19)
Age at first birth; mean (SD)	26.61 (4.35)	25.00 (3.77)	28.52 (4.41)	27.05 (3.88)
Marriage length at last birth in years; mean (SD)	4.91 (2.56)	8.27 (3.00)	4.82 (2.49)	8.09 (2.88)
Timing of marriage; n (%)
Married before first birth	19045 (84)	8882 (79)	16333 (83)	7470 (78)
Married before second birth	3663 (16)	1609 (14)	3439 (17)	1392 (15)
Married before third birth	-	766 (7)	-	689 (7)
SES at 35; n (%)
Farmer	848 (2)	515 (5)	630 (3)	575 (6)
Entrepreneur /self-employed	995 (4)	594 (5)	1481 (7)	793 (8)
Upper white-collar	3946 (17)	1943 (17)	4289 (22)	2099 (23)
Lower white-collar	9208 (41)	4165 (37)	4018 (19)	1865 (20)
Manual worker	3439 (15)	1771 (16)	6213 (31)	2950 (33)
Student	780 (3)	440 (4)	301 (2)	133 (1)
Unclassified	1787 (8)	1038 (10)	1275 (6)	653 (7)
Missing	2069 (9)	791 (7)	1565 (8)	492 (5)

aDuring 10-year follow-up. SES = Socioeconomic status; SD = standard deviation.

Measures

Register data on the birth dates of the participant and participants’ children (month and year) were used to calculate participant’s age at first birth and interbirth intervals (IBI; one birth interval in individuals with two children and two birth intervals in individuals with three children). For the analyses, interbirth intervals were categorized into 11categories, the upper limit of the first one being 18 months, the second 24 months, the third 30 months, etc. The upper limit of the first category, 18 months, was chosen because shorter birth intervals were very rare and if used as a comparison category, would make the statistical analyses non-robust. Date of marriage was used to calculate marital duration at the start of follow-up (birth of the second or third child, depending on the analysis). If marriage occurred after the first childbirth, marital duration at the last childbirth was calculated from the birth date of the first child plus six months (assumed minimal time for cohabitation before first childbirth). Timing of marriage (1 = Married before first child birth; 2 = Married after first but before second child birth; 3 = Married after second but before third child birth; only applicable in analyses on individuals with three children) was used as an additional covariate to account for the possible differences between unions with and without childbearing out of wedlock. Quadratic marital duration was included in all analyses to take into account the intrinsic curvilinear risk of divorce through time within unions [64].The participant’s birth cohort (in groups of five years) was included as a covariate to account for the increase in divorce rates during the study period, and participant’s sex to account for the different proportions of men and women in different socioeconomic groups. The effect of children’s sex on birth intervals and parental divorce risk was also examined, but no statistically significant associations between these variables were detected; child sex was thus dropped from the final analyses.

Participant’s SES at the age of 35 was classified according to Statistics Finland (1 = Farmer; 2 = Entrepreneur/Self-employed; 3 = Upper white-collar; 4 = Lower white-collar; 5 = Manual worker; 6 = Student; 7 = Retired; 8 = Unclassified; 9 = Missing). SES contains information about individuals’ main type of activity, occupation, occupational status, and industry and correlates quite strongly with yearly income [65]. SES also correlates with education, but provides a more detailed picture of an individual’s current social, cultural, and financial position, since some end up in high occupational positions despite not having a university degree, or in low occupational positions despite being highly educated. Information about SES was missing for 3385 and 1706 participants with two children and three children, respectively. About 39 per cent of people with a missing SES had vocational secondary school as their highest degree, and 38 per cent had lower or higher university degree. The individuals with missing information were coded as an additional category and included in the analyses.

Statistical analyses

The associations between interbirth intervals of children and parental risk of divorce were examined with single-event Cox regression. The dependent variable, i.e., timing of divorces was recorded in months, and follow-up started at the birth of the second child (in analyses on people with two children) or the third child (in analyses on people with three children). The Efron method was used to handle ties [i.e., simultaneous events in the dataset, 66]. The shortest birth interval category (intervals of max. 18 months) was used as the reference category. In models concerning individuals with three children, the first and second birth intervals were examined separately, mutually adjusted, and in interaction, to see if the two intervals had independent, additive, or interactive associations with the risk of divorce. Marital duration and quadratic marital duration at the last childbirth were used as covariates in all models, and thus the hazard ratios can be interpreted as the proportional hazard of a person experiencing a divorce given that the duration of marriage is kept constant. Participants were followed-up until divorce, ten years after the start of follow-up, the end of 2009, participant’s emigration or participant’s or partner’s death, whichever occurred first.

All models adjust for participant’s sex, birth cohort, and age at first reproduction, marital duration, quadratic marital duration, and timing of marriage. Additionally, to account for life history strategies, we tested whether or not the associations between birth intervals and risk of divorce differed by SES and AFR by running models with interaction terms between SES and interbirth interval and AFR and interbirth interval.

The proportional-hazards assumption inherent in Cox regression was tested by including an interaction between interbirth interval and analysis time in the Cox model, and on the basis of Schoenfeld residuals [67]. Due to the large sample size, the tests showed significant non-proportionality in the IBI in the analysis of participants with two children, and in the second IBI in the analysis of participants with three children. However, when examining the hazard ratios of different birth interval categories through time, no substantial non-proportionality was detected (i.e., the shortest birth intervals were always associated with the highest divorce risk, followed by the longer birth intervals in ascending order, even though at some points in time the risk ratios were more similar than at some other points). All analyses were carried out with Stata 15.1. [68].

Results

Participants with two children

Compared to upper white-collars (estimated marginal mean IBI 39.88 months), farmers and entrepreneurs and self-employed had shorter birth intervals (36.25, 39.22, and 38.27months, respectively). The birth intervals for lower white-collars (40.14 months), manual workers (39.96 months), and students (40.09 months) did not statistically differ from the birth intervals of upper white-collars.

Compared to individuals whose first two children were born at most 18 months apart, individuals whose children were more widely spaced had a lower divorce risk (Table 2, Model 1; for the associations between covariates and risk of divorce, see S1 Table). Adjusting for individual’s SES had little effect on the associations between IBI and divorce (Table 2, Model 2). Additional analyses with different IBI categories as the reference group showed that birth intervals up to 3–3.5 years were protective over shorter intervals and longer intervals provided no additional benefit (S2 Table).

Table 2

Adjusted hazard ratios (HR) from Cox regressions predicting parental risk of divorce by children’s interbirth interval in individuals with two children.

	Model 1		Model 2
	HR	95% CI	HR	95% CI
Interbirth interval
≤ 18 months	Ref.		Ref.
18.01–24	0.79	0.73, 0.87	0.80	0.73, 0.87
24.01–30	0.69	0.63, 0.76	0.70	0.64, 0.76
30.01–36	0.60	0.55, 0.67	0.61	0.55, 0.67
36.01–42	0.67	0.60, 0.75	0.68	0.61, 0.76
42.01–48	0.62	0.55, 0.71	0.62	0.55, 0.71
48.01–54	0.57	0.50, 0.66	0.57	0.50, 0.66
54.01–60	0.65	0.56, 0.77	0.65	0.56, 0.76
60.01–66	0.66	0.55, 0.79	0.65	0.55, 0.78
66.01–72	0.66	0.54, 0.82	0.66	0.54, 0.82
> 72	0.63	0.54, 0.75	0.63	0.53, 0.74

All p-values ≤.001.

Model 1 adjusted for birth cohort, sex, age at first reproduction, marriage length at second birth, quadratic marriage length at second birth, and timing of marriage; Model 2 additionally adjusted for SES.

CI = Confidence interval.

To illustrate the associations between the interbirth interval and risk of divorce, we ran a Cox regression model with wider IBI categories (1: ≤18 months; 2: 18.01–24; 3: 24–36; 4: 36–48; 5: > 48), and drew survival functions by IBI categories (Fig 1). Keeping all independent variables at their average, the predicted proportion of individuals who had divorced (vs. were still married) at the end of the follow-up period was 19 per cent (based on the survival function). For individuals with an interbirth interval of no longer than 18 months, the predicted proportion of divorced at the end of follow-up was 26 per cent, and for individuals with an interbirth interval of more than four years, 18 per cent (Fig 1). In other words, individuals with the shortest interbirth intervals were 1.49 times more likely to divorce during the next ten years following the second child’s birth than were individuals with the longest interbirth intervals.

Fig 1

Cumulative risk of divorce by interbirth interval categories in individuals with two children.

Curves are reversed survival functions from a Cox regression predicting parental risk of divorce, adjusting for parental birth cohort, sex, and age at first birth, marital duration and quadratic marital duration at second birth, timing of marriage, and SES.

Participants with three children

The effect of the first interbirth interval on the risk of divorce was smaller in three-child families than in two-child families (Table 3), but here also, a first interval of no more than 1.5 years was associated with an increased risk of divorce compared to longer birth intervals (note that Table 3 shows the results from models adjusted for SES). When adjusting for the second birth interval, the effect of the first interval was slightly strengthened (Table 3, Model 3). The length of the second birth interval, however, had no association with individuals’ divorce risk.

Table 3

Hazard ratios (HR) from Cox regressions predicting parental risk of divorce by interbirth intervals (IBI) in individuals with three children.

	Model 1			Model 2			Model 3
	HR	95% CI	p	HR	95% CI	p	HR	95% CI	p
1st IBI
≤ 18 months	Ref.						Ref.
18.01–24	0.79	0.70, 0.90	.000				0.79	0.69, 0.89	.000
24.01–30	0.74	0.65, 0.85	.000				0.73	0.64, 0.84	.000
30.01–36	0.70	0.60, 0.81	.000				0.68	0.59, 0.80	.000
36.01–42	0.85	0.72, 1.00	.052				0.83	0.70, 0.99	.042
42.01–48	0.78	0.64, 0.94	.010				0.76	0.61, 0.93	.008
48.01–54	0.89	0.72, 1.11	.313				0.87	0.69, 1.10	.243
54.01–60	0.79	0.60, 1.03	.087				0.76	0.57, 1.01	.063
60.01–66	0.72	0.52, 0.99	.042				0.69	0.49, 0.97	.032
66.01–72	0.84	0.58, 1.21	.343				0.80	0.54, 1.18	.267
> 72	0.85	0.68, 1.07	.167				0.82	0.61, 1.10	.183
2nd IBI
≤ 18 months				Ref.			Ref.
18.01–24				1.10	0.91, 1.33	.304	1.10	0.91, 1.33	.336
24.01–30				1.12	0.92, 1.35	.266	1.10	0.90, 1.34	.347
30.01–36				1.05	0.86, 1.28	.633	1.02	0.83, 1.25	.839
36.01–42				1.00	0.82, 1.24	.964	0.97	0.78, 1.21	.798
42.01–48				0.98	0.79, 1.22	.889	0.94	0.74, 1.18	.586
48.01–54				1.07	0.85, 1.33	.583	1.02	0.80, 1.30	.887
54.01–60				0.85	0.67, 1.09	.207	0.80	0.61, 1.05	.115
60.01–66				1.05	0.82, 1.35	.680	0.98	0.75, 1.30	.908
66.01–72				0.94	0.71, 1.24	.658	0.87	0.64, 1.19	.386
> 72				1.15	0.92, 1.45	.216	1.06	0.80, 1.41	.684

All models are adjusted for birth cohort, sex, age at first reproduction, marital length at 3rd birth, quadratic marital length at 3rd birth, timing of marriage, and socio-economic status. Model 1: First IBI only; Model 2: 2nd IBI only; Model 3: Both IBIs. CI = Confidence interval.

We further tested whether the second birth interval had differential associations with parental divorce risk depending on the length of the first interval by running a model including both interbirth intervals and their interaction (wider IBI categories along with all the covariates). No significant interactions were found (S3 Table).

Fig 2 illustrates the associations between the two interbirth intervals and risk of divorce, from a Cox regression model with wider IBI categories (adjusting for both interbirth intervals simultaneously as well as for SES). Keeping all independent variables at their average, the predicted proportion of individuals divorced at the end of the follow-up was 14 per cent. For individuals with a first IBI of no longer than 18 months (adjusting for the second IBI), the predicted proportion of divorced at the end of follow-up was 17 per cent, and for individuals with a first IBIof more than four years, 14 per cent (Fig 2a). In other words, the marriages of individuals with a short first IBI were 1.24 times more likely to dissolve than were the marriages of individuals with a long first IBI in the ten-year follow-up. The second birth interval was not associated with divorce risk (Fig 2b) among individuals with three children.

Fig 2

Cumulative risk of divorce by the first (a) and second (b) interbirth interval in individuals with three children.

Curves are reversed survival functions of a Cox regression predicting parental risk of divorce, controlling for parental birth cohort, sex, and age at first birth, marital duration and quadratic marital duration at third birth, timing of marriage, SES, and for both interbirth intervals simultaneously.

Differences in associations by SES and AFR

We then examined the interactions between SES and birth intervals and AFR and birth intervals to investigate the role of life history strategy in the association between birth intervals and divorce. In order to avoid numerous interactions between multiple birth interval and socioeconomic or AFR categories, we divided birth intervals into two classes (1.5 years and under, and over 1.5 years) for these analyses. None of the interactions among individuals with two and three children between SES and the length of interbirth intervals were significant (S4 Table), indicating that sibling spacing was similarly associated with divorce risk in all SES groups. Also the interactions between AFR and IBI among individuals with two and three children were mostly non-significant, indicating the sibling spacing was similarly associated with the risk of divorce regardless of the age of entry to parenthood (S5 Table). The only exception here was the interaction between IBI and entering parenthood at age 30–34, which was just significant (HR = 0.80, 95% CI = 0.65–1.00, p = .049), indicating that in this AFR group, longer IBI (more than 18 months compared to max. 18 months) was more strongly associated with lower risk of divorce compared to individuals who entered parenthood before turning 25 (HR = 0.62, 0.51–0.76 and HR = 0.78, 0.70–0.87, respectively). This interaction effect is in the opposite direction than what we hypothesized based on life history theory (weaker association in individuals with a slower life history strategy, i.e., later start of childbearing).

Sensitivity analyses

To see whether the arbitrary choice of relationship length for individuals who got married after the birth of the first child affected our results, we ran analyses where only the true length of the actual marriage was taken into account (regardless of the timing of marriage), and including only individuals who were married before the birth of the first child. The results were very similar to the main analyses (Fig 3).To see if our choice of study sample affected the results, we ran sensitivity analyses including all individuals regardless of their final number of children (censoring at the birth of the third child in the model where follow-up started at the second child, and at the birth of the fourth child in the model where follow-up started at the third child). In these analyses, the association between longer intervals and lower risk of divorce persisted, but was attenuated (Fig 3 and S6 Table). Note that this analysis strongly confounds higher number of children, lower risk of divorce, and shorter birth intervals—because individuals with more than two children with the same person very rarely divorce before the birth of the third child, and have substantially shorter first birth intervals (Table 1).

Fig 3

Hazard ratios (HR) from Cox regressions predicting the risk of divorce by the first IBI.

Reference category is max. 18 months, and showing HRs for the five next IBI categories. Results from main analyses (Table 2, Model 2, controlling for SES) and from different sensitivity analyses. Married before children: Model includes only individuals who were married at the birth of the first child. True length of marriage: Model controls for the length of marriage only, regardless of the timing of marriage (even if married after the birth of first child). All 2+ children individuals: Includes everyone who had at least two children.

Discussion

Short birth intervals between siblings have previously been shown to be associated (although not necessarily causally) with a range of negative perinatal and long-term health outcomes for parents and children [11,12] and more conflictual sibling relations [37,38]. Here, we showed in a large, nationally representative sample of Finns, that individuals whose children were more closely spaced had an increased risk of divorce, as hypothesized. In a ten-year follow-up, individuals with two children born no more than 1.5 years apart had a 49 per cent higher divorce risk compared to individuals with two children born more than 4 years apart. Consistent with previous research concerning the associations between number of children and parental divorce risk [58,64], individuals with three children were somewhat less likely to divorce than those with two children. But also among individuals with three children, a short birth interval between the first and second child was associated with an increased risk of parental divorce. Regardless of the length of the second birth interval, individuals with a first birth interval of no more than 1.5 years had 24 per cent higher divorce risk compared to individuals whose first birth interval was more than 4 years. The associations found in this study were little affected by adjusting for SES. Further, the associations were equally strong in all socioeconomic groups and among individuals who entered parenthood earlier and later.

We assumed that the higher divorce risk among parents of closely spaced children could stem from elevated daily stress between family members. We also assumed that a fast life history strategy could induce a correlation between the two among people with a lower SES or earlier entry to parenthood, irrespective of elevated daily stress. These mechanisms are, naturally, not mutually exclusive. We hypothesized that in the absence of a causal mechanism related to, for example, stress, the association between shorter intervals and higher divorce risk could still be present among individuals with a low SES or young AFR, but absent or less strong among the high SES or individuals entering parenthood later. The interaction analyses showed that shorter birth intervals were similarly correlated with a higher divorce risk in low as well as high socioeconomic groups and in those who started childbearing earlier vs. later. In other words, even among people who can be hypothesized to follow a slow life history strategy (i.e., people in high adulthood socioeconomic position or later entry to parenthood), closely spaced children were associated with an increased risk of divorce. Higher parental stress is one plausible pathway between the two phenomena.

However, in individuals with three children, the first birth interval was associated with divorce risk while the second birth interval was not (see Table 3 and Fig 2). On one hand, the alleged stress posed by first two closely spaced children appears to persist irrespective of the timing of the third child. On the other hand, a third child born within a short time period does not seem to increase the alleged parental stress—at least not in ways that would be detrimental to the marital union. It might be that the profound life changes associated with becoming a parent both for the individuals [69-71] and the couple relationship [72] present a particularly vulnerable period during which the marital union can be permanently shaken by a rapid birth of the next child. Above and beyond this risk the timing of the third birth seems to make no difference.

Given that our study is observational, one must, however, abstain from making causal inferences based on the association between children’s birth intervals and parental risk of divorce. Even though we controlled for several potential confounders and checked for potential moderating effects of life history strategy, it is possible that short birth intervals and divorce risk reflect the consequences of yet some other variables that were not included in our models. For example, parents with marital problems might attempt to overcome those problems by having more children faster. Further, some genetically varying tendencies such as impulsivity might be associated with both shorter pregnancy intervals and the likelihood of divorce [73]. Quasi-experimental studies based on exogenous variation in birth intervals would be needed to investigate the causality in the association between birth intervals and parental risk of divorce further.

One of the main strengths of the current study is the large, nationally representative dataset with highly reliable register data on dates of births, marriages, and divorces, spanning a period of 37 years. Among the limitations is the fact that the data may underestimate the proportion of children born to men, because not all biological fathers are registered. During the study period, however, the proportion of Finnish children without a registered father was only around 2 per cent [74]. A large majority of these children of non-registered fathers would have been born outside marriages—children born in marriages (around 80 per cent of the present study sample) are automatically registered to both members of the marital union. It is thus safe to say that the possible underestimation of children born to men has not severely affected the present analyses. A second caveat is that the data only had reliable data on divorces, while cohabitation was very common in Finland during the study period. Future studies should employ data with reliable information on all types of union dissolution.

The results of our study add to the growing body of evidence of negative outcomes associated with short birth intervals [75]. Nevertheless, with the rising age at first birth evident in many Western countries, short interbirth intervals are likely to become more common. Many welfare states, Finland including, also implement family benefits that encourage shorter intervals. The implementation of these types of family benefits seem to have decreased the average interbirth interval in Sweden, Australia, and Finland, above and beyond the effect of age at first birth [2,7-10]. The possible ramifications of short birth intervals may be further intensified by the limited social support from kin and peers that characterizes modern Western societies [76]. It has been suggested, and in some Western countries already implemented, that family planning services should expand their focus from how to prevent unwanted pregnancies to how to achieve the desired number of children. This includes, for example, advice on the optimal timing of wanted pregnancies [see, e.g., 77]. The present findings, alongside previous study results on short birth intervals, suggest that there is an increasing demand of such family planning knowledge in the broad sense, aimed at both policy makers and the general public.

Hazard ratios from Cox regressions predicting parental risk of divorce by all the covariates (mutually adjusted) in individuals with two and three children.

(PDF)

Click here for additional data file.

Results from Cox regressions predicting the risk of divorce by interbirth interval (IBI) between children in two-child families, with varying reference categories.

(PDF)

Click here for additional data file.

Hazard ratios (HR) from Cox regression predicting risk of divorce in individuals with three children by first and second birth interval and their interaction.

(PDF)

Click here for additional data file.

Hazard ratios from Cox regressions predicting the risk of divorce by interbirth (IBI) intervals, socioeconomic status (SES), and their interactions.

(PDF)

Click here for additional data file.

Hazard ratios from Cox regressions predicting the risk of divorce by interbirth (IBI) intervals, age at first reproduction (AFR), and their interactions.

(PDF)

Click here for additional data file.

Results from Cox regressions predicting the risk of divorce by birth intervals in all individuals regardless of their final number of children.

(PDF)

Click here for additional data file.

18 Nov 2019

PONE-D-19-28900

Shorter Birth Intervals between Siblings Are Associated with Increased Risk of Parental Divorce

PLOS ONE

Dear Dr Berg,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

We would appreciate receiving your revised manuscript by Dec 27 2019 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

David Meyre

Academic Editor

PLOS ONE

Journal Requirements:

1. When submitting your revision, we need you to address these additional requirements. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For more information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially sensitive information, data are owned by a third-party organization, etc.) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

3. We note that you have included the phrase “data not shown” in your manuscript. Unfortunately, this does not meet our data sharing requirements. PLOS does not permit references to inaccessible data. We require that authors provide all relevant data within the paper, Supporting Information files, or in an acceptable, public repository. Please add a citation to support this phrase or upload the data that corresponds with these findings to a stable repository (such as Figshare or Dryad) and provide and URLs, DOIs, or accession numbers that may be used to access these data. Or, if the data are not a core part of the research being presented in your study, we ask that you remove the phrase that refers to these data.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: No

Reviewer #2: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: PLOS One review

Summary: The current manuscript uses Finnish population data to examine the relation between interbirth intervals and divorce rates over a 37 year span of time with an average follow-up of 10 years. They find that shorter intervals are associated with increased divorce rates. The paper would be improved by a reorganization of the introduction and a greater consideration of alternative explanations.

Abstract:

1. Adjust language that it is established that short interbirth intervals are associated with adverse perinatal outcomes, as this is still somewhat debated.

Introduction:

1. The first paragraph of the introduction briefly provides some history of interpregnancy interval evolutionary theories, but would benefit from a more unbiased initial presentation of the studies and evidence. First, there have been some challenges to the assumption that shorter birth intervals are causally associated with adverse birth outcomes (see Ball et al., 2014, Class et al., 2017 for example). The WHO recommendations are outdated and based on 1 study that did not do an adequate job of controlling for confounding factors. Second, one can alternatively imagine that increased maternal education levels and workforce participation has contributed to delayed maternal childbearing and an elongation of interpregnancy intervals. Third, work has also shown that long interpregnancy intervals are similarly problematic – also associated with adverse birth outcomes (like a ‘U’ shaped curve). I think this issues can simply be addressed by including some of these alternate perspectives and citations.

As I continued to read the introduction, the above mentioned topics were discussed. Therefore, perhaps a reordering of the presentation may help to paint a less one-sided take to the evidence.

2. This claim is uncited: “Raising two or more children of similar age is even more demanding, and the elevated need to provide intensive care is likely to affect parental stress” and is an assumption. The citation that occurs a few sentences later is in regards to twins and does not support this claim. This claim needs citation as, it could alternatively be the case that having two children of similar age is beneficial for the parents as the parents are therefore more familiar with the developmental needs and abilities of a certain age group, they can participate in similar activities with their children rather than having to meet the needs of two different levels all the time, even daily hassles such as similar bedtimes, napping needs, and feeding abilities would be lessened. This is all to say that arguments can be made in either direction, but if they are made, they need to be cited or framed as hypotheses. This again, may be addressed by rearranging the presentation of the introduction.

Methods:

1. Can you explain the separation of Women and Men in Table 1? Is it the case that these might be the same children? Am I correct in understanding that the parental sex was included in models, but not the child’s sex?

2. It would be helpful to have percentages for all variables (e.g. divorces) in Table 1.

3. Was it possible to use interpregnancy interval as opposed to interbirth interval (thus accounting for short pregnancies influencing the possibility of shorter birth intervals)?

4. Please make more clear what your reference intervals are. As it is stated in the abstract, one might imagine that interbirth intervals over 4 years are the reference.

5. What is the reasoning behind the choice to look at the first birth interval and the second birth interval separately within the 3 child families? Why would one believe that those two intervals contribute to divorce differently? It seems that this possibility is not introduced until the discussion and renders investigating if there is a difference between those intervals as slightly unfounded.

6. Using different interval spacing models makes your findings a bit difficult to synthesize. I believe your findings would translate more cleanly to previous research if a consistent model was used throughout the paper and those categories were similar to what previous work has used (perhaps being interbirth rather than interpregnancy).

Discussion:

1. This reviewer appreciates the limitations of causal interpretations the authors provide. These explanations might deserve more attention as they appear stronger than the explanation of stress caused by having two kids close together causing divorce. The authors state that quasi-experimental studies are needed to “strengthen” the argument of a causal relation between short interbirth interval and divorce, but I don’t believe that this study has the ability to make that initial claim at all. Perhaps a quasi-experimental study is needed to challenge the explanation that people who have closely spaced children also have a higher likelihood of divorce (correlation, but not causal relation).

Minor:

1. Gender is used interchangeably with sex and should not be.

2. The abbreviation of SES is not defined when first used and then inconsistently used.

3. Line 330: Obstain not obtain.

Reviewer #2: This is in many ways a competent and interesting manuscript. However, I have concerns about how the research addresses the underlying hypothesis, which is that short birth intervals negatively affect family functioning and in turn increase subsequent divorce, in part due to maternal depletion: the analysis does not test this pathway. Second, I question the use of socioeconomic status as a proxy variable representing fast life-history strategies. The life-history strategy argument is not made clearly enough as a theoretical approach: the authors need to convince the reader that short birth intervals are part of a fast life-history strategy. Usually it is only applied to the timing of maturity and reproduction, not to short birth intervals. Under this hypothesis, low socioeconomic status couples are presumably focussed on offspring quantity over quality, and short birth intervals with subsequent divorce and remarriage could be part of this reproductive strategy.

It would be worth developing and tightening the theoretical framework and determining whether the FINNUNION data can more directly address any hypothesised mechanisms linking short birth intervals and the likelihood of divorce. This is not easy, as FINNUNION is specifically focussed on marriage and does not appear to contain relevant intermediary variables. One possibility is related to the idea that boys increase maternal depletion relative to girls. For example, women consume more calories during pregnancy with a male offspring (Tamimi et al. 2003). If depletion is a root cause of marital stress and subsequent divorce, then a boy-boy pair with short birth spacing should be the most depleting for mothers, and a girl-girl pair should be the least depleting. Perhaps instead of dropping infant sex from the analyses you could use the infant sex data to more directly address the maternal depletion-related hypothesis.

An alternative hypothesis that could be tested with the FINNUNION sample is that rather than being related to fast versus slow life-histories, short birth intervals followed by divorce could result from alternative male mating strategies. If, like in other nations, divorce results in an increased maternal childcare burden and reduced paternal involvement (Kalmijn 2015), short birth intervals followed by divorce could be a male mating strategy: have a few children closely spaced then move on to a new partner. This can be addressed using the Finnish data: controlling for age, do men who divorce after short birth intervals marry again sooner than men who divorce after longer birth intervals? This would entail an analysis of divorced individuals, with birth intervals predicting time from divorce to remarriage

Minor comments:

Summarize sensitivity analyses in the main part of the paper: readers will want to see what difference it makes to use particular subsamples and different model specifications. In STATA, there is an add-on for creating coefficient plots to graphically summarise hazard ratios from several separate statistical analyses (Jann 2014).

I don’t understand why birth interval was categorised – why not keep it as a continuous independent variable?

The gap of divorce by years since 2nd birth continues to widen between birth interval groups long beyond the period of high infant dependency. Is it still stressful to have a short birth interval when they are both in school?

References:

Tamimi et al, (2003). Average energy intake among pregnant women carrying a boy compared with a girl. BMJ 326:1245-6.

Kalmijn, M. (2015). Father-Child Relations after Divorce in Four European Countries: Patterns and Determinants. Comparative Population Studies, 40, n. 3.

Jann B. (2014). Plotting regression coefficients and other estimates. The Stata Journal 2011;14(4):708-37.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: David Waynforth

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

26 Dec 2019

Response to reviewers.

We thank the anonymous reviewer and Dr. David Waynforth for the time and effort they have invested in our study, and for the insightful comments and critique they have provided. Below are our point by point answers to their comments.

Reviewer #1: PLOS One review

Summary: The current manuscript uses Finnish population data to examine the relation between interbirth intervals and divorce rates over a 37 year span of time with an average follow-up of 10 years. They find that shorter intervals are associated with increased divorce rates. The paper would be improved by a reorganization of the introduction and a greater consideration of alternative explanations.

Abstract:

1. Adjust language that it is established that short interbirth intervals are associated with adverse perinatal outcomes, as this is still somewhat debated.

Author response #1: We have adjusted the language in the abstract.

Introduction:

1. The first paragraph of the introduction briefly provides some history of interpregnancy interval evolutionary theories, but would benefit from a more unbiased initial presentation of the studies and evidence. First, there have been some challenges to the assumption that shorter birth intervals are causally associated with adverse birth outcomes (see Ball et al., 2014, Class et al., 2017 for example). The WHO recommendations are outdated and based on 1 study that did not do an adequate job of controlling for confounding factors. Second, one can alternatively imagine that increased maternal education levels and workforce participation has contributed to delayed maternal childbearing and an elongation of interpregnancy intervals. Third, work has also shown that long interpregnancy intervals are similarly problematic – also associated with adverse birth outcomes (like a ‘U’ shaped curve). I think this issues can simply be addressed by including some of these alternate perspectives and citations.

As I continued to read the introduction, the above mentioned topics were discussed. Therefore, perhaps a reordering of the presentation may help to paint a less one-sided take to the evidence.

Author response #2: We have revised the beginning of the introduction in many ways: i) We deleted the reference to the WHO guideline; ii) we deleted the part referring to short intervals being risky from the first paragraph; iii) we made more clear that the idea that postponement of parenthood and higher education among women are associated with shorter birth intervals is based on research findings, not an assumption; iv) we mention the U-shaped association between birth intervals and health outcomes right at the beginning or the paragraph discussing these findings; and v) we have more explicitly stated that the causality between birth intervals and health outcomes is currently under debate.

2. This claim is uncited: “Raising two or more children of similar age is even more demanding, and the elevated need to provide intensive care is likely to affect parental stress” and is an assumption. The citation that occurs a few sentences later is in regards to twins and does not support this claim. This claim needs citation as, it could alternatively be the case that having two children of similar age is beneficial for the parents as the parents are therefore more familiar with the developmental needs and abilities of a certain age group, they can participate in similar activities with their children rather than having to meet the needs of two different levels all the time, even daily hassles such as similar bedtimes, napping needs, and feeding abilities would be lessened. This is all to say that arguments can be made in either direction, but if they are made, they need to be cited or framed as hypotheses. This again, may be addressed by rearranging the presentation of the introduction.

Author response #3: We have revised this part of the introduction to include the potential positive outcomes of tighter sibling spacing, and more clearly state the speculative / hypothetical nature of our assumption on short birth intervals and increased parental stress.

Methods:

1. Can you explain the separation of Women and Men in Table 1? Is it the case that these might be the same children? Am I correct in understanding that the parental sex was included in models, but not the child’s sex?

Author response #4: The data include individuals, not married couples, but since this is a random 11% sample of Finns, it is possible that a small fraction of the sample are individuals married to each other (i.e., same children). Yes, parental sex was included in the model, because it is known that men are more likely to divorce (a larger proportion of men than women never marry, and more men have multiple marriages than women). To control for this in a sample including both men and women, we included parental sex in the models. Children’s sex was not associated with parental divorce risk, so it was left out from the final models.

2. It would be helpful to have percentages for all variables (e.g. divorces) in Table 1.

Author response #5: We added the percentages in Table1, wherever they were lacking.

3. Was it possible to use interpregnancy interval as opposed to interbirth interval (thus accounting for short pregnancies influencing the possibility of shorter birth intervals)?

Author response #6: Unfortunately, our data do not contain information on gestational length, so we could only examine birth intervals, not interpregnancy intervals.

4. Please make more clear what your reference intervals are. As it is stated in the abstract, one might imagine that interbirth intervals over 4 years are the reference.

Author response #7: We have added a mention on the reference interval in the “Statistical analysis”-part of the Methods-section, and also state it now explicitly in all the tables. In the abstract, in our opinion, it would still be best to keep the comparison as it is, as XX% higher risk is more intuitive and easy to understand than XX% lower risk (naturally, this comparison would have stayed exactly the same even if we had had the longest interval as a reference category).

5. What is the reasoning behind the choice to look at the first birth interval and the second birth interval separately within the 3 child families? Why would one believe that those two intervals contribute to divorce differently? It seems that this possibility is not introduced until the discussion and renders investigating if there is a difference between those intervals as slightly unfounded.

Author response #8: To us, this was an intuitive choice because 3-child families have two birth intervals. We did not have a specific hypothesis about this, but feel looking at them separately is the most open way of examining the effects of the two intervals. We have now justified our choice explicitly: “In models concerning individuals with three children, the first and second birth intervals were examined separately, mutually adjusted, and in interaction, to see if the two intervals had independent, additive, or interactive associations with the risk of divorce.”

6. Using different interval spacing models makes your findings a bit difficult to synthesize. I believe your findings would translate more cleanly to previous research if a consistent model was used throughout the paper and those categories were similar to what previous work has used (perhaps being interbirth rather than interpregnancy).

Author response #9: The paper uses interbirth intervals as the predictor (not interpregnancy intervals), and we do not have information on interpregnancy. In studies with interpregnancy interval as the main predictor, the shortest category is usually max. 5 months (corresponding roughly to an interbirth interval of 14 months) – however, our shortest reference category is a bit longer (18 months) because we didn’t have enough cases in the shorter categories to maintain statistical power. Different limits for the highest birth interval category in different analyses were used because we wanted to keep as many separate categories for the longer intervals as possible to investigate the possibility of a U-shaped effect between birth intervals and divorce risk. Very long intervals were so rare in the first IBI for people with 3 children that the categorization here was different. However, we have now changed the categorization in all analyses to be similar, that is, highest category is now always longer than 72 months.

Discussion:

1. This reviewer appreciates the limitations of causal interpretations the authors provide. These explanations might deserve more attention as they appear stronger than the explanation of stress caused by having two kids close together causing divorce. The authors state that quasi-experimental studies are needed to “strengthen” the argument of a causal relation between short interbirth interval and divorce, but I don’t believe that this study has the ability to make that initial claim at all. Perhaps a quasi-experimental study is needed to challenge the explanation that people who have closely spaced children also have a higher likelihood of divorce (correlation, but not causal relation).

Author response #10: We have revised this bit of text, and it now reads: ”Quasi-experimental studies based on exogenous variation in birth intervals would be needed to investigate the causality in the association between birth intervals and parental risk of divorce.”

Minor:

1. Gender is used interchangeably with sex and should not be.

2. The abbreviation of SES is not defined when first used and then inconsistently used.

3. Line 330: Obstain not obtain.

Author response #11: Thank you, these mistakes are now corrected.

Reviewer #2: This is in many ways a competent and interesting manuscript. However, I have concerns about how the research addresses the underlying hypothesis, which is that short birth intervals negatively affect family functioning and in turn increase subsequent divorce, in part due to maternal depletion: the analysis does not test this pathway.

Author response #12: Please see Author response #14.

Second, I question the use of socioeconomic status as a proxy variable representing fast life-history strategies. The life-history strategy argument is not made clearly enough as a theoretical approach: the authors need to convince the reader that short birth intervals are part of a fast life-history strategy. Usually it is only applied to the timing of maturity and reproduction, not to short birth intervals. Under this hypothesis, low socioeconomic status couples are presumably focussed on offspring quantity over quality, and short birth intervals with subsequent divorce and remarriage could be part of this reproductive strategy.

Author response #13: We have made the point of short birth intervals being a part of fast life history more clear in the introduction: “In resource scarce and dangerous environments with a lower life expectancy, it is more optimal to invest in current reproduction rather than future reproduction because the future is uncertain, and in quantity rather than quality of offspring because of lack of resources (40–44). These features of faster reproductive strategy lead to shorter birth intervals between children (42).”

To accommodate the reviewers concern about SES being a good indicator of fast life history strategy, we conducted a new analysis examining the interaction between age at first reproduction and length of birth intervals with a similar assumption than the SES-interaction analyses (older age at entry to parenthood is indicative of slower life history and association between IBI and divorce should be weaker/absent among these people if life history alone would explain the association). The results for this additional analysis are in accordance with and confirm the SES-interaction analyses: no significant interactions were found (except for one just significant in the opposite direction than hypothesized). These new results are given in supplementary table S5, and summarized in the text.

It would be worth developing and tightening the theoretical framework and determining whether the FINNUNION data can more directly address any hypothesised mechanisms linking short birth intervals and the likelihood of divorce. This is not easy, as FINNUNION is specifically focussed on marriage and does not appear to contain relevant intermediary variables. One possibility is related to the idea that boys increase maternal depletion relative to girls. For example, women consume more calories during pregnancy with a male offspring (Tamimi et al. 2003). If depletion is a root cause of marital stress and subsequent divorce, then a boy-boy pair with short birth spacing should be the most depleting for mothers, and a girl-girl pair should be the least depleting. Perhaps instead of dropping infant sex from the analyses you could use the infant sex data to more directly address the maternal depletion-related hypothesis.

Author response #14: With this comment, we realize that we have been somewhat unsuccessful in explaining the assumed pathway from short birth intervals to parental stress and divorce. We do not assume that this pathway would result from maternal depletion during pregnancy, but rather the strains of everyday life related to bringing up children of similar age – affecting mothers as well as fathers. We have altered the introduction to make our point clearer and to state more openly the hypothetical nature of this stress-related pathway (see Author response # 3). For this reason, we think that testing the maternal depletion hypothesis the way suggested by the reviewer would not be compatible with the research questions in this study, and have not conducted them (also, the sexes of children were not associated with divorce risk when we tested them in initial analyses).

An alternative hypothesis that could be tested with the FINNUNION sample is that rather than being related to fast versus slow life-histories, short birth intervals followed by divorce could result from alternative male mating strategies. If, like in other nations, divorce results in an increased maternal childcare burden and reduced paternal involvement (Kalmijn 2015), short birth intervals followed by divorce could be a male mating strategy: have a few children closely spaced then move on to a new partner. This can be addressed using the Finnish data: controlling for age, do men who divorce after short birth intervals marry again sooner than men who divorce after longer birth intervals? This would entail an analysis of divorced individuals, with birth intervals predicting time from divorce to remarriage

Author response #15: This is an excellent idea, and definitely worth investigating – however, we feel that it would be a topic of a different study (we would be happy collaborate with the reviewer in conducting such a study). In this paper, we are mainly interested in establishing an association between birth intervals and risk of divorce, which is a new finding in itself, and be cautious about this association not necessarily being causal. In our opinion, adding analyses with follow-up starting at divorce would draw the reader’s attention away from the main point of the current study.

Minor comments:

Summarize sensitivity analyses in the main part of the paper: readers will want to see what difference it makes to use particular subsamples and different model specifications. In STATA, there is an add-on for creating coefficient plots to graphically summarise hazard ratios from several separate statistical analyses (Jann 2014).

Author response #16: We plotted the HRs from main analyses and sensitivity analyses in what is now Fig. 3, and added a section called “Sensitivity analyses” at the end of Results, summarizing the analyses.

I don’t understand why birth interval was categorised – why not keep it as a continuous independent variable?

Author response #17: Birth intervals are regularly categorized in literature examining this topic, and in order to keep our results comparable with the literature, we have categorized the birth intervals (see also Author response #9).

The gap of divorce by years since 2nd birth continues to widen between birth interval groups long beyond the period of high infant dependency. Is it still stressful to have a short birth interval when they are both in school?

Author response #18: Divorce processes in marital relationships are usually long and take years to develop. Stress related to bringing up closely spaced siblings in infancy may have long-term ramifications on the marital relationship that only manifest later (see also MS lines 351-354).

Author note: In addition to the changes requested by the reviewers, we have made a small number of changes to improve the clarity of the manuscript.

Submitted filename: BergV_IBI_divorce_plosone_ResponseToReviewers.docx

Click here for additional data file.

13 Jan 2020

Shorter birth intervals between siblings are associated with increased risk of parental divorce

PONE-D-19-28900R1

Dear Dr. Berg,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

David Meyre

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

17 Jan 2020

PONE-D-19-28900R1

Shorter birth intervals between siblings are associated with increased risk of parental divorce

Dear Dr. Berg:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr David Meyre

Academic Editor

PLOS ONE