The association between number of siblings and delinquent behaviour

ABSTRACT

This study examines the association between number of siblings and delinquency, adjusting for family relations and demographic variables. Data is based on a nationally representative school survey in Sweden consisting of approximately 25,000 youths. The results show a positive association for those having five or more siblings (IRR = 1.533, p = < .001), whereas one or two siblings is negatively associated with delinquency compared to those having no siblings. These results remain stable after adjusting for family relations. This study underscores the importance of further exploring the variation and direction of the association between the number of siblings and delinquency, as well as deepening our understanding of the various theoretical mechanisms through which the number of siblings is associated with delinquent behaviour.

KEYWORDS:

• Delinquency
• family size
• number of siblings
• nonlinear associations

Introduction

The number of siblings is considered an important factor in criminological theory, with the assumption that children from larger families, characterised by a greater number of siblings, are more prone to engage in delinquent behaviour compared to those from smaller families. Despite empirical support for this found by several studies, the association is often observed to be relatively weak. However, unlike other aspects of family dynamics such as parental monitoring, family relationships, and support, which have received more research attention (Hoeve et al. 2009, 2012; Kroese et al. 2021; Nilsson 2017; Racz and McMahon 2011; Svensson and Johnson 2022), number of siblings or family size has typically been treated as a secondary variable rather than the primary focus of study (e.g., Amato and Fowler 2022; Argys et al. 2006).

How and why the number of siblings is associated with delinquency have been discussed within a number of theoretical frameworks (Brownfield and Sorenson 1994; Collier and Mears 2022). One of the most discussed frameworks falls within social control theory, often focusing on family variables, such as parental monitoring and attachment to parents. While several studies discuss social control concepts to enhance our understanding of the relationship between the number of siblings and delinquency, only a limited few have conducted empirical tests to validate the assumptions of the theory (e.g., Brownfield and Sorenson 1994; Collier 2020; Sampson and Laub 1994).

Furthermore, most previous studies have examined the association between the number of siblings and delinquency using linear models. Such models anticipate a consistent change in delinquency with each additional sibling, overlooking the possibility that the relationship may be more complex.

The relationship between the number of siblings and delinquency has not received much attention since early studies were undertaken within the field. To gain a deeper understanding of the nature of this relationship, more research is needed. Against this background, the present study will further examine this association and expand upon previous research.

Background

Empirical research indicates that the association between the number of siblings and the development of delinquent behaviour in youth remains uncertain in terms of direction and causal impact. While some studies have identified a positive association (Farrington 1987, 2011; Farrington, Ttofi, and Piquero 2016; Fischer 1984; Mercer et al. 2016; Murray and Farrington 2010; Sampson and Laub 1994), other studies have found a negative association, or no significant effects (Argys et al. 2006; Brownfield and Sorenson 1994; Heck and Walsh 2000; Muola, Ndung’u, and Fredrick 2009; Touliatos and Lindholm 1980).

These studies do, however, differ in a range of aspects, which makes comparisons difficult. In some studies, the number of siblings is not the main focus, but only included as a control variable (e.g., Amato and Fowler 2022; Argys et al. 2006; Mercer et al. 2016). Investigating the number of siblings solely as a control variable poses a limitation, as it may oversimplify the complexity of family dynamics. Furthermore, the studies use different samples, including males only (e.g., Farrington, Ttofi, and Piquero 2016; Heck and Walsh 2000; Sampson and Laub 1994), known offenders (Heck and Walsh 2000), or children from rehabilitation homes (Muola, Ndung’u, and Fredrick 2009), making it challenging to generalize finding across broader populations.

Additionally, while some studies use official measures of delinquency (e.g., Farrington, Ttofi, and Piquero 2016; Mercer et al. 2016), others rely on self-reported delinquency measures (e.g., Brownfield and Sorenson 1994; Heck and Walsh 2000; Touliatos and Lindholm 1980). Although studies relying on self-reported offending are more likely to capture delinquent behavior, those employing self-reported delinquency measure vary considerably in terms of the reported number and types of delinquent acts. The differences in measurements, combined with the fact that many of these studies have relied on small, non-representative samples, restrict the generalizability of their findings, leaving us with less accurate comparisons of the estimated effects.

The limitations in measuring siblings, as identified by Collier (2020), involve, among other things, that some studies do not distinguish between siblings residing within and outside the household. This may have implications for the results, as non-resident siblings may not be directly related to the immediate family setting. Furthermore, even though some studies use a continuous measure for the number of siblings (e.g., Argys et al. 2006; Collier and Mears 2022; Sampson and Laub 1994), a considerable number of studies have used a rough, dichotomized measure, categorizing the number of siblings as either large or small family size (e.g., Heck and Walsh 2000). Such an approach may result in loss of important, risk-relevant information and hide potential non-linearities in the relationship between number of siblings and delinquency (Collier and Mears 2022).

Consequently, most studies thus far have assumed a linear relationship between the number of siblings and delinquency (e.g., Brownfield and Sorenson 1994; Lauritsen 1993; Mercer et al. 2016; Sampson and Laub 1994), usually finding fairly modest effect sizes. However, there are valid reasons to expect that the association between the number of siblings and delinquency is nonlinear. For example, studies outside the field of criminology have found evidence that the number of siblings has a nonlinear association with adolescent outcomes, such as educational achievement (Mogstad and Wiswall 2016) and mental health (Downey and Cao 2023). Furthermore, recent criminological research by Collier (2020) suggests that the association between the number of siblings and delinquency does not follow a simple and consistent pattern. Instead, the relationship between these variables may exhibit fluctuations or change in direction as the number of siblings increases. However, several datasets are examined in this study, and results are inconsistent across them (Collier 2020).

From a theoretical perspective, the relationship between number of siblings and delinquency has often been examined from the viewpoint of social control theory, which places primary emphasis on understanding the socialisation process within families. Social control theories explain why people abstain from crime, rather than why individuals engage in crime, and argue that delinquent behaviour is limited by controls, and when controls are absent or diminish, the risk of committing crime increases (Hirschi 1969). It is generally assumed that family dynamics may vary qualitatively based on the number of children in a family, given that the availability of parental resources is believed to diminish as the number of siblings increases. For instance, it is argued that the extent of direct forms of control, such as parental monitoring and supervision, as well as indirect forms of control, such as parent–child attachment, may decline in larger families. Consequently, this decline in controls is suggested to increase the likelihood of engaging in delinquent behaviour (Collier and Mears 2022; Sampson and Laub 1994; Tygart 1991). Based on this, family process variables are likely to play an important role in mediating the effects of the number of siblings on delinquent behavior. Including factors such as parental monitoring and attachment to parents into the analysis would likely reduce the association between the number of siblings and delinquency.

Even though the logic derived from social control theory is commonly used, there is a lack of studies that empirically examine these arguments. Only a few previous studies have examined the association between the number of siblings and delinquency, adjusting for social control variables, and the results presented are somewhat contradictory. For instance, Tygart (1991) found that a greater number of siblings predicts less parental control, which in turn leads to greater antisocial peer influences and, consequently, result in a higher likelihood of delinquency. In a similar vein, Sampson and Laub (1994) found that social control processes within the family, such as supervision, harsh discipline, and parent–child attachment, operate through the family size-delinquency relationship. Their findings suggest that children in large families experience lower levels of social control compared to children from small families, which in turn increase their risk of delinquency. In contrast, Brownfield and Sorenson (1994) did not find any support for social control theory, since the relationship between the number of siblings and delinquency is, to a large extent, explained by delinquent siblings, and argue that their findings are rather in line with social learning theory. Furthermore, studies have also found that children with two siblings experience higher levels of parental attachment, and reduced likelihood of delinquency, compared to children in one-child households (Collier 2020), which contrasts with the notions of social control theory.

Overall, these contradictory theoretical findings underscore the need for additional studies that explore the relationship between social control variables, the number of siblings, and delinquency. Additionally, findings also imply that it is not solely negative effects that can be anticipated from the number of siblings but also positive. For example, having multiple siblings may serve as a protective factor, as older siblings in larger families may assume parental roles and responsibilities in raising their younger siblings (Hirschi 1969). Within other fields, previous research has found that siblings can have a range of positive influences on factors such as well-being (Chen, Chen, and Wang 2021) and social skills (Downey, Condron, and Yucel 2015). However, when examining these associations, it is important to also consider the quality of the relationship between siblings (Jensen et al. 2023; Yucel and Downey 2015). While some sibling relationships have shown to provide comfort and promote healthy behaviours, other sibling relationships are more tension-filled, which has implications for various outcomes (Yucel and Downey 2015).

Sweden as a context

Most previous studies have been conducted in the United States. Thus, it is of importance to conduct research in other settings as well. An interesting context for family-related research is Sweden, as it is often seen as a forerunner in family demographic change (Ohlsson-Wijk, Turunen, and Andersson 2020). In Sweden, along with other Western industrialised nations, the concept of family has undergone significant transformations in recent decades. One notable change is the increased prevalence of family instability. Although the traditional model of a two-parent household with two children remains common, there has been a rise in split-family households (Andersson and Kolk 2015). Additionally, the substantial influx of immigrants in Sweden over the past years has also influenced family size patterns, resulting in rapid shifts in the country’s demographic composition and family dynamics (Carlsson 2022).

Nevertheless, the cultural perception of an ideal family size in Sweden often revolves around having two children. Cultural perceptions about ideal family size are important in this context because they may lead to heightened parental investment in children and greater societal support for families that conform to this presumed ideal size. Conversely, families with fewer or more than two children may receive comparatively less support (Hagawen and Morgan 2005; Sobotka and Beaujouan 2014).

This study

Against this background, we argue that the relationship between the number of siblings and delinquency is unclear and needs additional attention. The present study will improve upon problems identified in previous studies and address some of these limitations by incorporating a representative sample, examining the significance of social control variables, and exploring more nuanced models to better understand the relationship between the number of siblings in a household and delinquency. Using a large, pooled, nationally representative sample of Swedish youth, the aim of this study is to, in detail, further explore the relationship between the number of siblings and delinquency.

The present study has two overall research questions: (1) Does the number of siblings relate to delinquency, and is there evidence of a nonlinear association? (2) Does the association change when adjusting for two central family socialisation variables, parental monitoring and attachment, as well as key demographic factors including family structure and immigrant background?

Data and method

Participants

This study is based on four waves of a cross-sectional, nationally representative school survey of adolescents in year nine of compulsory education, age 15 on average, conducted by the Swedish National Council for Crime Prevention between 2003 and 2011. All of the surveys are based on systematic samples of schools with year-nine classes. The data were primarily collected in December. The principal of each school distributed the questionnaires along with information about the study to teachers, and students completed the questionnaires during lesson time in the presence of the teacher. The surveys included a total of 6,692 adolescents in 2003; 7,449 adolescents in 2005; 6,893 in 2008; and 6,490 in 2011. In these surveys, the non-response rate was calculated in relation to the respondents in the participating classes and amounted to 14% in 2003, 14% in 2005, 19% in 2008, and 15% in 2011. The four subsamples combined give a total sample of 27,524 adolescents. Following list-wise deletion of missing values, the following analyses are based on 25,132 cases.

Measures

Dependent variable

Self-reported delinquency is based on an overall measure of 20 items on criminal offending during the past 12 months covering, for example, violence, theft, and vandalism. For detailed information of the items, see Appendix Table A1. We will be using a variety scale that includes all 20 items from the survey, and which measures how many different types of offenses the youths report having committed. The variety scale is seen as the most appropriate measure of delinquency and in general it has higher reliability and validity than frequency measures (Bendixen, Endresen, and Olweus 2003; Sweeten 2012). The variety scale has a Cronbach’s alpha of .86. In some analyses, we also present measures of prevalence for overall delinquency (0 = no act/1 = one or more acts), violence (0 = no act/1 = one or more acts), and serious theft (0 = no act/1 = one or more acts).

Independent variables

Number of siblings is measured on the basis of an overall measure of the number of siblings the respondent is living with. This measure is based on the combination of two items: (1) number of brothers, and (2) number of sisters. This measure is coded as living with 0 siblings/1 sibling/2 siblings/3 siblings/4 siblings/5 or more siblings. A total of 22.5% have no siblings, 40.5% have one sibling, 24.6 % have two siblings, 7.9% have three siblings, 2.7% have four siblings and, 1.8% have five or more siblings. The number of siblings is included in the analyses as a dummy variable using the category ‘living with zero siblings’ as a reference category. Most previous studies have used a frequency scale of the number of siblings (e.g., Collier and Mears 2022; Sampson and Laub 1994) whereas some have used dummies (e.g., Heck and Walsh 2000). For our main analysis, we decided to use the dummy coded version of number of siblings as this will make the results clear so as to identify differences between the groups. However, we also performed analysis using a frequency scale of the number of siblings (presented in Figure A1, Appendix).

Family relations

We use two measures of family relations. Attachment to parents is a mean index based on four items: (1) Do you think that you usually get on well with your mum? (2) Do you think that you usually get on well with your dad? (response alternatives: never/not often/often/most of the time/always) (3) Can you usually talk about anything at all (e.g., problems) with your mum? (4) Can you usually talk about anything at all (e.g., problems) with your dad? (response alternatives: definitely not/not much/maybe (it depends)/probably/yes, definitely). The Cronbach’s alpha for the scale is .76. High scores on this scale indicate strong emotional bonds to parents. Parental monitoring is an additive index based on two items: (1) Do you have to be home by a certain time on the weekends? (2) Do your parents demand that you tell them in advance what you are going to be doing if you want to go out on e.g., a Friday night? (response alternatives: no, never/rarely/sometimes/often/yes, always). The Spearman’s coefficient for these two items is 0.44. High scores on the index suggest that respondents experience strong parental supervision.

Five demographic variables are included in the regression models. Gender is coded as 0 for girls and 1 for boys. Immigrant background is coded as (1) Born in Sweden to two Swedish-born parents: Native Swedish, i.e., no immigrant background; (2) Born in Sweden to one Swedish-born and one foreign-born parent: second generation mixed; (3) Born in Sweden to two foreign-born parents: second generation; (4) Born abroad: first generation. Immigrant background is included in the analyses as dummy variables using category (1) ‘Born in Sweden to two Swedish-born parents’ as reference category. Parental employment status is a measure of whether the mother and the father are employed. The variable is coded as 0 if both of the parents are employed and 1 if either the mother or the father or both are, for example, not employed, seeking work, or receiving disability pension or early retirement benefits. Split family is coded as 0 if the respondent is living with both biological parents and 1 if this is not the case. These variables are included in the study because previous research has found them to be associated with delinquent behaviour (e.g., Knaappila et al. 2019; Kroese et al. 2021; Pierce and Jones 2022; Svensson and Shannon 2021; Vasiljevic et al. 2023). They are assumed to be weakly associated with delinquency when family relation variables are included in the analysis.

Year of study represents the year when the study was conducted and is included in the form of dummy variables (year 2005, 2008, 2011) using 2005 as the reference category in the analyses.

Analytical strategy

First, we compared differences between the number of siblings in regard to the different variables using one-way analysis of variance and chi-square test. Second, we performed a number of negative binomial regression models using our delinquency variety scale as outcome. Negative binomial regression was used because our outcome is a skewed count variable with overdispersion (Cameron and Trivedi 1998; Hilbe 2011). The modelling was estimated in three stages. First, we examined whether there is an association between the number of siblings and delinquency. Second, we included our demographic variables, i.e., gender, immigrant background, split family, and not in employment. Third, we included our two family relations variables, i.e., attachment to parents and parental monitoring. In the regression analyses, having no siblings is used as a reference category. All models adjust for year of study. As the data are based on respondents clustered in schools, standard errors clustered at school level are presented. We present the regression coefficients as relative risk estimates of incidence rate ratio (IRR). In order to illustrate the modelling, the regression estimates were plotted with their 95% confidence intervals (CIs) (using the coefplot in Stata) (Jann 2014). To examine whether the association between the different sibling groups and delinquency are different between Model 2 and Model 3, we estimated a cross-model test using the suest command in Stata (Clogg, Petkova, and Haritou 1995; Mize, Doan, and Long 2019; Weesie 1999). Interval-scaled predictors were standardised prior to inclusion in the regression models. All statistical analyses were conducted in Stata version 17.

Results

Table 1 presents differences between the number of siblings in relation to the outcome and the independent variables. The results show that youth having five or more siblings reported a higher level of delinquency than the other groups. We also found those having no siblings reported higher levels of delinquency than those having one and two siblings. Furthermore, Table 1 shows that as the number of siblings increases, the proportion of respondents with a native background decreases, while the proportion of respondents with an immigrant background increases. Additionally, single-child households are more frequently found in split families. Finally, the results in the table show only a few significant differences between the number of siblings groups in relation to attachment to parents and parental monitoring. The respondents with no siblings reported significantly lower levels of attachment to parents compared to those having one sibling. In addition, children with no siblings reported significantly lower levels of parental monitoring in relation to those having one and two siblings.

Table 1. Description of number of siblings groups (means/%).

	Total sample (N = 25,132)	No siblings (n = 5,652)	1 sibling (n = 10,191)	2 siblings (n = 6,171)	3 siblings (n = 1,995)	4 siblings (n = 667)	5 or more siblings (n = 456)	F-test, Bonferroni chi-square
Delinquency variety (min,max: 0,20)	1.48	1.73	1.30	1.39	1.63	1.72	2.58	39.19***^a, b, e, g, h, i, j, k, l, n, o
Boy %	49.7	53.9	48.0	48.9	48.4	48.6	52.2	54.79*** ^a, b, c, d
Swedish %	77.8	75.0	81.3	79.5	71.2	67.5	56.4	321.75*** ^a, b, c, d, e, f, g, h, i, j, k, l, n, o
Second generation mixed %	9.4	11.2	9.0	8.8	8.9	8.5	9.2	27.38*** ^a, b, c, d
Second generation %	6.5	6.4	4.7	6.8	10.9	12.1	14.7	206.27*** ^a, c, d, e, f, g, h, i, j, k, l, n
Born abroad %	6.3	7.3	5.0	4.9	9.0	11.8	19.7	262.23*** ^a, b, c, d, e, g, h, i, j, k., ^l, ^m, n, o
Split family %	25.3	40.4	19.9	20.0	27.0	25.8	25.2	928.62*** ^a, b, c, d, e, g, h, i, j, k, l
Not in employment %	20.3	26.0	15.3	17.8	26.7	35.4	47.6	651.14*** ^a, b, d, e, f, g, h, i, j, k, ^l, ^m, n, o
Attachment to parents (min,max: 0,4)	2.69	2.67	2.71	2.69	2.68	2.62	2.70	3.34**^a
Parental monitoring (min,max: 0,8)	4.89	4.76	4.96	4.93	4.81	4.83	4.73	7.99***^a, b

Low values on attachment to parents and parental monitoring indicate weak bonds to parents and being weakly controlled by parents. Subscripts represent statistically significant (p < .05) Bonferroni comparison for mean values and chi-square for percentage.

^aNone versus one sibling. ^bnone versus two siblings. ^cnone versus three siblings. ^dnone versus four siblings. ^enone versus five or more siblings. ^fOne sibling versus two siblings. ^gOne sibling versus three siblings. ^hOne sibling versus four siblings. ⁱOne sibling versus five or more siblings. ^jTwo siblings versus three siblings. ^kTwo siblings versus four siblings. ^lTwo siblings versus five or more siblings. ^mThree siblings versus four siblings. ⁿThree siblings versus five or more siblings. ^oFour siblings versus five or more siblings.

*p < .05; **p < .01; ***p < .001.

Table 2 presents the IRR for Model 1 (only year of study included), Model 2 (demographic variables added), and Model 3 (family relations variables added) for the total sample. The results for the first model show that having one (IRR = .755, p = < .001) or two (IRR = .800, p = < .001) siblings is significantly negatively associated with delinquency in relation to having no siblings. No significant association was found for those having three or four siblings. For those reporting having five or more siblings, there is a significant positive association with delinquency (IRR = 1.533, p = < .001). The association between the number of siblings and delinquency shows no large differences when including the demographic variables in Model 2. Finally, in Model 3 the results are presented when adding attachment to parents and parental monitoring. The results show the same pattern as the second model, i.e., having one or two siblings is still significantly negatively associated with delinquency (IRR = .889, p = < .001 vs. IRR = .905, p = .002). Having five or more siblings shows a positive association (IRR = 1.423, p = < .001) with delinquency.

Table 2. The association between delinquency variety and number of siblings, full sample.

	Model 1		Model 2		Model 3
	IRR	p	IRR	p	IRR	p
1 sibling^a	.755	<.001	.882	<.001	.889	<.001
2 siblings^a	.800	<.001	.908	.002	.905	.002
3 siblings^a	.933	.127	.999	.993	1.014	.742
4 siblings^a	.979	.762	1.033	.630	1.057	.413
5 or more siblings^a	1.533	<.001	1.515	<.001	1.423	<.001
Boy			2.008	<.001	2.100	<.001
Second generation mixed^b			1.235	<.001	1.212	<.001
Second generation^b			1.390	<.001	1.407	<.001
Born abroad^b			1.262	<.001	1.264	<.001
Not in employment			1.253	<.001	1.161	<.001
Split family			1.474	<.001	1.366	<.001
Attachment to parents (z)					.674	<.001
Parental monitoring (z)					.802	<.001
Log pseudolikelihood	−39964.871		−39280.965		−38274.354
N	25,132		25,132		25,132

Negative binomial regression. IRR = incidence rate ratio. The p-values are calculated using robust standard errors clustered by schools. All models control for year of study.

^aNo siblings is the reference category.

^bNo immigrant background is the reference category.

To examine our second research question, a cross-model test was estimated to examine whether there are any significant differences between Model 2 and Model 3. The results show no significant differences between the number of siblings when control is held for attachment to parents and parental monitoring. For an illustration of the findings presented in the table, see Figure 1.

Figure 1. Association between number of siblings (ref: no siblings) and delinquency variety. IRR estimates presented with 95% CI 95. Negative binomial regression analysis.

In addition, we also performed the regression model including the number of sibling measure as a frequency measure. In the final model, the results show a weak positive association between the number of siblings and delinquency (IRR = 1.027, p = < .007). This gives support for including our dummy variables to examine the nonlinear association.

The results presented indicate the association to be nonlinear. According to this, we also performed Model 3 including a squared term of the number of siblings. The results show a clear nonlinear trend of the association between the number of siblings and delinquency (see Appendix Figure A1).

Sensitivity analyses

We also performed a sensitivity analysis, and estimated models using logistic regression estimating average marginal effects (AME) for prevalence of violence and serious theft as outcomes (Table A2 and A3 Appendix).¹ The findings are fairly stable over these outcomes.

The results show that having one and two siblings are both negatively associated with violence compared to those having no sibling. Having one sibling is negatively associated with serious theft compared to those with no siblings. For both violence and serious theft, having five or more siblings has the strongest association with the outcomes.

Discussion

The present study investigates the relationship between the number of siblings and delinquency. Previous studies have suggested that larger family size may increase the risk of delinquent behaviour, but findings have been inconsistent. This study addresses some limitations in previous research by using a large representative sample of Swedish youth.

The findings in this study show that having one or two siblings may have a protective effect against delinquent behaviour, while no association was found for youths with three or four siblings. However, after a certain threshold, such as five or more siblings, youths report higher levels of delinquency compared to other groups. Additionally, youths with no siblings show higher delinquency levels than those with one or two siblings. The findings remained stable after adjusting for parental monitoring and attachment as well as demographic variables. These results indicate a clear nonlinear result, suggesting that the association between the number of siblings and delinquency is neither simple nor consistent. These findings support previous research suggesting that the association between family size and delinquency can vary in different directions as the number of siblings increases (Collier and Mears 2022).

The findings indicate that the number of siblings becomes criminogenic primarily when it reaches higher numbers, suggesting that youths with a larger number of siblings (five or more) may exhibit increased susceptibility to engaging in criminal behaviour. In families with many children, available resources, such as parental time and attention, may be stretched thin. Consequently, this could lead to reduced parental supervision and increased exposure to negative peer influences. While parental monitoring, attachment to parents, and the demographic variables included are important factors to consider, the results in this study suggest that they may not fully capture the underlying mechanisms by which the number of siblings influences delinquency. The relationship between the number of siblings and delinquency is likely influenced by a combination of individual differences (e.g., self-control and moral values), sibling dynamics, parental capacity, and external environmental factors. Unraveling the complexity of these interactions would require a more nuanced exploration that goes beyond the factors used in this study.

With the exception of youths who have a large number of siblings, our study findings suggest that siblings may have a positive impact on the development of prosocial behavior. Previous research has demonstrated that sibling relationships play a significant role in promoting healthy development and overall well-being (Chen, Chen, and Wang 2021; Killoren et al. 2015; LeBouef and Dworkin 2021). These relationships can provide vital support and compensate for inadequate parenting or strained parent–child relationships in certain cases (Borchet et al. 2020; Milevsky 2022). However, as the number of siblings increases, the protective effect of having siblings appears to diminish, eventually reaching a tipping point at five children. At this tipping point, the advantages associated with having siblings transform into disadvantages in terms of the increased risk for delinquency. Our results differ from those found by Collier and Mears (2022), who did not find any conclusive thresholds effects. However, it is unlikely that there is a universal tipping point or specific numerical threshold that universally applies to the impact of the number of siblings on delinquency. The relationship between the number of siblings and adverse outcomes is influenced by various factors, including cultural and contextual variables (Björklund, Ginther, and Sundström 2007; Downey and Cao 2023).

Sweden, which stands out as having one of the most comprehensive social safety nets supporting families, ranks relatively low in terms of intergenerational solidarity compared to other Western countries and has one of the lowest rates of three-generational households in Europe (Georgas et al. 2006). Speculatively, this suggests that kin networks may be less developed within larger families in Sweden, potentially resulting in limited support for children. Subsequently, this could increase the risk for risky behaviours, including delinquency.

The findings of this study also indicate that individuals without any siblings have a higher risk of engaging in delinquent behaviour compared to those with one or two siblings. Previous research suggests that the advantages associated with growing up with siblings may outweigh the benefits of being an only child, receiving undivided attention (Jensen et al. 2023). Several factors may contribute to the increased risk among only children, including decreased socialisation opportunities due to limited exposure to positive sibling influences, and potentially inadequate parental supervision.

Furthermore, studies have indicated that only children tend to exhibit lower levels of self-control, poorer interpersonal skills, and a higher likelihood of externalizing problem behaviours compared to children with at least one sibling (Downey, Condron, and Yucel 2015). Most studies on only children have primarily concentrated on educational outcomes (e.g., Blake 1989; Downey 1995, 2001). It is, however, worth noting that research specifically focused on only children, particularly within the field of criminology, is limited. One study, however, found that individuals without siblings exhibited a higher risk of offending than those with non-offending siblings, suggesting that having non-offending siblings may act as a protective factor (Beijers et al. 2017).

Interesting to note, our descriptive data reveal that one-child households are more often split compared to households with one or more siblings, which may contribute to a higher susceptibility to delinquent behaviors among youths with no siblings. Previous research has found that family structure plays an important role in shaping individual outcomes (Cookston 1999; Svensson 2003; Vasiljevic, Svensson, and Shannon 2021).

Limitations

Several limitations need to be acknowledged in this study. Firstly, there is a limitation when it comes to contextual details. Some measures are relatively limited, and the study does not encompass a wide range of sibling- and family-related factors that could potentially influence the association between the number of siblings and delinquency. Consequently, important details that could be of significance in understanding the complexities of family dynamics might be overlooked. Families are intricate systems, and considering factors such as sibling constellations and relatedness (e.g., biological, half-siblings, or step-siblings), birth order, and age spacing is crucial as they have been shown to play a role in the association between the number of siblings and delinquency. Notably, research has highlighted the risk associated with having an older brother (Ardelt and Day, 2002) and the potential protective effects, particularly in certain groups, of having a sister (Killoren et al. 2015).

The study also lacks information regarding diverse family arrangements beyond whether adolescents live with both biological parents or not. Previous research has shown that adolescents in asymmetrical family arrangements, where either the mother or the father (but not both) has a new partner, tend to report higher levels of delinquency (Svensson and Johnson 2022). Additionally, studies indicate that children living in complex family structures with half- or step-siblings face a higher risk of negative life outcomes compared to those in family structures with only full biological siblings (Fomby and Osbourne 2017; Manning, Brown, and Stykes 2014).

Secondly, the study does not examine central, theoretical, individual-level indicators such as self-control (Gottfredson and Hirschi 1990) and moral values (Wikström et al. 2012). Individuals differ in their individual characteristics, and these characteristics can have a significant impact on behavior. The influence of the number of siblings on individuals may vary based on their levels of self-control and moral values. These need to be considered in further research.

Finally, the results in this study are based on several cross-sectional datasets. Despite the study being based on a large, nationally representative sample, it is not possible to draw conclusions about causal relationships. It might be valuable for further research to use data based on longitudinal design to be able to examine the time order of events and establish causality.

Despite these limitations, we argue that the main contribution of this study lies in its detailed examination of family size. The result of the present paper strengthens the notion that the magnitude and the direction of the number of siblings needs additional attention. Additionally, we also need to deepen the understanding of the theoretical mechanisms through which the number of siblings is associated with delinquent behaviour. To conclude, it is important to move away from simplistic models on the relationship between the number of siblings and delinquency and include more complex, and theoretically-based, measures to enhance understanding of this relationship.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Eva-Lotta Nilsson

Eva-Lotta Nilsson, Ph.D., is a researcher at the Department of Criminology at Malmö university, Sweden, where she also received her PhD. Her research interests center mainly around adolescent delinquency, with a special focus on family socialization processes.

Zoran Vasiljevic

Zoran Vasiljevic is a senior lecturer at the Department of Criminology at Malmö University, Sweden. He received his PhD in Criminology from Malmö University. His research interests cover a range of topics within crime and deviance, such as crime trends, segregation, and immigration and crime.

Robert Svensson

Robert Svensson is Professor at the Department of Criminology at Malmö University, Sweden. He received his PhD in Sociology from Stockholm University. His research interests span a range of topics in the field of crime and deviance, with a special focus on crime and deviance among adolescents.

Notes

1. Measure of violence includes the following items: carried a knife as a weapon; threated someone; assault; injured someone with a weapon. Measure of serious theft includes the following items: stolen a bicycle; moped/motorcycle; car theft; theft from car; stealing from pocket; snatched a bag.

 

Appendix

Table A1. Items measuring self-reported delinquency.

Intentionally destroyed a call box, streetlamp, window, somebody’s bicycle or some other thing that did not belong to you? Stolen something from a shop or department store? Used a forged ID Card? Cheated someone into giving you money? Stolen money from a coin-operated telephone or vending machine? Stolen a bicycle? Stolen a moped or motorcycle? Stolen a car? Stolen something out of a car? Stolen something out of somebody’s pocket? Snatched a bag, purse, or similar from somebody you did not know? Stolen something not included in the above questions? Broken into someone’s home or some other building? Bought something you knew was stolen? Sold something you knew was stolen? Taken a knife with you (as a weapon) when you went out? Threatened somebody with violence or with a weapon in order to obtain money or other valuables? Set fire to something valuable, e.g., a barn, a car, trees, or a building that did not belong to you without permission? Intentionally beaten somebody up (other than a member of your family) so you think or know he/she needed medical treatment? Intentionally injured somebody with a knife or other weapon?

Figure A1. Predictions of number of siblings and delinquency variety (including squared term of number of siblings). Results presented with 95% CI from negative binomial regression analysis. Controls included are: gender, second generation mixed, second generation, born abroad, not in employment, split family, attachment to parents, parental monitoring, year of study.

Table A2. The association between prevalence of violence and number of siblings.

	Model 1		Model 2		Model 3
	AME	p	AME	p	AME	p
1 sibling^a	−.057	<.001	−.031	<.001	−.030	<.001
2 siblings^a	−.050	<.001	−.026	<.001	−.026	<.001
3 siblings^a	−.017	.063	−.004	.624	−.004	.640
4 siblings^a	−.010	.473	−.002	.867	−.006	.626
5 or more siblings^a	.071	<.001	.056	.004	.054	.004
Boy			.125	<.001	.125	<.001
Second generation mixed^b			.044	<.001	.040	<.001
Second generation^b			.074	<.001	.078	<.001
Born abroad^b			.050	<.001	.050	<.001
Not in employment			.033	<.001	.025	<.001
Split family			.054	<.001	.038	<.001
Attachment to parents (z)					−.046	<.001
Parental monitoring (z)					−.027	<.001
N	25,001		25,001		25,001

AMEs from logistic regression. AME = Average marginal effects. The p-values are calculated using robust standard errors clustered by schools. All models control for year of study.

^aNo siblings is the reference category.

^bNo immigrant background is the reference category.

Table A3. The association between prevalence of serious theft and number of siblings.

	Model 1		Model 2		Model 3
	AME	p	AME	p	AME	p
1 sibling^a	−.046	<.001	−.021	.001	−.019	.004
2 siblings^a	−.032	<.001	−.010	.137	−.009	.169
3 siblings^a	−.019	.051	−.007	.461	−.006	.490
4 siblings^a	−.018	.277	−.008	.600	−.012	.442
5 or more siblings^a	.074	.001	.066	.001	.065	.001
Boy			.091	<.001	.089	<.001
Second generation mixed^b			.033	<.001	.029	<.001
Second generation^b			.068	<.001	.073	<.001
Born abroad^b			.034	.002	.034	.002
Not in employment			.024	<.001	.015	<.001
Split family			.067	<.001	.048	<.001
Attachment to parents (z)					−.050	<.001
Parental monitoring (z)					−.034	<.001
N	25,132		25,132		25,132

AMEs from logistic regression. AME = Average marginal effects. p-values are calculated using robust standard errors clustered by schools. All models control for year of study.

^aNo siblings is the reference category.

^bNo immigrant background is the reference category.