Sibship Size and Children’s Family Resources

Abstract

Despite ample evidence on the negative association between sibship size and educational attainment, relatively fewer studies have directly addressed the immediate effect of sibship size on children’s family resources. With data from a nationally representative survey in China, this study estimated the effects of sibship size on various family resources for Chinese adolescents between ages 10 and 15, with a particular interest in the possible role of gender on resource allocation. Consistent with prior research on educational attainment, sibship size was negatively associated with the amount of resources a child received from the family. Siblings who were closer in ages had equal or more negative effects on family resources a child received compared with widely spaced siblings. A more rapid decrease was observed in economic resources than in interpersonal resources. Furthermore, brothers showed greater negative effects than sisters. The current findings provided direct evidence for the resource dilution theory.

Keywords

family siblings family structure family relations/processes

Sibship size is among the most consistent predictors of educational attainment and intellectual development across different populations, with a general pattern of lower educational attainment associated with an increasing number of siblings (Downey, 1995, 2001; Jager, 2009; Lu & Treiman, 2008; Marteleto, 2010; Sen & Clemente, 2010; Silles, 2010). Several mechanisms have been proposed to account for such association, among which the resource dilution theory is the most widely accepted (Blake, 1981; Downey, 1995, 2001). Some recent studies have further suggested that when resources are diluted with larger family sizes, they may not be equally allocated among siblings, and children’s gender appears to function as an important factor in determining how much resources the child would obtain from the family (e.g., Chu, Xie, & Yu, 2007; Dayioglu, Kirdar, & Tansel, 2009; Jacob, 2011; Kang, 2011). While most existing studies of sibship size focused on distal outcomes such as educational attainment, the current study aimed to address the immediate outcomes of family resources, with a particular interest in how brothers and sisters might have differentiating effects on family resource allocation for Chinese boys and girls.

Resource dilution model offers a mathematically simple explanation for why educational attainment would be lowered with larger sibship size. As stated by Blake (1981, p. 422), “The more children, the more these resources are divided, and, hence the lower the quality of the output.” Blake’s statement encompasses at least two processes: first, larger sibship size leads to diluted resources; second, diluted resources lead to lower quality of children. A competing theory of the resource dilution model is the confluence model, which ascribes the negative effect of sibship size on educational attainment to a changing social, economic, and intellectual environment in the family with the birth of each additional child (Zajonc, 2001). According to the confluence model, younger siblings fare worse than older siblings due to lowered average intellectual environment. The confluence model is most applicable to intellectual development of children (de Haan, 2010). Common factors may also contribute to the association between sibship size and educational attainment. For example, parents who are intellectually advantaged may choose to have fewer children and their children also tend to have better educational performance (Downey, 2001).

So far, empirical evidence is in more favor of the resource dilution theory; however, the majority of such evidence focused on outcomes related to the quality of children, whereas little research directly targets children’s resources. For example, the most frequently studied outcomes include years of schooling (e.g., Blake, 1981; Chu et al., 2007) and academic achievement (e.g., Cheng et al., 2013; Kanazawa, 2012; Silles, 2010). Fewer studies have evaluated outcomes that directly measured resources. With a nationally representative sample of U.S. eighth graders, Downey (1995) examined the effects of sibship size on both economic resources (e.g., money saved for college, computer in home, etc.) and interpersonal resources (e.g., parent’s educational expectations, knowing eighth grader’s friends, etc.), and found consistent negative effects of sibship size on different parental resources, net of controls. Furthermore, the author noticed a non-linear pattern of sibship size effect on economic resource variables in which economic resources decreased sharply with the first one or two siblings and then leveled off with additional siblings, whereas the interpersonal resources steadily declined with additional siblings. Proponents of the dilution theory also hypothesized a more damaging effect from closely spaced siblings than widely spaced siblings because the former were more likely to compete for similar resources (Steelman, Powell, Werum, & Carter, 2002).

While Downey (1995) reported negative associations between sibship size and various parental resources, the study did not address whether resources might be disproportionately diluted among siblings. In fact, many recent studies have pointed out gender to be an important moderator of the association between sibship size and educational attainment. In general, a larger detrimental effect of sibship size tended to be observed on females than on males. For example, Lu and Treiman (2008) found that the negative effect of sibship size on years of education was only true for Chinese women, but not for Chinese men. Furthermore, gender of the sibling may also matter, and some studies have observed a stronger effect from male siblings than from female siblings. For example, Jacob (2011) reported a negative effect of older brothers on their sisters’ chance of graduation in Germany. Similar observations have been made in other cultures (e.g., Dayioglu et al., 2009).

The moderating effect of gender has been hypothesized to result from two possible mechanisms. First, some societies have strong patriarchal cultures, in which sons are considered permanent members of the families whereas daughters are not (Chu et al., 2007). Second, men on average are thought to have greater earning power, so more investment in male family members is likely to bring more economic returns to the family (Lu & Treiman, 2008). In either case, males would be in a more favorable position when resources are allocated among siblings.

Although much research has been dedicated to the role of children’s gender on outcomes of educational achievement (e.g., Butcher & Case, 1994; Conley, 2000; Dayioglu et al., 2009; Hauser & Kuo, 1998), evidence is scarce on outcomes directly measuring family resources. Distal outcomes such as years of education have the advantage of capturing accumulative effects of sibship through various mechanisms along the life course (Bonesronning & Massih, 2011), but a direct examination of family resources offers a clearer picture of the more immediate effect of family expansion, which may or may not extend to the distal outcomes.

The adolescent sample from China offers a particularly interesting context for the study of sibship size and the role of gender in this context. First, with the implementation of the one-child policy in China since the late 1970s, average family size significantly shrank, as indicated by an average total fertility rate of 1.7 in 2001 (Wang, 2003). It would be interesting to know whether the same pattern of sibship size effect observed with older Chinese population (e.g., Lu & Treiman, 2008) would also apply to the adolescent sample. The one-child policy has been strictly enforced for residents with urban household registration (hukou) and government employees, with only a few exceptions (e.g., when the first born has a disability etc.); rural residents are generally allowed to have a second child, especially when the first child is a girl (Hesketh, Lu, & Xing, 2005). Parents may also choose to have an additional child at the expense of their government positions or financial punishments.

In addition to the changing family sizes, a growing trend toward gender equality among the younger population may also have implications for the role of children’s gender on family resource allocation. Despite the traditional patriarchal and patrilineal culture in China, there has been some positive evidence of gender equality among Chinese youth, although with a relatively small sample (e.g., Tsui & Rich, 2002). The use of a nationally representative sample of Chinese adolescents would offer more evidence in this regard.

The current study had two main purposes. The first was to estimate the general effect of sibship size on various family resources with a nationally representative sample of Chinese adolescents between ages 10 and 15. Second, with the great wealth of sibling information from a family roster data set, the study further explored how sibling gender would affect the resource allocation for male and female children.

Method

Data Source

Data were from the nationally representative China Family Panel Studies (CFPS; Xie, 2012). CFPS is a longitudinal survey that follows all members of sampled families every 2 years and collects data on the socioeconomic, demographic, educational, and health aspects at the community, family, and individual levels (Xie, 2012). CFPS was initiated in 2010 and was the first of its kind in China.

CFPS samples were drawn through a multistage probability sampling process, where counties were the primary sampling units; communities were then sampled within counties, and the third and final sampling unit was household (Xie, 2012). Sampling frames for the first and second stages were from official administrative divisions, and within each stage, samples were selected with implicit stratification using administrative units and socioeconomic status as the main stratification variables. Because official household records from urban neighborhoods or villages may contain a large number of omissions due to internal migration, the final stage (i.e., household) sampling frame was constructed by CFPS personnel via map-address method. The CFPS sample bears an impressive resemblance to China’s 2010 Census in terms of its age and gender composition (Xu & Xie, 2013).

The 2010 baseline survey of CFPS collected information on 57,155 individuals from 14,960 households and 634 communities in China. Individuals below age 16 took the child questionnaires, whereas those aged 16 and above took the adult questionnaires. Children below age 10 were considered too young to directly participate in the survey, so their guardians (most often their parents) took the proxy form of child questionnaires.

Participants

A total of 3,463 children between ages 10 and 15 were in our sample. Those children were mostly born in years 1995 to 2000. Because about 1% of the total sample were interviewed in 2011 instead of 2010, the current sample also contained a small number of children (n = 46) who were born in 2001. Their average age was 12.47 (SD = 1.71) at the time of the interview, and there were more boys (53.8%) than girls. More than half of the sample (55.4%) lived in rural areas. About 30% of the sampled children’s mothers and 18% of their fathers received no formal schooling; 13% and 18% of the mothers and fathers had at least 12 years of education. After listwise deletion based on the core set of demographic variables, the analytical sample included 3,073 respondents. The two samples were largely similar in terms of age and gender compositions as well as parental education (see Table 1).

Table 1.

Descriptive Statistics of the Study’s Full and Analytical Samples.

Variables	Full sample (n = 3,463)	Analytical sample (n = 3,073)	Average sibship size
Age	12.47 (1.71)	12.47 (1.71)
Gender
Male	53.8%	53.1%	0.88 (0.89)
Female	46.2%	46.9%	1.10 (0.95)
Urban residency
Urban	44.6%	44.8%	0.67 (0.84)
Rural	55.4%	55.2%	1.24 (0.91)
Mother’s education
No schooling	30.7%	31.3%	1.51 (0.95)
Less than 12 years	56.8%	56.2%	0.84 (0.80)
≥12 years	12.5%	12.5%	0.33 (0.64)
Father’s education
No schooling	17.8%	17.9%	1.48 (1.01)
Less than 12 years	64.0%	64.0%	0.97 (0.87)
≥12 years	18.2%	18.2%	0.54 (0.78)
Yearly household income
Lowest quartile	22.3%	21.0%	1.14 (0.92)
2nd	31.2%	31.6%	1.05 (0.85)
3rd	26.9%	27.7%	0.94 (0.98)
Highest quartile	19.6%	19.7%	0.75 (0.95)
Number of siblings
0	30.7%	30.9%
1	48.1%	48.4%
2	15.2%	14.7%
3+	6.0%	6.0%

On average, the sampled children had .98 sibling, with a median of one. Average number of brothers and sisters were .47 and .51, respectively. Table 1 also shows that about 31% of the sampled children were from only child families and slightly less than half of the total sample had one sibling. The remaining 21% had two or more siblings.

Measures

Sibling information

CFPS provided very detailed family roster information, where a complete birth history was available for every family member. This formed the basis for the calculation of number of children. Sibship size was computed as the total number of children parents reported minus one. In the event of inconsistent reporting from both parents, the larger number would be taken. In addition to the total number of siblings, the study also distinguished between brothers and sisters, based on parent-reported gender information for each child. Birth order was calculated by comparing the birth year of the studied child to the entire age range of all his or her siblings. Birth order was then coded into a three-category (i.e., oldest, middle, youngest) variable. Note that the only child was recorded as the oldest child, and the middle child was only available in families with more than two children. Age spacing was measured by average difference in years between births of the respondent and his or her siblings.

Economic resources

Three variables of children’s economic resources were studied. Educational expenditure was reported by parents, reflecting family’s actual cost related to the surveyed child’s education during the past year. Items included tuition, books, private tutoring, lodging, transportation, and so forth. Based on values from Table 2, Chinese children between the ages of 10 and 15 had an average yearly educational expenditure of 1,308 RMB (equivalent to about US$190) in 2010, with a standard deviation (SD) of 2,253. Monthly allowance was self-reported, indicating whether the child received any monthly allowance and, if yes, the amount of it. In 2010, Chinese children aged 10 to 15 years reported an average monthly allowance of 36.06 RMB (equivalent to about US$5) with an SD of 59.44. Children’s self-reported nutritional food intake was measured by the total number of times the respondent consumed the following six types of food in a typical week during the past month: meat, fish/seafood, fresh vegetables/fruit, dairy, beans, and eggs. Based on Table 2, the average total weekly frequency was 22.77 times, with an SD of 13.03.

Table 2.

Descriptive Statistics of Child’s Eight Family Resources.

Type of resources	$\bar{X}$	SD	Minimum	Maximum
Economic resources
Educational expenditure (in RMB)	1,308.61	2,252.23	0	44,200
Monthly allowance (in RMB)	36.06	59.44	0	1,000
Weekly nutritional food intake	22.77	13.03	0	114
Interpersonal resources
Parents’ educational aspiration (years)	16.16	3.57	0	22
Parental attention	3.38	1.34	1	5
Parental supervision	19.33	4.73	6	30
Child-rated parents’ behavior	43.16	8.71	20	65
Family tutoring time (weekly hours)	2.17	4.13	0	54

Interpersonal resources

Five interpersonal resource variables were studied. Parental educational aspiration measured the highest level of education parents would hope the surveyed child to complete. Categorical responses were converted into years of education, ranging from 0 (no education) to 22 years (doctorate degree). On average, the parents of Chinese adolescents in 2010 had hoped their children to complete at least 16 years of education with an SD of 3.57 (see Table 2). Parental attention was measured by children’s responses to a question on the extent to which their parents would know where he or she is when he or she is not at home. A five-category response scale ranging from 1 (never) to 5 (always) reflected an increasing level of parental attention. In 2010, the average level of child-reported parental attention was 3.38, with an SD of 1.34. Parental supervision was computed based on a six-item scale. Parents were asked to rate the frequencies of their own behaviors in the following six aspects during the past year: giving up watching TV when the child is studying, discussing what happens with children, asking the child to finish homework, checking child’s homework, restricting the child from watching TV, and restricting the types of TV programs the child could watch. Scale scores were computed by summing up the item-level responses and the sum scores ranged from 6 to 30, with higher scores reflecting more frequent supervision. Cronbach’s alpha for the six-item scale was .70, indicating acceptable reliability coefficient. In 2010, parent-rated average parental supervision was 19.33, with an SD of 4.73. Child-rated parental behavior was computed based on frequencies of the following behaviors by parents/guardians: discussing with the child when he or she was doing something wrong, encouraging the child to make efforts, being gentle when talking to the child, encouraging the child to think independently, telling the child reasons when asking him or her to do something, liking to talk with the child, asking about what happened at school, checking homework, helping with schoolwork, telling stories, playing with the child, praising the child, criticizing the child, and attending parent-teacher meetings at school. We recoded all the 14 behaviors from 1 to 5 with higher scores associated with more frequent positive behaviors. Scale score was formed by summing all the item scores. To reduce administrative cost, this set of questions was only asked for the 11-year-olds. Cronbach’s alpha for the 14-item scale was .83. Average rating of parents’ behaviors was 43.16 with an SD of 8.71. Finally, family tutoring time was computed by adding up weekly tutoring hours for the surveyed child from all family members. In 2010, the average total family tutoring time within a week was 2.17 hours, with an SD of 4.13.

Parents’ education

Parents reported on their own education categories. In the case of non-response from individual report, data from family surveys would be used where a family respondent reported the highest degree of all family members. Father’s and mother’s education were both recoded into three categories: no schooling, some but less than 12 years of education, and 12 years of education or above (i.e., equivalent to at least a high school diploma).

Family income quartiles

The total family income included five sources: salary-based income including bonus and allowances; business-based income including farming and non-farming business; assets-based income including rent from land and other possessions as well as sales of the valuables; transfer-based income including government subsidies, pensions, and social securities; and other income including gifts from friends and relatives and all other types. Quartile scores were formed to allow for possible non-linear effects.

Children’s age, gender, urban status

In addition to the above variables, the study also controlled for children’s age in years, gender (i.e., male vs. female), urban status (urban vs. rural). Urban status was defined by the type of the community the child’s family was living in when they were sampled and may not be consistent with the type of household registration the child had.

Analyses

In order to address clustering effects due to multistage sampling and within-family clustering, data were analyzed using three-level models, where the total analytical sample of 3,073 children were nested within 2,656 households, which in turn were nested within 595 communities. All analyses were weighted by individual-level sampling weights, which accounted for both differential probabilities from multistage sampling as well as non-response rates; the final weights were also adjusted with post-stratification imposed on gender, age, and urban status to minimize sampling error (Xie, 2012). Analyses were carried out with SAS PROC MIXED (Singer, 1998) and proceeded in the following four main steps.

First, a linear effect of sibship size was examined on each of the eight outcomes by including the continuous sibship size, birth order of the sampled child, as well as all the control variables including children’s age, gender, urban residency, parental education, and family income. All outcomes were modeled as continuous variables. Second, the possible effect of age spacing was estimated by adding the average difference in ages between the respondent and his or her siblings to the models in Step 1. The age spacing analysis was done only within samples who had at least one sibling. Third, sibship size was modeled as a categorical variable instead of a continuous variable in order to allow for flexible forms of the sibship effects. The final stage of analyses used gender-stratified models to estimate the effects of number of brothers and sisters on the eight outcomes. The same set of control variables were used in the gender-stratified models as in the first step of analysis.

Results

The models controlled for a set of key demographic variables (i.e., urban status, age, gender, parental education, and family income) that are possible confounders between sibship size and family resources. The last column of Table 1 displays the average sibship size by each level of the key covariates. Not surprisingly, girls tended to have more siblings than boys, and families that had more siblings were more likely to be from rural areas and poorer. Furthermore, more educated parents, especially more educated mothers, tended to have fewer children.

Linear Effects of Sibship Size

The first three outcomes in Table 3 were economic resource outcomes. Sibship size showed consistent detrimental effects on each of them. On average, as number of siblings increased by one, yearly educational expenditure was reduced by 244.99 RMB (equivalent to about US$36), monthly allowance was reduced by about 7.04 RMB (slightly more than US$1), and weekly nutritional food intake was decreased by .88 time, net of birth order, demographic variables, and family socioeconomic conditions. A similar pattern was observed for interpersonal resources presented in the right panel of Table 2. Out of five studied interpersonal resources, four were negatively associated with sibship size. An additional sibling was associated with a reduction of parental educational aspiration by .38 year, less parental attention, less adult supervision, and lower frequencies of positive parental behaviors, but family tutoring time did not appear to be affected.

Table 3.

Linear Effect of Sibship Size on Eight Family Resources.

Covariates	Education expenditure	Monthly allowance	Nutritional food	Parental educational aspiration	Parental attention	Parental supervision	Child-rated parents’ behavior^a	Family tutoring time
Sibship size	−244.99***	−7.04***	−0.88**	−0.38**	−.11**	−0.36**	−2.18***	−0.08
Birth order
Youngest	−173.11*	3.63	0.27	0.05	.06	−0.14	1.88*	−0.37*
Middle	84.73	5.09	1.56**	0.03	−.22**	−0.34	2.07	0.13
Oldest	0	0	0	0	0	0	0	0
Rural	−672.62***	−0.87	−4.92***	−0.34	−.19**	−1.44***	−1.05	−0.57**
Age in years	126.45***	7.54***	0.38***	−0.04	.05***	−0.28***	NA	−0.50***
Male	16.78	−5.85**	−0.18	0.40*	−.48***	0.02	−1.25	0.44**
Mother’s education
No schooling	−739.27***	−11.89*	−4.72***	−0.74	−.47***	−2.15***	−5.88***	−2.50***
Less than 12 years	−598.00***	−7.85	−2.30**	0.12	−.23**	−0.80*	−2.41*	−1.71***
≥12 years	0	0	0	0	0	0	0	0
Father’s education
No schooling	−618.85***	0.48	−2.75***	−1.67***	−.15	−0.74*	−3.80**	−0.36
Less than 12 years	−516.24***	−0.84	−1.51*	−0.71*	−.08	0.31	−1.41	0.21
≥12 years	0	0	0	0	0	0	0	0
Family income quartiles
Lowest quartile	−756.74***	−15.84***	−5.19***	−0.59	−.11	0.43	−2.71*	−0.34
2nd	−636.87***	−9.69**	−4.10***	−0.41	−.09	0.43	−2.62**	−0.28
3rd	−502.19***	−4.39	−2.82***	−0.14	−.11	0.34	−0.81	−0.57*
Highest quartile	0	0	0	0	0	0	0	0

Note. Sibship size was treated as a continuous variable.

Only respondents whose age was 11 at the time of the survey answered questions related to this scale (n = 570).

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

Note that the effects of birth order were largely inconsistent across different resources. Out of the eight outcomes, youngest children fared the worst on only two resources (i.e., education expenditure and family tutoring) compared with either the oldest or middle-born children. In contrast, being the youngest child was associated with receiving more positive parental behaviors. In addition, being the middle child was associated with less parental attention, but they appeared to consume more nutritional food.

Among the other covariates, mother’s education presented the most consistent and also the strongest effects overall. Children with illiterate mothers had the least economic and interpersonal family resources, followed by children with mothers who had less than 12 years of education. Although father’s education also appeared to be important, its effects were less consistent and not as strong. The only exception was for the outcome of parental educational aspiration, where father’s education was a much stronger predictor than mother’s education.

In addition, children living in rural areas had fewer resources. Lower family income was associated with fewer economic resources but not interpersonal resources. While boys on average held higher educational aspiration from their parents and reported more family tutoring time, girls tended to receive more monthly allowance and more parental attention. Boys and girls were comparable in the remaining four outcomes.

Age Spacing

For all outcomes where sibship size had a significant effect, age spacing was estimated by adding average age difference between siblings and the results were presented in Table 4. Out of the seven outcomes, age difference showed positive effects for three outcomes (i.e., monthly allowance, parental educational aspiration, and parental attention) and null effects for the remaining four outcomes, while controlling for sibship size. Together with the negative effect of sibship size, the results indicated that closely spaced siblings had equal or greater negative effect on family resource allocation than widely spaced siblings.

Table 4.

Effect of Sibling Age Spacing on Family Resources.

Covariates	Education expenditure	Monthly allowance	Nutritional food	Parental educational aspiration	Parental attention	Parental supervision	Child-rated parents’ behavior^a
Sibship size	−135.63**	−7.25***	−.78*	−.55***	−.10*	−.44**	−1.77**
Average age difference^b	−9.82	1.70**	.11	.14***	.03**	−.01	−0.02

Note. Analysis was done on the subsample with at least one sibling. Family tutoring time was not modeled here because sibship effect was non-significant from Table 3. Sibship size was treated as a continuous variable. Other controlled variables were the same as analysis shown in Table 3.

Only respondents whose age was 11 at the time of the survey answered questions related to this scale (n = 570).

Average age difference was computed as average difference in years between births of the respondent and his or her siblings.

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

Non-Linear Effect of Sibship Size

In addition to the imposition of a linear trend, the study also allowed for non-linear effects of sibship size by modeling it as a categorical variable. Because families with more than four children were rare, this variable collapsed all categories with 3 or more siblings, which accounted for about 6% of the total sample. The results were presented in Table 5. Overall, a monotonic pattern of effects was observed with larger effects for more siblings; however, a strict linear relationship was not applicable in most cases. A differentiating pattern across economic and interpersonal resources was noted. While all economic resources showed an immediate reduction with one sibling, half of the interpersonal resource variables (i.e., parental educational aspiration and parental supervision) appeared to be unaffected until the child had two siblings.

Table 5.

Non-Linear Effect of Sibship Size on Eight Family Resources.

Sibship size	Education expenditure	Monthly allowance	Nutritional food	Parental educational aspiration	Parental attention	Parental supervision	Child-rated parents’ behavior^a	Family tutoring time
No sibling (reference)
One sibling	−509.74***	−9.09**	−1.18*	0.03	−.17*	−0.27	−3.16***	−.15
Two siblings	−698.19***	−16.82***	−1.97*	−0.44	−.25*	−0.53	−5.68**	−.04
Three or more siblings	−798.83***	−25.77***	−3.07**	−1.36**	−.37**	−1.29**	−7.06***	−.39

Note. Sibship size was modeled as a categorical variable. Other controlled variables were the same as analysis shown in Table 3.

Only respondents whose age was 11 at the time of the survey answered questions related to this scale (n = 570).

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

The Role of Gender

Table 6 displays results from gender-stratified analysis of number of brothers and sisters on the eight outcomes. For boys (results shown in upper panel of Table 6), there was a universal pattern of stronger or at least similar effects from male siblings than female siblings across all studied outcomes. Furthermore, the negative effect of sibship size on nutritional food intake became non-significant. Among girls, the results also presented a general pattern of stronger effects from brothers than from sisters for all economic resource outcomes; as for interpersonal resource outcomes, the findings were mixed. While parental attention and supervision were more affected by number of male siblings, parental educational aspiration and behaviors were more influenced by number of female siblings.

Table 6.

Gender-Stratified Analysis of Number of Brothers and Sisters on Family Resources.

	Education expenditure	Monthly allowance	Nutritional food	Parental educational aspiration	Parental attention	Parental supervision	Child-rated parents’ behavior^a	Family tutoring time
Boy
Number of brothers	−398.31***	−8.98**	−0.53	−.42	−.16*	−.85***	−2.23**	−.03
Number of sisters	−296.45***	−5.31*	−0.70	−.31	−.12*	−.88***	−1.36	−.06
Girl
Number of brothers	−374.35**	−11.10***	−1.56**	−.25	−.20**	−.52*	−2.34**	−.26
Number of sisters	−146.35**	−6.90*	−0.38	−.56**	−.03	−.03	−2.91***	−.13

Note. Number of brothers/sisters was treated as continuous variables. Other controlled variables were the same as analysis shown in Table 3.

Only respondents whose age was 11 at the time of the survey answered questions related to this scale (n = 570).

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

Discussion

The current study aimed to evaluate the effects of sibship size on various family resources for Chinese adolescents between ages 10 and 15, with a particular interest in whether brothers and sisters may present differentiating effects for Chinese boys and girls. Overall, the study found very consistent negative effects of sibship size on various economic and interpersonal resources a child would obtain from his or her family. Analyses further revealed that number of brothers posed greater threat than number of sisters in family resources available for an individual child. Of all the eight outcomes, only family tutoring time was not associated with sibship size. This may be explained by the fact that family tutoring time was computed by adding up contributing time from all family members, and children from larger families would have more resources in this regard. The same was not true for the other interpersonal variables that targeted more toward the parents or guardians.

The very consistent and negative effects of sibship on various family resources provide direct evidence for the resource dilution theory (Blake, 1981). Similar to Downey (1995), the current results based on the Chinese nationally representative adolescent sample provides evidence for not only a link between sibship size and economic resources but also interpersonal resources. Such negative effects were beyond what could be accounted for by parents’ educational background or family economic conditions. The consistent association between sibship size and various interpersonal resources is of particular concern as the amount of interpersonal resources, such as parental attention, is considered even more important than economic resources to a child’s overall well-being (Blake, 1981; Marjoribanks, Walberge, & Bargen, 1975). It is thus reasonable to expect that the negative effect of sibship may go beyond educational achievement.

The current findings on age spacing provide further evidence to resource dilution theory. As closely spaced siblings are more likely to compete for similar resources, resource dilution theory posited that siblings who were closely spaced posed greater threat to resources (Downey, 2001). The findings of either a null or positive effect of average age difference between births suggest that closely spaced siblings were either equally or more competitive in individual family resource allocation when compared with widely spaced siblings.

The current results also confirm prior findings that economic resources declined more rapidly with sibship size than interpersonal resources (Downey, 1995). Having one sibling would result in an immediate reduction in all studied economic outcomes, but most interpersonal resource outcomes would not be affected until at least two siblings were present. Given that interpersonal resources may function as more important determinants of intellectual ability than economic resources (Marjoribanks et al., 1975), this may explain why some studies did not find differences in educational achievement between children from one-child families and children from two-child families (Baydar, Hyle, & Brooks-Gunn, 1997; Blake, 1981).

With regard to gender effects on family resources, findings from this study presented mixed evidence among Chinese adolescents. On the one hand, the main gender effects from Table 2 are more in favor of gender equality, with null effects on half of the studied outcomes and mixed effects on the remaining half. In fact, when an interaction between gender and sibship size was added in an exploratory set of analyses (detailed results not presented here), most of the interaction effects were statistically non-significant, providing further support for a trend toward gender equality among this sample. On the other hand, results from a breakdown of siblings into brother and sisters (Table 6) have pointed out a clear and consistent pattern of greater influences from brothers than from sisters. A possible cohort effect may contribute to such seemingly contradictory findings. The current study focused on children between ages 10 and 15, most of whom were still covered by the compulsory educational law of China during the time of survey, but their siblings may go well beyond this age group. A stronger gender effect would be expected from college-bound children who are in need of greater financial investment from their family. Unfortunately, most of the studied outcomes were unavailable for children above age 15.

The mixed findings of the birth order effect was not consistent with a confluence model (Zajonc, 2001), which hypothesized a stronger negative effect for the later born than the earlier born. This is not surprising, given that the confluence model was the most applicable to outcomes measuring intellectual development (Downey, 2001).

While father’s education appears to be the most important predictor of educational achievement other than sibship size in some previous studies (e.g., Blake, 1981; Sen & Clemente, 2010), results from this study suggested that mother’s education may be even more important than father’s education on children’s resources. In fact, mother’s education presented the most consistent effects among all studied predictors across all outcomes. Such consistent negative effects, combined with the higher likelihood of more births from lower educated women, highlight the extremely vulnerable position of children from multiple-birth families with poor educated mothers.

The current study has several limitations. First, findings of this study are only generalizable to Chinese adolescents between ages 10 and 15. Most of those children are either in elementary schools or junior high schools and covered by China’s compulsory education law. A stronger effect of sibship size may be observed with older sample because significantly more amount of resources, especially economic resources, are required beyond the compulsory education stage. Unfortunately, the majority of the study’s variables are unavailable for children above age 15 because they were taking the adult questionnaires. Second, although a number of potential confounders were included in the model, it is still likely that some unmeasured confounders may contribute to the negative association between sibship size and various family resources.

In sum, the current study found consistent negative association between sibship size and children’s family resources, after controlling for a number of potential confounders including family income and parental education, providing direct evidence for the resource dilution theory. Furthermore, siblings who are more distant from one another in ages show lower or at least equal negative effects compared with more closely spaced siblings, which lends additional support to the resource dilution theory. While economic resources are significantly affected with only one sibling, many interpersonal resources remain unaffected until the child has more than two siblings. The negative effect of siblings on family resources is related to the gender of the sibling. In general, brothers present larger negative effects than sisters.

Footnotes

Acknowledgements

The data are from China Family Panel Studies (CFPS), funded by 985 Program of Peking University and carried out by the Institute of Social Science Survey of Peking University.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author Biography

Qiong Wu is an assistant professor in the Institute of Social Science Survey at Peking University, Beijing, China. She has extensive experiences on secondary data analyses of large-scale survey. Her research interests are quantitative methods in social sciences, educational achievement of children and adolescents, and cognitive decline of the elderly.

References

Baydar

Hyle

Brooks-Gunn

(1997). A longitudinal study of the effects of the birth of a sibling during preschool and early grade school years. Journal of Marriage and the Family, 59, 957-965.

Blake

(1981). Family size and the quality of children. Demography, 18, 421-442.

Bonesronning

Massih

S. S.

(2011). Birth order effects on young students’ academic achievement. The Journal of Socio-Economics, 40, 824-832.

Butcher

K. F.

Case

(1994). The effect of sibling sex composition on women’s education and earnings. The Quarterly Journal of Economics, 109, 531-563. doi:10.2307/2118413

Cheng

C.-C. J.

Wang

W.-L.

Sung

Y.-T.

Wang

Y.-C.

S.-Y.

C.-Y.

(2013). Effect modification by parental education on the associations of birth order and gender with learning achievement in adolescents. Child: Care, Health and Development, 39, 894-902. doi:10.1111/j.1365–2214.2012.01427.x

Chu

C. Y. C.

Xie

R.-r.

(2007). Effects of sibship structure revisited: Evidence from intrafamily resource transfer in Taiwan. Sociology of Education, 80, 91-113.

Conley

(2000). Sibship sex composition: Effects on educational attainment. Social Science Research, 29, 441-457. doi:10.1006/ssre.2000.0678

Dayioglu

Kirdar

M. G.

Tansel

(2009). Impact of sibship size, birth order and sex composition on school enrollment in urban Turkey. Oxford Bulletin of Economics and Statistics, 71, 399-426. doi:10.1111/j.1468-0084.2008.00540.x

Downey

D. B.

(1995). When bigger is not better: Family size, parental resources, and children’s educational performance. American Sociological Review, 60, 746-761.

10.

Downey

D. B.

(2001). Number of siblings and intellectual development: The resource dilution explanation. American Psychologist, 56, 497-504.

11.

de Haan

(2010). Birth order, family size and educational attainment. Economics of Education Review, 29, 576-588. doi:10.1016/j.econedurev.2009.10.012

12.

Hauser

R. M.

Kuo

H.-H. D.

(1998). Does the gender composition of sibships affect women’s educational attainment? The Journal of Human Resources, 33, 644-657. doi:10.2307/146336

13.

Hesketh

Xing

Z. W.

(2005). The effect of China’s One-Child Family Policy after 25 years. New England Journal of Medicine, 353, 1171-1176. doi:10.1056/NEJMhpr051833

14.

Jacob

(2011). Do brothers affect their sisters’ chances to graduate? An analysis of sibling sex composition effects on graduation from a university or a Fachhochschule in Germany. High Education, 61, 277-291. doi:10.1007/s10734-010-9377-8

15.

Jager

M. M.

(2009). Sibship size and educational attainment. A joint test of the confluence model and the resource dilution hypothesis. Research in Social Stratification and Mobility, 27, 1-12. doi:10.1016/j.rssm.2009.01.002

16.

Kanazawa

(2012). Intelligence, birth order, and family size. Personality and Social Psychology Bulletin, 38, 1157-1164. doi:10.1177/0146167212445911

17.

Kang

(2011). Family size and educational investments in children: Evidence from private tutoring expenditures in South Korea. Oxford Bulletin of Economics and Statistics, 73, 59-78. doi:10.1111/j.1468-0084.2010.00607.x

18.

Treiman

D. J.

(2008). The effect of sibship size on educational attainment in China: Period variations. American Sociological Review, 73, 813-834. doi:10.1177/000312240807300506

19.

Marjoribanks

Walberg

H. J.

Bargen

(1975). Mental abilities: Sibling constellation and social class correlates. British Journal of Social and Clinical Psychology, 14(2), 109-116.

20.

Marteleto

(2010). Family size and schooling throughout the demographic transition: Evidence from Brazil. Demographic Research, 23, Article 15.

21.

Sen

Clemente

(2010). Intergenerational correlations in educational attainment: Birth order and family size effects using Canadian data. Economics of Education Review, 29, 147-155.

22.

Silles

M. A.

(2010). The implications of family size and birth order for test scores and behavioral development. Economics of Education Review, 29, 795-803.

23.

Singer

J. D.

(1998). Using SAS PROC MIXED to fit multilevel models, hierarchical models, and individual growth models. Journal of Educational and Behavioral Statistics, 23, 323-355. doi:10.3102/10769986023004323

24.

Steelman

L. C.

Powell

Werum

Carter

(2002). Reconsidering the effects of sibling configuration: Recent advances and challenges. Annual Review of Sociology, 28, 243-269.

25.

Tsui

Rich

(2002). The only child and educational opportunity for girls in urban China. Gender and Society, 16, 74-92. doi:10.2307/3081877

26.

Wang

(2003). Estimation of TFR level with changing childbearing pattern in China: 1990-2000. Chinese Journal of Population Science, 4, 32-38.

27.

Xie

(2012). The user’s guide of the China Family Panel Studies (2010). Beijing, China: Institute of Social Science Survey, Peking University.

28.

Xie

(2013, August). The causal effects of rural-to-urban migration on children’s wellbeing in China (PSC Research Report No. 13-798). University of Michigan, MI: Population Studies Center.

29.

Zajonc

R. B.

(2001). The family dynamics of intellectual development. American Psychologist, 56, 490-496.