Childlessness is common in many modern Western societies (Buchanan & Rotkirch, Reference Buchanan and Rotkirch2013). For instance, 20% of Finnish women over the age of 40 have not given birth (Statistics Finland, 2011). These numbers are similar to those of many other Western countries (Hakim, Reference Hakim2005). Since it is unusual to have a first child after the age of 40, the majority of these women remain childless.

Studies have shown that childless individuals may give more support to their ascendants than parents do (e.g. Komter & Vollebergh, Reference Komter and Vollebergh2002; Albertini & Kohli, Reference Albertini and Kohli2009). This could be because they are trying to buffer themselves against social isolation (Wenger et al., Reference Wenger, Scott and Patterson2000). Even though childlessness has been an important topic in family sociology (e.g. Beck & Beck-Gernsheim, Reference Beck and Beck-Gernsheim2002), few studies have separated kin by the rate of genetic relatedness when researching helpers.

Kin selection theory predicts that individuals may increase their inclusive fitness by investing in their genetically related kin (Hamilton, Reference Hamilton1964). Individuals share on average 25% of their genes with their nieces, nephews, aunts and uncles. However, not only genetic relatedness but also the recipient's reproductive value may matter (Hughes, Reference Hughes1988). If someone tries to maximize his or her inclusive fitness, it is more effective to support kin in descending rather than ascending order. Therefore women who do not have children of their own may increase their inclusive fitness by increasing their siblings' reproductive success by in turn supporting their siblings' children (i.e. nieces/nephews). Previous studies support this supposition. Pollet and colleagues (Reference Pollet, Kuppens and Dunbar2006), for example, showed that childless women in Belgium reported more contacts with their nieces and nephews than mothers did (see also Essock-Vitale & McGuire, Reference Essock-Vitale and McGuire1985). Using historical data from the US, Pollet & Dunbar (Reference Pollet and Dunbar2008) found that childless couples were more likely to take care of their nieces and nephews than couples with children were.

Another view indicates that childless women may have more time and resources than mothers have, since childless women do not have the occasion to invest in their own children (Pollet et al., Reference Pollet, Kuppens and Dunbar2006). Thus, childless women may invest more than mothers in members of their kin, regardless of reproductive value. In contrast to this prediction, Pollet and colleagues (Reference Pollet, Kuppens and Dunbar2006) found no difference between childless women and mothers in their contacts with aunts and uncles.

Using data from the Generational Transmissions in Finland (Gentrans) project, the present study analyses whether childless women give more support than mothers to their nieces, nephews, aunts and uncles in contemporary Finland. In the spring of 2012 Statistics Finland conducted a representative survey of young adults via mail. The survey reached 1753 individuals born between 1962 and 1990 (mean=1976, SD=5.6). The present study compares childless women and mothers. Only women with at least one niece or nephew 10 years old or younger were included, since childcare is rarely provided to older children. After these exclusions the study sample included 601 women (childless women: n=187; mothers: n=414).

In the case of siblings' children, the survey asked whether respondents had looked after their nieces/nephews in the last 12 months. In the questionnaire investment in nieces/nephews was gathered separately for four of the respondents' oldest siblings, and the niece/nephew sets of the specific sibling. In the case of aunts/uncles, the Gentrans survey asked whether respondents had provided practical help to their aunts/uncles in the last 12 months. The respondents may have had four types of aunts/uncles (i.e. maternal aunts, maternal uncles, paternal aunts and paternal uncles) and the questions concerned whether the respondents had provided support to at least one member of a group. For example, in the case of maternal aunts the questionnaire asked whether respondents had provided practical help to any maternal aunt.

Separate analyses for nieces/nephews (phase 1) and aunts/uncles (phase 2) were conducted. For this purpose the datasets were reshaped for a long format so that the observations were those of the original respondents' niece/nephew sets (phase 1) and aunt/uncle groups (phase 2). Since the data were clustered, Stata's statistical software cluster option was used to calculate standard errors. Logistic regression was used to predict the kin investment. The results were determined by calculating the predicted probabilities of kin investment from the logistic regression models.

Since childless women and mothers may differ from each other, e.g. based on level of education and sociability (Abma & Martinez, Reference Abma and Martinez2006; Jokela et al., Reference Jokela, Alvergne, Pollet and Lummaa2011), and previous studies show that not all individuals invest equally in their kin (Michalski & Euler, Reference Michalski, Euler, Salmon and Shackelford2008; Pollet & Hoben, Reference Pollet, Hoben, Salmon and Shackelford2011), the present analysis controlled for several factors: respondent's birth year, education, partnership status, number of close relatives and lineage. In the case of nieces/nephews, the geographical distance to the niece/nephew set was also controlled for. Unfortunately, in the case of aunts/uncles the Gentrans survey did not collect information concerning this geographical distance.

Figure 1 shows that childless women have a greater probability of investing in their nieces/nephews than mothers (adjusted model: OR=0.62, SE=0.14, p=0.029) (−2LL=1032.9; χ²=39.4, df=9, p<0.0001; Nagelkerke's R ²=0.086). Figure 2 shows no significant difference between childless women and mothers in their probability of investing in aunts/uncles (adjusted model: OR=0.79, SE=0.21, p=0.388) (−2LL=962.0; χ²=20.5, df=8, p=0.009; Nagelkerke's R ²=0.040). In addition, those with more close relatives have greater odds of investing in their nieces/nephews and aunts/uncles. Respondents are more likely to invest in their nieces/nephews via sisters than brothers (Table 1).

Fig. 1. Women's investment in nieces/nephews (predicted probabilities and 95% confidence intervals).

Fig. 2. Women's investment in aunts/uncles (predicted probabilities and 95% confidence intervals).

Table 1. Predicting investment in nieces/nephews and aunts/uncles: results from two logistic regression models (odds ratios and 95% confidence intervals), Finland 2012

* p<0.05.

To conclude, childless women tend to invest more than mothers in their nieces and nephews but not their aunts and uncles. The results are in line with the reproductive value prediction (Hughes, Reference Hughes1988) and a previous study by Pollet and colleagues (Reference Pollet, Kuppens and Dunbar2006). The present study has certain limitations that highlight the importance of future research. First, due to the data limitations it was impossible to separate voluntary from involuntary childless women. However, it remains important to study whether voluntary and involuntary childless women differ from each other in the case of kin support. Second, future studies should also investigate whether those who have supported their nieces/nephews receive more support from them in their old age. Third, there is room for studies concerning the outcome of kin support.