1.1. Dolls and toy cars: strongly gendered play in all social categories

The composition of children’s toy cupboards is already clearly gendered by their first birthday [3], and boys and girls already play differently. These differences are confirmed at age 2 by the frequency of various kinds of play. Excepting water games (bath toys, water slides, or circuits), gender differences are observed for all other types of play (Table 1). Girls more frequently play with dolls (81%), draw pictures (73%), or play with soft (stuffed) toys (63%), while boys do so much less and more frequently play with toy cars (89%), balls (76%), or stacking toys (61%).

Some forms of play are more gendered than others, such as playing with dolls and toy cars, with a 60 percentage-point difference between the share of girls and boys who play with them daily or often. While significant, the gender differences are smaller for drawing and playing with soft toys (15 and 11 points, respectively, in favour of girls) and for ballgames (19 points, in favour of boys). They are even smaller for stacking games, more frequently played by boys (7 points), and for less widely played games such as jigsaw puzzles (5 points in favour of girls).

These differences observed among young children reflect the structural opposites known to exist in adults. The typical distinction between ballgames or toy cars (for boys) and dolls or cooking (for girls) exemplifies the contrast between outdoor and public activities (generally a masculine domain) and indoor and domestic activities (generally a feminine domain). But it especially illustrates a series of opposites that structure gender relations in the workplace [4]: boys’ play tends to develop their technical and building skills, and their ability to act upon objects (toy cars, stacking games, etc.), while girls’ play tends to focus on developing relational or care skills (symbolized by dolls), or physical appearance (dolls’ bodies and accessories).

These gender differences are observed across all social categories, from the bottom to the top of the social scale. In fact, the amplitude of the gender difference varies little by social category (2–3 percentage points). It is highest for playing with dolls (4 points), increasing from 60 percentage points in favour of girls among upper-class families to 64 points among middle-class families (Table 1).

1.2. Younger siblings’ play is less gendered

While gender differences are similar across all social classes, they vary according to the number of siblings and the birth order of the 2-year-old child. Regarding the two most gendered activities—playing with dolls and toy cars–the difference between girls and boys is much greater among children with no older siblings (either first-born or only children) than among younger siblings (Figure 1). For these two activities, the gender difference is 67 and 61 points, respectively, among first-borns, versus just 59 and 54 points among younger siblings.

This narrowing of the gender difference by birth order is explained mainly by the fact that younger siblings have a wider range of play activities than first-borns and are more inclined to engage in games typical of the other gender. For example, younger brothers play daily or often with dolls almost twice as frequently as first-born boys (24% vs. 13%), and 35% of younger sisters play daily or often with toy cars compared with just 28% of first-born girls. Why are younger siblings more inclined to play with toys typically associated with the other gender?

The sociological literature on the atypical trajectories and gender practices of women who engage in ‘masculine’ activities (such as football, boxing, or engineering) identifies two explanatory processes: first, a missing-boy effect whereby certain girls (with no siblings or sisters only) are encouraged and supported by their fathers in these activities, and second, for girls with siblings, a process of socialization by their brothers. The first process is not observed statistically in our study: first-born girls and those with only older sisters rarely play with toy cars. No particular missing-boy effect is observed (Figure 2). However, ‘diagonal’¹ socialization by older siblings explains the increase in gender-atypical play among younger siblings when older siblings are of a different gender. A sibling training effect of older siblings on younger siblings is observed, with the latter more often engaging in play generally associated with the opposite gender, either through imitation of their elders, or through direct play with them. Conversely, when older siblings are all the same gender as the younger sibling, his or her play is even more stereotypically gendered.

Hence, boys play with dolls much more frequently if they have at least one older sister, with a gender gap that narrows from 68 points for younger siblings with brothers only, to 52 points for those with sisters only. The pattern is symmetrical and similar for playing with cars. Girls play with cars more often if they have at least one older brother and much more often if they only have brothers; gender differences are smaller when there are siblings of both genders. These results show that it is the presence of an older sibling of the opposite gender (and its effect on brother–sister interaction), rather than status of younger sibling in itself, that reduces gender differences in the propensity to play with cars or dolls often or every day.

1.3. The sibling training effect on younger siblings through play

That said, the presence of an older sibling of the opposite gender is not sufficient to cause younger siblings to engage in gender-atypical play. If this were the case, play preferences would simply reflect the availability of the older sibling’s toys in the home. Older brothers and sisters need to play with their younger siblings every day or often for this sibling training effect to occur. For example, among younger sisters with at least one older brother,² those who play with him every day are almost twice as likely to play daily or often with toy cars as sisters who do not play with their brother, or only occasionally (46% vs. 26%, Figure 3). Likewise, younger brothers who play every day with their older sister(s) are 3.5 times more likely to play frequently with dolls than those who never play with their sister(s) (33% vs. 9%).

Thus, for boys and girls alike, only having siblings of the opposite gender and playing with them every day are major factors in the adoption of gender-atypical play. This sibling training effect exists across all social classes, and whatever the degree of parental involvement in the youngest children’s activities.

1.4. Less stereotypical play when the mother plays with the child

This does not signify, however, that parental involvement has no impact on children’s play but simply that siblings exert a stronger influence. For boys, the probability of playing almost daily with dolls increases by almost 7 points if the mother plays with him every day (vs. from occasionally or never), but by 14 points when he plays every day with his sister(s) (vs. never). The fact that siblings play a greater role than parents is mirrored, in similar proportions, among girls with their brothers and fathers.

Moreover, mothers tend to motivate gender-atypical play more strongly than fathers: for girls, playing daily or often with their mother increases the likelihood of playing with toy cars by 8 points, while for boys, playing daily or often with their father has no significant effect on their playing with dolls. This difference reflects the fact that fathers tend to engage in gender-typical play with their children, while mothers favour both gender-typical and atypical play.

***

Alongside parents and other adults, siblings thus have a major influence on small children’s gendered play, with older siblings exerting a sibling training effect on younger siblings via their preferred toys (dolls for girls and cars for boys). This may narrow or widen gender differences in play, depending on the sibling gender composition. When younger siblings have a majority of siblings of the opposite gender, their play is more diverse and more often resembles the gender-typical play of their older brothers and sisters. Conversely, boys with only older brothers or girls with only older sisters tend to engage in play that is more typical of their own gender (and of their older siblings). The sibling training effect is stronger among younger siblings because their parents are less available, so they more often emulate their elders in play that does not require adult supervision. With no older siblings, first-borns cannot, by definition, experience this sibling training effect. However, parents spend more time, on average, playing with them [5] and tend to favour less gendered, more ‘educational’ games such as jigsaw puzzles, drawing, stacking games, etc.

ELFE cohort data will enable us to track how these gendered play patterns at age 2 evolve over time. The observed gender differences may be affected by changes in children’ living conditions or their social environment, such as the birth of a new sibling or entry into nursery school.