Superhero toys and boys' physically active and imaginative play.

Abstract. The influence of superhero versus nonsuperhero toys on boys' physically active and imaginative play was studied in 29 dyads (n = 58 middle-class preschool boys; M age = 54.95 mos, SD = 5.28 mos). Each dyad participated in two play sessions: 1) superhero toys (media related) and 2) nonsuperhero (nonmedia related) toys. Dyads were observed for the frequency and quality of roles, themes, level of physical activity in their play, and aggression. Parents reported on their children's favorite television shows. Results indicated that dyads engaged in more character/fictive (media related) roles in the superhero condition and more frequent familial, occupational, exploration, and no roles in the nonsuperhero condition. Dyads were more physically active in the nonsuperhero than superhero condition, and no aggression or violence was observed in either play session. Findings are discussed in light of children's development and implications for educators and parents.

**********

The deregulation of children's television in 1984 led to an onslaught of marketing television programming with associated products (Levin & Carlsson-Paige, 1995b). The broadcasting media and the toy industry virtually bombarded youngsters with marketing for products associated with television shows and movies (e.g., action figures, other toys, games, lunchboxes, t-shirts, school bags) (Kline & Pentecost, 1990). The media's focus is on capturing and maintaining children's attention and extending this influence into their play. Parents, educators, and professionals have voiced concerns regarding the endorsement of promotional toys and the influence that they may have on young children's pretend play. They wonder about the influence of superhero versus nonsuperhero toys on boys' physically active and imaginative play. Some researchers and practitioners warn about the violent and aggressive nature of superhero play (Carlsson-Paige & Levin, 1990; Kostelnik, Whiren, & Stein, 1986; Kuykendall, 1995); however, little empirical data links superhero play and aggression. Rather, the literature consists of anecdotal reports of aggression based on educator surveys or studies that suffer from serious methodological problems (Boyd, 1997). Thus, the present study investigated the influence of superhero toys on the quantity and quality of 4-year-old boys' physically active and imaginative play, specifically examining 1) the frequency of superhero play, 2) themes and roles enacted, 3) level of physical activity, and 4) incidence of aggression.

Early childhood theorists argue that dramatic (pretend) play provides an important context within which children can express themselves, deal with the many complexities of their world, and develop cognitive, language, physical, and social skills (Johnson, Christie, & Wardle, 2005; Pellegrini & Bjorklund, 1998; Rubin, Fein, & Vandenberg, 1983). Fantasy or imaginative play contributes to social maturity and the construction of diverse aspects of cognition (Berk, 1994). In fact, dramatic play occurs daily in the lives of young children, as they are constantly involved in transformations and imitation of people, animals, and machines. Dramatic play encompasses children's ability to interact with the physical and social environment, engage in object transformations, and adopt roles to act out known stories, real or imaginary, as well as to develop new stories (Mellou, 1994).

Superhero Play: What Is It?

Superhero play is a special type of fantasy play that refers to "the active physical play of children pretending to be media characters imbued with extraordinary abilities, including superhuman strength or the ability to transform themselves into superhuman entities" (Boyd, 1997, p. 23). There is no allusion to aggression or violence, but rather the child is engaged in a pretend adventurous theme that may include physical activity. Boys, rather than girls, predominately engage in superhero play (Boyd, 1997; Pellegrini & Boyd, 1993) and, according to Bauer and Dettore (1997), find it appealing for several reasons: 1) children are primarily intrigued by the powers of superheroes, 2) superheroes demonstrate extraordinary strength and are capable of incredible accomplishments, 3) superheroes possess powers children may wish they had. Thus, superhero play may be appealing to young boys because it allows them to assume roles and explore different personas. Three dominant themes in superhero play have been identified: 1) capture and rescue, 2) submit or vanquish, and 3) attack or flee (Carlsson-Paige & Levin, 1990).

Children's involvement in superhero play has often concerned parents and educators (Greenberg, 1995; Gronlund, 1992; Levin & Carlsson-Paige, 1995a). According to Boyd (1997), the number of articles on the subject of superhero play from 1990-95 was twice that published from 1985-90. This increase may reflect, in part, educators' concern for the safety of children and for themselves, as well as worries about the violence permeating society and television and its effects on the themes of children's pretense. Television viewing has become the central activity in most North American homes (approximately 7.5 hours/day), and much programming contains frequent and excessive violence (Asamen & Berry, 1993; Simmons, Stralsworth, & Wentzel, 1999). Moreover, Waters (1993) found that 82 percent of the programs children watch contain at least some violence, and superhero shows, in particular, contain at least 32 violent acts per hour. These statistics may be the driving force behind the concerns adults express about superhero play (Carlsson-Paige & Levin, 1990).

Several studies have examined associations between superhero play and aggressive or violent actions. Boyatzis's (1997) parent survey indicated that the more parents reported girls watched the Power Rangers show, the more aggressive the girls were rated by parents and educators; associations for boys were in the same direction, but not significant. In a national survey of early childhood educators' experiences with the Teenage Mutant Ninja Turtles (Levin & Carlsson-Paige, 1995a), 97 percent of educators reported negative effects on behavior and interpersonal relations; specifically, they reported an increase in aggression and rough behavior. In an experimental study, Kaplan (1998) investigated whether children who watched the Power Rangers television show would engage in aggressive acts; he concluded that children who watched this show engaged in aggressive acts seven times more frequently than did those who did not. Boyatzis, Matillo, and Nesbitt (1995) reported that after watching only one episode of the Power Rangers show, boys committed aggressive acts seven times more frequently than control children. While the short-term experimental situation may have influenced the boys to imitate the Power Rangers by using karate chops and flying kicks, whether such acts were aggressive or merely examples of play fighting or rough and tumble play is not clear.

In a more controlled observational study, Boyd (1997) reported that only 2 out of 17 children in a preschool classroom exhibited superhero play over a one-month period. In her second sample of 16 children, only 5 percent were observed engaging in superhero play. Interestingly, in both samples, boys were the only superhero players and there were no incidences of physical harm or violence observed. These findings contradict previous work (Boyatzis et al., 1995; Carlsson-Paige & Levin, 1990) and suggest that superhero play may not always involve aggressive acts.

Yet, many educators and parents worry that by engaging in superhero play or by watching others being aggressive on television, children will become desensitized towards violence and become more tolerant of real-life aggression. In fact, some literature indicates that viewing violence on television is associated with increased aggression (see Beckman, 1997; Simmons et al., 1999, for reviews); however, it is not clear if superhero play has the same effect on children as more realistic television violence. Furthermore, Tobin (2000) criticized media research for analyzing the content of children's television shows and movies rather than investigating children's talk about television/movie text to determine their ability to distinguish reality from fantasy. Although a similar criticism can be directed to the findings reported above, further research is required before reaching any conclusions.

In sum, evidence of an association between superhero play and aggression is mixed at best. As Boyd (1997) argues, there is little valid data examining the association between violence, aggression, and children's superhero play, nor is there empirical data on the value of superhero play. In fact, Boyd (1997) criticized the extant work because it mostly relied on anecdotal reports and surveys of preschool/school-age educators (Boyatzis, 1997; Levin & Carlsson-Paige, 1995a) rather than on empirical data. We know little about the themes and roles of superhero play and if children, in fact, do engage in more aggressive acts as they play with superhero toys in a typical early childhood classroom. Finally, we need to consider the role of toys in influencing children's play.

Influence of Toys on Play

Toys have been conceptualized as being either low realism/structure (e.g., blocks) or high realism/structure toys (e.g., realistic dolls, superheroes, cooking pots) that may represent items in the real world (McLoyd, 1983). Toys do not independently elicit children's behavior, but rather play and language generally follow the themes suggested by the materials (Howe, Moiler, Chambers, & Petrakos, 1993), especially with functionally explicit or realistic toys such as a doctor's kit. By contrast, when children choose less realistic toys, their play is more varied but shorter in duration (Pellegrini, 1987).

Children also play longer and in more complex ways in same-gender groups and with gender-preferred toys (Rubin, Fein, & Vandenburg, 1981). Valued toys and social groups may elicit complex play and language because children are motivated to interact with peers around preferred toys. Boys are more attracted to fantastic games and warlike male characters (e.g., soldiers, pirates, outlaws) than girls are (McLoyd, 1983; McLoyd, Warren, & Thomas, 1984; Tobin, 2000; Wegener-Spohring, 1989), and they demonstrate greater enjoyment in their fantasy play (Pulaski, 1973).

Given the literature indicating that boys engaged in superhero play more frequently than girls, we chose to focus only on the play of boys with realistic toys (superhero figures versus more generic but realistic nonsuperhero figures). Males have been described as being physically active in their play, as well as engaging in more rough and tumble play than girls (Pellegrini & Boyd, 1993; Pulaski, 1973). Although Fein (1989) reported that physically active play accounted for only 10 percent of behavior in child care, boys who were engaged in superhero or rough and tumble play exhibited higher levels of physical play, wrestling, and physical assault on objects than girls (Maccoby, 1988).

The Present Study

The present study examined the influence of superhero toys (realistic character toys derived specifically from television or film superheroes) and nonsuperhero toys (generic realistic nonsuperhero characters) on the frequency and quality of 4-year-old boys' physically active and imaginative play. Specifically, we investigated 1) the frequency of superhero play, 2) the themes and roles enacted, 3) the level of physical activity, and 4) the frequency of aggression. Boys were placed in dyads with preferred play partners and were videotaped in a private space outside the regular classroom that allowed for physically active play. Each dyad was given two sets of play materials for two 8-minute play sessions: 1) superhero and 2) nonsuperhero figures.

First, taking into account boys' preferences for fantastic games (McLoyd, 1983) and warlike materials (Wegener-Spohring, 1989), we predicted that participants would engage in a higher number of character/fictive roles with superhero than with nonsuperhero toys. Second, based on the literature revealing that more varied themes and fantastic roles were elicited by highly realistic materials and that play and language tend to follow themes suggested by materials (Pulaski, 1973), it was predicted that there would be a higher number and a greater variety of themes observed in the nonsuperhero condition. Third, given that boys display frequent playful physical activity, including wrestling, jumping, and shouting (Maccoby, 1988), it was expected that the dyads would be involved in higher levels of physical activity with the superhero toys; we did not, however, expect to observe aggressive acts.

Method

Participants

Fifty-eight boys attending one of 12 English-speaking child care centers in a large bilingual (French, English) metropolitan area (population = 3,000,000) participated. The boys' M age = 54.95 mos (SD = 5.28 mos) and all attended child care full-time. All parents (80 percent mothers) completed a demographic questionnaire; years of post-secondary education (M = 4.10 yrs, SD = .79 yrs) and job descriptions indicated a middle-class sample (Hollingshead, 1975). Children came from bi- or trilingual homes (47 percent) and spoke English, French, and/or a third language.

Procedure

Once parental permission was obtained, the first author visited each classroom to familiarize herself with the setting and to establish rapport with the children. A peer nomination method was used to create dyads by pairing boys who nominated one another as someone they liked to play with a great deal (Asher, Singleton, Tinsley, & Hymel, 1979). Only boys with parental permission participated in the nomination process. If children nominated more than one peer, they were paired with the first child who nominated them. This procedure ensured that dyads were composed of preferred playmates.

First, the dyads participated in a 5-minute warm-up session in a separate room that ensured some privacy and included a clear space on the floor that allowed for physical activity. The warm-up session consisted of an open-ended game (farm set) that children could set up and play with as they wished. Once the boys were at ease with the surroundings, two counterbalanced, 8-minute play sessions (superhero or nonsuperhero materials) were introduced and videotaped. For example, the superhero characters were placed on the floor in random order and the dyad was told that they could play with the toys as they wished. After eight minutes, these materials were removed and the second set was introduced. The warm-up and two play sessions transpired on the same day and lasted approximately 20-25 minutes per dyad. Finally, the parental questionnaire, which assessed demographic information and television watching habits, was sent home, and parents returned it to the child care center director.

Materials

The researcher provided 22 figurines: 11 superhero figurines, such as Batman, Spiderman, and Power Puff girls, and 11 nonsuperhero or nonmedia-related figurines, such as generic female dolls (see Table 1). Each play condition included 1 female and 10 male figurines. The two conditions were balanced for the same number and types of figures--for example, the Spiderman superhero toy was matched with a comparable nonsuperhero character in terms of size and similar features.

Measures

Parental Questionnaire. Parents answered a short questionnaire assessing demographic information (e.g., number of children, level of education) and children's television viewing habits (e.g., number of hours of television watched weekly, favorite shows).

Play Measures. The following information was coded from the videotaped play sessions:

(1) Types of roles. The eight categories for types of roles enacted during pretense were based on a modification of Garvey and Berndt (1975) and McLoyd (1983). These roles were 1) character/fictive, 2) familial, 3) functional, 4) peripheral, 5) occupational, 6) exploration/negotiation, 7) other, or 8) no role (see Table 2 for definitions and examples). For each 15-s interval, the type of pretense role that each dyad engaged in for the majority of the interval was coded. If more than one role was displayed during the interval, the role with the longest duration was coded. If two roles of equal length occurred, the role that occurred first was coded. This procedure also was followed for the coding of themes and level of physical activity.

(2) Play themes. This measure assessed the type and variety of themes the dyads engaged in during the sessions. Categories of themes were derived from Garvey and Berndt (1975), McLoyd (1983), and Carlsson-Paige and Levin (1990) and were: 1) action/battling, 2) capture/rescue, 3) submit/vanquish, 4) attack/flee, 5) domestic, 6) nurturing, 7) building, 8) performing, 9) exploration/negotiation, 10) other, or 11) no theme (see Table 2 for definitions and examples). Themes were recorded for every 15-s interval during the play session and the theme observed for the majority of the interval was coded.

(3) Level of physical activity. This measure assessed the level of physical activity of the dyad during each 15-s interval by using a 5-point rating scale: 1 = no physical activity to 3 = moderate activity to 5 = high physical activity (see Table 2 for definition and examples).

(4) Frequency of physical and verbal aggression. The frequency of physical and verbal aggression was coded during each 15-s interval (see Table 2 for definitions and examples). A tally also was calculated for children's use of war materials during each 15-s interval (e.g., pretend guns, swords); however, the children were not given toy weapons.

Interrater Reliability

The first author and a second, independent person unfamiliar with the study's goals conducted interrater reliability on 5 of the 29 dyads (17 percent) for coding of themes, roles, physical activity, and aggressive acts. The independent coder was trained for 10 hours prior to coding. Interrater reliability for aggression was calculated by computing the sum of all agreements divided by the sum of agreements plus disagreements (96 percent). Reliability for roles and themes was calculated with Cohen's kappa (k = .81). Reliability for the physical activity ratings was calculated with Spearman-Brown Formula (r = .83), which is appropriate for rating scale data.

Results

Results for children's play will be presented as follows: 1) descriptive statistics; 2) comparing play (i.e., roles, themes) in the superhero and nonsuperhero sessions; and 3) associations of roles, themes, and aggression with type of favorite television program.

Descriptive Statistics

Means and standard deviations for the roles and themes are found in Tables 3 and 4. It was apparent that children engaged in a range of roles and themes. Although no incidents of physical or verbal aggression were observed in either play condition, a low frequency of play with aggressive play objects (e.g., pretend guns, swords) was observed (superhero condition M = .31, SD = 1.00, range 0 - 5; nonsuperhero condition M = .17, SD = .66, range 0 - 3).

Roles in the Superhero and Nonsuperhero Conditions

We predicted that a higher frequency of character/fictive roles would occur in the superhero than in the nonsuperhero toy condition. A paired sample t-test indicated a significant difference in the frequency of character/fictive roles in the two play conditions (see Table 3). Specifically, a higher number of character/fictive roles were observed with the superhero toys, thus providing support for the hypothesis. Next, paired sample t-tests detected that there were more familial/relational roles, occupational roles, exploration/negotiation roles, and no roles observed in the nonsuperhero than the superhero condition. Functional and peripheral roles did not differ in the two conditions. In sum, the pattern of findings supported hypothesis 1; in other words, that providing boys with superhero toys would be associated with a greater frequency of character/fictive roles. In contrast, more familial, occupational, exploration/negotiation roles, and a lack of roles were evident with the nonsuperhero toys.

Themes in the Superhero and Nonsuperhero Conditions

This hypothesis stated that the nonsuperhero condition would be characterized by a greater frequency of themes and a greater variety of themes. To test for frequency of themes, a series of paired sample t-tests were conducted between superhero and nonsuperhero themes (see Table 4). Findings revealed that the building theme occurred significantly more frequently in the nonsuperhero than in the superhero condition. Two trends emerged and suggested a greater occurrence of nurturing/caring themes and domestic/housekeeping themes in the nonsuperhero condition. Interestingly, there were no significant differences between the two play conditions in the remaining themes and, overall, the hypothesis was not supported.

To test for variety of themes, the total variety of themes for each dyad was calculated for both conditions. Using the total scores, a paired sample t-test indicated that there was no significant difference between the number of different themes observed in the two play conditions, t(28) = -.49, ns; superhero M = 3.48, nonsuperhero M = 3.62.

Physically Active Play in the Superhero and Nonsuperhero Conditions

This hypothesis predicted that dyads would exhibit higher levels of physically active play in the superhero condition. The total score for physical activity in each condition was calculated and tested by using a paired samples t-test. Contrary to the hypothesis, greater physical activity was rated during the nonsuperhero than the superhero condition, t(28) = 2.55, p < .01; superhero M = 45.34, nonsuperhero M = 49.58.

Physical and Verbal Aggression and Use of Aggressive Play Objects

First, the rate of physical and verbal aggression exhibited by the dyads and, second, the use of aggressive play objects were compared in the two conditions. Overall, no aggression was observed in either play condition. The findings for use of aggressive play objects (i.e., pretend guns, swords) indicated a low frequency overall and a nonsignificant difference between the two conditions, t(28) = .66, ns; superhero M = .31, nonsuperhero M = .17.

Favorite Type of Television Programming and Roles, Themes, and Physical Activity

Additional analyses were conducted to determine associations between the child's favorite type of television program and roles, themes, and level of physical activity. An individual score was calculated for each boy, based on his parents' report. Specifically, the child's favorite type of television program was calculated (superhero or nonsuperhero). Pearson correlations were employed to analyze the relationship between favorite type of television show and the frequency of 1) roles, 2) themes, and 3) level of physical activity. For the superhero condition, findings revealed no significant correlations between any of the roles and television programs that the children watched on a weekly basis, r(58) = -.10 to .11, ns. For the nonsuperhero condition, the only significant correlation indicated that children who preferred nonsuperhero programs were more likely to engage in peripheral roles, r(58) = -.29, p < .05. Next, correlations were conducted between the themes and favorite type of television show in both conditions. In the superhero condition, children who preferred superhero programming were more likely to engage in attack/battling themes when provided with superhero toys, r(58) = -.36, p < .01. In the nonsuperhero condition, boys who preferred superhero programs were more likely to engage in attack/battling themes with nonsuperhero toys, r(58) = -.48, p < .01. Finally, there were no significant correlations in either toy condition between type of television program and physical activity, rs(58) = -21 to .11, ns.

Discussion

The mass media has been postulated to play an instrumental role in children's play in relation to choice of toys, physical activity, pretend themes, and roles. Thus, the purpose of the present study was to investigate the influence of superhero versus nonsuperhero toys on boys' physically active and imaginative play. Furthermore, analyses examined the frequency of aggression and associations between types of preferred television programs and children's play.

Impact of Superhero Toys on Roles

The first hypothesis--that a higher frequency of character/fictive roles would occur in the superhero toy condition--was supported; specifically, superhero toys were associated with character/fictive and exploration/negotiation roles. Children's choices of roles may have been influenced by the media characters, which supports research indicating that older preschoolers are more interested in enacting roles that are less familiar (e.g., fictional and novel occupational roles) than in the typical domestic/familial roles (Garvey & Berndt, 1975). It is also possible that children were imitating the roles that they had associated with the characters from particular television programs. This argument is in line with Carlsson-Paige and Levin's (1987, 1990) work showing that children use television-based toys to imitate television images and behaviors. Alternatively, children may have been exploring the roles that they associated with these characters on a more personal level, in an attempt to build an understanding of life experiences (Piaget, 1962). Children may adopt points of view different from their own to learn how actions affect others. Clearly, such speculations require further study.

Gender also may be an important factor. Since the dyads were all male and they were friends, this may have encouraged the boys to engage in character/fictive roles, because they shared similar interests. According to Carlsson-Paige and Levin (1987, 1990), some girls were attracted to war and superhero play. Typically, however, boys were more interested in this type of play and in superhero toys (Levin & Carlsson-Paige, 1995a, 1995b). Clearly, our findings support Wegener-Spohring's (1989) work, which showed that warlike male heroes (e.g., soldiers, pirates) were extremely popular characters for males.

Interestingly, more familial, occupational, and exploration/negotiation roles were observed in the nonsuperhero condition, which supports Pulaski's (1973) research that less-structured but realistic toys allow greater freedom for children to manipulate and extend their play and engage in less repetition. In addition, our findings suggest that the more generic toys created an environment wherein children had to negotiate and establish a shared understanding regarding play, in particular regarding the possible roles and themes that may develop (Goncu, 1993). Without this kind of shared understanding, it may be challenging for children to move the scenario forward into pretense. This perspective supports Goncu's (1993) claims that imaginative play is an integrative tool influencing peer cooperation and understanding. In contrast, when presented with superhero figures that are accompanied by predefined media-scripted scenarios, the dyads focused mainly on superhero roles and engaged in significantly fewer types of roles.

Impact of Superhero Toys on Themes

Although it was expected that there would be a greater variety and frequency of themes in the nonsuperhero condition, this hypothesis was not fully supported. The building theme was observed more frequently with the nonsuperhero toys, and trends indicated that boys also engaged in more nurturing and domestic themes. Perhaps children were more creative with developing themes with the nonsuperhero toys (Pulaski, 1973), although boys did not engage in a significantly greater variety of themes in the nonsuperhero condition, as predicted.

Overall, the most popular theme that emerged was action/battling, regardless of the toys provided, which supports the literature (Wegener-Spohring, 1989). Apparently, these boys enjoyed battling types of themes, which may have been reinforced by the fact that they were interacting with a preferred male playmate. This finding was particularly evident for boys whose preferred television program was a superhero program, as we discuss below.

Impact of Superhero Toys on Boys' Level of Physical Activity

Contrary to our predictions, the level of physical activity was significantly higher in the nonsuperhero condition, suggesting that the boys engaged in more physical activity with the nonsuperhero than with the superhero toys. This finding was in contrast to previous literature indicating that superhero play was associated with aggressive and very physically active play (Boyatzis, 1997; Boyatzis et al., 1995; Carlsson-Paige & Levin, 1990; Kaplan, 1998). The level of physical activity observed in both conditions suggests that the boys engaged in a moderate amount of running, chasing, and mock fighting during play, perhaps reflecting the rough and tumble play characteristic of boys (Pellegrini, 1987). Certainly, this interpretation is in line with the finding that the predominant theme in both play sessions was action/battling. Perhaps the level of physical activity was moderate in both conditions as a result of the toys' novelty. In sum, the boys were physically active in both conditions, suggesting that these children may not necessarily need media-imbued characters to engage in high levels of physical activity.

Aggression and Superhero Play

Significantly, no physical and verbal aggression was recorded in either play condition, which supports Boyd's findings (1997). Apparently, superhero and physically active play does not necessarily result in aggression or violence and, in fact, may represent a subcategory of rough and tumble play (Pellegrini & Smith, 1998). These findings are contrary to many educators' views concerning superhero play (Boyatzis, 1997; Carlsson-Paige & Levin, 1990); further study is warranted to clarify why educators view superhero play as aggressive.

A small number of dyads employed aggressive pretend objects (i.e., sword, gun, weapons) in both play sessions, but there was no significant difference between the two conditions. Overall, the frequency was very low, although a slightly higher number of aggressive play objects was observed in the superhero condition and perhaps reflects the view that superhero play incorporates the use of weapons (Carlsson-Paige & Levin, 1987, 1990). Interestingly, no aggression ensued from the use of these imitative play objects as children incorporated imaginary swords or guns as props for the themes that they created, such as Star Wars (lasers) or cops and robbers (guns). This may be a case of imitation--namely, that children were at tempting to re-create Spiderman or Batman characters in their play. Nevertheless, boys used make-believe aggressive play objects in both conditions, suggesting that superhero toys alone do not necessarily encourage children to incorporate aggressive play objects into their pretense.

Impact of Television Viewing on Roles, Themes, and Physical Activity

Additional analyses conducted on type of television program viewing and roles, themes, and physical activity offer new insights. Interestingly, type of television program watched was not related to roles in the superhero session; in the nonsuperhero condition, however, a negative correlation suggested that the children who watched nonsuperhero television shows engaged in more peripheral roles. Furthermore, children who were reported to prefer superhero shows engaged in more attack/battling themes with the superhero toys, whereas children who preferred nonsuperhero television shows engaged in fewer attack/battling themes with the superhero toys. This pattern of findings also was evident in the nonsuperhero play condition and perhaps indicates that children's preference for types of television programming was associated with particular themes in their play. Perhaps television programs encourage children to adapt certain play themes to specific characters (Asamen & Berry, 1993; Beckman, 1997). Level of physical activity was not associated with the television programming. Overall, the type of television program that the child was reported to watch on a weekly basis was weakly associated with roles, themes, and physical activity.

Implications for Parents and Educators

Early childhood theorists as well as educators understand that play is one of the most important resources that children have for achieving emotional and intellectual equilibrium and growth (Carlsson-Paige & Levin, 1995). Given boys' interest in superhero play and the value of pretend play for children's development (Johnson et al., 2005), it seems shortsighted not to allow them opportunities to express themselves accordingly. Clearly, for some children, superhero play is an integral facet of their social, emotional, cognitive, and language development. Providing opportunities to engage in superhero play opens up a multitude of creative possibilities and allows children the freedom to explore their world with a sense of empowerment and control.

Our results have practical implications for parents and educators and illustrate that superhero toys may indeed influence children's choice of pretense roles, especially character/fictive roles. We note that play with superhero toys appeared to capture and sustain the young boys' attention throughout the play session, whereas the nonsuperhero toys elicited more nonplay behavior. This finding offers new evidence to suggest that superhero toys can serve to extend children's fantastic games, roles, and themes (Howe et al., 1993). Yawkey (1979) suggested that environmental materials provide depth and direction to role-playing and also serve to provide a context for the play. Therefore, we recommend that educators incorporate various play materials, such as superhero toys and props (e.g., capes, boxes, blocks), to encourage children's creativity in dramatic play (Mellou, 1994). Designating a specific corner in the classroom to highlight the children's interests in Batman or Spiderman also can elicit more in-depth and concrete forms of fantasy play. Such a center also should include opportunities for fine and gross motor activity through the direct manipulation of materials. Encouraging children to collaborate and build various structures to extend their play (e.g., building a cave for Batman using a box, or playing with blocks) can facilitate their social, cognitive, and motor development. These kinds of activities provide both boys and girls with excellent opportunities to engage in socializing and perspective-taking. Educators also can extend children's superhero play experiences to an outdoor area with sufficient space for running, jumping, and tumbling. With specific guidelines, rules, and supervision, children may increase cognitive flexibility, as well as develop more refined gross motor skills, as they take on the specific roles and themes associated with the play characters. Outdoor play also affords opportunities for children to release pent-up energy.

Educators can channel children's interests in superhero characters into different areas of the curriculum. For instance, educators can encourage children's interest in Batman by extending the play scenario into constructive and reflective play (e.g., creating a home for Batman, reflecting on what types of food Batman needs to eat to stay healthy, dealing with safety issues and aggression, or deciding what part of the world Batman lives in). Educators can intervene in the play and assist the children in expanding and elaborating their play (Carlsson-Paige & Levin, 1987; Levin & Carlsson-Paige, 1995a), while developing strategies to keep the play safe. Educators also can expand the curriculum by discussing related themes, such as fairy tales and folktales. By allowing children to explore different venues of expression, such as superhero play, children will learn to develop new ideas and change existing presuppositions (Levin & Carlsson-Paige, 1995b). Additionally, educators can use superhero themes and toys as a way to encourage children to develop concepts of right and wrong and good and bad, to resolve conflicts, and to communicate and share emotions (Bauer & Dettore, 1997).

In sum, by providing the children with support and encouragement, both educators and parents can find a middle ground where children are free to explore the world of make-believe and develop social skills and language by sharing roles and scripts with peers and taking on different perspectives.

Conclusions

Several limitations of the present study should be noted. The sample was small and composed of middle-class families; therefore, the results cannot be generalized to all male dyads. Furthermore, dyads were observed engaging in play for a short period of time and observations over a longer period of time might provide further information. However, the rich nature of the data and the controlled design of the study allowed for new perspectives on children's play.

In conclusion, previous literature on superhero play is sparse, although studies that have investigated specific aspects of this activity have provided valuable information. Our study has served to advance understanding of this phenomenon in a number of ways. Future research should study superhero play in children from different socioeconomic and cultural backgrounds, as well as compare boys and girls, so as to develop a deeper understanding of this type of play and the possible benefits that it may provide for the development of young children. Considering that the mass media plays a vital role in young children's lives, it is imperative that one understand both the strengths and limitations that it may have for pretend play. By encouraging superhero play in a safe environment, parents and educators can ensure that the messages children receive are positive and enriching.

Authors' Note: This work was supported by a grant from the Social Science and Humanities Research Council of Canada to the second author. We wish to thank the children, parents, and educators who made this project possible and Kathy Hall for reliability coding. Address correspondence to Nina Howe, Education Department, Concordia University, 1455 de Maisonneuve West, Montreal, PQ, Canada H3G 1M8; E-mail: nina.howe@education. concordia.ca.

References

Asamen, J., & Berry, G. (1993). Children and television: Images in a changing sociocultural world. London: Sage.

Asher, S. R., Singleton, L. C., Tinsley, B. R., & Hymel, S. (1979). A reliable sociometric measure for preschool children. Developmental Psychology, 15, 443-444.

Bauer, K. L., & Dettore, E. (1997). Superhero play: What's a teacher to do? Early Childhood Education Journal, 25(19), 17-21.

Beckman, J. (1997). Television violence: What the research says on its effect on young children. Winnetka, IL: The Winnetka Alliance for Early Childhood.

Berk, L. E. (1994). Vygotsky's theory: The importance of make-believe play. Young Children, 50(1), 30-39.

Boyatzis, C.J. (1997). Of Power Rangers and V-Chips. Young Children, 52(7), 74-79.

Boyatzis, C. J., Matillo, G., & Nesbitt, K. M. (1995). Effects of the Mighty Morphin Power Rangers on children's aggression with peers. Child Study Journal, 25(1), 45-55.

Boyd, B.J. (1997). Teacher response to superhero play: To ban or not to ban? Childhood Education, 74, 23-28.

Carlsson-Paige, N., & Levin, D. E. (1987). The war play dilemma: Balancing needs and values in the early childhood classroom. New York: Teachers College Press.

Carlsson-Paige, N., & Levin, D. E. (1990). Who's calling the shots?." How to respond effectively to children's fascination with war play and war toys. Philadelphia: New Society Publishers.

Carlsson-Paige, N., & Levin, D. E. (1995). Viewpoint #4--Can teachers resolve the war-play dilemma? Young Children, 50(5), 62-63.

Fein, G. G. (1989). Mind, meaning, and affect: Proposals for a theory of pretense. Developmental Review, 9, 345-363.

Garvey, C., & Berndt, R. (1975). The organization of pretend play. (ERIC Document Reproduction Service No. ED 114 891)

Goncu, A. (1993). Development of intersubjectivity in social pretend play. Human Development, 36, 185-198.

Greenberg, G. (1995). Making friends with the Power Rangers. Young Children, 50(6), 60-61.

Gronlund, G. (1992). Coping with ninja turtle play in my kindergarten classroom. Young Children, 48(1), 21-25.

Hollingshead, A. B. (1975). Four factor index of social status. Cambridge, MA: Cambridge University Press.

Howe, N., Moller, L., Chambers, B., & Petrakos, H. (1993). The ecology of dramatic play centers and children's social and cognitive play. Early Childhood Research Quarterly, 8, 235-251.

Johnson, J. E., Christie, J. F., & Wardle, F. (2005). Play, development and early education. Boston: Pearson.

Kaplan, P. (1998). The human odyssey. Pacific Grove, CA: Brooks & Cole.

Kline, S., & Pentecost, D. (1990). The characterization of play: Marketing children's toys. Play and Culture, 3, 235-255.

Kostelnik, M., Whiren, A., & Stein, L. (1986). Living with He-Man: Managing superhero fantasy play. Young Children, 41(4), 3-9.

Kuykendall, J. (1995). Is gun play ok here??? Young Children, 50(6), 56-59.

Levin, D. E., & Carlsson-Paige, N. (1995a). The Mighty Morphin Power Rangers: Teachers voice concern. Young Children, 50(6), 67-72.

Levin, D. E., & Carlsson-Paige, N. (1995b). Can teachers resolve the war play dilemma? Young Children, 50(5), 62-63.

Maccoby, E. E. (1988). Gender as a social category. Developmental Psychology, 24, 755-765.

McLoyd, V. C. (1983). The effects of the structure of play objects on the pretend play of low-in come preschool children. Child Development, 54, 626-635.

McLoyd, V. C., Warren, D., & Thomas, E. A. (1984). Anticipatory and fantastic role enactment in preschool triads. Developmental Psychology, 20, 807-814.

Mellou, E. (1994). A theoretical perspective on the relationship between dramatic play and creativity. Early Child Development and Care, 100, 77-92.

Pellegrini, A. D. (1987). Rough-and-tumble play: Developmental and educational significance. Educational Psychologist, 22, 23-43.

Pellegrini, A. D., & Bjorklund, D. F. (1998). Applied child study. Mahwah, NJ: Erlbaum.

Pellegrini, A. D., & Boyd, B. (1993). The role of play in early childhood development and education: Issues in definitions functions. In B. Spodek (Ed.), Handbook of research on the education of young children (pp. 105-121). New York: Macmillan.

Pellegrini, A. D., & Smith, P. K. (1998). Physical activity play: The nature and function of a neglected aspect of play. Child Development, 69, 577-598.

Piaget, J. (1962). Play, dreams and imitation in childhood. New York: Norton.

Pulaski, M.A. (1973). Toys and imaginative play. In J. L. Singer (Ed.), The child's world of make believe (pp. 75-103). New York: Academic Press.

Rubin, K. H., Fein, G. G., & Vandenberg, B. (1981). Pretend play in childhood: An integrative review. Child Development, 52, 1095-1118.

Rubin, K. H., Fein, G. G., & Vandenberg, B. (1983). Play. In P. H. Mussen (Ed.), Handbook of child psychology (Vol. 4, pp. 694-775). New York: Wiley.

Simmons, J., Stralsworth, K., & Wentzel, H. (1999). Television violence and its effects on young children. Early Childhood Education Journal, 26(3), 149-152.

Tobin, J. (2000). "Good guys don't wear hats". Children's talk about the media. New York: Teachers College Press.

Waters, H. F. (1993). Networks under the gun. In K. L. Bauer & E. Dettore (Eds.), Superhero play: What's a teacher to do? Early Childhood Education, 25(1), 17-21.

Wegener-Spohring, G. (1989). War toys and aggressive games. Play and Culture, 2, 35-47.

Yawkey, T. D. (1979). Role play and language growth: Strategies for imaginative experiences. (ERIC Document Reproduction Service No. ED 170 050)

Amy Parsons 

Nina Howe

Concordia University

Table 1
Toys in the Superhero and Nonsuperhero Conditions

Superhero Toys                  Nonsuperhero Toys

Green Goblin              (1)   Male doll           (1)
Spiderman                 (1)   Male doll           (1)
Mad Hatter                (1)   Male doll           (1)
Batman                    (1)   Male doll           (1)
Robin                     (1)   Male doll           (1)
GI-Joe doll               (1)   Male doll           (1)
GI-Joe enemy              (1)   Male doll           (1)
Power Puff Girl           (1)   Female doll         (1)
Goku (Dragon Ball Z)      (1)   Male doll           (1)
Crillin (Dragon Ball Z)   (1)   Male doll           (1)
Power Ranger              (1)   Male doll           (1)

Table 2

Definitions and Examples of Roles, Themes, Physical Activity, and
Aggression

Roles

(1) Character/fictive: Child assigns or enacts role where figure's
    main characteristic is stereotypical or fictive, involving combat
    (e.g., Spiderman, Batman) or saying "I'm Batman."
(2) Familial/relational: Child assigns or enacts a familial role (e.
    g., mother) or a nonfamilial relationship (e.g., friend comes over
    to play), using a figure.
(3) Functional: Child assigns or enacts a role organized by object or
    situation (e.g., child says object is a plane, so he/she makes it
    fly).
(4) Peripheral: Child assigns a role that is discussed but not enacted
    (e.g., imaginary friend, or assigns role of mother to peer but is
    not enacted, or says "Let's fly" or "Let's be ...").
(5) Occupational: Child assigns or enacts occupational role to a
    figure (e.g., doctor, teacher). (6) Other role: Child assigns or
    enacts a role to a figure not described above (e.g., animal).
(7) Exploration/negotiation role: Child assigns or explores the
    possibility of roles for the figure (e.g., "Well, he could play
    the doctor" or "Maybe he is the bad guy").
(8) No role: Dyad does not assign or enact a role during play sequence.

Themes

(1) Action/battling/averting threat or danger: Child uses figure to
    describe or enact dangerous acts, fighting, killing, dying
    sequence (e.g., Spiderman fights with enemy and entangles him/her
    in his web until the villain dies).
(2) Capture/rescue: Child uses figure to describe or enact a sequence
    that involves plots, threats, good vs. evil, but no killing (e.g.,
    Spiderman captures Green Goblin).
(3) Submit/vanquish: Child uses figure to describe or enact a sequence
    involving submission or defeat (e.g., Batman gives himself up to
    Mad Hatter).
(4) Attack/flee: Child uses figure to describe or enact sequence where
    figure uses force to succeed or to flee (e.g., Spiderman uses
    powers, such as his web, to overthrow enemy; to escape, says,
    "You're going to get it," "Come and get me," or "Catch me if you
    can").
(5) Domestic/housekeeping: Child uses figure to describe or enact a
    cooking, cleaning, shopping, and familial sequence (e.g., mother/
    father cooking dinner).
(6) Nurturing/caring: Child uses figure to describe or enact an act of
    nurturing or caring for individual or pet (e.g., sister cares for
    a sick sibling).
(7) Building: Child uses figure to describe or enact theme involving
    building with a purpose (e.g., building a cave for Batman).
(8) Performing: Child uses figures to describe or enact artist, singer,
    and actor/actress (e.g., uses figure to depict a pop star).
(9) Other theme: Child uses figure to describe or enact a theme
    sequence not described above (e.g., girl goes skating with a
    friend).
(10) Exploration/negotiation: Child uses figure in order to describe
    or explore possible themes (e.g., "Why don't we play house").
(11) No theme: Dyad does not describe or enact a theme during the play
    sequence.

Physical Activity (ratings)

1 = no or low physical activity: Boys are not using gross (i.e.,
children sitting quietly) or fine motor skills (i.e., figures are not
manipulated)

3 = moderate physical activity: Mostly fine motor movement and
manipulation of characters. Little gross motor play occurs and the
boys change location infrequently (i.e., standing up, changing
sitting position, taking step forward or backwards).

5 = high physical activity: Frequent whole body movement (i.e.,
running, swinging, wrestling, tumbling, zooming, kicking, hopping,
sliding) and very active fine motor movements (i.e., manipulating
figures in an active way).

Physical and Verbal Aggression

(1) Verbal: speaking or yelling in an insulting or threatening manner
with the intent to harm.
(2) Physical: hitting, shoving, kicking, and tripping other children
with the intent to harm. Aggressive Play Objects: use of pretend guns,
swords, and other weapons in play.

Table 3
T-tests Between Roles in Superhero and Nonsuperhero Conditions (n = 29)

                     M       (SD)       t      df   p

Character/fict Sup   29.41   (2.28)    4.19    28   .01
  Nonsuperhero       25.17   (5.67)
Familial/relat Sup    0      (0)      -2.17    28   .04
  Nonsuperhero         .79   (1.97)
Occupational Sup     0       (0)      -2.20    28   .04
  Nonsuperhero         .55   (1.35)
Explo/nego Sup        1.48   (1.90)   -1.80    28   .08
  Nonsuperhero        2.45   (2.81)
No Role Sup            .79   (1.21)   -2.77    28   .01
  Nonsuperhero        2.10   (2.74)
Peripheral Sup         .003   (.19)   -1.44    28   ns
  Nonsuperhero         .17    (.47)
Functional Sup         .24    (.95)   -1.03    28   ns
  Nonsuperhero         .76   (2.68)

* All analyses are 1-tailed

Note. Sup = Superhero, Character/fict = Character/fictive,
Familial/relat = Familial/relational,
Explo/nego = Exploration/negotiation

Table 4
T-tests Conducted Between Themes in Superhero and Nonsuperhero
Conditions (n =29)

                        M (SD)         t       df   p

Building Sup         .006    (.26)   -1.72     28   .05
  NonSuperhero       .90    (2.73)
Nurt/caring Sup      .003    (.19)   -1.44     28   .08
  NonSuperhero       .10     (.31)
Domest/house Sup    0      (0)       -1.49     28   .07
  NonSuperhero       .28    (1.00)
Act/battl Sup      22.76    (4.98)     .475    28   ns
  NonSuperhero     22.24    (5.60)
Performing Sup      0      (0)       -1.00     28   ns
  NonSuperhero       .003    (.19)
Capt/resc Sup       2.90    (3.59)    1.24     28   ns
  NonSuperhero      2.03    (3.74)
Submit/van Sup       .14    (0.58)    1.28     28   ns
  NonSuperhero      0      (0)
Attack/flee Sup      .59     (.98)    -.538    28   ns
  NonSuperhero       .72    (1.67)
Explor/negot Sup    3.66   (2.91)      .044    28   ns
  NonSuperhero      3.62   (3.71)
No Theme Sup        1.90   (2.45)     -.272    28   ns
  NonSuperhero      2.07   (2.69)

* All Analyses are 1-tailed.

Note. Sup = Superhero, Nurt/caring = Nurturing/caring,
Domest/house = Domestic/ housekeeping, Act/battl = Action/battling,
Capt/resc = Capture/rescue, Submit/van = Submit/vanquish,
Explor/negot = Exploration/negotiation.