Sexy Dolls, Sexy Grade-Schoolers? Media & Maternal Influences on Young Girls’ Self-Sexualization
Concern is often expressed that mass media contribute to the early sexualization of young girls; however, few empirical studies have explored the topic. Using paper dolls, we examined self-sexualization among sixty 6–9 year-old girls from the Midwestern United States; specifically self-identification, preference, and attributions regarding sexualized dress. Based on simultaneous maternal reports, we also investigated potential risk factors (media consumption hours, maternal self-objectification) and potential protective factors (maternal television mediation, maternal religiosity) for young girls’ sexualization. Findings support social cognitive theory/social learning theory and reveal nuanced moderated effects in addition to linear main effects. Girls overwhelmingly chose the sexualized doll over the non-sexualized doll for their ideal self and as popular; however, dance studio enrollment, maternal instructive TV mediation, and maternal religiosity reduced those odds. Surprisingly, the mere quantity of girls’ media consumption (tv and movies) was unrelated to their self-sexualization for the most part; rather, maternal self-objectification and maternal personal religiosity moderated its effects.
KeywordsSexualization Self-objectification TV mediation Gender roles/Schema Mother-daughter relationship Mass media
Dolls may act as a thermostat for the sexualization of young girls in U.S. society. On the heels of 20th century criticism of the anatomically questionable Barbie doll came the 21st century Bratz doll—an adolescent-figured doll modeling sexy clothing and make-up on huge eyes and plump lips (Levin and Kilbourne 2008). Despite the ubiquity of these sexualizing messages targeted towards young girls, it is surprising that there remains a dearth of scientific knowledge on early sexualization, including self-sexualization (Durham 2008). For this reason, the 2007 American Psychological Association (APA) Taskforce on the Sexualization of Girls called for more research on the prevalence of early sexualization, and factors that both contribute to and buffer girls from this trajectory (p. 42). The present study, conducted in response to the APA’s call, explored young girls’ self-identification, preference, and attributions regarding sexualized dress, and investigated some potential risk factors (media consumption, maternal self-objectification) and potential protective factors (maternal television mediation, maternal religiosity) for early sexualization.
This paper adheres to the APA’s 2007 definition of sexualization as the act of being sexualized or sexualizing oneself (i.e., self-sexualization), which includes reducing physical attractiveness to sexiness, valuing someone based solely on sex appeal, or treating someone as a sexual object rather than as a person. For developmental reasons, we focus on two of three main contributors to sexualization listed by the APA Task Force (2007): for cultural contributions, we address media consumption, and for interpersonal contributions we address maternal variables. In addition, we examine two of the nine approaches listed by the APA as buffers against girls’ sexualization: TV mediation, and religion. Young girls spend a great deal of time at home compared to older children; therefore, social learning in the home, both from mothers and from television, was expected to be particularly important for early sexualization. In addition, media and maternal modeling are two of the earliest sexualization influences that most young girls have, and this learning precedes other influences such as peers and the development of their own intrapsychic attitudes. Although this study focuses on U.S. families (and all studies cited use U.S. samples unless otherwise noted), early sexualization research may be relevant in other countries, particularly because U.S. media is widely exported.
Early Sexualization of Girls
The turn of the new millennium has spawned an intriguing phenomenon: the sexy little girl…with preternaturally voluptuous curves, and one whose scantily clad body gyrates in music videos, poses provocatively on teen magazine covers, and populates cinema and television screens around the globe…she is Lolita (Durham 2008, p. 22).
Many authors identify mass media as the primary culprit for early sexualization of young girls because of the increasing sexual content in TV programming, music videos, movies, magazines, and advertising (Kunkel et al. 2005; Malamuth and Impett 2001), and the hypersexualized portrayal of women in videogames (Behm-Morawitz and Mastro 2009). However, most research on the predictors and consequences of sexualization to date has been conducted with adolescent girls and women. This body of correlational and experimental work demonstrates that teenage girls learn facts about sex from TV, whether accurate or not (Donnerstein and Smith 2001). Congruently, adolescent and college-age women who consume more sexualized mass media have more sexualized attitudes (Bryant and Rockwell 1994; Dill et al. 2008), more sexual intentions and behaviors including earlier debut (Brown et al. 2006; Pardun et al. 2005), and compromised self-efficacy and body-related attitudes (Behm-Morawitz and Mastro 2009; Daniels 2009; Grabe and Hyde 2009). Moreover, although women sometimes enjoy sexualization (e.g., social benefits of attractiveness), it also brings significant liabilities including greater body shame, objectification by others, and unwanted sexual advances (Liss et al. 2011). Substantial evidence that media is both sexualized and sexualizing for teenage girls and women, taken together with the knowledge that regular TV viewing begins before 3 years old and peaks at 12 years (Comstock and Scharrer 2001), make it particularly important to study the association between media consumption and sexualization for young girls and the potential moderators of this association (APA 2007).
Social Cognitive Theory of Gender Development
The social cognitive theory of gender development and differentiation (Bussey and Bandura 1999; Perry and Bussey 1979) helps to explain how young girls learn gender-related behaviors, attitudes, and preferences from the outside world. According to this theory, gender development is facilitated by modeling observed gender-linked behavior (e.g., a young girl imitates her mother in putting on lipstick), by learning from the consequences of one’s gender-linked behavior (e.g., the mother rewards this feminine act by purchasing the young girl her own lip balm), and by direct instruction on gender-linked behavior (e.g., the young girl’s mother provides step-by-step directions on how to apply lip balm). Young children are especially likely to adopt gender-linked behaviors when their role models’ behaviors are rewarded or go unpunished (Bandura 1965). In these instances, the child becomes motivated to seek similar positive outcomes by imitating the rewarded behavior (e.g., the young girl uses lip balm in hopes of achieving the same social attention or personal satisfaction from looking pretty as does her mother). The child may also learn the underlying rules governing the modeled behavior to produce novel actions that are likely to also be rewarded (e.g., the young girl chooses to wear more stereotypically feminine clothes after discerning that looking pretty in general is socially and personally rewarding).
Moreover, young children selectively attend to same-gender models. For example, in a classic study of 84 eight year-old girls and boys, children who were shown a film with a same-gender and other-gender model were significantly more likely to imitate the stated food preferences of the same-gender model (Perry and Bussey 1979). More recent research on children’s reactions to media concurs. For example, Hoffner (1996) found that among 7 to 12-year-olds, most boys and approximately half of the girls reported a same-gender character as their favorite TV character, and girls were more likely to express the desire to emulate the behaviors of female than male characters. Parents play a particularly salient role in their young children’s gender role development as the first same-gender models young children have (a British study found that babies as young as 10 months pay significantly more attention to models of the same gender: Kujawski and Bower 1993), and as the primary source of instruction and feedback about gendered behavior (e.g., the young girl previously described would care more about her mother’s opinion of her lip balm application: Bussey and Bandura 1999). Additionally, there is a significant positive association between adolescent girls’ body image and eating behaviors and their perceptions of their mothers’ body/eating behaviors and attitudes (Benedikt et al. 1998; Cooley et al. 2008). This supports the idea that girls learn gender roles, attitudes towards their body, and possibly sexualized attitudes and behaviors primarily from their mothers. Maternal influences on sexualization may be particularly strong for young girls because their developmental stage requires high levels of direct mother–daughter involvement, which allows daughters more social learning opportunities. For this reason, the current study investigates several maternal influences as moderators of the likelihood of early sexualization.
Risk Factors for Early Self-Sexualization of Girls
There is compelling experimental and correlational evidence that young children learn and imitate the behaviors of film and cartoon characters, particularly same-gender models (Anderson et al. 2010; Bandura et al. 1963). The contemporary 6–11 year old child in the United States views a weekly average of 22 hr of TV and over 28 hr of combined media (TV, movies, and video games) (McDonough 2009), and there has been an exponential increase in sexual TV content since the mid-‘70s (Kunkel et al. 1996). Currently, over 70 % of all TV programming and 77 % of prime-time TV programming includes sexual material—roughly six scenes per hour of sexual talk or physical acts (Kunkel et al. 2005). This is especially concerning as childhood TV viewing patterns morph into adolescent and adult patterns in terms of show and time slot preferences (Comstock and Scharrer 2001). Ward (1995) found that approximately 30 % of the most viewed child and adolescent TV programs contain sex talk which highlights physical attractiveness and objectifies women’s bodies. A study from the Parents Television Council (2010) paints an even bleaker picture: among the 25 most popular shows for 12-17 year-olds, 86 % of female actors in sexy prime-time scenes are high school age, more depictions of sexual behavior in these shows contain underage teen girls (47 %) than adult women (29 %), and girls in these shows typically display positive or neutral responses to their sexualization.
Sexualized female media characters are especially alluring models for young girls. First, children initially select screen character models based on similarities including gender, then later based on wishful identification with an older character (i.e., a desire to resemble that person: Comstock and Scharrer 2007). Second, wishful identification for girls is strongly based on physical attractiveness (Reeves and Greenberg 1977). Third, female media characters often display ultra feminine behavior, which is likely to appeal to younger girls who characteristically hold rigid gender stereotypes (Bussey and Bandura 1999; Perry and Bussey 1979). Therefore, it is reasonable to expect that young girls may also be influenced by sexualized media images, including sexy clothing.
Mothers give their daughters important cues about desirable appearance and body image, which may have implications for self-sexualization. A meta-analysis of 43 studies on the development of gender schemas found a reliable association between parents’ gender schemas and their children’s gender self-concepts (Tenenbaum and Leaper 2002). Similarly, research on the “thin ideal” reveal that a mother’s opinion about her own body and messages regarding physical appearance are associated with her daughter’s perceived importance of being thin and body-related behaviors including dieting and disordered eating (Hill et al. 1990; Levine et al. 1994). In addition, girls with stronger internalized body ideals such as the thin ideal are more likely to respond positively to pictures of sexually objectified women (Murnen et al. 2003). Given that mothers model important values and behaviors regarding body ideals and related behaviors, it is likely that mothers who have a more objectified self-view may predispose their daughters to a more objectified self-view and more sexualized preferences.
Protective Factors Against Early Self-Sexualization of Girls
Maternal TV Mediation
The parenting literature demonstrates the strong protective effects of parental involvement against a variety of maladaptive developmental outcomes (Jeynes 2005; Sieverding et al. 2005). One particular type of parental involvement, TV mediation, has been found to have a protective effect against a range of risky youth behaviors including sexual behavior. TV mediation refers to parental involvement in a child’s TV viewing and has three components: restrictive mediation (forbidding certain programs or amount of viewing hours), instructive mediation (talking about specific things in the show with an instructive goal in mind), and social coviewing (the parent informally viewing TV with their child with no specific goal in mind; Valkenburg et al. 1999). In a sample of over 1,000 adolescents, Fisher et al. (2009) found that parental mediation of sexual TV content was negatively correlated with adolescents’ reports of engagement in sexual behaviors including oral and vaginal intercourse. This was particularly true for the teenagers of parents who implemented restrictive mediation techniques. In addition, adolescents who discuss or watch TV with their parents (instructive TV mediation and social coviewing) have better ‘critical viewing’ skills and lower sexual activity, and for girls in particular, increased body- and self-esteem (Peterson et al. 1991; Schooler et al. 2006). Similarly, among children who watch TV clips portraying stereotypical gender roles, those who receive instructive TV mediation express greater acceptance of non-traditional gender roles than do those who do not (Nathanson et al. 2002). Given the protective effects of TV mediation against risky sexual behavior for adolescents and against traditional gender role endorsement for young children, it is plausible for it to also have a protective effect against the self-sexualization of young girls.
Parental values such as religiosity may also be important environmental moderators to consider, given that family religiosity may be a protective factor against adolescent early sexual debut and multiple sexual partners (Manlove et al. 2008; Zaleski and Schiaffino 2000). Many religions including Christianity value modest dress and actively discourage sexualized clothing (Mahaney et al. 2007). In addition, religious belief has been linked to higher self-acceptance of one’s appearance. Stronger religious values are significantly related to greater body satisfaction and less dieting (Kim 2006), and religiosity is negatively correlated with anxiety about aging appearance (Homan and Boyatzis 2009). Religion may also serve as a buffer against media’s negative effects on youth. Ward (2004) found that among African American youth with high media consumption, those who were highly religious had higher self-esteem than those who were less religious. Maternal religiosity may be protective against early sexualization of young girls via modeling (i.e., religious mothers are more likely to model body acceptance and modesty) or instruction (i.e., religious mothers may instruct their daughters regarding religious values and modest dress). That said, the relationship between religiosity and sexual behavior is a complex one, as certain Christian evangelical traditions, including virginity pledges, have been found to be associated with early sexual debut (Rostosky et al. 2003).
Hypothesis 1.Prevalence of self-sexualization. Based on social learning, girls will choose the sexualized doll more often than the non-sexualized doll as their actual self (what they look like) and ideal self (what they would prefer to look like), as more popular, and as a play toy.
Hypothesis 2.Risk factors for early self-sexualization. Higher media consumption (TV and movies) and maternal self-objectification will increase girls’ odds of choosing the sexualized doll by providing girls greater opportunities to learn sexualized behaviors and preferences modeled by media and mothers.
Hypothesis 3.Protective factors for early self-sexualization. Higher maternal TV mediation, maternal personal religiosity, and maternal instructive religiosity will decrease girls’ odds of choosing the sexualized doll. This hypothesis is based on the knowledge that parental TV mediation is a protective factor against risky sexual behaviors among adolescents, family religiosity may be a buffer against early sexual debut, and same-gender parental instruction regarding values may be particularly salient for young girls.
Hypothesis 4.Moderation effect of maternal variables on the association between media consumption and self-sexualization. Maternal self-objectification will interact with media consumption to increase the odds that girls will choose the sexualized doll. That is, exposure to high levels of two risk factors for sexualization may be disproportionately worse than exposure to high levels of only one. On the other hand, maternal TV mediation, maternal personal religiosity, and maternal instructive religiosity, will each interact with media consumption to decrease the odds that girls will choose the sexualized doll. That is, exposure to high levels of a protective factor for sexualization may act as a buffer against the self-sexualizing effects of high media consumption.
Sixty girls ranging from 6 to 9 years (M = 7.78, SD = .98) and 47 of their mothers participated in this study. Most girls (78 %) were recruited from two public grade schools in the Midwestern United States (average of 85 % free and reduced lunch) and the remaining 22 % were recruited from a local dance studio (approximately 85 % public school attendance). The average temperature in the Midwestern region of data collection was 59 °F, and girls were observed to be dressed primarily in long-sleeved shirts/sweatshirts and jeans/sweatpants (covered by a jacket/coat when going outdoors). The majority of mothers in the sample self-identified as White (77 %) and the remainder self-identified as Black or African American (13 %), Hispanic or Latina (6 %), and Multiracial (4 %). In addition, 91.5 % of mothers described their religious views as Christian, 4.3 % as non-religious or secular, 2.1 % as Buddhist, and 2.1 % as Unitarian Universalist. Chi-squared analyses showed no significant differences in race or religion based on recruitment location.
Measures and Materials
Actual Self: “Which doll do you think looks most like you?”
Ideal Self: “If you could look like one of these two dolls, which one would you like to look like?”
Popularity Attribution: “I’m going to read you a story. Listen carefully, because I will ask you [a question] about the story afterwards. ‘Leila is the most popular girl in school. She has many friends, and many people want to sit next to her at the lunch table.’ Which doll is Leila?”
Play Preference: “Which doll would you like to play with?”
Eight dolls total were created (4 pairs of sexualized/non-sexualized) using an online “doll maker” (Dollz Mania 2008) targeted towards young girls (see Appendix). To avoid introducing confounds, each pair of dolls was created with identical skin tone (light brown), facial features, eye color, hairstyle, and body posture, and differed only in their attire. One doll was dressed in revealing, sexualized or “sexy” clothing that was skin-tight and revealing (e.g., a low cut shirt with midriff showing, and short jean shorts), whereas the other was dressed in stylish but non-sexualized clothing (e.g., a “v” neck sweater, belt, and cargo pants). A different pair of dolls was used for each of the four questions and the left/right presentation order of the sexy/non-sexy dolls was alternated on each successive question to avoid response sets.
Mothers were asked to complete the following measures.
Daughter’s Media Consumption
Each mother reported the number of weekly hours of media (defined as TV and movies) viewed by her participating daughter. Weekly TV and movie consumption hours was considered an adequate proxy for sexualized media consumption for the purposes of this study based on the most recent content analysis findings that the vast majority of TV programming (overall and prime time) contains sexualized material, mostly acted by adolescent girls in popular teen and tween shows (Kunkel et al. 2005; PTC 2010).
The 8-item Body Surveillance subscale of the Objectified Body Consciousness Scale (McKinley and Hyde 1996) was used to measure each mother’s attention to her own clothing and the importance placed on looking good (Cronbach’s α = .87). Sample items include “I often worry about whether the clothes I am wearing make me look good” and “During the day, I think about how I look many times.” Mothers responded to items on a 6-point scale from 1 (Strongly Disagree) to 6 (Strongly Agree). Per scoring instructions “Not Applicable” was counted as missing data and a subscale mean was created for mothers who had ≤25 % missing data (all but four).
Television Mediation Scale
Valkenburg et al.’s (1999) 15-item Television (TV) Mediation Scale was used, which comprises three 5-item subscales: restrictive mediation (α = .73, e.g., “How often do you forbid your child to watch certain programs?”); instructive mediation (α = .77, e.g., “How often do you point out why some things actors do are bad?”); and social coviewing (α = .74, e.g., “How often do you watch together just for fun?”). Parents were instructed to respond to each item on a 4-point Likert scale of 1 (Never) to 4 (Often) and to answer in reference to the participating daughter. Subscale means scores were created. To maximize the internal reliability of the measure, one item was removed from the restrictive mediation subscale (“How often do you say to your child to turn off TV when she is watching an unsuitable program?”, α including item = .63), and one item was removed from the social coviewing subscale (“How often do you watch together because of a common interest in a program?”, α including item = .67).
Mothers reported their personal religiosity (“How important is religion to your daily life?”) and their instructive religiosity (“How important do you think it is to teach your children your religious values?”) using a 3-point Likert scale of 1 (Not important) to 3 (Very important).
Girls (30 % participation) received parental permission and mothers (24 % participation) gave consent. Child-reported measures were administered orally in group format. Each girl was given a 4-page stapled packet with one question and corresponding doll pair on each page. Girls were seated at least two seats apart and were instructed to circle their doll choice after each question was read aloud and to focus on their own packet rather than looking at other participants’ choices. It was reinforced that “there is no right or wrong answer” and that participants could stop at any time. Mothers completed their questionnaires at home.
Means and standard deviations of main study variables for girls who selected the sexualized versus non-sexualized doll
Hypothesis 1: Prevalence of Early Self-Sexualization
Observed numbers (%) of girls choosing the sexualized versus non-sexualized dolls (n = 60)
Public School (n = 47)
Dance Studio (n = 13)
33 (70 %)
14 (30 %)
0 (0 %)
13 (100 %)
36 (77 %)
11 (23 %)
5 (38 %)
8 (62 %)
39 (83 %)
8 (17 %)
4 (31 %)
9 (69 %)
27 (57 %)
20 (43 %)
5 (38 %)
8 (62 %)
Hypotheses 2, 3, & 4: Risk, Protective, and Moderating Factors of Self-Sexualization
Hierarchical binary logistic regression analyses were employed to predict the probability that girls would choose the sexualized dolls based on various risk and protective factors in isolation and in interaction. In addition to being the most appropriate analytic strategy to handle the dichotomous dependent variables (i.e., doll choice; Agresti 2007), logistic regression also maximized the power of analyses to detect interaction effects by avoiding breaking the sample into smaller groups, which would be required by MANOVAs. Separate regression analyses were conducted for each dependent variable (i.e., Actual Self, etc.). To assess Hypothesis 2 (i.e., higher Media Consumption and Maternal Self-Objectification will increase girls’ odds of choosing the sexualized doll), and Hypothesis 3 (i.e., maternal TV mediation and maternal religiosity will decrease girls’ odds of choosing the sexualized doll), each continuous predictor and the covariate (Recruitment Location) were entered in step/block 1 of regression analyses. To assess Hypothesis 4 (i.e., each maternal variable will interact with Media Consumption to either increase or decrease girls’ odds of will choosing the sexualized doll), 2-way interaction terms were created using standardized variables and entered in step/block 2 of each regression analysis. For theoretical reasons (i.e., construct differences between the three forms of TV mediation) and statistical reasons (i.e., to avoid problems associated with multicollinearity), we conducted separate regression analyses for each TV mediation variable and each maternal religion variable, and standardized variables before creating interaction terms.
Oneway univariate analyses of variance (ANOVAs) were used to further investigate significant interactions after dichotomizing step 1 variables via median split. An α of .05 was employed in all analyses except those containing Maternal Self-Objectification, in which case α = .06 was used due to the lower n in these analyses. Effect sizes from means comparisons are reported as Cohen’s d (M1–M2/pooled S). Results for Recruitment Location mirrored findings from chi-squares, thus they are not repeated below.
Hierarchical logistic regression with maternal self-objectification predicting doll choice
Girls’ media consumption
Girls’ media consumption × maternal self-objectification
Hierarchical logistic regression with maternal restrictive mediation predicting doll choice
Girls’ media consumption
Maternal restrictive mediation
Girls’ media consumption × maternal restrictive mediation
Hierarchical logistic regression with maternal instructive mediation predicting doll choice
Girls’ media consumption
Maternal instructive mediation
Girls’ media consumption × maternal instructive mediation
Girls’ media consumption
Maternal instructive mediation
Girls’ media consumption × maternal instructive mediation
Hierarchical logistic regression with maternal personal religiosity predicting doll choice
Girls’ media consumption
Maternal personal religiosity
Media consumption × maternal personal religiosity
Girls’ media consumption
Maternal personal religiosity
Media consumption × maternal personal religiosity
Girls’ media consumption
Maternal personal religiosity
Media consumption × maternal personal religiosity
Hierarchical logistic regression with maternal instructive religiosity predicting doll choice
Girls’ media consumption
Maternal instructive religiosity
Media consumption × maternal instructive religiosity
Although there was no main effect of Maternal Self-Objectification on doll choice, there was a significant interaction between girls’ Media Consumption and Maternal Self-Objectification on Popularity Attribution, B = 3.76, p = .055 (see Table 3). ANOVAs investigating the interaction indicated that, as expected, among girls with high weekly media consumption, those who chose the sexy doll as more popular had significantly higher maternal self-objectification scores (M = 3.95, SD = .84) than those who did not (M = 2.88, SD = .70), F(1, 20) = 9.8, p = .005, Cohen’s d = 1.48. However, there were no significant differences in maternal self-objectification for girls who had low media consumption.
Maternal Television Mediation
There was a main effect of Maternal Instructive Mediation on girls’ Actual Self in the hypothesized direction (B = −2.23, OR = .11). See Table 5. That is, for each one point increase in mothers’ instructive mediation on the 4-point scale, the odds that a girl would self-identify as the sexy doll decreased by 89 %. There were no other significant effects of maternal mediation on girls’ doll choice.
There was a significant interaction between girls’ Media Consumption and Maternal Personal Religiosity on Ideal Self (B = −1.16, OR = .31). See Table 6. ANOVAs indicated that, as expected, among girls with high weekly media consumption, those who chose the non-sexualized doll as their ideal had significantly higher maternal personal religiosity scores (M = 1.88, SD = .35) than those who chose the sexualized doll as their ideal (M = 1.20, SD = .68), F(1, 21) = 6.86, p = .016, d = 1.26. An unexpected finding also emerged from these analyses. Among girls with low weekly media consumption, those who chose the sexualized doll as their ideal had marginally higher maternal personal religiosity scores (M = 1.69, SD = .48) than those who chose the non-sexualized doll as their ideal (M = 1.25, SD = .71), F(1, 22) = 3.24, p = .086, d = .73. There was also a hypothesized main effect of Maternal Personal Religiosity on girls’ Play Preference (B = −1.33, OR = .26). That is, for each one point increase on the 3-point religiosity item the odds that a girl would choose the sexy doll as their preferred play toy decreased by 74 %. In addition, there was a main effect of maternal instructive religiosity on girls’ Popularity Attribution (B = −2.06, OR = .13). See Table 7. That is, for each one point increase on the 3-point maternal instructive religiosity item, the odds that a girl would choose the sexy doll as popular decreased by 87 %. There was no significant interaction.
Concerned community members and scholars alike echo observations regarding the early sexualization of girls in modern society and the possible contribution of mass media (Durham 2008; APA 2007). The current study was conducted in response to the call from the 2007 APA Task Force on the Sexualization of Girls for more research on the prevalence of sexualization and related risk and protective factors for girls. Using paper dolls, we examined self-sexualization of young girls in terms of self-identification, preference, and attributions regarding sexualized dress, and investigated media consumption hours as a primary risk factor moderated by certain maternal influences. Findings were rich and nuanced: three of four hypotheses (#1, 3, & 4) were at least partially supported with mainly medium to large effect sizes. Young girls overwhelmingly chose the sexualized doll over the non-sexualized doll for their ideal self and as popular; however, dance studio enrollment, maternal instructive TV mediation, and maternal religiosity (both personal and instructive) greatly reduced the odds of sexualization. Surprisingly, the mere quantity of girls’ weekly media consumption was unrelated to their self-sexualization for the most part; rather, maternal self-objectification and personal religiosity moderated its effects. High maternal personal religiosity was both a protective factor and a risk factor based upon girls’ level of media consumption.
Sexy Dolls, Sexy Grade-Schoolers? Yes, for Many But Not for All
Our findings indicate that there is reason to be concerned about the early sexualization of girls. The vast majority of young girls in our study recruited from the public schools were sexualized in terms of what they wanted to look like and their attributions of popularity to sexy appearance. In fact, girls’ preference for the skimpily dressed sexy doll occurred in the face of cold Midwestern weather during early spring data collection! These findings are consistent with research on the sexualization of adolescents which has found that teen girls are sexually objectified (Grabe and Hyde 2009) and identify with sexualized media persona (Gordon 2008). Findings also support prior research with elementary school children showing that sexiness boosts popularity among girls (Adler et al. 1992). The pressure young girls feel to be sexy in order to be popular may be part of why they prefer to look sexy; “girls anticipate that they will accrue social advantages, such as popularity, for buying into the sexualization of girls (i.e., themselves), and they fear social rejection for not doing so” (APA 2007, p. 18). On the other hand, despite preferring to look like the sexualized doll and perceiving such a girl to be more popular, young girls were not significantly more likely to want to play with it. Granted, girls in the study may have begun to outgrow doll play or may have been confused about how to play with a paper doll. Notwithstanding, this finding is interesting in light of the prevalence of sexy dolls sold in stores, such as Bratz dolls. Doll franchises claim that they produce sexualized dolls because that is what young girls want to play with; however, our findings argue otherwise.
Dance studio enrollment emerged as a protective factor against young girls’ sexualization. Although unexpected, this finding supports prior research among adolescent and adult women: street dancers have higher body appreciation than non-dancers (Swami and Toveé 2009), experienced dancers report more positive body image than less experienced dancers (Lewis and Scannell 1995), and aesthetic sport athletes (e.g., diving, figure skating) have more positive body image than non-athletes (meta-analysis of 78 studies: Hausenblas and Downs 2001). One possible explanation is that girls and women involved in physical activities are less prone to sexualization because they become aware that their bodies can be used for other purposes besides looking sexy or attractive for others (APA 2007). Daniels’ (2009) experiment with adolescent and college-age girls demonstrated that viewing images of performance athletes versus sexualized athletes or models produces a significantly less objectified self-view (i.e., greater focus on physicality than beauty, and fewer body-focused self-descriptions overall). This may elucidate why none of the young dancers in this study chose the sexualized doll for their actual self whereas most of the public school girls did (in actuality, none of the girls during data collection were dressed like the sexy doll). On the other hand, one longitudinal study found that young girls with aesthetic sport participation had more concerns about weight at age 5 and at age 7 compared to girls with non-aesthetic or no sport participation (Davison et al. 2002). Thus, while dance enrollment may offer young girls some protection from early sexualization, it may simultaneously present an increased risk of weight concerns.
An alternative explanation for the lack of sexualized choices among young dancers in this study is that their mothers were significantly less likely to objectify their own appearance. Regression analyses revealed that especially for girls with high media consumption, having a less self-objectified mother is a protective factor against early sexualization. The latter finding will be discussed in more detail below.
Is Media Consumption the Primary Risk Factor for Early Self-Sexualization?
Whereas our findings lead us to share others’ (Durham 2008; Levin and Kilbourne 2008) concerns regarding the prevalence of early sexualization, and we agree with their assertion that parental TV involvement is an important protective factor, we do not find media consumption to be the primary culprit for early sexualization of young girls. According to our study, the quantity of TV and movies watched is not, in and of itself, a risk factor for young girls’ sexualized self-views, ideals, attributions, or play toy preferences. Rather it is the interaction between media hours and maternal self-objectification that creates vulnerability for early sexualization—daughters with high media consumption whose mothers have a more objectified view of their bodies are at greater risk for equating sexiness with popularity. High media consumption may provide young girls a predisposition towards early sexualization which is only realized for those whose mothers display reinforcing self-objectifying attitudes and behaviors. Alternatively, girls of highly self-objectifying mothers may model their mother’s self-objectified attitudes and behavior, and effectively begin to self-sexualize and self-objectify in the presence of myriad reinforcing images afforded by high media consumption. Longitudinal research beginning even earlier in life is needed to investigate these potential developmental pathways.
Fortunately for young girls with high media consumption, a second maternal variable—high personal importance of religion to mothers—acts as a protective factor against their sexualization. Mothers who see religion as personally important may be more likely to model higher body-esteem and communicate values such as modesty, lessening the likelihood that their daughters will prefer to look sexy or play with sexy looking doll, despite what they see on TV/movies.
Inasmuch as high media consumption alone is not the singular culprit for young girls’ sexualization, results also reveal that low media consumption is not a silver bullet. Rather, young girls with low media consumption may be more likely to attribute popularity to sexiness, and if their mothers are highly religious, they may also be more likely to idealize sexiness. This pattern of results may reflect a case of “forbidden fruit” or reactance, whereby young girls who are overprotected from the perceived ills of media by highly religious parents (through restriction of TV/movie viewing time) begin to idealize the forbidden due to their underexposure. This interpretation resonates with the findings of an experimental study by Krcmar and Cantor (1997) in which children who thought their choice of TV program was limited (either because of a restrictive TV rating or because of their parent’s negative appraisal of it) were more likely to show positive affect when discussing the show. Alternatively, counterintuitive findings of higher odds of self-sexualization in the presence of lower media consumption may reflect a parental reaction to limit some daughters’ TV consumption after observing highly sexualized attitudes/behaviors. Creative experimental research would allow further exploration of this potential cause-and-effect relationship.
Discrepancies between results of this study and Fisher et al.’s (2009) findings regarding the usefulness of TV restriction may be due to the differences in sample age (young girls versus adolescent girls, respectively). For adolescents, higher parental monitoring and restriction of certain activities is associated with less risky sexual cognitions and less sexual intentions (Sieverding et al. 2005). According to our findings, this is not the case of young girls.
Taken together, it appears that the media’s effect on young girls’ self-sexualization acts in combination with maternal characteristics, rather than in isolation. Moreover, the mere restriction of young girls’ media may not be an effective solution; overprotective parenting strategies may backfire if girls feel that their personal choices are limited due to their mothers’ religious values and/or their own lack of TV exposure.
Buffering Against Early Self-Sexualization: The Protective Effect of Mothers as Teachers
When parents, through their religious or ethical practices, communicate the message that that other characteristics are more important than sexuality, they help to counteract the strong and prevalent message that it is only girls’ sexuality that makes them interesting, desirable, or valuable (APA 2007, p. 38).
Thus, reducing the likelihood of early sexualization of girls requires a more active parenting approach than simply restricting TV/movies. Mothers who find it important to teach values and help their daughters apply these values in real life situations, such as when watching TV, may be more effective buffers against their daughters’ self-sexualization by preparing them to successfully navigate the onslaught of sexualizing messages out in the world.
Limitations and Recommendations for Future Research
Notwithstanding the modest sample size, the current study yielded a wealth of interesting findings with mostly medium to large effect sizes. Nevertheless, this study had some limitations worth discussing. We used a global measure of TV and movie consumption based on evidence that sexual content is pervasive across TV genres (Kunkel et al. 2005). However, this may have masked important differences in the risk posed by different genres such as educational children’s programming (e.g., Sesame Street) versus prime time TV (e.g., Two and a Half Men). In addition, prior studies have found that sexualized media consumption in particular, rather than general media consumption, is associated with early sexual activity among adolescents (Brown et al. 2006). Therefore, future work should measure the quality (i.e., content), quantity, and nature (e.g., setting) of sexualized media use, and include other types of media consumption such as internet and video games. Future studies can also examine the mechanisms by which maternal variables are protective against girls’ early sexualization (e.g., what is the active ingredient in mothers’ instructive mediation that makes it helpful?).
It is possible that despite the advantages of the chosen data analytic strategy (regression maximizes the n across analyses unlike ANOVA) some null or marginally significant findings were due to insufficient power. Therefore, replication of this study with a larger sample including once which intentionally oversamples girls enrolled in dance or other aesthetic sports may be beneficial. Results of this study may be most applicable to predominantly White, low to middle income young girls from Christian families in the Midwestern United States; therefore, replication with young girls from other racial/ethnic groups, income levels, religions, regions, or countries will build on these novel findings. In addition, future researchers may consider including interesting moderators such as socioeconomic status, whose inclusion was not possible in the current study given the already complex nature of the research questions. For example, recruitment location (public school vs. dance studio) may have served as a proxy for socioeconomic status (SES). Future research can directly investigate the potential association between SES and early sexualization.
Precautions were taken to minimize the potential for peer pressure in the group administration procedure: children were seated at least two seats apart and instructed to focus on their own papers while the examiner walked around to provide individual assistance as needed. However, it is impossible to completely eradicate the potential for socially desirable responding in self-report research (e.g., even when interviewed individually, children may perceive examiner pressure to choose a given response). The forced choice method used intentionally limited the range of responses young girls could give in order to present the simplest format for their developmental stage (see Cramer and Anderson 2003). It is possible that some girls had more equal preferences for dolls or nuanced attitudes than this method could capture. Therefore, investigation into the degree of self-identification/preference (e.g., using a Likert scale) and perceived reasons for doll choices (e.g., asking girls why they choose the doll they do) may be a valuable next step for future studies. Follow-up research might explore other aspects of sexualized appearance (e.g., body type/posture), behavior (e.g., peer interactions), and attributions (e.g., intelligence). Although our dolls task operationalized “sexualization” as wearing skin tight, revealing clothing (while holding all other physical features constant), future studies might employ other operationalizations such as seductive body posturing. Investigation of the prevalence of actual sexualized behaviors among young girls and the developmental consequences of early sexualization are particularly important avenues to pursue.
Our study is the first to examine young girls’ self-sexualization and attitudes/preferences regarding sexualized dress. With the notable exception of girls enrolled in dance classes, young girls overwhelmingly demonstrate a sexualized view of their desired selves and equate sexiness with popularity. Overall media consumption (TV and movies) is not on its own a risk factor for early sexualization; rather, mothers play an important contributory or buffering role in this process. High media consumption in the presence of high maternal self-objectification or low maternal religiosity puts girls at greater risk for early sexualization (double jeopardy); however, so does low media consumption in the presence of high maternal religiosity (forbidden fruit). On the other hand, maternal instruction about TV shows and the importance placed on teaching daughters religious values buffers girls from early self-sexualization.
- American Psychological Association, Task force on the Sexualization of Girls. (2007). Report of the APA Task Force on the Sexualization of Girls. Retrieved from http://www.apa.org/pi/women/programs/girls/report.aspx
- Anderson, C. A., Shibuya, A., Ihori, N., Swing, E. L., Bushman, B. J., Sakamoto, A., …Saleem, M. (2010). Violent video game effects on aggression, empathy, and prosocial behavior in Eastern and Western countries. Psychological Bulletin, 136, 151–173. doi:10.1037/a0018251;10.1037/a0018251.supp.
- Brown, J. D., L’Engle, K. L., Pardun, C. J., Guo, G., Kennevy, K., & Jackson, C. (2006). Sexy media matter: Exposure to sexual content in music, movies, television, and magazines predicts Black and White adolescents’ sexual behavior. Pediatrics, 117, 1018–1027. doi:10.1542/peds.2005-1406.PubMedCrossRefGoogle Scholar
- Bryant, J., & Rockwell, S. (1994). Effects of massive exposure to sexually oriented prime-time television programming on adolescents' moral judgment. In D. Zillmann, J. Bryant, A. C. Huston (Eds.), Media, children, and the family: Social scientific, psychodynamic, and clinical perspectives (pp. 183–195). Hillsdale, NJ England: Lawrence Erlbaum Associates, IncGoogle Scholar
- Comstock, G., & Scharrer, E. (2001). The use of television and other film-related media. In D. G. Singer & J. L. Singer (Eds.), Handbook of children and the media (pp. 47–72). Thousand Oaks: Sage.Google Scholar
- Comstock, G., & Scharrer, E. (2007). Media and the American child. Burlington: Elesvier.Google Scholar
- Dempster, A. P., Laird, N. M., & Rubin, D. B. (1977). Maximum likelihood from incomplete data via the EM algorithm. Journal of the Royal Statistical Society B: Methodological, 39, 1–38.Google Scholar
- Dollz Mania (2008). Dollz Mania ChaZie Dollmaker. Retrieved from http://dollzmania.net/ChaZieMaker.htm
- Donnerstein, E., & Smith, S. (2001). Sex in the media: Theory, influences, and solutions. In D. G. Singer & J. L. Singer (Eds.), Handbook of children and the media (pp. 289–308). Thousand Oaks: Sage.Google Scholar
- Durham, M. G. (2008). The Lolita effect: Why the media sexualize young girls and what we can do about it. New York: Overlook Hardcover.Google Scholar
- Hogan, M. J. (2001). Parents and other adults: Models and monitors of healthy media habits. In D. G. Singer & J. L. Singer (Eds.), Handbook of children and the media (pp. 663–680). Thousand Oaks: Sage.Google Scholar
- Kunkel, D., Cope, K. M., & Colvin, C. (1996). Sexual messages on family hour television: Content and context. Menlo Park: Kaiser Family Foundation.Google Scholar
- Kunkel, D., Eyal, K., Finnerty, K., Biely, E., & Donnerstein, E. (2005). Sex on TV 4: A biennial report to the Kaiser Family Foundation. Menlo Park: Kaiser Family Foundation.Google Scholar
- Levin, D. E., & Kilbourne, J. (2008). So sexy so soon: The new sexualized childhood and what parents can do to protect their kids. New York: Ballantine Books.Google Scholar
- Levine, M., Smolak, L., Moodey, A., & Shuman, M. (1994). Normative developmental challenges and dieting and eating disturbances in middle school girls. International Journal of Eating Disorders, 15, 11–20. doi:10.1002/1098-108X(199401)15:1<11::AID-EAT2260150103>3.0.CO;2-Q.PubMedCrossRefGoogle Scholar
- Mahaney, C., Bradshaw, J., & Whitacre, N. (2007). Fashion and following the savior. Crosswalk.com: The intersection of Faith and Life. Retrieved from http://www.crosswalk.com/spirituallife/women/11535955/page3/
- Malamuth, N. M., & Impett, E. A. (2001). Research on sex in the media: What do we know about effects on children and adolescents? In D. G. Singer & J. L. Singer (Eds.), Handbook of children and the media (pp. 269–288). Thousand Oaks: Sage.Google Scholar
- McDonough, P. (2009). TV viewing among kids at an eight year high. The Nielsen Company. Retrieved from The Nielson Company website: http://blog.nielsen.com/nielsenwire/media_entertainment/tv-viewing-among-kids-at-an-eight-year-high/
- Parents Television Council (2010). Sexualized teen girls: Tinseltown’s new target: A study on teen female sexualization in prime-time TV (December Special Report). Retrieved from http://www.parentstv.org/FemaleSexualization/Study/Sexualized_Teen_Girls.pdf
- Valkenburg, P., Krcmar, M., Peeters, A., & Marseille, N. (1999). Developing a scale to assess three styles of television mediation: ‘Instructive mediation,’ ‘restrictive mediation,’ and ‘social coviewing’. Journal of Broadcasting & Electronic Media, 43, 52–66. doi:10.1080/08838159909364474.CrossRefGoogle Scholar