Notes on Self-Socialization and Parent Socialization in Toddlers’ Gender-Typed Appearance
Background and Theoretical Framework
Topic: Early gender development focusing on gender-typed appearance in toddlers and how it relates to children’s own gender concepts and to mothers’ gender attitudes.
Self-socialization theories (e.g., Martin, Ruble, & Szkrybalo, 2002) propose that as children become aware of gender categories, they are motivated to adhere to gender stereotypes to fit “the self” and to maintain self-esteem. Key ideas:
Understanding gender categories leads children to seek gender information and to conform to stereotypes to be good members of their gender group.
Early gender labeling of self and others predicts higher levels of gender-typing (e.g., play and toy preferences).
The process can be bidirectional: gender concepts may influence behaviors (e.g., appearance) and experiences (appearance) may reinforce gender concepts.
Early milestones include producing and recognizing gender labels and later gender constancy (gender remains the same over time).
Socialization theories emphasize environmental influences (parents, peers, media) on gendered beliefs and behaviors:
Parents often provide gender-stereotypical toys and may respond differently to gender-typical vs. counter-stereotypical play.
More traditional gender attitudes in parents have been linked to more traditional gender cognitions and behaviors in children.
Gender-typed appearance as a construct:
Appearance can be a stable indicator of gender-typing because children often wear the same outfit across a day, signaling gendered presentation.
Appearance is tightly linked to constructions of gender in society and can influence others’ perceptions of gender.
Prior work (e.g., Halim et al., 2013a, 2014) shows that appearance-typical clothing becomes more gender-typed over time and relates to gender concepts like gender stability and gender importance.
Significance of toddler-age focus:
Toddlerhood is when gender labeling and gender-typed dress begin to emerge, making it a critical period to examine the self-socialization–socialization dynamic in a bidirectional, early window.
Cross-cultural relevance:
To test generalizability beyond Euro-American samples, this study includes Dominican American, Mexican American, and African American toddlers, acknowledging diverse cultural and linguistic contexts.
Key Concepts and Definitions
Gender labeling: production and understanding of gender terms for self and others (e.g., labeling self as boy/girl; using gender terms in language).
Gender-typed appearance: observed clothing and appearance that signal gender-typical presentation (e.g., color choices, formal wear, fabrics, patterns, and gender-specific accessories).
Self-socialization: children’s active role in shaping their gender identity through cognitive representations and expectations about gender, which then influence their behaviors and preferences.
Parent socialization: parental attitudes and behaviors that convey gender norms (e.g., traditional gender-role attitudes).
Gender concepts: knowledge and beliefs about gender categories, including gender labeling, gender stability/constancy, and gender importance.
Covariate: child’s total word production (vocabulary) used to control for overall language development when examining gender labeling effects.
Research Questions and Hypotheses
Primary aim: Examine factors associated with toddlers’ gender-typed appearance at age ~2 years, focusing on:
Child contributions: gender labeling (production and recognition).
Parent contributions: mothers’ gender-traditional attitudes.
Hypotheses:
H1 (Self-socialization): Toddlers’ production and recognition of gender labels would be associated with greater gender-typed appearance.
H2 (Socialization): Mothers’ more traditional gender attitudes would be associated with greater gender-typed appearance in children.
H3 (Bidirectionality/interaction): The association between gender labeling and appearance might be moderated by mothers’ gender traditionalism (i.e., stronger labeling–appearance links when mothers hold more traditional attitudes).
Exploratory angle: Test whether effects differ by ethnicity (Dominican American, Mexican American, African American) or by child gender.
Methods
Participants:
N = 175 toddlers (87 girls, 88 boys), mean age M = 2.09 years, SD = 0.12 (range 1.88–2.45).
Ethnicity: 63 Dominican American, 59 Mexican American, 53 African American.
Language: interviews conducted in dominant language (English, Spanish, or both); household income M = $26,744, SD = $22,990; education levels varied (37% < high school; 35% HS/GED; 28% some college).
Setting: home visits in a large Northeastern city, part of a larger longitudinal study on culture and school readiness.
Procedure:
In-person assessment with measures for child language, gender concepts, and gender-typed appearance; mothers completed interview measures in English/Spanish as appropriate.
Measures
Predictors and covariates
Children’s gender labeling (composite, 0–3): three criteria integrated into a single additive score:
1) Mother-reported child self-reference to gender (e.g., “Me boy/girl,” “I am a boy/girl”) in English and/or Spanish (0 = No, 1 = Yes; 56.1% answered Yes).
2) Child produced both boy and girl in English and/or Spanish during a broader vocabulary assessment (CDI) (0 = No, 1 = Yes; 41.0% produced both).
3) Gender pointing task: 3 trials with photo pairs (1 boy, 1 girl); child pointed to the gender; scoring 0 = fewer than 3 correct, 1 = all 3 correct; 21.4% fell into the 1 category.Composite score ranges 0–3. Language dominance was assessed to select English vs. Spanish CDI version; auxiliary probing if there was bilingual production.
Intercorrelations: productive gender labeling from CDI correlated with mother’s report of child self-labeling (r(171) = .34, p < .001) and with passing the gender pointing task (r(168) = .16, p = .038).
Covariate: Children’s total word production (vocabulary): summed across English and/or Spanish (Fenson et al.); used to control for broader language ability.
Group differences observed: girls produced more words than boys; Dominican American and African American children produced more words than Mexican American children.
Mothers’ gender traditionalism (composite): derived from two scales assessing gender roles and gender pressure on child (0–3 scale per item; α = .71).
Items include agreement with statements about marital roles and child rearing (e.g., “A man should help in the house, but housework and child care should be mainly a woman’s job”) and attitudes toward girly toys for girls (reverse-coded).
Also included a measure of gender pressure on the child: agreement with statements about being upset if the child acts like the other gender or wants to play with the other gender’s toys (0 = Not at all true to 3 = Very true; α = .68).
The two scales were combined into one gender-traditionalism composite. Across the sample, overall levels were somewhat low but variable (means below 1 for 54% of cases; 37% between 1 and 1.99; 9% ≥ 2).
Outcome: Children’s gender-typed appearance
Coding based on video observations of mother–child interactions from the larger longitudinal study.
Coding scheme: presence (0 = Absent, 1 = Present) of 8 gender-typed appearance aspects (for girls) or 5 for boys, aggregated into a total score:
Shared across genders: color (e.g., pink for girls; dark colors for boys), formal wear (e.g., party dresses for girls; ties for boys), fabric/fit (lace/satin for girls; baggy pants for boys), patterns/logos (hearts for girls; camouflage for boys).
Additional gender-specific items: girls – dresses/skirts, jewelry, hair accessories, trendy items (e.g., knee-high boots); boys – sports-themed wear (e.g., jerseys, tracksuits).
Reliability: κ values ranged from .77 to .98 across coders.
Score interpretation: sum across items; higher scores indicate greater gender-typed appearance.
Range details (example): a girl wearing a pink dress with hearts would score high on gender-typed appearance; a boy in baggy pants with a baseball tee would score moderately.
To enable fair comparison across genders, gender-typed appearance scores were z-scored within gender: GenderTypedAppearancez = (AppearanceScore − meanbygender)/SDby_gender.
Data handling and centering:
All continuous predictors were mean-centered before entering the regression models.
Data Analyses
Analytic plan: Hierarchical multiple regression predicting toddlers’ gender-typed appearance from predictors and covariates, with a test of interaction effects.
Step 1: Enter main predictors and covariates:
Predictors: children’s gender labeling, mothers’ gender traditionalism.
Covariates: child gender (coded boys as reference), ethnicity (two dummy variables with Dominican American as reference), and children’s total productive vocabulary (covariate).
Continuous variables centered: gender labeling, traditionalism, vocabulary.
Step 2: Add interaction term: GenderLabeling × GenderTraditionalism.
Equations (in LaTeX):
Step 1 model:
ext{GenderTypedAppearance} = eta0 + eta1( ext{GenderLabeling}) + eta2( ext{GenderTraditionalism}) + eta3( ext{ChildGender}) + eta4( ext{DominicanAmerican}) + eta5( ext{MexicanAmerican}) + eta_6( ext{Vocabulary}) + oldsymbol{
u}Step 2 model (adds interaction):
ext{GenderTypedAppearance} = eta0 + eta1( ext{GenderLabeling}) + eta2( ext{GenderTraditionalism}) + eta3( ext{ChildGender}) + eta4( ext{DominicanAmerican}) + eta5( ext{MexicanAmerican}) + eta6( ext{Vocabulary}) + eta7( ext{GenderLabeling} imes ext{GenderTraditionalism}) + oldsymbol{
u}The key interpretation is in the standardized regression coefficients (betas) and their p-values.
Descriptive and supplemental checks:
Examined potential interactions by gender and by ethnicity (supplemental analyses).
Centering and interpretation of main effects in the presence of interactions were followed.
Results
Step 1: Main effects (predicting z-scored gender-typed appearance)
Gender labeling significantly predicted gender-typed appearance after accounting for vocabulary, child gender, and ethnicity:
β = .28, t(165) = 2.91, p = .004.
Mothers’ gender traditionalism did not predict gender-typed appearance:
β = 0.00, t(165) = 0.04, p = .969.
Child gender, ethnicity, and vocabulary showed the following general patterns (control variables):
Vocabulary had a negative trend (larger vocabulary associated with lower appearance score in this model) but was not robust (p ≈ .064).
No significant main effects of Mexican American or African American ethnicity relative to Dominican American on appearance after accounting for other factors.
Model fit: Step 1 explained a meaningful portion of variance in gender-typed appearance beyond covariates (exact R^2 not provided here).
Step 2: Interaction (GenderLabeling × GenderTraditionalism)
Interaction effect: β = .03, t(164) = 0.43, p = .668.
Change in R^2: ΔR^2 ≈ .001 (0.1%), indicating no meaningful improvement with the interaction term.
Supplemental analyses (by gender and ethnicity):
No significant gender or racial/ethnic interactions with gender labeling and mothers’ gender traditionalism predicting gender-typed appearance (Table 3 results summarized as non-significant interactions).
Overall interpretation of results:
Consistent with a self-socialization account: toddlers’early conceptual understanding of gender categories (as indexed by gender labeling) relates to more gender-typed appearance by age ~24 months.
Mothers’ gender traditionalism did not predict toddlers’ gender-typed appearance in this cross-sectional design, suggesting that the child’s own gender knowledge may exert a stronger influence on appearance at this very early stage.
The association between labeling and appearance appears robust across ethnic groups in this sample, though the strength and direction of other associations may vary by group, as suggested by descriptive differences in labeling by ethnicity (e.g., Dominican American girls showing higher labeling than other groups).
Discussion and Implications
The findings support self-socialization as an early mechanism in gender-typed appearance:
By age 24 months, children’s awareness and use of gender labels are linked to more gender-typed dress and appearance, suggesting that gender concepts begin to influence how children present themselves physically in everyday life.
This aligns with prior work linking gender labeling to gender-typed play and to other gender cognitions (e.g., gender stability, gender identification) in older preschoolers (3–6 years) and extends it to toddlers.
Cross-cultural relevance:
The association held in a diverse sample including Dominican American, Mexican American, and African American toddlers, underscoring generalizability beyond Euro-American samples and across language contexts.
Implications for theory:
Supports the notion that gender self-concepts (self-socialization) are active early in development and can shape tangible appearance-based expressions of gender without requiring explicit parental traditionalism.
The null effect for maternal gender traditionalism in predicting appearance at this age suggests that parental influence on appearance may be less pronounced at 2 years or potentially detectable only via other measures (e.g., actual parent behaviors, language, or longitudinal change).
Practical and ethical considerations:
Early gender-typed appearance reflects social and cultural norms; understanding its development can inform discussions about gender socialization, parental guidance, and the balance between child autonomy and parental influence.
The study highlights the importance of considering cultural and linguistic contexts when assessing gender development and appearance in early childhood.
Limitations and Future Directions
Limitations:
Cross-sectional design at a single wave limits causal inference about directionality (bidirectional influence between labeling and appearance is possible but not testable here).
Measures of mothers’ gender traditionalism relied on self-reports which may be susceptible to social desirability bias; two related scales were combined into a single composite, potentially masking distinct aspects of traditionalism.
The study did not include fathers or other caregivers; paternal attitudes and behaviors may have unique influences on toddlers’ appearance.
Direct data on toddlers’ clothing input (i.e., whether the child influences outfit choices) were not collected; future work should assess child input and preferences more explicitly.
Future directions:
Longitudinal, multi-wave designs to assess directionality and mechanism of the bidirectional relationship between gender labeling and gender-typed appearance.
Cross-cultural and cross-language replication with larger samples to test potential moderation by ethnicity, language, and cultural norms.
Include fathers and other caregivers to parse familial contributions to gender-typed appearance over time.
Investigate how early gender labeling and appearance relate to later developmental milestones (e.g., gender constancy, rigidity in dress) and to transitions in appearance across early childhood.
Explore whether appearance changes precede shifts in gender labeling or vice versa, and whether appearance acts as a feedback mechanism that strengthens gender identity.
Takeaway Messages
By around 24 months, toddlers’ emerging gender concepts (as indexed by gender labeling) are meaningfully associated with gender-typed appearance, supporting self-socialization theories at an early developmental stage.
Mothers’ gender traditionalism did not predict toddler gender-typed appearance in this sample, suggesting that children’s own gender knowledge may be a more proximal driver of appearance at age 2.
The observed association appears to generalize across diverse ethnic groups (Dominican American, Mexican American, African American), highlighting the potential universality of early self-socialization processes in gender appearance.
The directionality and mechanisms of the labeling–appearance relation warrant longitudinal investigation to determine causal pathways and potential reciprocal effects.
Key Figures and Tables (described)
Table 1: Means and standard deviations for key variables by gender and ethnicity; highlights that gender labeling varies by gender and ethnicity (e.g., African American girls show lower labeling than Dominican American and Mexican American girls).
Table 2: Hierarchical regression results for Step 1 and Step 2 predicting toddlers’ gender-typed appearance; key finding: gender labeling is a significant predictor in Step 1 (β ≈ .28, p ≈ .004); mothers’ traditionalism non-significant (β ≈ 0, p ≈ .97); interaction non-significant (β ≈ .03, p ≈ .668).
Table 3: Supplemental regression results testing interactions by gender and by ethnicity; none reached significance (no robust interaction effects).
References (selected for context)
Fagot, B. I., & Leinbach, M. D. (1989, 1993). On gender labeling and gender-typed behaviors.
Gelman, S. A., Taylor, M. G., & Nguyen, S. P. (2004). Mother-child conversations about gender.
Halim, M. L., Ruble, D. N., Tamis-LeMonda, C. S., Zosuls, K. M., & Shrout, P. E. (2013a, 2013b, 2014). Studies on appearance rigidity and gender concepts in early childhood and cross-cultural samples.
Martin, C. L., Ruble, D. N., & Szkrybalo, J. (2002). Cognitive theories of early gender development.
Zosuls, K. M., Ruble, D. N., Tamis-LeMonda, C. S., Shrout, P. E., et al. (2009, 2014). Acquisition and use of gender labels in infancy and toddlerhood.
Other foundational works cited in the article (e.g., Kohlberg, Tajfel & Turner) for foundational theories.
Note
The content above is a comprehensive, study-focused set of notes drawn from the provided transcript. It captures major and minor points, methods, results, interpretations, and implications, and includes defined equations and LaTeX-formatted expressions where appropriate.
Background and Theoretical Framework
Topic: This research delves into early gender development, specifically investigating how gender-typed appearance in toddlers correlates with their developing gender concepts and their mothers’ gender attitudes. The study bridges self-socialization and broader socialization theories to understand early gender expression.
Self-socialization theories (e.g., Martin, Ruble, & Szkrybalo, 2002) posit that as children cognitively grasp gender categories, they become intrinsically motivated to align with gender stereotypes. This adherence serves to affirm their sense of self within their gender group and maintain self-esteem. Key ideas include:
The fundamental understanding of one's own gender and the existence of gender categories propels children to actively seek and process gender-relevant information. This cognitive drive leads them to conform to established gender stereotypes as a means of becoming competent and accepted members of their respective gender groups.
Early abilities such as accurately labeling themselves and others by gender are strong predictors of higher levels of gender-typing in subsequent behaviors, such as preferred play activities and toy choices.
This process is theorized to be bidirectional: a child's evolving gender concepts can directly influence their observable behaviors (like chosen appearance), and in turn, their experiences with these behaviors (e.g., being perceived in a gender-typical way due to appearance) can reinforce and solidify their internal gender concepts.
Crucial early developmental milestones in this framework involve the child's ability to produce and recognize gender labels, eventually progressing to understanding gender constancy (the knowledge that one's gender remains stable over time despite changes in appearance or activities).
Socialization theories emphasize the profound influence of external environmental factors (such as parents, peers, and media) on the development of gendered beliefs and behaviors:
Parents frequently engage in gender-stereotypical practices, often by providing toys traditionally associated with their child's gender and by offering differential responses to gender-typical versus counter-stereotypical play. For instance, a parent might praise a girl for playing with dolls more than for playing with trucks.
Research consistently shows that more traditional gender attitudes held by parents are significantly linked to more traditional gender cognitions and behaviors observed in their children. This suggests that explicit or implicit parental transmission of gender norms is influential.
Gender-typed appearance as a construct:
Appearance is considered a particularly stable and salient indicator of gender-typing. This is partly because children often wear the same outfit for extended periods throughout a day, consistently signaling their gendered presentation to others.
Appearance is intricately linked to societal constructions of gender, holding significant cultural meaning. It strongly influences how others perceive a child's gender and can shape subsequent social interactions and expectations.
Previous scholarly work (e.g., Halim et al., 2013a, 2014) has demonstrated that appearance-typical clothing becomes increasingly gender-typed as children age and that it relates significantly to other key gender concepts, such as gender stability (understanding gender as consistent over time) and gender importance (how salient gender is to a child's identity).
Significance of toddler-age focus:
Toddlerhood (around age 2) is a pivotal developmental period because it marks the nascent emergence of gender labeling abilities and the initial stages of gender-typed dress. This makes it an ideal, critical early window to empirically examine the dynamic interplay between self-socialization and external socialization, particularly in a potentially bidirectional manner.
Cross-cultural relevance:
To assess the broader generalizability of these theoretical perspectives beyond predominantly Euro-American populations, this study purposefully includes toddlers from diverse cultural and linguistic backgrounds, specifically Dominican American, Mexican American, and African American communities. This approach helps determine if the observed patterns are universal or culturally specific.
Key Concepts and Definitions
Gender labeling: Refers to the developing cognitive and linguistic ability of children to both produce (e.g., say "I'm a boy/girl") and comprehend gender terms for themselves and others (e.g., understanding who is a "boy" or a "girl" in their environment).
Gender-typed appearance: Encompasses observable characteristics of clothing, accessories, and overall presentation that are culturally understood to signal a typical gender presentation. Examples include specific color choices (e.g., pink for girls, dark colors for boys), types of formal wear, characteristic fabrics (e.g., lace/satin for girls, denim for boys), specific patterns (e.g., hearts for girls, camouflage for boys), and gender-specific accessories.
Self-socialization: This refers to the active and internal process by which children construct and shape their own gender identity. It is driven by their cognitive representations and expectations about what it means to be a boy or a girl, which subsequently influence their preferences, behaviors, and self-presentation.
Parent socialization: This concept describes the various parental attitudes, beliefs, and overt behaviors that explicitly or implicitly convey gender norms and expectations to their children, thereby influencing the child's gender development. An example is holding traditional gender-role attitudes.
Gender concepts: This is a broader term referring to a child's overall knowledge and beliefs concerning gender categories. It includes sub-components such as gender labeling (identifying gender), gender stability (gender is constant over time), gender constancy (gender is constant regardless of appearance changes), and gender importance (the personal relevance or salience of gender to the child).
Covariate: In this study, the child's total word production (overall vocabulary size) is used as a covariate. It is included in the statistical models to control for general language development, ensuring that any observed effects of gender labeling are not merely due to differences in broader linguistic ability.
Research Questions and Hypotheses
Primary aim: The central objective of this research is to meticulously examine the specific factors associated with toddlers’ gender-typed appearance at approximately 2 years of age. The investigation focuses on two main categories of contributions:
Child contributions: Specifically, the study explores the role of toddlers’ gender labeling abilities, encompassing both their production and recognition of gender terms.
Parent contributions: The study also investigates the influence of mothers’ gender-traditional attitudes on their children’s gender-typed appearance.
Hypotheses:
H1 (Self-socialization): Based on self-socialization theories, it was hypothesized that toddlers’ greater facility in both producing and recognizing gender labels would be significantly associated with a higher degree of gender-typed appearance.
H2 (Socialization): Drawing from socialization theories, it was hypothesized that mothers holding more traditional gender attitudes would be associated with a greater prevalence of gender-typed appearance in their children.
H3 (Bidirectionality/interaction): An exploratory hypothesis suggesting a potential moderating effect. It was posited that the association between gender labeling (child contribution) and appearance might be stronger when mothers also held more traditional gender attitudes (parent contribution), thus indicating an interaction effect.
Exploratory angle: Beyond the primary hypotheses, the study also aimed to explore whether the hypothesized effects might differ significantly based on the child's specific ethnic group (Dominican American, Mexican American, or African American) or the child's own gender (boy vs. girl).
Methods
Participants:
The study cohort consisted of N = 175 toddlers, with a nearly equal distribution of 87 girls and 88 boys. The mean age of the toddlers at the time of assessment was M = 2.09 years, with a standard deviation of SD = 0.12 years, and an age range spanning from 1.88 to 2.45 years.
The sample was ethnically diverse, including 63 Dominican American, 59 Mexican American, and 53 African American children, reflecting a key aspect of the study's cross-cultural relevance.
Language: Extensive measures ensured that interviews and assessments were conducted in the family's dominant language, which could be English, Spanish, or a combination of both. Economic diversity was also present, with the mean household income being M = , SD = . Parental education levels varied considerably, with 37% of mothers reporting less than a high school education, 35% having a high school diploma or GED, and 28% having some college education.
Setting: The research was conducted through home visits in a large Northeastern city within the United States. These assessments were part of a broader, ongoing longitudinal study focused on culture and school readiness.
Procedure:
Data collection involved in-person assessments during a home visit. These assessments included direct measures of child language proficiency, specific gender concepts, and observations used to code gender-typed appearance.
Concurrently, mothers completed various interview measures, which were administered in either English or Spanish as per their preference and linguistic ability.
Measures
Predictors and covariates
Children’s gender labeling (composite score, ranged 0–3): This composite score was meticulously constructed by integrating three distinct criteria, each scored as 0 (absent) or 1 (present):
1) Mother-reported child self-reference to gender: Mothers were asked if their child spontaneously used gender labels to refer to themselves (e.g., saying “Me boy/girl” or “I am a boy/girl”) in either English and/or Spanish. of mothers reported their child doing so.
2) Child produced both boy and girl during vocabulary assessment: During a broader vocabulary assessment using the MacArthur-Bates Communicative Development Inventories (CDI), children were scored if they produced both the terms "boy" and "girl" in English and/or Spanish ( produced both).
3) Gender pointing task: Children participated in a task comprising 3 trials, where they were presented with photo pairs (one boy, one girl) and asked to point to the gender. This was scored 0 if they correctly identified fewer than 3, and 1 if they correctly identified all 3. achieved a score of 1.
These three binary scores (0 or 1) were summed to create a composite gender labeling score ranging from 0 to 3. Language dominance was assessed beforehand to ensure the appropriate English or Spanish CDI version was used, and auxiliary probing was employed for bilingual children. The intercorrelations among the components revealed that productive gender labeling from the CDI correlated moderately with mother’s report of child self-labeling ( r(171) = .34, p < .001 ) and weakly with passing the gender pointing task (), indicating they tap related but somewhat distinct aspects of gender knowledge.
Covariate: Children’s total word production (vocabulary): This measure, derived from the summed vocabulary across English and/or Spanish versions of the CDI, served as a crucial covariate. Its inclusion in the models controlled for broader language ability, ensuring that any effects attributed to gender labeling were not confounded by general linguistic proficiency. Group differences were observed, with girls generally producing more words than boys, and Dominican American and African American children producing more words than Mexican American children.
Mothers’ gender traditionalism (composite score): This composite was constructed from two distinct scales, each scored on a 0–3 point scale, with an acceptable internal consistency reliability ().
The first scale assessed agreement with statements regarding marital roles and child-rearing practices (e.g., “A man should help in the house, but housework and child care should be mainly a woman’s job”). Some items were reverse-coded (e.g., attitudes toward girly toys for girls) to ensure consistent directionality of traditionalism.
The second scale measured gender pressure on the child, inquiring about mothers’ agreement with statements expressing upset if their child acts like the other gender or shows preference for the other gender’s toys (scored 0 = Not at all true to 3 = Very true; ). Both scales were combined into a single gender-traditionalism composite, which allowed for a broad assessment of maternal traditionalism. Across the sample, overall levels of traditionalism were somewhat low but exhibited variability, with means below 1 for of cases, between 1 and 1.99 for , and for of cases.
Outcome: Children’s gender-typed appearance
Coding for gender-typed appearance was meticulously performed based on video observations of naturalistic mother–child interactions, recorded as part of the larger longitudinal study. Each video was approximately 10 minutes long, capturing typical daily interactions.
A comprehensive coding scheme was developed to identify the presence (0 = Absent, 1 = Present) of specific gender-typed appearance aspects. Eight aspects were coded for girls and five for boys, aggregated into a total score:
Shared across genders (but with gender-specific manifestations): color (e.g., prevalence of pink, purple, or pastel colors for girls; dark colors like blue, brown, or black for boys), formal wear (e.g., elaborate party dresses or ruffled skirts for girls; ties, dress shirts, or blazers for boys), fabric/fit (e.g., lace, satin, ruffles, or fitted styles for girls; baggy pants, denim, or utilitarian fabrics for boys), and patterns/logos (e.g., hearts, flowers, or princesses for girls; camouflage, superheroes, or sports logos for boys).
Additional gender-specific items: for girls, these included wearing dresses/skirts, jewelry (e.g., necklaces, earrings), hair accessories (e.g., bows, headbands), and trendy items popular for girls (e.g., knee-high boots, tutus). For boys, additional items included sports-themed wear (e.g., jerseys, tracksuits with athletic motifs) and specific masculine accessories (e.g., baseball caps).
Inter-rater reliability was strong, with Kappa () values ranging from to across different coders, indicating excellent agreement. The score calculation involved summing the presence/absence of these items, with higher scores signifying a greater degree of gender-typed appearance. For example, a girl wearing a pink dress adorned with hearts and a matching hair bow would score high, while a boy in baggy denim pants and a baseball-themed tee would score moderately. To facilitate equitable comparison across genders, the raw gender-typed appearance scores were transformed into z-scores in a gender-specific manner: This standardization accounts for the inherent differences in the number and type of items coded for each gender, allowing for interpretations of relative gender-typicality.
Data handling and centering:
All continuous predictor variables (gender labeling, traditionalism, vocabulary) were mean-centered prior to their inclusion in the regression models. This practice helps to reduce multicollinearity among predictors and makes the interpretation of main effects more meaningful, especially when interaction terms are included.
Data Analyses
Analytic plan: Hierarchical multiple regression was employed as the primary statistical method to predict toddlers’ gender-typed appearance. This approach allowed for the systematic examination of main effects of the predictors and covariates, followed by a test of interaction effects in a subsequent step.
Step 1: The main predictors and covariates were entered into the regression model:
Predictors: Children’s gender labeling (mean-centered) and mothers’ gender traditionalism (mean-centered).
Covariates: Child gender (coded as a dummy variable, with boys serving as the reference group), ethnicity (represented by two dummy variables, with Dominican American as the reference group), and children’s total productive vocabulary (mean-centered) to control for general language ability.
Step 2: An interaction term, GenderLabeling GenderTraditionalism, was added to the model to assess whether the association between gender labeling and appearance was moderated by mothers’ gender traditionalism.
Equations (in LaTeX) outlining the regression models:
Step 1 model (main effects):
ext{GenderTypedAppearance}z = eta0 + eta1( ext{GenderLabeling}) + eta2( ext{GenderTraditionalism}) + eta3( ext{ChildGender}{ ext{boy vs girl}}) + eta4( ext{Ethnicity}{ ext{DomAmer vs MexAmer}}) + eta5( ext{Ethnicity}{ ext{DomAmer vs AfAmer}}) + eta6( ext{Vocabulary}) + ext{ extepsilon} Here, eta0 represents the intercept, and eta1 through eta6 are the regression coefficients for the respective predictors and covariates, indicating their unique contribution to gender-typed appearance when other variables are held constant. represents the error term.Step 2 model (adding the interaction):
ext{GenderTypedAppearance}z = eta0 + eta1( ext{GenderLabeling}) + eta2( ext{GenderTraditionalism}) + eta3( ext{ChildGender}{ ext{boy vs girl}}) + eta4( ext{Ethnicity}{ ext{DomAmer vs MexAmer}}) + eta5( ext{Ethnicity}{ ext{DomAmer vs AfAmer}}) + eta6( ext{Vocabulary}) + eta7( ext{GenderLabeling} imes ext{GenderTraditionalism}) + ext{ extepsilon}
In this model, eta_7 specifically tests the interaction effect. The key interpretation of the results focuses on the standardized regression coefficients () and their associated p-values to determine statistical significance.
Descriptive and supplemental checks:
Additional exploratory analyses were conducted to examine potential interactions by child gender and by ethnicity (as summarized in Table 3 of the original study). The process of centering variables was carefully followed, ensuring that main effects could be robustly interpreted even in the presence of (or absence of) interaction terms.
Results
Step 1: Main effects (predicting z-scored gender-typed appearance)
Gender labeling emerged as a statistically significant positive predictor of gender-typed appearance. This effect remained robust even after controlling for the child’s vocabulary, gender, and ethnicity:
Standardized regression coefficient eta = .28 , associated t-statistic , and a highly significant p-value . This indicates that toddlers with more developed gender labeling abilities tended to exhibit a higher degree of gender-typed appearance.
In contrast, mothers’ gender traditionalism did not significantly predict children’s gender-typed appearance:
Standardized regression coefficient eta = 0.00 , t-statistic , and a non-significant p-value . This suggests that, in this cross-sectional snapshot, parental traditional attitudes, as measured, did not directly influence the child's gender-typed appearance at this early age.
Control variables (child gender, ethnicity, and vocabulary) showed the following general patterns:
Vocabulary (total word production) displayed a negative trend, suggesting that a larger vocabulary was somewhat associated with a lower gender-typed appearance score in this specific model, though this trend was not statistically robust ().
There were no significant main effects of Mexican American or African American ethnicity relative to the Dominican American reference group on gender-typed appearance after accounting for other factors. This implies that, when controlling for gender labeling, maternal traditionalism, and vocabulary, the average level of gender-typed appearance did not differ significantly across these ethnic groups.
Model fit: Step 1 of the regression analysis collectively explained a meaningful portion of the variance in gender-typed appearance above and beyond the covariates (an exact value was not explicitly provided in the excerpt but implied as substantial).
Step 2: Interaction (GenderLabeling GenderTraditionalism)
The interaction effect between gender labeling and mothers’ gender traditionalism was not statistically significant:
Standardized regression coefficient eta = .03 , t-statistic , and a non-significant p-value . This indicates that the strength of the association between a child's gender labeling and their gender-typed appearance does not significantly vary based on the mother's level of gender traditionalism.
The change in R-squared () upon adding the interaction term was approximately (), confirming that the interaction term did not contribute any meaningful additional explanatory power to the model.
Supplemental analyses (by gender and ethnicity):
Further exploratory analyses investigated potential interactions of the main predictors with child gender and with ethnicity. As summarized in Table 3, none of these supplemental interaction terms reached statistical significance, suggesting that the observed associations between gender labeling and appearance were consistent across boys and girls, and across the three diverse ethnic groups sampled.
Overall interpretation of results:
The findings strongly align with a self-socialization account: toddlers’ emergent conceptual understanding of gender categories, as indexed by their ability to label gender, is significantly related to their presentation of a more gender-typed appearance by approximately 24 months of age.
Mothers’ gender traditionalism did not predict toddlers’ gender-typed appearance in this cross-sectional design. This suggests that at this very early developmental stage, the child’s own budding knowledge and internal representations of gender may exert a more direct and stronger influence on their physical appearance choices (or parental influences on those choices) than explicit maternal attitudes.
Crucially, the observed association between gender labeling and appearance appears robust across diverse ethnic groups (Dominican American, Mexican American, and African American) within this sample. This highlights the potential generalizability of early self-socialization processes, though descriptive differences in gender labeling by ethnicity (e.g., Dominican American girls reportedly showing higher labeling scores than other groups) still warrant consideration in broader contexts.
Discussion and Implications
The findings provide compelling support for self-socialization as an early and active mechanism in the development of gender-typed appearance:
By the tender age of 24 months, children’s growing awareness and proficient use of gender labels are demonstrably linked to their adoption of more gender-typed dress and appearance styles. This suggests that even at this early stage, nascent gender concepts are beginning to significantly influence how children choose to present themselves physically in their everyday lives, aligning with their internal understanding of gender.
This discovery broadens and extends prior research that had primarily linked gender labeling to gender-typed play and other gender cognitions (e.g., gender stability, self-identification with gender) in older preschoolers (typically ages 3–6 years). This study successfully demonstrates these links in a younger, toddler population, pushing back the timeline of self-socialization's observable effects.
Cross-cultural relevance:
A significant strength of this study is that the observed association between gender labeling and gender-typed appearance held true across a diverse sample that included Dominican American, Mexican American, and African American toddlers. This underscores the generalizability of these self-socialization processes beyond typically studied Euro-American samples and across varying linguistic contexts, suggesting a potentially universal developmental mechanism.
Implications for theory:
The study bolsters the theoretical notion that gender self-concepts, rooted in self-socialization processes, become active and influential remarkably early in development. These internal concepts can shape tangible, appearance-based expressions of gender without necessarily requiring explicit or measurable traditional gender attitudes from parents at this specific age.
The null effect observed for maternal gender traditionalism in predicting appearance at age 2 is noteworthy. It suggests that parental influence on a child's gender-typed appearance may be less direct or pronounced at this very early age. Alternatively, it could indicate that parental influence might manifest through different, unmeasured pathways (e.g., specific parental behaviors, subtle linguistic cues, or implicit reinforcement) or that its effects become more detectable through longitudinal observation as children age.
Practical and ethical considerations:
Understanding the early development of gender-typed appearance is crucial for informing broader discussions about gender socialization, effective parental guidance, and the delicate balance between fostering a child's autonomy in self-expression and the influence of parental and societal norms. Early gender-typed appearance is often a reflection of complex social and cultural norms.
The inclusive nature of the study highlights the paramount importance of considering diverse cultural and linguistic contexts when assessing gender development, gender expression, and appearance choices in early childhood. This approach facilitates a more nuanced and culturally sensitive understanding of child development.
Limitations and Future Directions
Limitations:
Cross-sectional design: The study's design, capturing data at a single point in time, inherently limits the ability to draw definitive causal inferences regarding the directionality of effects. While a bidirectional influence between gender labeling and appearance is theoretically plausible, this design cannot empirically test such dynamic relationships.
Measures of mothers’ gender traditionalism: The reliance on mothers’ self-reports for assessing gender traditionalism poses a potential limitation due to the susceptibility to social desirability bias, where individuals may report attitudes they perceive as more socially acceptable. Furthermore, combining two related but potentially distinct scales into a single composite might have masked more nuanced aspects or differential impacts of specific dimensions of traditionalism.
Absence of other caregivers: The study exclusively focused on mothers’ attitudes and did not include data from fathers or other primary caregivers. Paternal attitudes, as well as the beliefs and behaviors of grandparents or extended family members, may exert unique and significant influences on toddlers’ appearance and gender development.
The study did not collect direct data on toddlers’ agency in clothing choices (i.e., whether the child actively influences or expresses preferences for their outfits). Future research should endeavor to explicitly assess child input and preferences regarding appearance to more fully capture the self-socialization aspect.
Future directions:
Longitudinal, multi-wave designs are strongly recommended to precisely assess the directionality and underlying mechanisms of the proposed bidirectional relationship between gender labeling and gender-typed appearance. Such designs would allow for tracking developmental trajectories and potential reciprocal effects over time.
Cross-cultural and cross-language replication with larger, more diverse samples is vital to thoroughly test for potential moderation by ethnicity, specific language backgrounds, and varying cultural norms. This would strengthen the understanding of generalizability and cultural specificity.
Future studies should actively include fathers and other key caregivers to comprehensively parse the multifaceted familial contributions to gender-typed appearance and gender development over time.
Investigate precisely how early gender labeling and appearance relate to later developmental milestones in gender understanding, such as the achievement of gender constancy, potential rigidity or flexibility in dress and toy preferences, and subsequent transitions in appearance across the broader span of early childhood.
Explore whether changes in appearance precede shifts in gender labeling or vice versa, and critically examine whether gender-typed appearance itself functions as a feedback mechanism that reinforces and strengthens a child’s emergent gender identity.
Takeaway Messages
By approximately 24 months of age, toddlers’ developing gender concepts, as effectively indexed by their ability to label gender, are significantly and meaningfully associated with their expression of gender-typed appearance. This finding robustly supports self-socialization theories taking effect at a very early developmental stage.
Mothers’ gender traditionalism, as measured in this study, did not predict toddler gender-typed appearance in this specific sample. This suggests that children’s own burgeoning gender knowledge and self-concept may function as a more direct and proximal driver of their appearance choices or expressions at age 2.
The observed association between gender labeling and gender-typed appearance appears to generalize consistently across diverse ethnic groups, including Dominican American, Mexican American, and African American toddlers. This highlights the potential universality of early self-socialization processes in shaping gender expression through appearance.
Due to the cross-sectional nature of the study, the precise directionality and underlying mechanisms of the labeling–appearance relationship warrant further rigorous longitudinal investigation to delineate conclusive causal pathways and potential reciprocal effects.
Key Figures and Tables (described)
Table 1: This table presents the means and standard deviations for all key study variables, disaggregated by both child gender and ethnicity. Crucially, it highlights descriptive variations, such as how gender labeling scores differ across genders and ethnic groups (e.g., African American girls descriptively showed lower gender labeling scores compared to Dominican American and Mexican American girls, indicating important baseline differences).
Table 2: This table details the hierarchical regression results for both Step 1 (main effects) and Step 2 (interaction effect) predicting toddlers’ z-scored gender-typed appearance. The central finding showcased here is that gender labeling is a significant positive predictor in Step 1 ( eta ext{ extasciitilde} .28, p ext{ extasciitilde} .004 ). Conversely, mothers’ traditionalism was found to be non-significant ( eta ext{ extasciitilde} 0, p ext{ extasciitilde} .97 ), and the interaction term in Step 2 also proved non-significant ( eta ext{ extasciitilde} .03, p ext{ extasciitilde} .668 ).
Table 3: This table summarizes the supplemental regression results that explicitly tested for interaction effects by child gender and by ethnicity. It indicated that none of these potential interaction terms reached statistical significance, reinforcing the main finding that the primary associations hold across these demographic subgroups.
References (selected for context)
Fagot, B. I., & Leinbach, M. D. (1989, 1993). Influential work on the acquisition of gender labels and their relationship to the emergence of gender-typed behaviors in young children.
Gelman, S. A., Taylor, M. G., & Nguyen, S. P. (2004). Research focusing on mother-child conversations and their role in the transmission of gender concepts.
Halim, M. L., Ruble, D. N., Tamis-LeMonda, C. S., Zosuls, K. M., & Shrout, P. E. (2013a, 2013b, 2014). A series of studies investigating appearance rigidity, the development of gender concepts in early childhood, and cross-cultural examinations of these phenomena.
Martin, C. L., Ruble, D. N., & Szkrybalo, J. (2002). A seminal review and theoretical framework outlining cognitive theories that explain early gender development and the child's active role.
Zosuls, K. M., Ruble, D. N., Tamis-LeMonda, C. S., Shrout, P. E., et al. (2009, 2014). Foundational research on the acquisition and subsequent use of gender labels during infancy and toddlerhood.
Other foundational works cited in the article (e.g., Kohlberg's cognitive-developmental theory of gender, Tajfel & Turner's social identity theory) provide the broader theoretical underpinnings for the study.