Gender Differences in Emotional Closeness Between Preschool ...

4 downloads 139 Views 133KB Size Report
Tell us about your experience on SpringerLink and you could win a MacBook Pro. Take survey. Download PDF · Sex Roles. June 1998 , Volume 38, Issue 11, ...
Sex Roles‚ Vol. 38 ‚ Nos. 11/12 ‚ 1998

Gender Differences in Em otion al Closeness Between Preschool Ch ildren an d Th eir 1 Mothers Joyce F. Ben enson ‚2 Dean n a Mor ash ‚ an d Harriet Petrakos McG ill University

The emotional closen ess between m others and their children was exam in ed to determ ine if m other¯daughter pairs exhibit closer emotional relation ships than m oth er¯son pairs. Forty-on e m other¯child predom in an tly Cau casian dyad s were videotaped durin g a semistructured play setting. The children were 4 (10 m ales an d 10 fem ales) an d 5 years of age (10 m ales an d 11 fem ales). Em otion al closeness was defined as physical proxim ity‚ m utual eye contact‚ and global level of enjoym ent. Results dem on strated that com pared to boys‚ girls were ph ysically closer to their m others ‚ engaged in m ore m utual eye contact with their m others‚ and were rated higher on global enjoym ent. Results are con sistent with Chod orow ’s theory that gen der iden tity developm ent is related to the emotion al relation ships that develop between children and their m others. The de ve lopme nt of ge nde r identity in childre n has long be en an are a of curiosity and de bate . Diffe rent theoretical vie wpoints ‚ emphasizing either environme ntal or biological influe nces‚ have be en utilize d to e xamine the factors crucial in e liciting gende r-type d diffe re nce s be tween males and females. Although the controve rsy continue s‚ it is currently assumed that en1

Support for this project was provide d to the first author by a Faculty of Graduate Studies and Rese arch Award from McGill University. We are grateful to Prof. David McClelland who made this study possible ‚ and to teachers Mary Mackie and JoAnne Pre ssman ‚ and the children and their mothers ‚ at the Community Nurse ry School. Thanks also to Gerry Delli Quadri ‚ Carol Franz ‚ Patrick Glaude ‚ Indra Hardy‚ Jessica Kudlats‚ Caroline McLeod ‚ Ruth Jacobs ‚ Stephen Kelner ‚ Dan Mroczak ‚ Josie Morello‚ Z ainab Muhsin‚ Loren Naidoo ‚ Eugene Taylor ‚ and Catherine Zygmuntowicz for their careful coding of the videotape s. 2 T o whom co rr e sponde nce should be add re sse d at De par tm e nt of E ducatio nal an d Counselling Psychology‚ McGill Unive rsity‚ 3700 McTavish Street ‚ Montreal ‚ Q uebec ‚ Canada H3A 1Y2; e-mail: Be nenson@ Education.McGill.ca. 975 0360 ¯0025/98/0600 ¯0975$15.00/0

Ó

1998 Plenum Publishing Corporation

976

Ben en son et al.

vironme nt and biology inte ract (for a re vie w‚ se e Golombok & Fivush ‚ 1994) . Evide nce continue s to accumulate clarifying the influe nce of biology on diffe re ntial ge nde r-type d de ve lopme nt (e .g.‚ Be re nbaum & Snyde r ‚ 1995; for a revie w‚ se e Ehrhardt & Me ye r-Bahlburg ‚ 1981) . Few explanations ‚ howe ve r‚ specify how the e nvironme nt influe nce s gende r-type d behavior (for revie ws‚ see Lytton & Romney‚ 1991; Maccoby & Jacklin ‚ 1974; Ste rn & Karrake r‚ 1989) . We propose that the focus of past socialization research has bee n too narrow. We pre dict that an examination of more subtle factors ‚ spe cifically the de ve lopme nt of early emotional re lationships betwee n childre n and their mothers‚ may provide e vide nce of differe ntial expe rience s for male s and female s. Maccoby and Jacklin ( 1974) conducte d an e xtensive revie w of the literature addre ssing ge nde r differences. Afte r e xamining the re search‚ they conclude d that the re was little evide nce to support the hypothe sis that parents treat boys and girls diffe re ntly during the first five ye ars of life . More recently‚ Lytton and Romney ( 1991) re viewed 172 studie s involving pare ntal socialization of childre n in North America and othe r We ste rn countrie s. The revie wers atte mpted to distinguish be tween differe nt areas of socialization by clustering variable s from se parate studie s unde r gene ral categorie s such as “ disciplin ary strictne ss ‚” “ e ncourage me nt of ge nde r-type d activitie s‚” or “ discourage ment of aggre ssiveness.” The y conclude d that the only significant differences were that pare nts encourage stereotypical gender-type d toy and activity play and that pare nts use gre ater physical punishment and e ncourage achie vement more in boys vs. girls. Anothe r important revie w (Ste rn & Karrake r‚ 1989) e xamine d ge nde r socialization of infants. All of the studie s manipulate d the gende r labe l of the infant. The authors found that although half of the studie s re porte d at le ast one main effe ct involving gende r manipulation ‚ the findings were inconsiste nt and only 18% of all the variable s conside red across a varie ty of studie s were significantly influe nce d by gende r labe ling. Therefore ‚ the researche rs cautione d against assuming that infant gende r labe ling studie s provide strong evide nce for the influe nce of early ge nde r role socialization. In general‚ these literature reviews provide little support for the influence of social factors on the developme nt of childre n’s gende r identity. It is important to ask‚ howe ve r‚ whe the r the se findings provide e vide nce against the influe nce of socialization factors‚ or whether the y simply refle ct the fact that most of the studie s did not capture more subtle aspe cts of pare ntal treatment that may diffe r according to a child ’s gende r (Lytton & Romne y‚ 1991) . Conside ring the focus of many of the studie s‚ the latter explanation see ms probable . In most of the studie s conside red by Stern and Karrake r (1989) ‚ participants were reacting to photos or vide otape s of infants. The reviewers acknowle dged that it may not be possible to dete ct

Gend er Differen ces in Em otion al Closen ess

977

gender role socialization unde r such artificial circumstance s. Furthe rmore ‚ studie s that examine superficial processes such as differe ntial toy se lection or encourage ment of achie vement cannot be expe cted to fully e xplain subtle but pe rvasive differe nces between the ge nders. A greater emphasis must be place d on how relational processes are involve d in the emergence of gender differences—spe cifically‚ on how the emotional relationship between pare nts and childre n may have a differe ntial effect on male s and female s. In support of the importance of e xamining re lationships ‚ Maccoby ( 1990) sugge sted that unde rstanding of gender differences is best accomplishe d by studying relationships ‚ not by analyzing individual behaviors. Recent reports by clinicians emphasize the importance of mutuality for girls ’ and women’s relationships (e.g.‚ Jordan ‚ 1991; Mille r‚ 1976). In contrast ‚ boys ’ and men’s relationships are believed to be drive n more by desires for power or status (e.g.‚ Tanne n‚ 1990) . Further‚ equality and empathy are believed to characterize females’ more than males’ relationships ‚ which are believed to be more hierarchical and assertive (e.g.‚ Gilligan & Wiggins ‚ 1988). It may be that it is qualitie s of the emotional relationship between two individuals then that best differentiate females’ from males’ relationships. O ne of the few the orie s that incorporate s e motional re lationships and gende r-type d de ve lopme nt into a cohere nt framework has bee n propose d by Chodorow (1978) . She theorizes that be cause almost unive rsally the primary care give r of childre n is a fe male ‚ girls and boys will form differe nt emotional relationships with their care give r base d on their ge nde r. Spe cifically ‚ betwee n the age s of 3 and 5 years ‚ as childre n be come more aware of the importance of ge nde r ide ntity‚ boys must renounce the ir close ties with their primary caretake rs‚ who are fe males. Boys learn to diffe re ntiate the mse lve s from the ir close st e motional re lationship and be come more separate and distinct. This allows the m to develop a diffe rent ge nde r identity than that of their mothers. In contrast ‚ girls can maintain their primary emotional bond with their mothers as they de ve lop their gende r identitie s as fe male s. Since there is re lative ly little re search on the role of emotional de ve lopme nt as relate d to gende r typing ‚ there is little evide nce to support or reje ct Chodorow ’s theory. The small amount of evide nce that doe s e xist ‚ howe ve r‚ supports he r hypothe sis that the re are diffe re nce s in the e motional relationships betwee n mothers and daughte rs vs. sons. An early study by Sears‚ Maccoby ‚ and Le vin (1957) found that on global ratings scales mothers indicate d highe r le ve ls of warmth toward their infants if the infant was a girl. In anothe r study‚ Rose n and Brigham (1984) use d a questionnaire to e xamine young childre n ’s affective re sponse s toward their pare nts. They found that girls tende d not to e xpre ss ne gative fe e lings toward their

978

Ben en son et al.

mothers as re adily as boys. In addition ‚ girls score d highe r on depende ncy fe e lings toward their mothers than boys. More recent work has reve ale d that the re may be a ge nde r influe nce in pare nts ’ use of emotion words in conve rsations with their childre n. In a serie s of longitudinal studie s (Adams ‚ Kue bli ‚ Boyle ‚ & Fivush ‚ 1995; Kuebli ‚ Butle r‚ & Fivush ‚ 1995) ‚ Fivush and her colle ague s analyze d the narrative s of pare nts and childre n when the childre n were first 40 the n 70 months. Transcripts of the pare nt ¯child conve rsations were code d for e motion words. Results indicate d that regardle ss of the age of their childre n ‚ pare nts use d a gre ater numbe r and varie ty of emotion terms with daughte rs than with sons. In the study by Adams et al. ( 1995) ‚ the re sults furthe r re vealed that this diffe re nce did not appe ar to influe nce childre n at 40 months ‚ because boys and girls at this age use d approximate ly the same numbe r and varie ty of e motion words. At 70 months ‚ however‚ although both girls and boys continue d to use an equal numbe r of e motion words ‚ girls used a greate r varie ty of emotion terms than boys. In the second study (Kue bli et al.‚ 1995) conducte d only with mothe rs‚ mothe rs used a gre ater numbe r of negative e motion words when talking with daughte rs than sons. Through an incre ase d e mphasis on e motions ‚ particularly sensitive and painful emotions that may enge nde r e mpathy‚ mothe rs may be nurturing a closer re lationship with their daughte rs than with their sons. Although these studie s be gin to illuminate the subtle process of diffe rential emotional socialization of daughte rs vs. sons ‚ the y are limite d by the ir re liance on verbal measure s. Studie s have consiste ntly found that one of the few consiste nt gende r diffe re nce s is that fe male s are more verbal than male s (Maccoby and Jacklin ‚ 1974) . The re sults of these studie s may the re fore be confounde d by the fact that female s are more verbally fluent than males. The purpose of the pre sent study was to inve stigate the e motional closene ss be tween pre school childre n and their mothers using non-ve rbal measure s. Base d on Chodorow ’s the ory‚ we hypothe sized that mothe r ¯ daughte r dyads would show more emotional close ness than mothe r¯son dyads by displaying increase d levels of proximity ‚ by more e ye contact ‚ and by exhibiting highe r le ve ls of enjoyme nt.

METHOD Participants The participants were 41 mothe r¯child dyads recruited from two preschool classes in the Boston are a. Twenty of the childre n were 4 ye ars of

Gend er Differen ces in Em otion al Closen ess

979

age : 10 males (M = 58.30 months; SD = 2.63) and 10 female s (M = 56.90 months; SD = 2.92) ‚ and 21 were 5 years of age : 10 male s (M = 65.40 months; SD = 2.01) and 11 females (M = 65.82 months; SD = 2.86) . The dyads were 90% white and 10% Asian (n = 4) ‚ English spe aking ‚ and from middle to uppe r-middle socioe conomic backgrounds. Conse nt forms were sent home with childre n from two classe s at one pre school. All childre n in both classes re ceive d pare ntal conse nt to participate . Procedu re Mothe r¯child pairs were vide otape d in a structure d play se tting for 15 minute s in a room in the pre school. The vide o came ra was hidde n from view and locate d 10 fee t from a table and two chairs whe re the mother and child were seate d as the y playe d toge ther. The play mate rials were in three se parate boxe s locate d 20 fe et away and consiste d of transportation vehicle s‚ dolls and furniture ‚ and medium-size d blocks. The play paradigm and mate rials were base d on those created by Erikson (1951) and use d by Cramer and Hogan ( 1975) with pre school childre n to examine gende r differe nces in childre n ’s conce ptualizations of relationships. The paradigm consiste d of having childre n sit at a table and play with the toys. Toys were available that were attractive to childre n of e ach gende r. Thus‚ dolls and furniture were chosen to be attractive to girls and transportation vehicle s and blocks were sele cted for their attractive ne ss to boys. Because childre n ’s actual relationships with their mothe r was the focus ‚ mothers were se ated at the table with the ir childre n. At the be ginning of the play session ‚ the expe rimente r told the child that he or she could choose any of the toys from the boxe s and play with the m at the table . The mother was asked to remain at the table and to allow the child to re trie ve the boxe s and se lect the toys. The expe rimenter cautione d against placing the boxe s on the table and aske d the pair to remain se ated as much as possible . The n the e xpe rime nte r told the childre n to play with whate ver they wishe d for 15 minute s and have lots of fun. The expe rimenter the n left the room and the vide o camera recorde d the next 15 minute s of inte raction be tween the mothe r and child. Afte r 15 minute s‚ the e xpe rime nt re ente red the room and thanke d the pair for the ir participation. Coding The vide otape s were coded in five separate passe s by five inde pe nde nt sets of coders. The tape s were first coded for the child ’s proximity to the

980

Ben en son et al.

pare nt on a se cond by se cond basis during minute s 0 ¯2:59 and minute s 7:00 ¯ 9:59 for a total of 360 se conds. These 6 minute s‚ the first 3 minute s and 3 minute s toward the end of the se ssion ‚ were chose n specifically in an attempt to accurate ly sample the inte raction. Fagot (1985) re porte d that behaviors fluctuate across and within sessions. It was hope d that the se two times would balance any behavior fluctuations and increase the stability of the obse rvations. Proxim ity. Proximity was defined as the close st distance from the pare nt (de fine d as full body from fee t to he ad but not including arms) to the child (de fine d as head and torso —from bottom to shoulde rs but not including fe et‚ legs ‚ or arms). A nume rical assignme nt was give n for e ach code d se cond according to the following criteria: If the distance be tween the pare nt and the child was fewer than thre e inche s‚ a 1 was coded; if the distance betwee n the pare nt and the child was greate r than 3 inche s but fewer than 12 inche s‚ a 2 was coded; if the distance betwee n the pare nt and the child was more than 12 inche s and the child was at the table ‚ a 3 was code d; if none of the above crite ria were met‚ a four was coded. O ne coder‚ who had ne ve r see n the vide otape s be fore and was blind to the hypothe sis of the study‚ was traine d to code the entire data set according to the se crite ria. After the coding was comple te d ‚ a mean score for each pair across the 360 seconds was compute d. The re liability of this code was de te rmined by comparing the means obtaine d by the blind coder with the totals of an expe rimente r who also code d the data. The correlation betwee n the code rs was r(39) = .95. Eye Con tact. Mutual e ye contact was measure d by indicating the presence or abse nce of e ye contact for each 10-se cond block within minute s 0 ¯2:29 and 7:00 ¯ 9:59. The final score for e ach mother¯child was the percentage of blocks during which e ye contact occurre d. Two coders who had never see n the vide otape s and were unfamiliar with the hypothe ses of the study were inde pende ntly traine d to use this code ‚ and each code d the entire data set. The two coders’ score s were ave rage d for each mother¯child pair. The corre lation betwee n the two coders’ ratings was r(37) = .84. Enjoym ent of Child and Enjoym ent of Mother. Degre e of enjoyme nt le ve l was coded for both the child and the mother separate ly using a modified version of a global rating scale of enjoyme nt develope d by Sears et al. (1957). The scale re quire d the coders to se lect a rating from 1 to 6 that best characte rize d the le vel of enjoyme nt displaye d. A rating of 1 indicate d some negative expre ssion and/or paying little attention ‚ while a rating of 6 indicate d positive expre ssion and/or paying atte ntion almost the whole time . The ratings were made by five inde pe nde nt coders who had ne ve r vie wed the vide otape s be fore and who were blind to the hypothe ses of the study. The code rs watched minute s 3:00 ¯5:59 of the vide otape s the n rate d the

Gend er Differen ces in Em otion al Closen ess

981

inte raction on the 6-point scale . The se minute s were chose n be cause they were differe nt than the minute s code d for proximity and eye contact ‚ and he nce might capture a diffe re nt part of the mothe r ¯child inte raction. Mothe r and child were code d se parate ly by visually covering the noncode d subje ct. For e xample ‚ whe n the code rs we re assessing the le vel of enjoyme nt of the child ‚ the expe rimenter covere d the mothe r on the vide o monitor for the e ntire thre e minute s that the code rs were watching the tape . This ensure d that the code rs focused on the e xpre ssion and atte ntion leve l of the code d membe r of the pair. Howe ver‚ be cause verbal behavior of the child was conside re d important ‚ the mothe r’s and child ’s ve rbalizations were not maske d. To obtain a score for le vel of enjoyme nt‚ the five coders’ ratings were average d. Using the Spe arman ¯Brown formula ‚ effe ctive re liability for childre ns’ e njoyme nt and mothe rs’ enjoyme nt were R = .68 and R = .79 ‚ re spe ctively. Enjoym ent of Dyad. Degre e of enjoyme nt level was also coded for the mother and child as a dyad using the same enjoyme nt scale as whe n the mother’s and child ’s be haviors were coded separate ly. The ratings for enjoyme nt of the dyad were made by a ne w se t of code rs who had never viewed the vide otape s be fore and were unaware of the hypothe se s of the study. Se ven coders inde pende ntly rate d de gree of enjoyme nt of the dyad afte r watching minute s 3:00 ¯5:59 of the vide otape s. To obtain a score for le ve l of e njoym ent ‚ the se ve n code rs’ ratings were ave rage d. Using the Spe arman ¯Brown formula ‚ e ffective re liability for enjoyme nt of the dyad was R = .85.

RESULTS Three sets of analysis were conducte d. First ‚ Pe arson product-mome nt corre lation coefficie nts were compute d to e xamine the relationship be tween the de pe nde nt variable s. Second ‚ data were analyze d by a multiple analysis of variance (MANOVA) using gende r and age of the child as the inde pende nt variable s. Because the analysis reve ale d no significant effe ct of age on any of the variable s‚ the data were collapse d across age . Finally ‚ one -taile d t-tests were used to compare the ge nde rs on e ach depe nde nt variable . Table I displays the corre lation coe fficie nts among the five measure s of e motional closene ss and betwee n the de pende nt variable s and age . As expe cted ‚ e njoyme nt of the dyad was significantly correlate d with both enjoyme nt of the child and e njoyme nt of the mothe r. None of the othe r variable s were significantly correlate d. The se results sugge st that the proximity ‚

982

Ben en son et al. Table I. Correlations Between Measures of Emotional Closeness Measures

Proximity Eye Contact Enjoyment of Child Enjoyment of Mother Enjoyment of Dyad

Proximity 1.00

Eye Contact

Enjoyment of Child

Enjoyment of Mothe r

- 0.05

- 0.13

- 0.08

1.00

0.21

0.06

1.00

Enjoyment of Dyad

Age

- 0.11

0.05 0.21

- 0.15

0.30

0.33

a

- 0.22

1.00

0.66 b

- 0.04

1.00

- 0.11

a

p < .05 p < .0001.

b

eye contact ‚ and e njoyme nt variable s were measuring differe nt aspe cts of emotional e njoyme nt or close ness. A MANOVA with gende r and age as the inde pe nde nt variable s and the five de pe nde nt variable s yielde d a significant e ffect of ge nde r‚ F(5‚31) = 5.93 ‚ p < .001 ‚ but no significant e ffect of age or inte raction be tween gende r and age ‚ Fs < 1. To examine e ach of the depe nde nt variable s for the pre dicted ge nde r differences‚ one -taile d t te sts were conducte d. Table II displays the means and standard de viations for both girls and boys on each of the de pende nt variable s. As displaye d in the table ‚ there were significant gende r difference s on the measure s of proximity ‚ eye contact ‚ enjoyme nt of the child ‚ and enjoyme nt of the dyad. Compare d to boys ‚ girls exhibite d close r le vels of proximity to the ir mothe rs‚ e ngage d in more mutual e ye contact and were rate d highe r on the global enjoyme nt scale. In addition ‚ the mother¯child dyad was rate d as exhibiting highe r le ve ls of enjoyme nt whe n girls interacte d with their mothe rs than when boys interacte d with their mothe rs. Despite this finding ‚ there was no significant diffe re nce be tween rate d e njoyme nt leve ls of mothers who inte racted with sons vs. daughte rs‚ despite the high corre lation be tween ratings of mothers’ le ve ls of enjoyme nt and dyads ’ le ve ls of enjoyme nt‚ r( 37) = 66.

DISCUSSION Results from the current study provide d evide nce that in the context of a semistructure d play setting ‚ girls were more e motionally involve d with the ir mothe rs than were boys. The se re sults are consiste nt with Chodorow ’s (1978) the ory that during the preschool ye ars when gende r identity is be-

Gend er Differen ces in Em otion al Closen ess

983

Table II. Means and Standard Deviations of Me asure s of Emotional Closeness Girls Measures b

Proximity Eye Contact Enjoyment of Child Enjoyment of Mother Enjoyment of Dyad

Boys a

M

SD

n

2.01 0.10 4.83 4.53 3.80

0.36 0.08 0.71 0.91 1.03

21 21 20 21 21

M

SD

n

t

p

2.24 0.06 4.47 4.74 3.25

0.53 0.05 0.55 0.57 0.95

20 18 18 18 20

1.69 1.77 1.75 0.86 1.73

< .05 < .05 < .05 n.s. < .05

a

Sample size s vary slightly when parent or child was off scree n and measure was not code d. Lower scores indicate gre ater proximity.

b

lie ved to de ve lop ‚ mothe rs and daughte rs expe rience close r e motional relationships than mothers and sons. Due to limitations of the study‚ howe ve r‚ une quivocal support for Chodorow ’s the ory still re mains to be de te rmine d. As this study examine d only the inte ractions betwee n childre n and mothers‚ it is impossible to asce rtain whe the r increase d emotional closene ss is e xclusive to the mothe r¯daughte r re lationship or whe ther fe male s are simply closer to all adults. Furthe rmore ‚ this study only e xamine d the inte ractions of mothers and childre n in the conte xt of a semistructure d play se tting. It is possible that in a different context ‚ boys would appe ar to be more e motionally involve d with their mothe rs. Finally‚ this study is limite d by the homoge ne ity of the sample . Replication with pare nts and childre n from othe r socioe conomic and e thnic groups would pe rmit ge ne ralization of the results. Despite these limitations ‚ the study provide d consiste nt evide nce across several inde pende nt measure s that in a semistructure d context ‚ girls were emotionally close r than boys were to their mothe rs. Furthe r‚ the se tting was de signe d to be ecologically valid in that childre n participate d in their school in a room in which the y were very familiar and were able to choose from a varie ty of toys that are known to be appe aling to childre n of preschool age . In addition toys were provide d that are known to be appe aling to childre n of each gende r— dolls and doll furniture for girls ‚ and transportation ve hicle s and blocks for boys. Furthe r evide nce for the e cological validity of the context is that the childre n appe ared to e njoy the play session ‚ fre que ntly re fusing to stop playing at the e nd of the 15-minute period. The reasons for the lack of corre lation be tween the de pe nde nt measure s are unknown. It is like ly‚ however‚ that individuals e xhibit e motional closene ss in diffe rent ways ‚ and conseque ntly‚ some measure s are reliable indicators of close ness for some dyads and not othe rs. In support of this hypothe sis ‚ many code rs were nee ded to rate global measure s of enjoyme nt ‚

984

Ben en son et al.

because coders differed wide ly in the cue s they utilize d to rate enjoyme nt. In addition ‚ initial atte mpts to code de gre e of emotional close ness directly prove d futile ‚ be cause agre e ment be tween coders on indicators of closene ss was e ven lower than for e njoyme nt. Despite the lack of correlation be tween measure s of e motional close ness‚ however‚ the ge nde r diffe re nce s were appare nt on all measure s e xce pt mothers’ le vels of e njoyme nt. Pre vious studie s have faile d to e lucidate ways in which pare nts differentially treat boys and girls that would e xplain the developme nt of separate gende r identitie s. Results from the current study sugge st that it is the e motional relationship that develops be tween childre n and mothe rs that may le ad to diffe rential ge nde r socialization. It is uncle ar ‚ howe ver‚ whethe r parents re inforce diffe re nt behaviors in daughte rs vs. sons or whe ther girls and boys active ly re spond to pare nts in differe nt ways. O ne study (Robinson ‚ Little ‚ & Biringe n ‚ 1993) showe d that daughte rs who initially created more share d state s (e .g.‚ mutual smiling or laughing) with their mothers at 18 months had mothe rs who matche d the ir e xpre ssion more ofte n at 24 months. The se re sults indicate that diffe re nce s in e motional close ness may be due to emotions e manating from the childre n. In the curre nt study‚ mothers rarely le ft the ir chairs during the 15-minute play session ‚ so proximity was de te rmined by the childre n. In addition ‚ it was almost always the childre n who de te rmine d whe n mutual eye contact occurre d‚ be cause mothers looke d at their childre n more than childre n looke d at mothers; childre n focused more on the toys. Finally‚ global ratings of enjoyme nt showed girls to be enjoying be ing with the ir mothe rs more than boys; Mothe rs‚ howe ve r‚ exhibite d no pre fe re nce for interaction with daughte rs vs. sons. The se results strongly sugge st that it was the childre n who were driving the interaction rathe r than mothe rs. O the r studie s (Adams et al. ‚ 1995; Kue bli e t al.‚ 1995) ‚ howe ve r‚ have found that pare nts appe ar to socialize their daughte rs vs. sons to use e motions diffe rently. More research is nee ded to de te rmine the direction of effe cts. Future re se arch should also include fathe rs to de te rmine whether the findings in the curre nt study apply just to mothe rs‚ thereby supporting Chodorow ’s theory‚ or whe the r girls are emotionally closer than boys to both mothers and fathe rs. Finally‚ the emotional re lationship be tween parents and childre n from othe r populations in different contexts dese rve s furthe r study. In summary‚ this study adds to the growing possibility that ge nde r ide ntity de ve lopme nt is a comple x proce ss intimate ly re late d to the de ve lopme nt of early e motional relationships. As personality develops throughout the pre school years‚ boys and girls may e xpe rience diffe rent type s of e motional relationships with their mothe rs. The influe nce of this primary re lationship undoubte dly affe cts how childre n le arn to vie w the world and ultimate ly

Gend er Differen ces in Em otion al Closen ess

985

define the mselves. More research is nee ded to analyze the emotional relationships be tween pare nts and childre n in order to elucidate more cle arly the influe nce of social factors on gende r identity de ve lopme nt.

REFERENCES Adams ‚ S.‚ Kuebli‚ J. ‚ Boyle ‚ P.‚ & Fivush ‚ R. ( 1995) . Gender differences in parent child conversations about past emotions: A longitudinal inve stigation. Sex Roles‚ 33 ‚ 309-323. Berenbaum ‚ S. A.‚ & Snyder ‚ E. (1995). Early hormonal influences on childhood se x-typed activity and playmate prefere nces: Implications for the deve lopment of sexual orientation. Developm ental Psychology‚ 31 ‚ 31-42. Chodorow‚ N. ( 1978) . The reproduction of mothering: Psychoanalysis and the sociology of gender. Be rkeley: University of California Press. Cramer ‚ P.‚ & Hogan ‚ K. A. ( 1975) . Sex differences in ve rbal and play fantasy. Developm ental Psychology‚ 11 ‚ 145-154. Ehrhardt ‚ A. A.‚ & Meyer-Bahlbe rg ‚ H. F. L. (1981) . Effects of prenatal sex hormones on gende r-relate d be havior. Science ‚ 211 ‚ 1312-1318. Erikson ‚ E. H. (1951). Se x differe nce s in the play configurations of preadole scents. Journal of Orthopsychiatry ‚ 2 ‚ 667-692. Fagot ‚ B. (1985). A cautionary note: Parent ’s socialization of boys and girls. Sex Roles‚ 12 ‚ 471-475. Gilligan ‚ C.‚ & Wiggins ‚ G. (1988) . The origins of morality in early childhood relationships. In C. Gilligan‚ J. V. Ward ‚ J. M. Taylor‚ & B. Bardige (Eds.) ‚ Mapping the moral dom ain. Cambridge ‚ MA: Harvard University Pre ss. Golombok‚ S.‚ & Fivush ‚ R. (1994) . G ender developm ent. Ne w York: Cambridge University Press. Jordan ‚ J. V. (1991) . The meaning of mutuality. In J. V. Jordan ‚ A. G. Kaplan ‚ J. B. Miller‚ I. P. Stiver ‚ & J. L. Surrey (Eds.) ‚ Wom en ’s growth in connection . Ne w York: Guilford. Kuebli‚ J.‚ Butler ‚ S.‚ & Fivush ‚ R. ( 1995) . Mother-child talk about past emotions: Relations of maternal language and child gender ove r time. Em otions and Cognition ‚ 9‚ 265-283. Lytton ‚ H.‚ & Romney‚ D. M. (1991) . Pare nts’ sex-differe ntiated socialization of boys and girls: A me ta-analysis. Psychological Bulletin ‚ 109 ‚ 267-296. Maccoby ‚ E. E. ( 1990) . Ge nder and relationships. American Psychologist ‚ 45 ‚ 513-520. Maccoby ‚ E. E. ‚ & Jacklin ‚ C. N. ( 1974) . The psychology of sex differences. Stanford‚ CA: Stanford University Press. Miller‚ J. B. (1976) . Toward a new psychology of wom en. Boston: Beacon Press. Robinson‚ J.‚ Little ‚ C.‚ & Biringen ‚ Z. (1993) . Emotional communication in mother-toddler re lationships: Evide nce for e arly gender differentiation. Merrill-Palm er Quarterly‚ 39 ‚ 496515. Rose n‚ A.‚ & Brigham ‚ S. (1984) . Sex differences in affe ctive re sponse on the Bene -Anthony test. Journal of Personality Assessm ent‚ 48 ‚ 520-524. Sears ‚ R. R.‚ Maccoby ‚ E. E.‚ & Le vin ‚ H. (1957) . Pattern s of child rearing. Ne w York: Harper & Row. Stern ‚ M.‚ & Karraker ‚ K. (1989) . Sex stereotyping of infants: A review of gende r labeling studies. Sex Roles ‚ 20 ‚ 501-522. Tanne n‚ D. (1990) . You just don ’t understand . New York: Ballantine.