Sex Roles, Vol. 41, No s. 11/12, 1999
In¯ uen ces o f Social Expectations of G ender, G ender Stereotypes, and Situation al Co nstraints on Child ren ’s To y Cho ices Tarja R aag 1 Colby College
Child ren, w hose eth nic /racial b ackgro unds (prim arily cau casian ) an d household com positio ns (prim arily tw o-paren t hom es) re¯ ected lo cal populatio n statistics, w ere v id eotap ed p layin g w ith toy dish es an d to ols. The am ount of tim e spent w ith each toy w as calcu lated to determ in e w hether this v aried as a function of child ren’ s perceptio ns of social expectatio ns of gen der, aw areness of gen der stereotyp es, an d situ atio nal constrain ts (no in form atio n, gen der-typ ed in form atio n unrelated to the toys, gen der-typ ed toy lab els). In study 1, the toy choices of girls an d bo ys w ith percep tio ns of hav in g one or m ore fam iliar people w ho though t cross-gen der-typ ed play w as ``bad ,’ ’ w ere in ¯ uenced by gen der-typ ed toy lab els. How ev er, o nly boys w ith perceptio ns of h av in g one o r m ore people w ho though t cross-gen der-typ ed play w as ``bad ,’ ’ w ere som ew hat in ¯ uenced by gen der-typ ed in form atio n unrelated to the toys. Fu rth erm ore, in study 2, boys’ (but not girls’ ) perceptio ns of h av in g one or m ore people w ho though t cross-gen der-typ ed play w as ``bad ,’ ’ in depend ent of an aw aren ess of gen der stereotyp es predicted the am ount of tim e boys spent w ith cross-gen der toys. The discu ssio n high ligh ts the utility of m easu res of child ren’ s perceptio ns of o thers’ social expectatio ns of gen der in gen der research. A dditio nally, the discussio n high ligh ts the com plex relatio nship s betw een such perceptio ns, situ atio nal constrain ts, an d different socializ atio n th at girls an d boys experience in the dom ain of toy play. A rich source of e mpirical and the oretical work is available on the topics of ge nde r and ge nde r de ve lopme nt. Research informe d by cognitive de velopmental and ge nde r sche ma pe rspe ctives has de monstrate d the importance of cognition in ge nde r and gende r de ve lopm ent (B e m, 1981; Carter & 1
To whom correspondence should be addre ssed at De partme nt of Psychology, Colby College, Waterville, Maine , 04901. e-mail:
[email protected] du. 809 0360-0025/99/1200-0809$1 6.00/0 Ó
1999 Plenum Publishing Corporation
810
Raag
Levy, 1988; Kohlbe rg, 1966; Markus, Crane , B ernste in, & Siladi, 1982; Martin & Halve rson, 1981, 1983; Slaby & Frey, 1975) . Howeve r, the orists also highlight the nee d for deve lopm e ntal re search to be gin carefully e valuating the comple xity of the cognitive schemas and proce sse s that contribute to gende r and ge nde r de ve lopm ent (Fagot, 1995; Martin, 1993) . A dditionally, re se arche rs point to the necessity of de termining the nature of the re lationship be twee n deve loping ge nde r cognition and childre n’ s ge nde r be haviors ( Bussey & B andura, 1992; Carte r, 1987; Roopnarine & Mounts, 1987) . The work of B usse y and B andura (1992) has cle arly demonstrate d the utility of me asure s of childre n’ s se lf-e valuative re actions (fe e lings one has about one se lf as a conse que nce for behavior) in predicting gende r-re late d be haviors. In a re cent line of rese arch, Raag and he r colle ague s (Raag, B issoon, Randolph, Rackliff, & Le tizia, 1998; Raag & Rackliff, 1998) have found that childre n’ s pe rceptions of familiar pe ople ’ s reactions to crossgender-type d play (pe rceive d social constraints of ge nde r) inte ract with situational ge nde r constraints to shape childre n’ s ge nde r-re late d toy choice s. Spe ci® cally, boys who pe rceive that the y have social constraints of ge nde r show more gende r-type d and le ss cross-ge nder-type d toy choice s whe n e xplicit ge nde r-type d toy labe ls are applie d to toys. This line of research is continue d in the studie s pre se nte d here. Study 1 e xtends the work of Raag and her colle ague s by comparing e ffe cts of two diffe re nt type s of situational constraints of ge nde r on the toy choice s of boys and girls who re porte d having z ero or one or m ore familiar pe ople who thought cross-ge nde r-type d play was bad ( i.e ., zero or one or more pe rceive d social constraints of ge nde r). Study 2 assesses whe ther the effe cts of childre n’ s pe rceive d social constraints of gende r on toy choice s re ¯ e ct childre n’ s aware ne ss of gende r ste reotype s more ge ne rally rathe r than an aware ne ss of othe rs’ expe ctations of ge nde r speci® cally. 2 STU D Y 1 Rese arch has de monstrate d the powe r of situational factors in shaping gender be haviors (B radbard & E ndsle y, 1983; B radbard, Martin, E ndsle y, & Halve rson, 1986; Hartup, Moore , & Sage r, 1963; Jacklin & Maccoby, 1978; Kinsman & B erk, 1979; Monte mayor, 1974; Raag & Rackliff, 1998; Se rbin, Conne r, B urchardt, & Citron, 1979; Trautne r, 1995) . Pe rhaps eve n situational inform ation that highlights the importance of ge nde r without spe ci® cally highlighting the activitie s in which childre n will participate will shape childre n’ s gender-re late d be haviors. In study 1, childre n e xpe rie nce d e ithe r 2
This was originally suggeste d in e arly anonymous re views of this line of rese arch.
Childre n’s Toy Cho ices
811
task-re late d ge nde r-type d information , task-unre late d ge nde r-type d information, or no ge nde r-type d information prior to play. It was e xpe cted that both type s of situational inform ation about ge nde r would in¯ ue nce childre n with pe rceived social constraints by making the ir toy choice s more ge nde r ste reotype d than the be havior of childre n with no pe rceive d social constraints. B oys’ ge nde r role s within a play domain are thought to be more rigidly de ® ne d than girls’ gender role s (Singe r & Singe r, 1990) . Thus, one might e xpe ct boys to be slightly more se nsitive to varie d type s of cue s signaling the importanc e of ge nde r. A lthough it was pre dicte d that both girls and boys would re spond to ge nde r-type d situational information about the toys with which the y would play, boys were e xpe cted to be in¯ ue nce d more than girls by situational cue s of ge nde r that were not re late d to toys.
ME THOD S Particip an ts Name s of participants came from lists of name s of all 4- to 5-ye ar-old childre n born in the county within which this rese arch was conducte d. This information was obtaine d from the State of Maine De partm ent of Human Se rvices. Pare nts of all childre n for whom phone numbe rs and addre sse s could be found were contacte d ® rst through the mail, and the n by te le phone , to invite the m to participate . The sample include d 50 girls (mean age 5 4.7 years, range 4.2 to 5.6) and 57 boys (me an age 5 4.7 ye ars, range 4.2 to 5.5) . With re spe ct to racial/ e thnic identity, 1% of pare nts de scribe d their childre n as A sian; 1% of pare nts de scribe d the ir childre n as A sian, black, and white ; 3% of pare nts de scribe d the ir childre n as Native A merican; 2% of pare nts describe d the ir childre n as Native -A merican and white; 91% of pare nts described the ir childre n as white ; and 2% of pare nts did not provide any inform ation re garding the ir childre n’ s racial/ e thnic backgrounds . E ight pe rcent of childre n came from one-pare nt home s; 89% of childre n came from two-pare nt/guardian home s; 1% of childre n came from home s with two pare nts/guardians and e xte nde d family; and 2% of que stions on type of house hold were not answere d. This sample of childre n had spe nt an ave rage of 90% of the ir live s in their curre nt house hold (minimum amount of time 20%, maximum amount of time 100%). The primary care take rs of all of the childre n were the childre n’ s biological mothe rs. The mothe rs’ mean numbe r of ye ars of e ducation was 14.62 ye ars (range 12 to 26 years) . Mothe rs’ partne rs’ (for those mothe rs with partne rs) mean numbe r of ye ars of e ducation was 14.59 ye ars (range 9 to 22 ye ars). O ur sample
812
Raag
appare ntly had a highe r education le vel than is found in the local population of persons ove r the age of 25 years (79% of the local population of adults have a high school de gree or highe r). With the e xce ption of e ducational leve l and pe rcent of time that childre n had live d in the ir curre nt house hold (for which we have no local population statistics) , the se pe rcentage s are comparable to those found in the local population. Te stin g Pro cedure s an d Toy Stim ulu s D e scrip tio ns Childre n were te sted individually in a lab play room . E ach child was ® rst introduce d to the play room by an e xpe rime nte r while pare nts remaine d in anothe r room . In the playroom , the e xpe rime nte r de monstrate d a toy dish set (dishe s, dish draine r, table with lace table cloth) , and a toy tool se t ( tools, toolbox, she lf with a pe gboard) in randomly dete rmine d orde rs to each child to ensure that all childre n saw e ach toy prior to play. Howe ver, childre n were not allowe d to touch the toys during toy demonstrations. Toy de monstratio ns involve d showing childre n that the dishe s and tools functione d as large puzzle s. O utline s of the dishe s we re drawn onto the table and outline s of the tools were drawn onto the she lve s so that the se toys could be ® tte d onto the ir matching shape s. Following toy de monstration s, some childre n re ceived ge nde r-ste re otype d inform ation about the toys (the se dishe s are for girls, and the se tools are for boys) , some childre n were give n ge nd er-ste re otype d inform ation about ite ms (childre n’ s clothe s) unre late d to the toys, and some childre n re ceive d no additional inform ation. The childre n who re ceive d gende rste reotype d information unre late d to the toys were told ``Look ove r he re . B e fore you came , I was picking out clothe s for girls and boys. The se clothe s (pointing to a pink shorts se t) are for girls, and the se clothe s ( pointing to a blue shorts se t) are for boys.’ ’ Then the expe rime nter picke d up the clothe s and remove d the m from the playroom. Prior to the ir arrival to the lab, childre n had bee n random ly assigne d to one of the thre e be twee nsubje cts groups using a random number table . Childre n were the n vide otape d playing alone with the toys for 3 minute s. Childre n were told that the y could play as the y wante d with any of the toys. Following play sessions, childre n and pare nts were inte rvie wed separate ly. E v alu atio ns of Toy Stim uli These toys are unique in how similar the y are to one anothe r. The table and she lf are similar in color and size . The dishe s and tools are similar
Childre n’s Toy Cho ices
813
in color (reds, gree ns, ye llows, blue s), size , numbe r of obje cts available , number of moving parts, and amount of activity they e licit. It was assume d that the toy sets diffe red in that they could be se e n as fe minine (dishe s) or masculine (tools) . It has bee n sugge ste d that some obje cts have a history of gende r typing (Martin & Halve rson, 1981) . A dditionally, dishe s and tools are use d in ge nde r-role inve ntorie s to re pre se nt fe minine (dishe s) and masculine (tools) ite ms (E de lbrock & Sugawara, 1978) . Thus, e ve n without gender-type d labe ls, toy dishe s and tools may have a ge nde r-type d history for many childre n. Using childre n’ s pe rceptions of the toys as a validity che ck is complicate d by the fact that most childre n teste d with these toys have like d both toy sets (Raag et al., 1998; Raag & Rackliff, 1998) . It might be expe cted that if childre n like the toys, the y will not state that e ithe r toy se t is for the othe r se x (see also Martin, E isenbud, & Rose, 1995) . None the le ss, an inde pe nde nt sample of childre n was inte rvie wed regarding their perceptions of the toys. The sample , re cruite d using the main data colle ction proce dure s, include d 39 girls (mean age 5 4.9 ye ars) and 26 boys ( mean age 5 4.8 years) . Pare ntal reports of childre n’ s racial/ e thnic ide ntitie s, type of house hold, time in house hold, and pare ntal/ guardian e ducation le vel were comparable to the primary sample in study 1. These childre n were aske d whe the r the y thought each toy set (dishe s, tools) was for girls, boys, or both girls and boys, whe ther the y like d the toys (do you like this dish set? do you like this tool se t? ) , and which of the toys was the ir favorite ( if you had to pick, which toy se t would be your favorite ? ). Most childre n like d the dishe s (38 girls and 24 boys) , and the tools (37 girls and 25 boys) . A small number of childre n e ithe r did not like the dishe s ( 1 boy) or the tools (1 girl and 1 boy) , or were unce rtain about whethe r the y like d the dishe s (1 girl and 1 boy) or the tools (1 girl) . Most childre n (27 girls and 20 boys) thought the dishe s were for both girls and boys (1 girl and 2 boys thought the dishe s were for boys, 10 girls and 4 boys thought the dishe s were for girls) . Most childre n (25 girls and 18 boys) thought the tools were for both girls and boys (2 girls and 1 boy thought the tools were for girls, 11 girls and 7 boys thought the tools were for boys) . Se x of child and se le ction of favorite toy we re not inde pe nde nt of one anothe r, x 2 (2, N 5 63) 5 14.8 p , .0006. Girls were more like ly than e xpe cted by chance to choose dishe s (rathe r than tools) as the ir favorite toy se t. B oys were more like ly than expe cted by chance to choose tools (rathe r than dishe s) as the ir favorite toy se t (Table I). A me asure of favorite toy is not e xplicitly a validity che ck of whe ther childre n pe rceive each toy se t to be for girls, boys, or both. Howeve r, a child’ s se le ction of he r or his favorite toy may indire ctly re ¯ e ct the toy’ s ge nde r-type d history. This may
814
Raag Tab le I. Girls’ and B oys’ Selections of Favorite Toy Favorite Toy Se x of Child Girls B oys
Tools
Dishe s
B oth
7 ( 2 3.2) 15 (3.2)
23 (3.7) 4 ( 2 3.7)
7 ( 2 .8) a 7 (.8)
Note. Post hoc ce ll contributions for this chi-square analysis are in pare nthese s. a Two girls did not answer the question.
be a particularly use ful measure in a case whe re toy stimuli are quite similar on many dime nsions, and where most childre n like the toy stimuli that are use d.
CHILD INTE R V IE W QU E STIONS FOR STU D Y 1 Childre n were aske d about the ir pe rceptions of familiar othe rs’ social expe ctations of the ir cross-ge nde r-type d play. Spe ci® cally, all childre n were aske d: ``What would your say if you playe d with (othe r gender) ’ s toys? Would she /he say that this is good, bad, or it doe sn’ t matte r? ’ ’ A ll childre n were aske d about all of the following source s, if the se sources were available : mothe r, fathe r, daycare worker/babysitte r, brothe r(s) /sister(s), and be st frie nd( s). Childre n without particular source s did not have data re garding those pe ople ’ s thoughts about cross-ge nde r play.
D ata Han dlin g an d D ata A nalyse s For data on pe rceptions of siblings’ re actions to cross-ge nde r-type d play, the sibling close st in age who was at le ast 3 years old (and not an adult) was use d in analyse s. The focus of inte re st was on siblings similar in age who might spe nd time playing with the childre n in this study. For data on be st friends’ re actions to cross-ge nder-type d play, the ® rst name mentione d was use d in analyse s. B e st frie nds who were family pets, stuffed anim als, or othe r imaginary frie nds were not use d in analyse s. The amount of time spe nt playing with the dishe s, with the tools, and wande ring anywhe re e lse in the room was de te rmine d for e ach child. Pre sumably, the amount of time spe nt at the dishe s and at the
Childre n’s Toy Cho ices
815
tools re ¯ e cted ge nde r-re late d be haviors of girls and boys. The depe nde nt variable s (time at dishe s and time at tools) were analyze d using se parate 2 (se x of child) 3 3 (inform ation provide d prior to play) 3 2 (pe rceive d social e xpe ctations of cross-ge nde r-type d play) analyse s of variance . Howe ve r, be cause time at dishe s and time at tools are not inde pe nde nt of one anothe r, a B onfe rroni adjustm e nt of alpha le ve ls to .025 (as sugge ste d in B ray & Maxwe ll, 1985) was use d to de te rmine signi® cance of e ffects in both analyse s of variance . In the analyse s, ``pe rceive d social expe ctations of cross-ge nde rtype d play’ ’ was de ® ne d as a two-le ve l factor. Childre n who re porte d having one or more source s who thought cross-ge nde r-type d play was ``bad’ ’ were cate gorize d se parate ly from childre n who pe rceived that the y had no source s who thought cross-ge nder-type d play was ``bad.’ ’ Table II shows the number of girls and boys who re porte d having zero or one or more familiar othe rs who thought cross-ge nde r-type d play was ``bad.’ ’
R E SU LTS E ffe cts of Pe rceiv ed So cial E xp ectatio ns an d Situ atio nal Constrain ts of G e nder on Toy Choice s Tim e at D ish es The 2 (se x of child) 3 3 (inform ation provide d prior to play) 3 2 (pe rceive d social expe ctations of cross-ge nde r-type d play) analysis of variance of time spent at the dishe s yielde d a signi® cant main e ffect of se x of child, F(1,95) 5 19.03, p , .0001, a two-way inte raction e ffe ct of se x of child 3 informatio n about the toys, F(2,95) 5 4.91, p , .009, and a three way inte raction e ffe ct of sex of child 3 information about the toys 3
Tab le II. Freque ncy of Children W ho Reported Having Ze ro or One /More Source s Who Think Cross-Gende r-Type d Play is B ad Se x of Child Perce ive d Social Constraints of Gende r Z e ro source s O ne or more source s
Girls
B oys
28 22
21 36
816
Raag
pe rceive d social e xpe ctations of cross-ge nder-type d play, F(2,95) 5 4.63, p , .01. The main e ffect and two-way inte raction can be e xplaine d within the context of the thre e -way inte raction. Figure s 1 (girls’ data) and 2 (boys’ data) show this three -way inte raction. Simple e ffects Fs indicate d that for both girls and boys who perceive d that the y had no familiar sources who thought cross-ge nde r-type d play was ``bad,’ ’ type of information provide d prior to play (none , ge nde rste reotype d inform ation about the toys, ge nde r-ste re otype d information unre late d to the toys) made no diffe re nce to the amount of time spe nt with the dishe s. The inte raction be twee n se x of child and inform ation provide d be fore play showe d up for childre n with pe rceptions of having one or more familiar sources who thought cross-ge nde r-type d play was ``bad.’ ’ Spe ci® cally, for girls with pe rceive d social constraints of ge nde r, the signi® cant e ffe ct of type of inform ation provide d prior to play, F(2,95) 5 6.16, p , .01, was due to more play with the dishe s (M 5 148 seconds, SD 5 33) when ge nde rtype d inform ation about the toys was re ceive d, than whe n no information (M 5 65 se conds, SD 5 40) or ge nde r-type d inform ation unre late d to the toys (M 5 68 seconds, SD 5 81) was received ( Sche ffe ’ s critical diffe re nce s were 75, p , .03, and 78, p , .04, re spe ctive ly) . For boys with pe rceive d social constraints of gende r, the signi® cant e ffe ct of type of inform ation provide d prior to play, F(2,95) 5 5.51, p , .01, was slightly diffe rent. It was due to le ss play with the dishe s (M 5 15 se conds, SD 5 25) whe n ge nde r-type d information was re ceived about the toys than whe n no inform ation (M 5 81 se conds, SD 5 45) was re ceive d (Scheffe ’ s critical diffe re nce 5 43, p , .002) . Inte re stingly, ge nde r-type d information unre late d to the toys re sulte d in boys playing with the dishe s at an amount that was inte rme diate (M 5 39 se conds, SD 5 49) to the amount of play with the dishe s shown in the no information condition (M 5 81 se conds, SD 5 45) and the gende r-type d toy condition (M 5 15 se conds, SD 5 25) . Simple e ffects F te sts also re ve ale d a nonsigni® cant tre nd (give n the more conse rvative alpha le vels) toward a ge nde r diffe rence for childre n with no perceive d social constraints who re ceive d gende r-type d inform ation not re late d to the toys, F(1,95) 5 5.91, p 5 .05, and a signi® cant ge nde r diffe re nce for childre n with perceive d social constraints who re ceive d ge nde r-type d information about the toys, F( 1,95) 5 27.81, p , .001. Thus, a gender diffe re nce occurre d whe n childre n had both perceived social constraints and situational social constraints of gende r that we re explicitly about the toys. Finally, for girls with e xplicit gende r-type d toy labe ls, the re was a nonsigni® cant trend (give n the more conservative alpha leve ls used he re )
Fig. 1. Mean numbe r of se conds girls spe nt playing with the dish set as a function of both information about the toys (no information, ge nde r-typed information unre lated to the toys, or gende r-type d information about the toys), and pe rce ptions of having eithe r zero or one or more familiar source s who think cross-ge nde r-typed play is ``bad.’ ’
Childre n’s Toy Cho ices 817
818
Raag
toward an effe ct of pe rceive d social e xpe ctations of cross-ge nde r-type d play, F( 1,95) 5 6.15, p 5 .05.
Tim e at Tools The 2 (se x of child) 3 3 (inform ation provide d prior to play) 3 2 (pe rceive d social expe ctations of cross-ge nde r-type d play) analysis of variance of time spent at the tools yie lded a signi® cant main effe ct of se x of child, F(1,95) 5 13.80, p , .0001. O ve rall, boys spe nt more time at the tools than girls did. Give n the more conse rvative alpha leve ls use d he re , the two-way inte raction e ffect of se x of child 3 inform ation about the toys, F(2,95) 5 3.74, p 5 .027, and the three -way inte raction e ffe ct of sex of child 3 inform ation about the toys 3 perceive d social e xpe ctations of cross-ge nde r-type d play, F(2,95) 5 3.47, p 5 .03, were nonsigni® cant tre nds. Howe ve r, the thre e -way inte raction is shown in Figs. 3 (girls’ data) and 4 (boys’ data) for the reade r’ s inte rest. The data for time at tools mirror those for time at dishe s.
D ISCU SSION For both girls and boys with pe rceptions of having zero familiar sources who thought cross-ge nde r-type d play was ``bad,’ ’ the type of information provide d in conjunction with toy de monstration s made no diffe re nce to play with dishe s. It was only when childre n had pe rceptions of having one or more familiar source s who thought cross-ge nde r-type d play was ``bad’ ’ that diffe re nt type s of situational inform ation diffe re ntially in¯ uence d girls’ and boys’ toy choice s. Girls and boys with pe rceive d social constraints showed the most dive rgent be havior patte rns (for girls, a lot of play with dishe s; for boys, little play with dishe s) in a conte xt with e xplicit gender-type d inform ation about the toys, re sulting in the large st ge nde r diffe rence in behaviors. These ® ndings are consiste nt with the e arlie r ® ndings of Raag and her colle ague s (Raag e t al., 1998; Raag & Rackliff, 1998) . A dditionally, it was found that for girls with pe rceive d social constraints, ge nde r-type d inform ation not relate d to the toys functione d in the same way that no inform ation functione d. For boys with pe rceived social constraints, gender-type d inform ation not about the toys resulte d in boys’ play with dishe s at an amount that was inte rme diate to that se e n in the no-inform ation and ge nde r-type d toy conditions. A lthough it was pre dicte d that boys would be more sensitive to gende r-type d inform ation not about
Childre n’s Toy Cho ices
819
the toys, it was not e xpe cted that this information would have no e ffe cts on girls’ toy choice s. The ® nding that ge nde r-type d inform ation unre late d to the toys slightly in¯ ue nce d the amount of time boys with perceive d social constraints spe nt with ``cross-ge nde r-type d’ ’ toys ( see Fig. 2) , but not the amount of time girls with perceive d social constraints spent with ``crossgender-type d’ ’ toys (se e Fig. 3), ne eds e xplaining. Ge nde r-type d information unre late d to the toys may signal the importance of ge nde r in a situation. The pre se nt re sults sugge st that boys may be more se nsitive to varie d type s of situational cue s about the importance of ge nde r. B y being se nsitive to e ve n subtle inform ation about ge nde r boundarie s, and by avoiding crossgender-type d toys and activitie s, boys can avoid pote ntial negative social conseque nces for their behaviors. A lthough the re sults for girls were une xpe cted, the y are not ne cessarily contradicto ry to research in this area. B ussey and Pe rry (1982) showe d that boys are more like ly to re je ct cross-ge nde r be havior than girls. A dditionally, be cause preschool girls do not as readily receive negative social conse que nce s as boys do for cross-ge nde r be havior (Fagot, 1977) , girls may learn that only gende r-type d information e xplicitly about play activitie s within which childre n are involve d matte rs as a warning for avoiding ge nde r boundarie s. Raag and Rackliff (1998) originally sugge sted that a combination of both situational constraints of ge nde r and aware ne ss of social e xpe ctations of ge nde r contribute to the actual manife stations of ge nde r behaviors. Thus, ge nde r-re late d toy choice s are shape d by situational and pe rceived social constraints rathe r than sole ly by liking of toys. This line of re se arch measure s childre n’ s aware ne ss of othe rs’ social e xpe ctations of ge nde r as a face t of ge nde r sche mas. It is conceivable , howeve r, that the measure of ``aware ne ss of social expe ctations of ge nde r’ ’ use d he re actually re¯ e cts childre n’ s aware ness of ge nde r ste reotype s more ge ne rally. Those childre n who re port having one or more pe rceive d social constraints may simply be more aware of ge nde r ste re otype s. A n aware ne ss of ge nde r ste re otype s in turn may in¯ ue nce toy choice s. A lte rnative ly, re gardle ss of an aware ne ss of social e xpe ctations of ge nde r, an aware ness of ge nde r ste re otype s may be a more critical factor in de termining toy choice s. E ithe r way, the inde pe nde nt contribution of ge nde r stereotype aware ne ss ne e ds to be asse sse d.
STU D Y 2 The purpose of study 2 was to examine how childre n’ s response s to que stions about familiar othe rs’ social expe ctations of cross-ge nde r-type d
Fig. 2. Me an numbe r of se conds boys spe nt playing with the dish se t as a function of both information about the toys (no information, ge nde r-typed information unre late d to the toys, or ge nder-typed information about the toys), and perce ptions of having eithe r zero or one or more familiar source s who think cross-ge nde r-typed play is ``bad.’ ’
820 Raag
Fig. 3. Mean numbe r of seconds girls spe nt playing with the tool set as a function of both information about the toys (no information, ge nde r-typed information unre late d to the toys, or ge nder-typed information about the toys), and perce ptions of having eithe r zero or one or more familiar source s who think cross-ge nde r-typed play is ``bad.’ ’
Childre n’s Toy Cho ices 821
Fig. 4. Me an number of seconds boys spent playing with the tool se t as a function of both information about the toys (no information, ge nde r-typed information unre late d to the toys, or ge nder-typed information about the toys), and perce ptions of having eithe r zero or one or more familiar source s who think cross-ge nde r-typed play is ``bad.’ ’
822 Raag
Childre n’s Toy Cho ices
823
play contribute d to the amount of time childre n spe nt playing with crossgender toys inde pe nde nt of childre n’ s aware ne ss of (own and othe r) ge nde r ste reotype s. More spe ci® cally, it is possible that childre n’ s response s to que stions about familiar othe rs’ reactions to cross-ge nde r-type d play are re ¯ e ctions of childre n’ s aware ness of ge nde r ste re otype s, or an aware ne ss of ge nde r stere otype s is more important than perceived social constraints in pre dicting ge nde r-re late d toy choice s. If e ithe r alte rnative is true , the n an aware ne ss of familiar othe rs’ social expe ctations of cross-ge nde r-type d play should not inde pende ntly contribute to the amount of time childre n spe nd playing with ge nde r-type d toys once the effe cts of ge nde r ste re otype aware ne ss are re move d.
ME THOD S Particip an ts, Testin g Procedures, an d Toy Stim uli Recruitme nt methods, te sting proce dure s, and toy stimuli were the same as those use d in study 1 with the e xce ption of the addition of an asse ssment of gende r stereotype aware ne ss. The sample include d 26 girls (mean age 5 4.8 ye ars) and 24 boys (mean age 5 4.6 ye ars). With re spe ct to racial /e thnic ide ntity, 1% of pare nts de scribe d the ir childre n as black, 2% of pare nts de scribe d the ir childre n as Native A merican, and 97% of pare nts de scribe d their childre n as white. E ight pe rcent of childre n came from one-pare nt home s, 90% of childre n came from two-pare nt/guardian home s, and 2% of childre n came from homes with two pare nts/guardians and exte nde d family. This sample of childre n had spe nt an ave rage of 86% of the ir live s in the ir curre nt house hold (minimum amount of time 19%, maxim um amount of time 100%) . The primary caretake rs of all of the childre n were the childre n’ s biological mothe rs. The mothe rs’ mean number of ye ars of e ducation was 13.95 years (range 12 to 20 ye ars). Mothe rs’ partne rs’ (for those mothe rs with partne rs) mean numbe r of years of e ducation was 14.28 ye ars (range 9 to 20 ye ars). Pare ntal de scriptions of childre n’ s racial/ ethnic identitie s, type of house hold, time in curre nt house hold, and pare ntal /guardian e ducation leve ls were comparable to study 1. B ecause it has alre ady be en shown that the toy choice s of childre n with pe rceive d social constraints be come more gender type d in conte xts whe re toys are labe le d as gende r type d, only data for childre n who re ceive d e xplicit ge nde r-ste re otype d informatio n about the toys ( the se dishe s are for girls, and these tools are for boys) were analyze d he re . A dditionally, be cause childre n’ s toy e valuations were comparable to othe r studie s, the se data are not pre se nte d he re .
824
Raag
Child Inte rv ie w Que stio ns for Stu dy 2 The sex role le arning inde x (SE RLI) deve lope d by E de lbrock and Sugawara (1978) was administe re d to childre n to obtain measure s of ge nde r ste reotype aware ne ss. The SE RLI include s a task in which childre n distribute picture s of traditionally ``feminine ’ ’ and ``masculine ’ ’ obje cts, one at a time , into thre e cate gorie s (for girls only, for boys only, for both girls and boys) . Childre n receive se x-role discrimination score s for the ir own ge nde r (SRD-O wn) and for the othe r ge nde r (SRD-Othe r). Highe r se x-role discrimination scores (SRD-O wn and SRD-O the r) are inte rpre te d as a gre ate r aware ne ss of ge nde r ste re otype s. Childre n were also aske d about their pe rceptions of familiar othe rs’ social e xpe ctations of the ir cross-ge nde r-type d play using the same que stions as in study 1. Time at ``cross-ge nde r’ ’ play was use d in analyse s as the crite rion variable , and was de® ned for girls as time spent with tools, and for boys as time spent with dishe s.
R E SU LTS Table III pre se nts (for girls and boys se parate ly) the means, standard de viations, and range s associate d with sex-role discrimination scores (SRDO wn, and SRD-O the r), and the amount of time spent at ``cross-ge nde r’ ’ play. Table III also pre sents fre que ncie s of girls and boys who reporte d
Table III. De scriptive Statistics for Criterion V ariable and Predictors Se x of Child V ariable s
Girls
Boys
Criterion variable Time at ``cross-ge nder’ ’ play (in seconds) Me an (SD) 63 (62) 28 (30) Range 0 ± 180 0 ± 106 Pre dictors SRD-own ge nder (SERLI) Me an (SD) 86 (14) 89 (15) Range 50± 100 40± 100 SRD-other ge nde r (SE RLI) Me an (SD) 74 (24) 76 (19) Range 10± 100 30± 100 Whether or not children have one /more sources who think cross-gende r play is ``bad’ ’ Yes 11 12 No 15 12
Childre n’s Toy Cho ices
825
having zero or one or more familiar othe rs who thought cross-ge nde r-type d play would be ``bad.’ ’ Following de scriptive analyse s, se parate nonhie rarchical multiple re gre ssion analyse s were use d for girls’ and boys’ data to pre dict the amount of time spe nt at ``cross-ge nder’ ’ play. A warene ss of own-ge nde r stere otype s, aware ne ss of othe r-gender ste re otype s, and aware ne ss of othe rs’ social e xpe ctations of cross-ge nde r-type d play were use d as pre dictors. Table IV pre sents the zero-orde r corre lations for predictor and criterion variable s for girls and boys. A s might be e xpe cted, time at ``crossgender’ ’ play is signi® cantly ne gative ly corre late d with pe rceive d social constraints (numbe r of familiar othe rs who thought cross-ge nde r play would be ``bad’ ’ ), for both girls and boys. The more pe rceive d social constraints childre n have , the le ss time they spend at ``cross-ge nder’ ’ play and the le ss pe rceive d social constraints childre n have , the more time the y spe nd at ``cross-ge nde r’ ’ play. For girls’ data (Table V ), the ove rall re gre ssion mode l was not signi® cant (p 5 .42) , nor did any of the pre dictor variable s contribute signi® cantly to the mode l. Thus, with re spe ct to the que stion addre sse d in this study, once e ffe cts of aware ne ss of ge nde r stereotype s were re move d, perceive d social constraints no longe r in¯ uence d the amount of time girls spe nt at ``cross-ge nde r’ ’ play. For boys’ data (Table V ) , the ove rall re gression mode l was signi® cant (p 5 .04) , and both the variable s of aware ne ss of othe rgender ste re otype s (p 5 .04) and aware ness of othe rs’ pe rceptions of crossgender-type d play (p 5 .02) contribute d signi® cantly to the mode l. Thus,
Tab le IV . Z ero-O rde r Correlations A mong Criterion and Predictor Variable s for Girls (N 5 26) and B oys (N 5 24) V ariable SRD-O wn Girls B oys SRD-O ther Girls B oys Perceive d social constraints Girls B oys
``Cross-Gender’ ’ Play
SRD-O wn
SRD-O ther
.006 ( p 5 .49) .03 ( p 5 .46) .07 .37) 2 .35 ( p 5 .05) (p 5
2 .34
(p 5
.04) 2 .40 ( p 5 .03)
.58 .001) 2 .09 ( p 5 .34)
(p 5
.07 .37) .25 ( p 5 .12) (p 5
2 .09
(p 5
.33)
2 .11 ( p 5 .30)
826
Raag
Table V . Results of the Nonhie rarchical Re gre ssion Pre dicting the A mount of Time Spent at ``Cross-Gende r’ ’ Play
Predictors Girls’ data Mode l SRD-O wn SRD-O ther Pe rceive d social constraints B oys’ data Mode l SRD-O wn SRD-O ther Pe rceive d social constraints
R
2
Standard Re gre ssion Coef® cient
.12
Ð .01 .03
2 .34 .32
df 3, 22 25 25
F/t
p
.97 (F ) .04 (t) .14 (t)
ns ns ns
25
2 1.70 (t)
ns
.11 2 .39
3, 20 23 23
3.20 (F ) .60 (t) 2 2.10 (t)
.04 ns .04
2 .47
23
2 2.50 (t)
.02
Ð
e ve n afte r effe cts of aware ness of ge nde r ste re otype s were re move d, pe rceive d social constraints still signi® cantly in¯ ue nce d the amount of time boys spe nt at ``cross-ge nde r’ ’ play.
G E NE R A L D ISCU SSION Study 1 showed that for childre n with perceive d social constraints of gender, gende r-type d situational information about toys with which childre n will play in¯ uences childre n’ s gender-re late d toy choice s. Situational cue s of gender that are unre late d to toys with which childre n will play are more like ly to e nhance boys’ than girls’ avoidance of cross-ge nde r-type d toys. Study 2 showe d that whe n ge nde r-type d toy labe ls are applie d, the re lationship be twe e n pe rceptions of othe rs’ thoughts of cross-ge nde r play and time childre n spend at ``cross-ge nder’ ’ play is more complicate d than originally assum ed in this line of re search. For girls, once e ffects of ge nde r ste reotype aware ne ss are re move d, no relationship is found be twe en pe rceptions of othe rs’ thoughts of cross-ge nde r play and time spent at ``crossgender’ ’ play. For boys, on the othe r hand, pe rceptions of othe rs’ thoughts of cross-ge nde r play, inde pende nt of aware ne ss of gende r stereotype s, pre dicts amount of time at ``cross-ge nde r’ ’ play within a conte xt whe re toys are gender type d. B e cause most childre n te ste d like the se toys and se e the toys as inte nde d for both girls and boys, the re sult of e nhance d avoidance of the dishe s (cross-ge nde r-type d toys) by boys is probably most simply e xplaine d by looking to situational and pe rceived social constraints. A lthough the combine d e ffe cts of situational constraints of gender and
Childre n’s Toy Cho ices
827
pe rceive d social constraints of ge nde r are appare nt in both girls’ and boys’ toy choice s, the effe cts also diffe r for girls and boys. Martin and Halve rson (1987) sugge sted that gender is more salie nt for boys than girls. Pe rhaps incre ased ge nde r salie ncy for boys is a function of the conte xts whe re childre n’ s ge nde r role s have be e n studie d thus far. Conve rge nt e vide nce from se ve ral source s highlights the idea that some A me rican boys may have a smalle r range of culturally acce ptable ge nde r be haviors within play conte xts (as sugge ste d by Singe r & Singe r, 1990) . Thus, cross-ge nde r behavior during play may not have the same me aning for girls and boys, and may not be e valuate d and re acted to in the same way. More ge nerally, conseque nces for girls’ and boys’ ge nde r-re late d behaviors may be conte xt spe ci® c. If girls and boys have diffe re nt conse que nce s for their ge nde r be haviors within a give n context, the y may de velop diffe re nt motivations for behavior within the same conte xt. Raag and Rackliff (1998) originally propose d that the conse que nce s for gender-re late d behaviors that have be e n inte grate d into childre n’ s ge nde r schemas may be used by childre n in circumstance s whe re ge nde r is highlighte d as important, but whe re immediate /dire ct conse que nce s for ge nde rre late d be haviors are abse nt. A social cognitive pe rspe ctive of gende r, on the othe r hand, highlights ``e xternal social sanctions’ ’ as the main re gulators of gende r be haviors primarily in e arly de velopm ent. This theore tical pe rspe ctive propose s that, ultimate ly, childre n deve lop the ir own personal standards for ``appropriate ’ ’ ge nde r behavior. A t that point, se lf-sanctions be come critical in re gulating ge nde r be havior. E vide nce supporting this has be en note d e arlie r in this pape r. Spe ci® cally, B usse y and B andura (1992) found that the self-e valuative re actions of childre n (comparable in age to the present sample ) pre dicte d their ge nde r-re late d behaviors. Howe ve r, ge nde r knowle dge did not pre dict ge nde r-relate d be haviors in that study. A t ® rst glance , this appe ars to be inconsiste nt with the prese nt ® ndings showing the importance of boys’ aware ness of othe r-gende r ste re otype s in boys’ toys choice s. Perhaps not all face ts of ge nde r ste re otype aware ne ss are equally critical in predicting behavior. Give n e vide nce of boys’ greate r avoidance of cross-ge nde r be havior, pe rhaps aware ness of othe r ge nde r ste re otype s is more critical for predicting boys’ ge nde r-relate d be haviors than for pre dicting girls’ ge nde r-re late d be haviors. A dditionally, the present data highlight the importanc e of e valuating the combine d in¯ ue nce s of childre n’ s pe rceptions and situational factors in ge nde r re se arch. It should be highlighte d that childre n’ s re sponse s to que stions about social e xpe ctations of ge nde r in the prese nt work may in fact re¯ e ct inte rnalize d personal standards of ge nde r, and the se personal standards may he lp to re gulate ge nde r behavior in conte xts with situational constraints of ge nde r. In othe r words, childre n may answer que stions about familiar othe rs’
828
Raag
social e xpe ctations of ge nde r using their own pe rsonal standards of ge nde r. A lte rnative ly, once childre n have inte rnalize d personal standards of gende r, childre n’ s be havior may continue to be guide d by an aware ne ss of othe rs’ social e xpe ctations of ge nde r. Finally, it is also possible that the combine d e ffe cts of an aware ness of othe rs’ social expe ctations of ge nde r, an anticipation of exte rnal social sanctions (within situations whe re ge nde r has be en highlighte d as important) , inte rnalize d pe rsonal standards of ge nde r, and aware ne ss of gende r ste re otype s have comple x inte rre lationships that in¯ ue nce behavior. B e cause e xternal social sanctions are quite re al for some childre n, it make s se nse that the y would continue to in¯ uence childre n’ s be havioral choice s e ven after pe rsonal standards of ge nde r have be en de ve lope d, particularly in situations whe re childre n are aware that gende r is important. Whe n the situational value of ge nde r is made salie nt, childre n who have e xpe rie nce d cle ar conseque nces for ge nde r-relate d be haviors may have he ighte ne d expe ctations of social sanctions for crossing ge nde r boundarie s. Fu ture D ire ctio n s Se ve ral issue s should be addre sse d in future work. First, it is important to focus on whe ther se lf-e valuative re actions, aware ne ss of social sanctions, and aware ne ss of social e xpe ctations are unique or ove rlapping constructs. It ne eds to be de te rmine d how the se constructs diffe rentially inte ract with situational constraints to shape gender be haviors. A dditionally, it is critical to dete rmine what type s of expe rience s (comments about ge nde r, joke s, or primarily direct conse que nce s) are necessary and suf® cie nt for shaping ge nde r be haviors. What type s of e xpe rie nce s become inte grate d into pe rceptions of othe rs’ social e xpe ctations and into personal standards of ge nde r? Se cond, the diffe re ntial in¯ ue nce s of diffe re nt sources’ (pare nts’ / guardians’ , teachers’ , siblings’ , and pe ers’ ) social e xpe ctations of crossgender-type d be haviors ne ed to be de te rmine d. Cle arly, childre n have pe rceptions of othe rs’ expe ctations of gende r. These pe rceptions may re¯ e ct a wide range of socializatio n e xpe rience s and social conse que nce s (both direct and indire ct) for be havior. A dditionally, measure s of childre n’ s pe rceptions of social e xpe ctation pre ssure s should include pe rceptions of e xpe rime nters’ social e xpe ctations. Such pe rceptions may furthe r add to situational pressures that childre n fe e l re garding ge nde r be havior. Third, although not statistically signi® cant, and not pre dicted, it should be note d that in study 1, boys ( see Fig. 2) and girls (se e Fig. 3) who re ceive d no ge nde r-type d information showe d more cross-ge nde r play whe n the y had more perceive d social constraints . It might be e xpe cted that in a condition with no immediate situational constraints, childre n with zero perceive d
Childre n’s Toy Cho ices
829
social constraints and childre n with one or more pe rceive d social constraints would show e qual le ve ls of cross-ge nde r play. Interestingly, the une xpe cted patte rn has be e n appare nt in boys’ behavior in othe r sample s as well ( Raag e t al., 1998; Raag & Rackliff, 1998) . A lthough it is critical to not make more out of nonsigni® cant diffe re nce s than is appropriate , pe rhaps this patte rn could be indicative of a ``re bound e ffe ct’ ’ to the re moval of immediate situational constraints of gende r for childre n who norm ally e xpe rie nce social constraints of gende r. Such childre n, whe n give n the opportunity, may play more with cross-ge nde r toys than childre n without perceive d social constraints be cause they are fre e from ge nde r constraints the y norm ally e xpe rie nce . A lthough not te stable with this sample , it is pre dicted that such a ``re bound e ffe ct’ ’ (if it is re al) would show up as signi® cant with sample s that had extre mely high le ve ls of pe rceived social e xpe ctation pressures. Fourth, future work on the inte ractive e ffects of situational and cognitive factors that in¯ ue nce ge nde r behaviors ne e ds to turn to more varie d conte xts of study. For instance , it is predicte d that contexts involving e ating be havior, conte xts highlighting physical appe arance s, and conte xts involving se lf-prote ction might reve al ke y gende r-role e xpe ctations and limits for girls. A dditionally, conte xts highlighting emotions might re ve al ke y ge nde rrole expe ctations and limits for boys. These are not conte xts that have re ceive d much attention in re se arch focuse d on ve ry young childre n although the y are conte xts whe re gende r diffe re nce s are appare nt for adole scents and adults. Preliminary work (Raag, B anos, & Puza, 1999) has found that pre school girls and boys do perceive diffe rent social e xpe ctations pre ssure in some of the se domains. The multiface te d nature of ge nde r appe ars to be the dire ction in which many theorists are moving (Fagot, 1995; Martin & Halve rson, 1987) . A CK NOWLE D G ME NTS This study was supporte d by Colby Colle ge Social Scie nce Division Grants 01-2252 and 01-2279. I would like to thank Jessica L. Banos, France s L. B erry, Colle e n J. B urnham , Heathe r L. Hunte r, Jennife r J. Lavigne , Me lanie A . Puza, Cecilia K. Stashwick, and Leah D. We st for the ir help with this rese arch. I would also like to thank all of the pare nts and childre n who participate d in this study. R E FE R E NCES B e m, S. L. (1981) . Gende r-schema theory: A cognitive account of sex-typing. Psycho logical Rev iew , 88, 354 ± 364.
830
Raag
B radbard, M. R. & E ndsley, R. C. (1983) . The effe cts of sex-type d labe ling on pre school children’ s information-see king and re te ntion. Sex Roles, 9, 247± 260. B radbard, M. R., Martin, C. L., E ndsley, R. C., & Halverson , C. F. (1986) . In¯ uence of sex ste reotypes on children’ s e xploration and memory: A compe te nce ve rsus pe rformance distinction. D ev elo pm ental Psycholo gy, 22, 481± 486. B ray, J. H., & Maxwe ll, S. E. (1985) . Mu ltiv ariate A nalysis of Variance. Newbury Park, California: Sage . B usse y, K., & B andura, A. (1992) . Self-regulatory mechanisms gove rning ge nder de ve lopme nt. Ch ild Dev elo pm en t, 63, 1236± 1250. B usse y, K., & Perry, D. G. (1982) . Same -sex imitation: The avoidance of cross-sex mode ls or the acce ptance of same -se x models? Sex Ro les, 8, 773± 784. Carte r, D. B . (1987) . In D. B ruce Carter (E d.), Current con cep tions of sex roles an d sex typing: Th eory an d research. New York: Prae ge r. Carte r, D. B ., & Levy, G. D. (1988) . Cognitive aspe cts of e arly sex-role de velopme nt: The in¯ uence of ge nder sche mas on preschoolers’ me mories and pre fe re nce s for sex-type d toys and activities. Child D ev elop ment, 59, 782± 792. E de lbrock, C., & Sugawara, A . I. (1978) . A cquisition of se x-type d pre fe rence s in pre schoolage d children. D ev elopm ental Psycho logy, 14, 614± 623. Fagot, B . I. (1977). Consequence s of moderate cross-ge nder be havior in pre school children. Ch ild Dev elo pm en t, 48, 902 ± 907. Fagot, B . I. (1995) . Psychosocial and cognitive de terminants of e arly gende r role de ve lopme nt. In R. C. Rosen (E d.), A nn ual rev iew of sex research , 6, 1 ± 31. Hartup, W. W., Moore, S. G., & Sage r, G. (1963) . A voidance of inappropriate sex-typing by young children. Jo urnal o f Con sulting Psych olo gy, 27, 467 ± 473. Jacklin, C. N., & Maccoby, E . E . (1978) . Social be havior at thirty-three months in same -sex and mixe d-sex dyads. Child Dev elop m en t, 49, 557± 569. Kinsman, C. A . & B erk, L. E . (1979) . Joining the block and house ke eping are as: Changes in play and social behavior. You ng Children, 35, 66± 75. Kohlbe rg, L. (1966) . A cognitive -deve lopmental analysis of children’ s sex-role concepts and attitudes. In E . E . Maccoby (E d.) Th e d ev elop ment of sex d ifferen ces. Stanford: Stanford Unive rsity Press. Markus, H., Crane , M., B e rnste in, S., & Siladi, M. (1982) . Se lf-schemas and ge nder. Jou rnal o f Personali ty an d Social Psych olo gy, 42, 38± 50. Martin, C. L. (1993) . Ne w directions for investigating children’ s ge nder knowledge. D ev elo pm ental Rev iew , 13, 184 ± 204. Martin, C. L., E ise nbud, L., & Rose , H. (1995). Children’ s gende r-based reasoning about toys. Child D ev elop ment, 66, 1453 ± 1471. Martin C. L., & Halverson , C. F. (1981) . A schematic proce ssing mode l of sex-typing and ste reotyping in children. Ch ild D ev elo pm ent, 52, 1119± 1134. Martin C. L., & Halve rson, C. F. (1983) . The effects of se x-typing sche mas on young children’ s me mory. Ch ild D ev elo pm ent, 54, 563± 574. Martin C. L., & Halve rson, C. F. (1987) . The roles of cognition in se x-role acquisition. In D. B ruce Carte r (E d.), Current co ncep tions of sex roles an d sex typin g: Theory an d research . Ne w York: Praeger. Monte mayor, R. (1974) . Children’ s performance in a game and their attraction to it as a function of sex-type d labels. Child D ev elop ment, 45, 152± 156. Raag, T., B anos, J. L., & Puza, M. A . (1999) . B oys’ an d girls’ p erception s o f social pressures in different d om ains. Poster prese nted at the biennial mee ting of the Socie ty for Re search in Child De ve lopment, A lbuquerque , Ne w Me xico. Raag, T., B issoon, C., Randolph, A ., Rackliff, C. L., & Le tizia, A . (1998) . The in ¯ uen ce o f imm ediate an d learn ed so cial con straints o n preschoo lers’ gend er-typ ed p lay. Poster prese nted at the annual confere nce of the Association for W omen in Psychology, B altimore , Maryland. Raag, T. & Rackliff, C. L. (1998) . Preschoolers’ aware ne ss of social expe ctations of gende r: Re lationships to toy choice s. Sex Roles 38, 685 ± 700. Roopnarine, J. L., & Mounts, N. S. (1987) . Curre nt theore tical issues in sex roles and se x
Childre n’s Toy Cho ices
831
typing. In D. Bruce Carter (E d.), Cu rrent conceptio n s of sex ro les an d sex typin g: Theory and research . Ne w York: Praeger. Se rbin, L. A ., Conner, J. M., B urchardt, C. J., & Citron, C. C. (1979) . E ffects of pee r pre sence on sex-typing of children’ s play. Jou rnal o f E xperim en tal Child Psych olo gy, 27, 303 ± 309. Singer, D. G., & Singe r, J. L. (1990) . The high se ason of imaginative play. In D. G. Singe r, & J. L. Singe r (E ds.), The ho use of m ake-believ e: Play and the dev elo pin g im agin ation . Cambridge : Harvard University Pre ss. Slaby, R. G., & Fre y, K. S. (1975) . Developme nt of ge nder constancy and sele ctive attention to same -sex mode ls. Child D ev elopm ent, 46, 849± 856. Trautner, H. M. (1995). B oys’ and girls’ play be havior in same -sex and opposite-se x pairs. Jou rnal o f G en etic Psych olo gy, 156, 5 ± 15.