Your Research Proposal is a six- to seven-page plan for a new study on your research topic. Incorporate at least four scholarly/peer-reviewed journal articles in addition to the course text to support

17, No. 2, 1U-120 BIRTH ORDER AND MOTHER-CHILD INTERACTION IN AN ACHIEVEMENT SITUATION * MARY K.

ROTHBART2 Stanford University Hypotheses about birth order and the antecedents of adult achievement were in- vestigated by observing mother-child interactions in a laboratory setting. Subjects were fifty-six 5-year-old boys and girls and their mothers; half of the children had a same-sex sibling who was approximately 2 years older than himself, and half had a same-sex sibling, 2 years younger. Mothers supervised their child's perform- ance on five tasks, and mother-child interaction data were examined as they related to birth order and sex of child.

No differences were found in the amount of time mothers spent interacting with their children.

In terms of the quality of interac- tion, however, mothers gave more complex technical explanations to firstborn children.

They also exhibited greater pressure for achievement and greater anxious intrusiveness into the performance of the firstborn, with these findings accentuated in the mother's behavior toward the first girl.

Most research on the effects of birth order has concentrated on identifying personality characteristics that vary as a function of ordi- nal position. Although such research often con- cludes with hypotheses about differing social- ization experiences for children of different ordinal positions, only a few studies have at- tempted to observe the actual behavior of parents toward children of differing birth order.

The present study presents a test of hypotheses about differential socialization of first- and later-born children in an achievement situation.

It is part of a larger study in which mothers' interactions with firstborn and second-born • children were observed in a structured inter- action setting.

The general finding that the firstborn tends to gain greater eminence in school and later life 1 This article is based upon a doctoral dissertation submitted to the Department of Psychology of Stanford University.

The research was financed in part by Public Health Service Predoctoral Fellowship 5 Fl MH-20, 971-02 from the National Institute of Mental Health.

The author wishes to express special thanks to her dis- sertation advisor, Eleanor Maccoby, and to committee members Robert Sears and Alberta Siege], for their generous contribution of ideas, to Phyllis Plainer and John Masters, for help in collecting and analyzing data, to Nicholas Anastasiow, Research Director of the Palo Alto School District, for his cooperation, and to Myron Rothbart, who helped in every phase of the research.

Finally, the author wishes to thank the Palo Alto moth- ers and children, who so willingly participated in this project.

2 Requests for reprints should be sent to the author, who is now at the Department of Psychology, Univer- sity of Oregon, Eugene, Oregon 97403.

(Altus, 1966; Sampson, 1965; Schachter, 1963), though recently disputed (Bayer, 1966), has led to questions about the kinds of parental expectations and pressures for success exerted upon the firstborn as compared with the second born.

In addition, numerous studies have con- sidered possible IQ differences between first- born and later-born children (Sampson, 1965).

These studies often have had conflicting re- sults; in infant tests (Bayley, 1965; Cushna, 1966), slight differences in favor of firstborns have been found. Bayley, however, reports that these differences are small, and seem to have no cumulative effect or relation to later intelligence test scores.

What, then, are other variables that might prompt the firstborn to higher academic levels in later life?

It has been suggested that greater parental pressures are directed toward the firstborn's achievement and acceptance of re- sponsibility (Davis, 1941; McArthur, 1956; Rosen, 1961), that the firstborn is often given the role of parent surrogate (Sutton-Smith, Roberts, & Rosenberg, 1964), that the parents talk and interact more with the firstborn (Bossard, 1945), and pay more attention to the firstborn (Koch, 1954). Rosen (1961) and Phillips (1956) have proposed that since parents have no frame of reference in their expectations for the firstborn, they tend to overestimate his ability more than the second born's, setting higher standards for his performance.

Work done on the relation of birth order and eminence thus suggests numerous questions for 113 114 MARY K.

ROTHBAET this stud)':

Do parents exert greater pressure on their firstborn to excel in his work?

Do par- ents have higher or more unrealistic expecta- tions for their first child's performance?

Do parents interact more with the first child?

Arc parents more likely to praise and criticize the performance of the first child than the second?

When a parent instructs his child, are his ex- planations more complex and at a higher level for the firstborn than the second?

In the pres- ent study these questions were investigated through an observation of mothers supervising their firstborn or second-born children in the performance of a variety of tasks.

METHOD The experimenter compared mother-child interactions for 5-year-old firstborn and second-born boys and girls from two-child, same-sex families.

In a 2 X 2 design, sex of child and birth order were the independent vari- ables.

Mothers were asked to supervise their children in the performance of five different tasks, two of which involved explanations by the mother.

Subjects Subjects were kindergarten children from the Palo Alto Unified School District and their mothers. Sub- jects were chosen from two-child families only, where both children were of the same sex and there was an ap- proximate age difference of 2 years between the subject and his sibling. Half of the subjects had a 3-year-old younger sibling; half had a 7-year-old older sibling.

Of a total of 56 subjects, 30 were girls and 26 were boys.

Distributions were matched with respect to age of the subject and age difference with the sibling. Approxi- mately the same proportion of children in each group (67%) came from professional homes.

One important variable on which groups were not matched was age of mother, with mothers whose second-born children were 5 years old, as would be expected, significantly older than mothers whose firstborn children were 5 years old.

Mothers were introduced to the study by a letter from the school district director of research, who de- scribed the purpose of the study as investigating differ- ences between children of different birth orders.

After the mothers had received the letter, sessions were scheduled by telephone.

Of over 60 mothers asked to take part in the study and pretest, only 2 were unable to participate.

Procedure All interaction sessions were held in a specially equipped trailer provided by Stanford University's Laboratory of Human Development. Since subjects came from all parts of the city, the trailer was located at four different schools, providing the same situation for all subjects.

The trailer was divided into three rooms, with two large rooms at either end and a small central observation room containing one-way mirrors and recording equipment.

Tasks were performed in one large room; children played with toys in the other large room while the mother was given instructions.

In the case of six firstborn boys and seven firstborn girls, the younger sibling was also present in the toy room, since the experimenters offered baby-sitting to participating mothers.

Two experimenters, the author and another woman experienced in working with children, were pres- ent for all sessions. When the mother and child arrived, introductions were made, and subjects were shown the experimental room.

One experimenter then invited the child to come into the playroom for a short time.

The mother was then seated at a card table, and in- structions were given to her by the experimenter.

The instructions, developed with the help of nine pretest subjects, began:

What we'd like to try to do as soon as \jiame of child~] comes back in is to set up and record several simple situations that you probably go through with him fairly often.

Most of the situations will involve you describing or explaining something to him, or supervising his work on a problem.

Of course, it's best if you can be as natural as possible, and act toward him as you usually do.

The mother was then introduced to the live tasks she was to engage in with her child.3 The situations were as follows:

Conversation.

The mother was asked first of all to ask her child what he had seen in the other room, to "get an idea of his usual conversation." The playroom con- tained a special toy (a set of dump trucks), along with a workbench, coloring book, paper, crayons, and two children's books.

The mother did not know what the playroom contained.

The conversation was designed to put subjects at ease at the beginning of the session, to provide an indication of the number of questions asked the child by his mother, and the specificity and correct- ness of the child's recall.

Cartoons.

The mother was asked to show her child two four-frame "Peanuts" cartoons by the artist Charles Schulz, and to explain to him what was happening in the cartoons.

The cartoons offered a set of somewhat ambiguous stimulus materials that the mother could structure a great deal or not at all.

The mother could simply read the cartoon, she could point out emotions displayed by the cartoon characters, and for one car- toon, she could draw a moral lesson.

It was thus possible to get a measure of the complexity of the mother's ex- planation, along with the amount of time she spent explaining the cartoons.

Picture.

Mothers were given a picture of 20 zoo ani- mals and were asked, initially, to show the child the picture for a 3-minute period. During this time, the mother, or child, or both could name the animals aloud.

The mother was then to turn the picture over for 3 minutes and ask the child to remember as many of the animals as he could.

She was allowed to prompt him as much as she wished.

The mother was also asked to esti- mate the number of animals her child would be likely to remember, after being told that the average child of his 3 For detailed instructions see Rothbart (1967). BIRTH ORDER AND MOTHER-CHILD INTERACTION 115 age would remember 10 animals.

This situation proved to be a powerful measure of the mother's pressure on the child to perform well. Most children attempted to stop the task at some point during the recall period, and a measure was made of pressure exerted on the child to continue.

Explanation.

Mothers were given a simple diagram of the workings of a water tap, along with an extremely complicated written description of how a water tap works.

The language was technical, and included some information about water pressure that was tangential to the explanation.

The mother was asked to show the diagram to the child, and explain to him in her own words how the water tap worked. This situation allowed a measure of complexity and length of the mother's explanation, along with her use of praise and criticism.

Puzzle.

The last situation was a difficult geometric puzzle, on which the mother was asked to supervise her child.

She was given a model of how the puzzle should look when it was completed, and told that she could help the child as much as she wanted without actually show- ing him the solution to the puzzle.

The time allowed for working on the puzzle was 6 minutes, after which the .experimenter made sure that the child successfully com- pleted the puzzle if he had not yet done so. The mother was also asked to estimate how quickly her child would he likely to solve the puzzle on a 6-point scale ranging from "very slowly and with difficulty" to "very rapidly." After the mother was given the initial instructions, she was provided with a shorter written version of the oral instructions, and the child was returned to the room to begin the tasks.

The experimenter remained in the room with the mother and child to answer the mother's questions about procedure and to make sure the correct order of tasks was followed.

The entire session was tape-recorded.

After the tasks were completed, the child was given paper construction materials, and the mother, a ques- tionnaire to complete during a 15-minute period while the experimenter was out of the room.

(These data have not yet been analyzed.) The mother was also asked if she and the child would straighten the room before the experimenter returned.

Two brooms, one large and one child-size, were conspicuously located in the corner of the room, and the experimenter noted the extent of the child's involvement in the cleanup. When the 15 min- utes had elapsed, the experimenter returned to ask the mothers a few questions about rules she had for her children, and the session was then terminated.

(Inter- view data have also not yet been analyzed.) Coding Procedures and Reliabilities Coding for data analysis was based for the most part on tape recordings of the five tasks completed by the mother and child.

This coding was blind, with all tape recordings coded by the author and 21 recordings coded independently by a second rater for reliability.

In addi- tion, length of interactions was timed, and the child's level of performance on the conversation, picture, and puzzle was rated.* The average reliability coefficient * For coding categories and additional reliabilities, sec Rothbart (1967).

was .92; the only achievement scale with a reliability of .80 or below was the mother's structuring of the puzzle task, where r = .79.

The basic statistical analysis for all variables was a two-way analysis of variance (Sex X Birth Order).

In addition, items on the same general variable, for example, mother's pressure for achievement, were in- tercorrelated.

When correlations were significant at the .01 level, standard scores of these measures were added to form larger scales, and analyses of variance, applied to scale scores.

When individual variables failed to correlate or ATs were unequal, separate analyses were carried out.

RESULTS Results are summarized in Tables 1, 2, and 3. For all measures, the meaning of a score cor- responds to the name of the scale.

For example, a high score on the pressure for achievement variable indicates strong pressure for achieve- ment.

In data analysis, it was first necessary to determine whether performance of firstborns and second borns differed significantly on the tasks assigned them.

If there were actual dif- ferences in the success or failure of first- and second borns, these differences might have, in turn, differentially affected their mothers' behavior, greatly complicating an interpreta- tions of differences on maternal variables.

Measures of the child's performance, in- cluding the number of playroom objects de- scribed in the conversation, number of zoo animals recalled, and time to complete the puzzle, failed to correlate with each other.

The direction of recall differences was the same, with higher means for first children, but no significant differences were found (see Table 1).

The time required to solve the puzzle showed a significant interaction (F = 5.34, df = 1/51, p < .05), with firstborn girls and second-born boys taking longer to solve the puzzle.

The puzzle was fortunately the last task assigned, so differential child performance on the puzzle could not have affected the mother's behavior toward her child on any subsequent tasks.

The first general maternal variable examined was the mother's estimate of her child's per- formance.

It was thought that this variable would be reflected in (a) the mother's response to a direct question about her child's likely performance on a concrete task; (b) a measure of complexity of the mother's explanation, with mothers who have higher estimates of their child's performance using more complex and 116 MARY K, ROTHBART TABLE 1 MEANS AND STANDARD DEVIATIONS FOR VARIABLES DESCRIBING CHILD'S BEHAVIOR Area and Variable Performance Conversation Picture Puzzle time Adoption of responsibility Cleanup involvement Girls Firstborn M 4.91 9.60 3.38 3.43 SD 4.08 4.63 1.33 1.28 Second born M 3.92 8,53 2.87 3.21 SD 3.01 3.02 1.42 1.25 Boys Firstborn M S.80 8.38 2.42 3.30 SD 3.39 4.73 .90 1.95 Second born M 4.33 7.46 3.35 2,38 SD 2.87 4.14 1.45 1.43 detailed concepts in their explanations; and (c) the mother's structuring of tasks for her child.

If the mother of the second born is more aware of her child's level of understanding, she might be expected to give him a more structured in- troduction to a task.

Mothers estimated their child's performance for both the picture and puzzle tasks.

For the picture estimate, there was a barely significant birth-order effect (F = 3.31, df = 1/48, p < .10), with mothers tending to have a higher estimate of the first child's performance; how- ever, a stronger interaction (F = 5.55, df = 1/48, p < .05) indicated that the birth- order difference was contributed by the mother's higher estimate for the first girl than for the second girl (see Table 2). The direction of estimates was reversed for the puzzle, with higher estimates for firstborn boys and second- born girls, and no indication of a birth-order effect.

The puzzle differences were not sig- nificant.

Two measures were made of the mother's complexity of explanation.

For the water tap explanation, a count was made of the number of technical terms used by the mother, for ex- ample, "water pressure," "valve," and "cy- lindrical stopper." For the cartoons, a count was made of the number of different features of the picture described, for example, sizes of cars, number of pieces of the snowman, and of interpretations made by the mother (e.g., "The little dog is copying the big dog.").

The two measures of complexity failed to correlate sig- nificantly with each other, but the direction of differences for both measures was the same:

mothers tended to give a more complex de- scription to the first child.

This difference was only significant for the water tap explanation OF = 4.35, r//= 1/52, p < .05).

Measures were also made of the mother's structuring of tasks, that is, the amount of in- formation she gave the child about what would happen in a task and what he would be ex- pected to do.

Structuring for the picture and the puzzle surprisingly failed to correlate.

There were no significant differences found in struc- turing the picture task, or agreement in direc- tion of findings for the picture and puzzle.

For structuring the puzzle, a strong interaction was found (F = 9.61, df = 1/51, p < .01). Moth- ers structured the puzzle task more for firstborn girls and second-born boys, those groups that in fact performed less well on the puzzle than the other two groups.

A second general variable thought to corre- late with later achievement of the child was the sheer amount of interaction of the mother with the child.

This was measured by the amount of time spent in the conversation and explana- tion and the number of questions asked by the mother in the conversation, cartoons, and ex- planation tasks. Since time spent in conversa- tion and explanation correlated .47, these vari- ables were combined, and analysis showed no differences for either the grouped or separate data.

Standard scores of number of questions were also combined, yielding only a barely sig- nificant sex difference (F = 3.35, df = 1/52, p < .10), with mothers tending to ask more questions of their daughters than of their sons (see Table 2).

The difference probably proposed most often about achievement socialization of firstborn and later-born children involves the prediction of greater pressures for success and greater BIRTH ORDER AND MOTHER-CHILD INTERACTION 117 parental involvement in the firstborn's per- formance.

The first direct measure of pressure for achievement was pressure for naming, a rating of the speed at which zoo animals were named by the mother and child, that is, wheth- er there was leisurely discussion or whether animals were named rapidly and repeated several times before the end of the period.

Pressure for remembering rated the mother's reaction to the child's desire to stop trying to recall the names of animals—whether she al- lowed him to stop or insisted that he continue until the time was up.

The two scales were posi- tively correlated with each other (r = .30, p < .05), and both showed highly significant birth- order effects (pressure for naming, F = 8.87, df = 1/51, p < .01; pressure for remembering, F = 8.05, df = 1/38, p < .0.1), with mothers exerting more pressure on firstborns than on second borns, Ns for the pressure for remember- ing variable were smaller, since not all subjects attempted to stop working on the task.

The pressure for remembering variable also showed a significant interaction (F = 9.76, df = 1/38, p < .01), with greater differential pressure for girls than for boys (see Table 2).

More indirect measures of achievement variables were amount of time spent in conver- sation and explanation, and the number of ques- tions asked by the mother.

As described above, no differences were found in interaction time, while mothers showed a tendency to ask more questions of their daughters than their sons, with no birth-order effects.

After doing all coding, raters made an overall judgment of the extent to which the mother seemed to intrude herself into the child's per- formance in a worried or anxious way.

Raters achieved good agreement (r = .88) on this variable, and it yielded a significant birth- order effect (F = 5.92, df = 1/51, p < .05), with mothers showing higher anxious intru- siveness for firstborns than for second borns.

A significant sex difference was also found (F = 5.45, df = 1/51, p < .05), with mothers ex- erting more anxious intrusiveness on girls than on boys (see Table 2). The intrusiveness vari- able was included in intercorrelations for both pressure and help-giving variables.

It corre- lated with conversation time (r = .34, p < .01); the number of questions asked in the con- versation, cartoons, and explanation (rs = .38, .26, and .41, ps < .01, .05, and .01, re- spectively; the mother's help giving on the picture (r — .48, p < .01); and on the puzzle (r = .41, p < .01), even though these last two measures failed to correlate with each other.

TABLE 2 MEANS AND STANDARD DEVIATIONS FOR VARIABLES DESCRIBING MOTHER'S BEHAVIOR Area and variable Expectations Picture estimate Puzzle estimate Explanation complexity Cartoon complexity Picture structuring Puzzle structuring Length of interaction Conversation and explanation time Number of questions Pressure for success Pressure for naming Pressure for remembering Anxious intrusiveness Praise and criticism Praise Tells child he is correct Tells child he is incorrect Girls Firstborn M 12.13 3.27 8.47 5.27 1.93 2.47 4.06 6.98 3.07 2.90 3.33 1.27 19.87 3.87 SD 3.93 .96 5.98 1.75 1.10 .74 2.16 3.11 .70 .32 1.50 1.66 8.04 4.85 Second born M 9.07 3.60 5.40 5.20 2.07 1.80 4.23 5.98 2.13 1.50 2.47 1.40 19.20 2.07 SD 1.83 1.05 2.77 2.48 1.10 .86 1.92 1.67 1.13 .71 1.13 1.82 9.10 2.15 Boys Firstborn M 10.25 3.50 7.00 5.92 2.00 1.92 3.38 5.27 2.75 2.17 2.50 1.82 17.17 1.33 SD 2.83 .80 4.30 3.04 1.21 .79 1.47 2.22 .87 .71 1.00 1,82 10.47 .98 Second born M 11.00 3.00 5.85 5.00 1.92 2.54 4.17 5.49 2.15 2.20 1.85 .84 19.31 2.77 SD 2.31 1.22 2.61 2.61 1.11 .66 1.17 1.64 1.14 .79 .90 1.02 9.59 3.06 118 MARY K.

ROTHBART TABLE 3 CHI-SQUARE FOR MOTHER'S CRITICISM 01 CHILD Reaction Mother criticizes Mother does not criticize Girls First- born 8 7 Second born 2 13 Boys First- born 2 9 Second boru 4 9 Note.—x2 = 6.67, p < ,10.

A related variable was the extent to which the mother evaluated and reinforced her child's ongoing performance through the use of praise and criticism.

If the mothers were more in- volved in.

their firstborn's performance, they might be expected to give him more praise and criticism. Since the number of occurrences of praise and criticism were relatively small, they were initially combined over all tasks except the conversation.

The direction of differences for use of criticism was opposite to that of use of praise.

The data, however, were confounded by the fact that the second-born girl and firstborn boy had actually performed better on the puz- zle task than the firstborn girl and second-born boy.

When praise and criticism data from the puzzle task were discarded, the differences for praise were in the same direction, that is, for more praise to the second-born girl and first- born boy, but not significant.

The frequencies of use of criticism were so low that a chi-square test was performed to see whether the four groups differed according to the mothers' use of criticism (see Table 3).

There was a slight tendency for mothers of firstborn girls to use criticism more than the other groups (x2 = 6.67, p < .10, two-tailed).

A closely related measure, this one taken over all situations except the puzzle, was the number of times the mother told her child that he was correct or incorrect. There were no sig- nificant differences in telling the child he was correct, but a significant interaction was found for telling the child he was incorrect (F — 4.31, df = 1/51, p < .05), with the mother more likely to tell the firstborn girl or the second- born boy that he had done something wrong.

Finally, in an attempt to measure the moth- er's encouragement of her child's responsible behavior, a rating was made of the child's in- volvement in cleaning up the room. Ratings of cleanup were made from the experimenter's descriptions, with a scale ranging from the mother doing all of the cleanup (low score) to the child doing all of the cleanup (high score).

The cleanup scale yielded a significant inter- action (F = 8.39, df = 1/47, p < .01), with a larger difference between firstborn and sec- ond-born boys in the direction of firstborn boys participating more in the cleanup.

DISCUSSION The present study showed no tendency for mothers to interact more with firstborns, or to have a general overestimation of the firstborn's ability. Indeed, there was a notable lack of cor- relation among measures intended to assess the mother's estimate of her child's ability.

The mother's behavior appears to be affected more by the nature of the particular task than by some general notion of her child's ability.

This finding may not be surprising in view of the fact that subjects were 5 years old at the time of the study; the mothers would have had sufficient time to observe that their children were more successful at some tasks than at others.

It seems likely that any generalized overestimate of the firstborn child's ability would be more evident in the mother's behavior toward the infant or very young child than toward the 5 year old, The strongest birth-order difference among the estimate measures was the mother's tend- ency to give a more complex technical explana- tion to the firstborn; on the less technical car- toons, differences were in the same direction, but not significant. Apparently, even if the mother does not show a generalized overesti- mate of the first child's ability, she nevertheless uses more complex language with the firstborn than the second-born child. Whether she is overrating the firstborn child's ability to under- stand or simply providing him with better in- tellectual stimulation than the second-born child is a question for further research.

In considering measures of maternal pres- sure for success, a comparison of this study's findings with those of Hilton (1967) is of value.

She gave mothers differential information about their child's success or failure in order to compare mothers' interactions with firstborn and second-born children on a puzzle task.

Hilton reported greater maternal interference BIRTH ORDER AND MOTHER-CHILD INTERACTION 119 with firstborn and only children (grouped to- gether) on several variables:

mothers of first- born and only children were rated as "more involved," were more likely to initiate work on the puzzle task, and gave more task-oriented suggestions and direct help to the firstborn child.

While a detailed analysis of maternal help-giving data from the present study is not completed, the measure of overall anxious intrusiveness is highly relevant to the question of maternal interference with the firstborn.

Mothers of firstborns were rated as more intrus- ive than mothers of second borns, with mothers also more intrusive toward girls than boys.

Another finding which may be related to inter- ference is that mothers exerted more pressure for achievement on the firstborn in the picture task, just as mothers of firstborns were more likely to initiate the puzzle task in Hilton's study. Results of both studies suggest that mothers are more intrusive into the perform- ance of firstborn than second-born children, although the present study found this accentu- ated for the firstborn girl.

It is possible that the mother's greater interference with the first- born provides such a readily accessible source of support that the firstborn may depend more on others for support in achievement situa- tions.

The firstborn's later success in school may be mediated to some extent by his depen- dency on others for setting standards for his performance.

A major difference in.

the findings of the two studies, however, is that while Hilton failed to find consistent differences in the mother's be- havior as a function of the sex of the child, re- sults of the present study favor the view that maternal behavior toward the first- or second- born child is also influenced by his sex.

For ex- ample, the firstborn girl seems to have evoked the most extreme responses from the mother.

The pressure for remembering variable showed greater differential pressure on the firstborn girl as compared to the second-born girl, in addition to a birth-order difference; the mother gave a higher estimate of her firstborn girl's performance on the picture; the mother was more likely to tell the first girl she was incorrect, and showed a tendency to be more likely to criticize her; and the mother showed most anxious intrusiveness toward the firstborn girl.

There is a striking similarity between these findings and those of Cushna (1966), reported by Sutton-Smith and Rosenberg (1970).

Cushna's subjects were 16-19-month-old chil- dren from middle-class families. When mothers were asked to determine their children's per- formance on a number of tasks, mothers were found to be more involved in influencing the performance of firstborn children, but in differ- ent ways for boys and girls. Mothers were more supportive and cautious in directing their boys, but more demanding, exacting, and in- trusive toward their firstborn girls.

The second-born girl seems to be shown both less criticism and less pressure to achieve than the firstborn girl Her mother expected less from her on the picture task, and less often told her that she was incorrect.

The firstborn boy shared with the firstborn girl greater maternal pressure for achievement, but only on the first part of the picture task; the birth-order differ- ence on the pressure for remembering variable was contributed chiefly by the mother's pres- sure on the firstborn girl.

The mother gave her firstborn boy a more complex explanation and showed greater anxious intrusiveness toward the firstborn than the second-born boy, but showed no greater intrusiveness toward the firstborn boy than toward the second-born girl.

The direction of differences was also for the firstborn boy to be told he was incorrect less often than was the second-born boy.

The mother's pressure for the firstborn boy's achievement and intrusiveness into his per- formance thus seems somewhat tempered in comparison with her pressures on the firstborn girl.

Finally, the second-born boy seems to have received less pressure and interference from the mother than the firstborn boy, but was also more likely to be told that he was incorrect than was the firstborn boy.

It is interesting to speculate about the more extreme pressure and intrusiveness exhibited toward the firstborn girl in this study.

One possibility is that of all ordinal positions, the mother most closely identifies with the first- born girl, since the firstborn girl most resembles the mother at any given time. This could lead to both greater pressure for achievement and to less satisfaction with the child's performance, since it is difficult to fulfill high expectations.

The mother may also feel something of an at- traction toward the firstborn boy that would 120 MARY 1C.

ROTIIBAKT temper her behavior toward him, and a sense of rivalry toward the firstborn girl, in a role re- versal of the Oedipal triangle.

Her feelings, both positive and negative, may be less ex- treme toward the second-born girl or boy.

It was found in a previous study (Rothbart & Maccoby, 1966) that parents tended to re- spond more permissively toward a child's voice of the opposite sex; this differential re- action may be stronger for the firstborn in the family, and less accentuated in the parent's reaction to the later born.

An additional finding of interest is that while firstborn boys were involved in the cleanup to as great an extent as either first- or second-born girls, the second-born boy was less likely to be involved in the cleanup.

Cleaning up, while a measure of adoption of responsibility, is an activity that is also highly sex typed in the feminine direction.

The differences found may reflect a ceiling effect in the behavior of girls which could mask a higher adoption of responsi- bility for the firstborn to be found in a more neutral (or masculine) activity.

In addition, the findings may reflect the greater masculine identification of the second-born boy in an all- male family that has been reported elsewhere (Rosenberg & Sutton-Smith, 1964).

It should be noted that the findings of the present study are very limited.

Only two-child, same-sex families were involved, with only moiher-child interaction observed.

The value of observing fathers interacting with their chil- dren seems apparent from the Sex X Birth Order interactions found in this stud)'.

We might expect that in such research, the first- born boy would occupy the extreme position occupied by the firstborn girl in the present study.

It would also be of value to include cross-sex sibling pairs and families of a larger size than two children, although with needed controls for size of family, sex, and spacing of siblings, such research becomes very difficult.

REFERENCES ALTUS, W.

D.

Birth order and its sequelae. Science, 1966, 151, 44-49.

BAYER, A. E.

Birth order and college attendance.

Journal of Marriage and Family Living, 1966, 28, 480-484.

BAYLEV, N.

Comparisons of mental and motor test scores for ages 1-15 months by sex, birth order, race, geographical location, and education of parents.

Child Development, 1965, 36, 379-411.

BOSSARD, J. H. S.

Family modes of expression.

A men- can Sociological Review, 194S, 10, 226-237.

CUSHNA, B.

Agency and birth order differences in very early childhood. Paper presented at the meeting of the American Psychological Association, New York, September 1966.

DAVIS, A.

American status systems and the socialization of the child. American Sociological Review, 1941, 6, 345-354.

HILTON, I.

Differences in the behavior of mothers to- ward first- and later-born children. Journal of Per- sonality and Social Psychology, 1967, 7, 282-290.

KOCH, H. L. The relation of "Primary Mental Abili- ties" in five- and six-year olds to sex of child and characteristics of his sibling.

Child Development, 1954, 25, 209-223.

McARTHUR, C.

Personalities of first and second chil- dren.

Psychiatry, 1956, 19, 47-54.

PHILLIPS, E. L.

Cultural vs.

intropsychic factors in childhood behavior problem referrals.

Journal of Clinical Psychology, 1956, 12, 400-401.

ROSEN, B. C.

Family structure and achievement moti- vation. American Sociological Review, 1961, 26, 574-585.

ROSENBERG, B. G., & SUTTON-SMITH, B.

Ordinal posi- tion and sex-role identification.

Genetic Psychology Monographs, 1964, 70, 297-328.

ROTHBABT, M.

K.

Birth order and mother-child inter- action. (Doctoral dissertation, Stanford University) Ann Arbor, Mich.:

University Microfilms, 1967.

No.

67-7961.

ROTHBART, M.

K., & MACCOBY, E. E.

Parents' differ- ential reactions to sons and daughters.

Journal of Personality and Social Psychology, 1966, 4, 237-243.

SAIIPSON, E. E. The study of ordinal position: Ante- cedents and outcomes.

In B.

Maher (Ed.), Progress in experimental personality research.

Vol.

2. New York:

Academic Press, 1965.

SCHACHTER, S.

Birth order, eminence, and higher edu- cation. American Sociological Review, 1963, 28, 757-767.

SUTTON-SMITH, B., ROBERTS, J.

M., & ROSENBERG, B. G.

Sibling associations and role involvement.

Merrill-Palmer Quarterly of ^Behavior and Develop- ment, 1964, 10, 25-38.

SUTTON-SMITH, B., & ROSENBERG, B. G. The sibling.

New York: Holt, Rinehart & Winston, 1970.

(Received January 5, 1970)