quality of parental caregiving and security of...

Developmental Psychology1993. Vol. 29. No. 5.358-367

Copyright 1993 by the American Psychological Association, Inc.0012-1649/93/S3.00

Quality of Parental Caregiving and Security of Attachment

Karen Schneider Rosen and Fred Rothbaum

The assumption that qualitative differences in parental behavior are associated with attachmentsecurity was examined in a sample of 62 children (M age =21.5 months) who were seen separatelywith their mothers and fathers. Multiple measures of parental caregiving were used, including 2qualitative behavioral rating scales and a self-report measure of attitudes and beliefs about childrearing. Analyses of the relation between these measures and maternal and paternal Strange Situa-tion classifications of attachment security revealed effects only for mothers and only with 1 parentmeasure. These results add to a fairly impressive body of evidence indicating inconsistent and oftenweak associations between parental behavior and attachment security. Conceptual and method-ological issues relevant to the parenting-attachment association were identified.

One of the assumptions pervading attachment theory andresearch is that the security of the attachment relationship re-flects, in large part, variations in maternal behavior (Ains-worth, Bell, & Stayton, 1974; Ainsworth, Blehar, Waters, &Wall, 1978; Bretherton, 1985). Early findings on the effects ofmaternal deprivation (Bender & Yarnell, 1941; Goldfarb, 1943;Rutter, 1979; Skodak & Skeels, 1949) and findings regardingthe quality of attachment in populations of maltreated infants(Carlson, Cicchetti, Barnett, & Braunwald, 1989; Egeland &Sroufe, 1981; Schneider-Rosen, Braunwald, Carlson, & Cic-chetti, 1985; Schneider-Rosen & Cicchetti, 1984) provide evi-dence of the consequences of extreme variation in maternalbehavior for the developing attachment relationship. Ains-worth's early studies (e.g., Ainsworth & Bell, 1969; Ainsworth &Wittig, 1969; Stayton & Ainsworth, 1973) exploring variationsin maternal responsiveness within the normal range and theirassociation to patterns of attachment strengthened the assump-tion that maternal behavior contributes significantly to theearly attachment relationship.

Although this assumption does not preclude the possibilitythat caregivers are influenced by their infants (Bell, 1968;Belsky & Rovine, 1987; Chess & Thomas, 1982; Crockenberg,1986; Kagan, 1984; Rosen & Chmiel, 1992; Sroufe, 1985;Thompson, 1986), it does implicate a critical influence of themother on the organization of attachment relationships. Find-

Karen Schneider Rosen, Department of Psychology, Boston Col-lege; Fred Rothbaum, Department of Child Study, Tufts University.

This study was supported by faculty research grants from the au-thors' respective universities.

We gratefully acknowledge the staff of the Family and Child Devel-opment Project for their invaluable contributions to the collection andcoding of the data reported here. Denise Rao was especially helpful inproviding critical insights into many theoretical issues and assistancein data analyses. We also appreciate the efforts of Mona Beatty, Ste-phen Bowen, Karen Braunwald, Susan Chmiel, Amy Curcio, SandraCohen, Carol DeVirgilio, Kristin Keuter, Carla Picariello, ColleenPreston, Cara Regan, Jeanne Ruckert, and Kristen Spada.

Correspondence concerning this article should be addressed toKaren Schneider Rosen, Department of Psychology, Boston College,Chestnut Hill, Massachusetts 02167.

ings based on the Strange Situation procedure are used contin-ually to draw conclusions about the contributions of mothers(Ainsworth, 1973; Blehar, Lieberman, & Ainsworth, 1977; Er-ickson, Sroufe, & Egeland, 1985; Grossman, Grossman,Spangler, Suess, & Unzner, 1985; Waters, 1978), even thoughthe procedure does not directly assess maternal behavior. Thereis relatively less research on the connection between maternalbehavior and the quality of attachment than on other issues inattachment theory, such as the developmental consequences ofearly attachment patterns. This is surprising, given that attach-ment theorists have described the experiential origins of theattachment relationship even more explicitly than they have thedevelopmental outcomes (Ainsworth, 1973; Ainsworth, Bell, &Stayton, 1971,1974; Bowlby, 1969,1988).

Where the association between parenting and attachmenthas been examined, there are mixed results. That is, there isinconsistent support in the literature for an association be-tween dimensions of maternal responsivity or sensitivity andattachment classifications. Some research appears to confirmthe association (Ainsworth et al., 1978; Grossman et al., 1985;Lewis & Feiring, 1989; Matas, Arend, & Sroufe, 1978; Smith &Pederson, 1988). However, there is also empirical work that hasyielded low magnitude or nonsignificant relations (Frodi,Bridges, & Grolnick, 1985; Goldberg, Perrotta, Minde, &Corter, 1986; Isabella, Belsky, & von Eye, 1989; Lyons-Ruth,Connell, Zoll, & Stahl, 1987; Maslin & Bates, 1983; Miyake,Chen, & Campos, 1985). Even within the same study, there isevidence that both confirms and fails to support the contentionthat maternal behavior is critical in determining the quality ofattachment (e.g., Bates, Maslin, & Frankel, 1985; Belsky, Rovine,& Taylor, 1984; Egeland & Farber, 1984; Londerville & Main,1981).

Given the extant research findings, there are some who haveconcluded that the role of maternal responsivity is not as strongas was originally proposed (e.g., Lamb, Thompson, Gardner, &Charnov, 1985; Lamb, Thompson, Gardner, Charnov, & Estes,1984; see Goldsmith & Alansky, 1987, for a meta-analysis). Webelieve that it is premature to draw such a conclusion beforeexamining some of the methodological constraints of currentempirical work. For example, the operational ization of the re-

358

PARENTAL CAREGIVING AND ATTACHMENT 359

sponsivity construct has not been uniform across studies. Manydifferent measures of maternal behavior have been used. Be-cause the critical behaviors that constitute responsivity havebeen denned inconsistently by different researchers, the realassociation between maternal behavior and the quality of at-tachment may be obscured.

Whereas certain methodological shortcomings may lead tounderestimates of the parenting-attachment relationship,others may lead to overestimates. Some researchers, while rec-ognizing the value of using multiple measures of parental care-giving behavior, emphasize findings only for those measures orfor those subsamples yielding significant differences betweenattachment classifications (Egeland & Farber, 1984; Isabella etal., 1989; Pettit & Bates, 1989). When strong effects are ob-tained (Ainsworth et al., 1978; Matas et al., 1978), it is not alwaysclear that the parent and attachment-security measures werecoded independently, thereby raising the possibility of halo ef-fects.

In addition, instructions accompanying qualitative measuresof maternal caregiving do not prohibit raters from consideringthe child's behavioral response to the mother's intervention(e.g., Ainsworth et al., 1978; Matas et al, 1978). Thus, in assess-ing whether a parent's behavior is responsive, raters may takeinto consideration such positive responses as a decrease in cry-ing or an immediate compliance to a parental request. Ourposition is that raters should be explicitly instructed to not con-sider the child's reaction to a parent's intervention because do-ing so introduces a confound. Children who are socially compe-tent or temperamentally easy and who respond positively totheir mother's behavior may make what the mother did look"better" or "more successful" than children who are more dif-ficult to manage or whose behavior does not change or becomesworse as a result of the mother's intervention. In other words,consideration of the child's reaction in assessing the parent canbuild in a positive relationship between measures of child be-havior (such as the Strange Situation assessment of attachment)and measures of maternal behavior.

There are several other limitations of extant research. First,few studies to date have examined paternal responsivity andattachment security (Easterbrooks & Goldberg, 1984). Thus, weknow little about the association between paternal behaviorand attachment. Second, in most studies, only a single measure-ment approach typically has been used to evaluate parentalbehavior. Using only one measure is particularly problematic,because in most cases there has been little independent evi-dence of the reliability and validity of the measure. Third, moststudies examine parental behavior in one situation as opposedto multiple situations (Wachs, 1987). These problems and limi-tations in existing work may explain why there is inconsistentsupport for the association between parental behavior and at-tachment classifications.

In the present study, the association between several contem-poraneous measures of maternal and paternal behavior andpatterns of attachment was assessed in children during the sec-ond half of the second year. This is an important periodof developmental transition (Cicchetti & Beeghly, 1991;Schneider-Rosen & Wenz-Gross, 1990; Sroufe, 1985) duringwhich the attachment relationship remains critical (Bretherton

& Waters, 1985; Greenberg, Cicchetti, & Cummings, 1991;Schneider-Rosen, 1991). There are conflicting findings regard-ing the association between caregiving and attachment duringthis period (Frodi et al., 1985; Matas et al, 1978). Concurrentassessments of both parental behavior and attachment securityallow for an examination of their association at the point wherethe child is assuming greater autonomy and independencewhile simultaneously negotiating these developmental changesinto the attachment relationship.

The Strange Situation was used to assess the quality of attach-ment. The parent-child dyads also participated in several inter-active exchanges in which parental behavior was evaluated.Two qualitative observational measures of parenting were used:one was designed to assess parental Quality of Assistance dur-ing problem-solving tasks (Sroufe, Matas, & Rosenberg, 1981),and one was the Parental Acceptance Coding Scheme, appro-priate for evaluating parental behavior in a variety of interactivesituations and designed explicitly so as to not evaluate the par-ent's sensitivity or responsiveness on the basis of the child'sreaction to the parent's behavior (Rothbaum & Schneider-Ro-sen, 1988). A self-report measure of parental attitudes and be-liefs, the Child-Rearing Practices Report (Block, 1981), wasused because it provides information about a variety of aspectsof caregiving. Multiple measures were used so that results couldnot be attributed to the reliance on a single measure of parent-ing. All of the measures had prior evidence of reliability andvalidity.

The goal of this study was to further assess the associationbetween parental caregiving and security of attachment. Sev-eral questions and hypotheses were examined. First, does theassociation between caregiving and attachment vary as a func-tion of the measure of caregiving? Variations in this associationmay provide important clues as to those aspects of caregivingthat are particularly relevant to the emerging attachment rela-tionship; alternatively, similarities in this association across thedifferent measures of caregiving will increase our confidencein the findings obtained.

Second, are the caregiving measures associated with one an-other? These associations were of interest, because they pro-vided evidence of the convergent validity of the measures usedin this study. We expected to find associations between thecaregiving measures, although the behavioral measures (Qual-ity of Assistance and Parental Acceptance) were expected to bemore highly associated with one another than either was withthe self-report measure (see Holden & Edwards, 1989).

Third, is there concordance between attachment securitywith mothers and fathers? Because there is evidence that bothsupports (Belsky, Garduque, & Hrncir, 1984; Grossman, Gross-man, Huber, & Wartner, 1981; Lamb, Hwang, Frodi, & Frodi,1982; Main & Weston, 1981) and refutes (Fox, Kimmerly, &Schafer, 1991; Lamb, 1978; Owen & Chase-Lansdale, 1982) theidea that maternal and paternal attachment classifications areindependent of each other, the association between ratings ofattachment to mothers and fathers was explored.

Finally, is there a difference in the association between paren-tal caregiving and attachment security for mothers and fathers?Few theorists have addressed this comparison in research todate.

360 KAREN SCHNEIDER ROSEN AND FRED ROTHBAUM

Method

Subjects

The subjects for this study were 62 children (29 girls and 33 boys)who were part of a larger sample of children participating in the Fam-ily and Child Development Project at Boston College. Children wereseen with their mother and father in two separate 2-hr laboratory visits,scheduled with an average of 2 weeks between visits. The procedureswere identical in the two sessions. The order of visits with mother andfather was counterbalanced to diminish possible order effects. Themean age of the sample was 21.5 months (range = 17.8 to 25.6 months).Twenty-seven (44%) of the children were firstborn.

The children and their parents were recruited through city hall birthrecords. AH families were contacted through letters describing the re-search and requesting voluntary participation. Of the 82% who re-sponded, 85% (n = 70) agreed to participate. Separate appointmentswere scheduled for mothers and fathers to come in with their child.Scheduling difficulties precluded the inclusion of eight families. Thechildren were all White and were from intact middle- to upper-middle-class families (Hollingshead, 1975).

Measures and Procedure

All parent-child dyads were seen in a 15 X 20 ft (4.57 X 6.09 m)carpeted playroom, and all experimental procedures were videotapedfrom behind a one-way mirror. The procedures in which the subjectsparticipated were identical across the two laboratory visits.

Quality of attachment. The initial laboratory assessment was theStrange Situation (Ainsworth & Wittig, 1969), a standardized proce-dure for evaluating the quality of the parent-child attachment relation-ship. The child's exploration of the novel environment in the presenceof the parent, the child's reaction to separation from the parent, andthe reunion with the parent were observed. Individual differencesboth in the use of the caregiver as a secure base from which to explorethe environment and in the ability to derive comfort from the caregiverwhen stressed were assessed throughout the seven 3-min episodes.

We rated the interactive behaviors of proximity-seeking, contactmaintenance, avoidance, resistance, search, and distance interactionby using a modified scoring system appropriate for this older age group(Schneider-Rosen et al., 1985). We classified children into one of threeattachment groups: anxious avoidant (Group A), secure (Group B), andanxious-resistant (Group C). Five children displayed behavioral char-acteristics representative of a disorganized (Group D) attachment pat-tern (Main & Solomon, 1986). Because of the small number of thesecases, forced classifications were used to fit them into traditional cate-gories.

All Strange Situations were coded from videotapes by a trained grad-uate student. One half of the videotapes (n = 31 mother-child and n =31 father-child tapes) were scored by a second trained rater toestablishreliability. Interrater reliability (number of agreements over number ofagreements and disagreements) was 86% (Cohen's kappa = .72). Forthose cases in which disagreements occurred, a final rating was madethrough discussions between the two raters and a third trained coder.

Situations for scoring parental caregiving behavior. Qualitativeaspects of maternal and paternal behavior were assessed during fivesituations: a free-play session, a reading period, a clean-up period, andan easy and a difficult problem-solving task. After the Strange Situa-tion, a new box of toys was brought into the playroom and the childrenand their parents were observed in a free-play situation. The parent wastold to behave as he or she would when playing with his or her child athome. After5 min, the parent was instructed to tell his or her child thatit was time to read a book together. During this time, the parent wastold to keep his or her child from playing with the toys.

After the 5-min reading period, the parent told the child that it was

time to clean up the toys and worked together with his or her child toreturn all the toys to the toy box. After this 3-min clean-up period, thetoys were removed from the room and a table and chair were broughtin. The child was seated at the table and was presented with two differ-ent problems to solve in a standard order. The first problem, a puzzle,was relatively easy for the child and was presented for 3 min; the secondand more difficult problem, a complicated shape-sorting task, waspresented for 5 min. The specific puzzle and shape-sorting tasks dif-fered for the visits with mothers and fathers, and their order of presenta-tion was counterbalanced. For all of the problems, the parent wasinstructed to allow the child to work independently for 1 min and thento assist the child in an effort to help him or her solve the problemduring the time remaining.

Parental acceptance. The Parental Acceptance Coding Scheme(Rothbaum & Schneider-Rosen, 1988) was used tocode parental behav-ior in the five situations described earlier. The Parental AcceptanceCoding Scheme was developed to more clearly identify the interrelateddimensions of sensitivity and responsiveness and to provide a measureof the quality of the parent-child relationship with a specific focus onparental behavior. The Parental Acceptance measure was developedby drawing from the conceptual models and scoring systems of Ains-worth et al. (1978) and Sroufe et al. (1981). It provides a comprehensiveassessment of the different aspects of parental sensitivity described onboth theoretical and empirical grounds (Ainsworth et al., 1978;Bowlby, 1988; Matas et al., 1978) as relating most strongly to attach-ment security. The measure does not evaluate other characteristics ofparenting, such as parental style (demanding, restrictive, directive,anxious, or talkative) or methods (teaching, rewards, or humor), whichare not considered to be important to the developing attachment rela-tionship (Ainsworth etal., 1978; Rothbaum & Schneider-Rosen, 1988).

The scheme extends beyond existing measurement systems in sev-eral important ways, (a) Unlike the measure developed by Ainsworth etal. (1978), which was appropriate for parents of infants and involvedglobal ratings, the Parental Acceptance measure is appropriate for par-ents of children beyond the first year of life and entails more behav-iorally based ratings of sensitivity and responsiveness, (b) In contrast tothe Sroufe et al. (1981) measure, which assessed parental behavior in ateaching situation, the Parental Acceptance measure is designed foruse in multiple situations, (c) An explicit part of the scoring system is tonot evaluate the parent's sensitivity or responsiveness on the basis ofthe child's reaction to the parent's behavior. The reliability and validityof the Parental Acceptance measure have been demonstrated in pre-vious concurrent and longitudinal studies with parents of preschoolersand school-age children (Bishop & Rothbaum, in press; Rothbaum,1986,1988; Rothbaum & Crockenberg, 1992; Rothbaum, Rosen, Pott,& Beatty, 1992).

Parental acceptance is conceptualized as the parent's ability to inte-grate the child's needs with the parent's own needs. Parents are accept-ing when they act as if their children's needs are in harmony with theirown and when they are responsive to their children's needs. By con-trast, rejecting parents act as if their own and their children's needs arein opposition to one another and that responsiveness to the child'sneeds implies a thwarting of their own needs. For a fuller descriptionof the acceptance-rejection construct, see Rothbaum and Schneider-Rosen (1988).

Although the parent's behavior is the object of assessment in thiscoding scheme, the child's behavior is important to evaluate insomuchas it "sets the stage" for the parent's subsequent behavior. We mustconsider the child's behavior to determine (a) the developmental levelat which the child is functioning and (b) the specific needs that arebeing expressed in the interaction. However, ratings of the parent arenot based on the child's reactions to the parent's behavior, because thiswould confound assessments of parent and child.

Using the detailed instructions delineated in the Parental Accep-


tance Coding Scheme, minute-by-minute analyses of parent-child in-teraction yielded scores on the three main aspects of acceptance: ap-proval, availability, and structure. Both Ainsworth et al. (1978) andSroufe et al. (1981) discussed components of these three aspects, but weelaborate on and describe these in the Parental Acceptance CodingScheme so as to include behaviors relevant to a wide range of situationsand to children during toddlerhood and early childhood. Qualitativecriteria for evaluating the three aspects of acceptance are also de-scribed in detail in the scoring manual. The aspects and their criteriaare as follows: approval (evaluative affect and evaluative content); avail-ability (physical presence, acknowledgment, and participation); andstructure (helpfulness of guidance, explanations, motivation setting,consistency, and noncoercion).

Coders rated each minute of interaction using videotapes and verbaltranscripts. They rated each criterion on a scale of 1 to 5, using descrip-tions anchored at the high end (rating of 5) by accepting behaviors thatreflect responsiveness to and harmony with the child's needs and at thelow end (rating of 1) by rejecting behaviors that reflect nonresponsive-ness and opposition to the child's needs. For example, the criterionnoncoercion was assigned a rating of 5 for behaviors such as "giving thechild a sense of choice" and a rating of 1 for behaviors such as "thwart-ing the child's sense of choice" (e.g., by using force). The score for eachaspect was determined by averaging across the minute-by-minute rat-ings for the criteria relevant to that aspect. The overall acceptancescore was computed by averaging across ratings for the three aspects.

A group of four trained raters scored each parent-child dyad. Theseraters were unaware of the attachment classifications and of all otherdata. Raters were trained to evaluate parental behavior in response tothe immediately preceding child behavior. Training continued untileven small disagreements between raters were judged not to be causedby the confounding of parental behavior and child's responses to thatbehavior. Each situation was scored by pairs of raters; each raterworked independently but met regularly to discuss and resolve discrep-ancies. The average overall reliability between raters within pairs (i£.,raters scoring the same situation) was r = .80, p < .001. As in otherstudies using measures to capture various aspects of caregiving behav-iors (Ainsworth et al., 1978, Crowell & Feldman, 1988, Easterbrooks &Goldberg, 1984), the different aspects of Parental Acceptance werehighly intercorrelated (rs > .90, p < .001). Therefore, only the raters'final Parental Acceptance scores, computed by averaging the variousaspects, were analyzed.

Parental quality of assistance. A second measure of parental care-giving behavior used in this study was developed by Sroufe et al. (1981).The original system, which is widely used and well constructed withsubstantial evidence of reliability and validity, rated mothers on twodimensions: Supportive Presence and Quality of Assistance. Crowelland Feldman (1988) and Easterbrooks and Goldberg (1984) found ahigh correlation between the two dimensions (rs = .90 and .71, p < .001for the two studies, respectively). Sroufe (personal communication,September 1988) also reported that the two dimensions are highlycorrelated and in recent work (Pianta, Sroufe, & Egeland, 1989) usedonly the Quality of Assistance scale to assess parental behavior inparents of 2-year-old children. We selected the Quality of Assistancescale for the purposes of this study because it identifies criteria that arethe most different from the parental acceptance measure. The Qualityof Assistance scale was used to assess behavior in the two problem-solving situations because it was developed for that type of situation.

Quality of Assistance refers to the parent's ability to provide thechild with minimal assistance and allow for the child's maximal explo-ration while also responding sensitively to the child's cues for directionby helping him or her see relations between actions required to solvethe problem. The coding manual delineates several major criteria andcorresponding subcriteria for coding Quality of Assistance, and it pro-vides detailed instructions for assigning ratings. In this study, we inter-

preted the instructions to mean that coders should not take children'sreactions into account when coding the parents. This required adapt-ing some of the scales so that the child's responses could not influenceevaluation of the parent's preceding behavior. The final score rangedfrom 1 to 7, with higher scores indicating a greater degree of sensitivityand the capacity to provide helpful, well-timed instructions in a nonin-trusive, clear, comprehensible, and positive manner. Because theSroufe et al. (1981) coding system was originally developed for use with24-month-old children, data were coded for only a subsample of 31children who ranged in age from 23 to 25.6 months. One coder, un-aware of the attachment classifications and parental acceptancescores, coded all the tapes and a second rater scored half of the tapes(15 mother-child tapes and 16 father-child tapes). The Pearson corre-lation between the two raters was r= .93, p < .001.

Child-rearing practices report (CRPR). At the end of each labora-tory visit, the CRPR (Block, 1981) was administered to the parents.This Q-sort measure requires parents to sort 91 items into seven pilesdepending on how descriptive each statement is of his or her attitudes,practices, values, and goals with respect to child rearing. Based on eachparent's sorting of the items into piles varying in their descriptivenesswith regard to their child, each item received a score from 1 (leastdescriptive with regard to their child) to 7 (most descriptive). Twenty-one different item clusters are then obtained from the CRPR (Roberts,Block, & Block, 1984). For each cluster, higher scores indicate more ofthe dimension in question (e.g., a higher score on the Negative AffectToward Child cluster indicates more negative affect). Evidence of thereliability and validity of the CRPR has been reported by Block,Block, and Morrison (1981); Roberts et al. (1984); Susman, Trickett,lannotti, Hollenbeck, and Zahn-Waxler (1985); and Trickett and Sus-man (1988).

Thus, the critical variables used to explore the hypotheses delin-eated earlier were maternal and paternal attachment classifications,Parental Acceptance scores, Quality of Assistance ratings, and scoreson the 21 CRPR clusters.

Results

Attachment Classifications

In the mother-child dyads, 16% (n = 10) were anxious-avoidant (Group A), 73% (n = 45) were secure (Group B), and11% (n = 7) were anxious-resistant (Group C). In the father-child dyads, 13% (n = 8) were anxious-avoidant (Group A), 69%(n = 43) were secure (Group B), and 18% (n = 11) were anxious-resistant (Group C). These proportions are comparable withthose obtained in prior research on mother-child dyads in mid-dle-class populations (Ainsworth et al., 1978).

Relation Between Quality of Attachment and Measuresof Parental Caregiving

It was expected that there would be significant associationsbetween attachment classifications and the various measures ofparental caregiving. Separate analyses were conducted for chil-dren with their mothers and fathers. Analyses of variance (AN-OY\s) revealed no significant differences within the three at-tachment categories in Parental Acceptance scores or in Qual-ity of Assistance ratings. Analyses of variance on the 21 CRPRclusters revealed only one significant effect for mothers and one


for fathers. Because there were 21 possible effects, obtainingone significant effect could have been due just to chance.1

We conducted an ANOVA to compare the secure group(Group B) with the insecure groups—anxious avoidant (GroupA) and anxious-resistant (Group C)—separately for mothersand fathers.2 Mothers of securely attached children had signifi-cantly higher Parental Acceptance scores (M = 3.21) than didmothers of insecurely attached children (M = 2.93), /(60) =1.86, p < .05, one-tailed. For fathers, although Parental Accep-tance scores for the secure dyads were higher than those for theinsecure dyads (Ms = 2.82 and 2.75, respectively), the differ-ences were not significant.

There were no significant differences between the secure andinsecure groups on the Quality of Assistance measure for eithermothers or fathers. Similarly, the number of significant differ-ences between the secure and insecure dyads on the CRPRclusters was not significantly greater than chance for mothers orfor fathers.

Thus, mothers of securely attached children did not appearto offer better Quality of Assistance in problem-solving tasks,nor did they differ from mothers of insecurely attached chil-dren in their reported attitudes and beliefs, although they hadmarginally higher Parental Acceptance scores than didmothers of children who were insecurely attached. Fathers ofchildren who were securely attached did not have higher Paren-tal Acceptance scores, did not offer better Quality of Assis-tance, and did not differ from fathers of insecurely attachedchildren in their reported attitude and beliefs as measured bythe CRPR.

Relations Between Measures of Parental Caregiving

Multiple measures were used to examine the quality of paren-tal caregiving. It was expected that the ratings from the differ-ent measures would be associated with each other, thereby in-creasing our confidence in the findings. Results for mothersand fathers were considered separately.

For mothers, Parental Acceptance correlated significantlywith Quality of Assistance, r(29) = .63, p < .01. Parental Accep-tance scores also correlated significantly with the clusters ofSuppression of Sex, r(60) = .35, p < .01, and Negative AffectToward Child, r(60) = -.35, p < .01. Although Quality of Assis-tance scores correlated significantly with only one CRPR clus-ter (Suppression of Sex, r(29) •= -.55, p < .01)—a finding thatcould be due to chance—it is interesting to note that this onecluster also correlated significantly with Parental Acceptance.

For fathers, Parental Acceptance and Quality of Assistancescores were also significantly correlated, r(29) = .42, p < .05.There were significant correlations between Parental Accep-tance scores and three of the CRPR clusters: Rational Guid-ance of Child, r(60) = .40, p < .01; Control by Anxiety Induc-tion, r(60) = -.35, p < .01; and Suppression of Aggression,r(60) = .36, p < .01. There were no significant correlations be-tween Quality of Assistance scores and the CRPR clusters.

The pattern of results indicates that the Parental Acceptancemeasure and the Quality of Assistance measure are evaluatingsimilar aspects of parenting. Although the observational mea-sures bear only a weak relation to the CRPR self-report mea-sure, the significant correlations between Parental Acceptance

and the CRPR were not likely due to chance. Also, all of thesignificant correlations were in the predicted direction for bothmothers and fathers. Thus, the observational measures alwaysrelated positively to clusters that, on the basis of item inspectionand prior empirical findings (e.g., Trickett & Susman, 1988), wehad a priori designated as positive caregiving clusters. Finally,higher Parental Acceptance scores for fathers appear to be asso-ciated with control-related attitudes and beliefs, whereas formothers, higher Parental Acceptance scores appear to be asso-ciated with more emotionally related concerns. These findingsprovide evidence Of the convergent validity of these measureswith the present sample.

Concordance Between Maternal and PaternalAttachment Classifications

The concordance between mother-child and father-child se-cure and insecure attachment classifications was examined andrevealed a significant association, x20, N = 62) = 8.75, p < .01.Children behaved similarly in the Strange Situation when ob-served with their mothers and fathers on two separate occasionsseparated by a 2-week time interval. The phi coefficient (Hays,1981, pp. 555-558) was also computed to determine thestrength of association for dichotomous variables. The associa-tion between classifications of children with their mothers andtheir fathers was significant (<t> = .38, p < .01).

Relations Between Measures of Maternal and PaternalCaregiving

Because children were seen with both their mothers and fa-thers, it was possible to examine differences between parentson measures of Parental Acceptance, Quality of Assistance,and parental attitudes and beliefs on the CRPR. Analyses re-vealed that mothers (M= 3.13) had higher Parental Acceptancescores than fathers (M= 2.80), /(61) = 4.50, p < .001, but therewas no significant difference between mothers and fathers inQuality of Assistance. On the CRPR, fathers scored signifi-cantly higher on Emphasis on Achievement, /(61) = 4.30, p <.001, and Parental Inconsistency, r(61) = 2.88, p < .01.

The correlation between maternal and paternal caregiving,as assessed by either of the observational measures, was notsignificant. However, there were several significant correlations

1 Using the binomial expansion (Hays, 1981), we computed the prob-ability of obtaining 1, 2, 3, and so on significant effects out of the 21possible effects. We selected a significance level of p < .01 in analyzingthe 21 individual CRPR clusters because of the large number of clus-ters and, thus, the large number of effects tested. The probability ofone p < .01 effect out of the 21 possible is p < .20. To further decreasethe likelihood of a Type I error, we treated findings involving theclusters as nonsignificant (and these were not reported) unless therewere two or more p < .0! effects; the probability of obtaining two ormore p < .01 effects out of 21 possible is p < .05.

2 Although there is empirical evidence identifying different patternsof maternal behavior associated with the anxious-avoidant (Group A)and anxious-resistant (Group C) groups, there is also a great deal ofinconsistency even in that literature (e.g., Belsky et al., 1984; Bohlin,Hagekull, Germer, Andersson, & Lindberg, 1989; Lyons-Ruth et al.,1987).


between mothers' and fathers' CRPR clusters: Enjoyment ofParental Role, r(60) = .48, p < .01; Suppression of Sex, r(60) =. 38, p<.01;and Control by Guilt Induction, r(60)=. 50, p<. 01.These results highlight the differences between self-report andobservational measures of parenting and suggest that similaritybetween parents may emerge more in the realm of attitudesthan behaviors.

Discussion

Attachment theorists emphasize the importance of the earlyinteractional history of the caregiver and the child in determin-ing the security of the attachment relationship (e.g., Ainsworthet al., 1978; Bowlby, 1969). They acknowledge that infant char-acteristics (such as temperament) and parental characteristics(such as responsiveness) interact in complex ways to affect thequality of the relationship and thus patterns of Strange Situa-tion behavior. However, child characteristics are discussed insuch a way as to minimize their importance, and parental re-sponsivity is accorded a more significant role (e.g., Goldberg etal., 1986). In part because of the parent's greater cognitive andsocioemotional competence, more emphasis is placed on paren-tal rather than child contributions. Thus, the parent's accommo-dation to his or her child's distinct behavioral style is assumedto be largely responsible for individual differences in the qual-ity of the attachment relationship.

In this study, mothers of securely attached children weremore responsive to their children's needs, as measured by theParental Acceptance Coding Scheme, than were mothers ofinsecurely attached children. This pattern did not emerge forfathers. Although there are several other studies that havefound higher quality caregiving in dyads that are classified assecurely attached, the magnitude of the effects in most of thesestudies, as in the present one, is small (see Goldsmith &Alansky, 1987, for a review). Also, as in this study, prior findingsare often inconsistent in that they do not apply for all relevantcaregiving measures. Even though there are a few studies thatshow large differences between attachment groups (e.g., Ains-worth et al., 1978; Grossman et al., 1985; Lewis & Feiring, 1989;Matas et al., 1978; Smith & Pederson, 1988), the majority ofstudies do not. Support for hypotheses from attachment theoryis often gained from secondary analyses, for example, frompost hoc analyses of subgroups, such as girls or boys (e.g., Ege-land & Farber, 1984) or borderline attachment groups (e.g.,Goldberg et al., 1986). Moreover, many studies that fail to rejectthe null hypothesis may exist, yet they remain unpublishedbecause of the "file drawer phenomenon," that is, data report-ing no associations are unlikely to be published. The results ofthe present study, together with the inconsistent data from pre-vious studies, support the conclusion that there is only a mod-est association between parental caregiving and the quality ofthe attachment relationship.

We can be reasonably certain that the modest associationsare not due to lack of reliability or validity of the caregivingmeasures. The use of different measures, each of which hasindependent evidence of reliability and validity, contributes toour confidence in the findings. Also adding to this confidenceare the significant interrelationships between the measures andthe fact that two of them (the Parental Acceptance and Quality

of Assistance measures) provide a comprehensive assessment ofparental behaviors believed to be critical to the emerging at-tachment relationship. These measures focus on general quali-ties of parental responses to children's specific needs, while alsorelying on behaviorally based criteria. Although the CRPR wasnot designed to predict attachment security, many of its factorsare relevant to sensitive and responsive caregiving. These ad-vantages of the caregiving measures used in this study contrib-ute to our conclusion that the modest relations between paren-tal caregiving and security of attachment is not the result ofinadequate measurement.

There are several points to consider when interpreting themodest associations found between caregiving and attachmentsecurity. First, we observed parents and children in a laboratorysituation. It is possible that highly socialized, middle-class fami-lies view the laboratory as a context for displaying their "best"behavior and, as such, draw on a repertoire of socially accept-able behaviors that consequently (a) reduces the range of vari-ability and (b) does not accurately capture differences in daily,private, and more intimate behavioral patterns. It is importantto observe children and their parents in the home, in additionto the laboratory, to determine whether there are differences inthe strength of parenting-attachment associations in the twosettings.

Second, the duration of the observation periods may be im-portant in influencing the findings. The study that yielded thelargest association between parenting and attachment (Ains-worth et al., 1978) not only relied on home observations, butalso included extensive observations of the parent-child dyadover several months and in a variety of situations. Wachs (1987)provided evidence that the length and diversity of situations inwhich dyads are observed contribute to the predictive validityof measures of caregiving.

Third, it is possible that the specific nature of the situation inwhich parental responsiveness is evaluated may influence thestrength of the obtained findings. Although we adopted theapproach taken by Ainsworth et al. (1978), Matas et al. (1978),and others and observed parental behavior in situations involv-ing parent-child interaction in general, it is possible that ratingsof parental responsiveness in situations that evoke distress (e.g.,separation-reunion sequences) will yield stronger associationswith attachment classifications. This is certainly an issue thatwarrants further examination.

We examined other differences between studies that mightshed light on the inconsistency in the literature on the parent-ing-attachment connection. Although studies vary in samplesize, age of child, socioeconomic status, use of qualitative orquantitative measures of parental behavior, and reliance on thesecure versus insecure attachment distinction or the three tra-ditional attachment groups, none of these variables helped toexplain the inconsistency. Interestingly, even studies using thesame measure report large differences in results. For example,the lack of association between Quality of Assistance scoresand attachment security found in this study parallel those ob-tained by Bates et al. (1985) and Easterbrooks and Goldberg(1984), although they stand in contrast to the very strong resultsobtained by Matas et al. (1978). One factor that might explainthis inconsistency is that when scoring parents' Quality of As-sistance, raters are allowed to consider child responses (e.g., one


subscale of the Quality of Assistance scale requires ratings ofthe "effectiveness" of the parent's behavior). A liberal tendencyto rely on child responses may contribute to strong relation-ships between parenting and attachment. To the extent thatstudies incorporate child behavior within their measures ofcaregiving (cf. Isabella & Belsky, 1991), we suspect that they willobtain strong caregiving-attachment associations. In our study,raters were explicitly instructed not to evaluate parental Qualityof Assistance on the basis of the child's reactions to the tasks orto the parent. Because we obtained no associations to attach-ment security with these instructions, there is reason to suspectthat other studies obtaining weak findings (e.g., Bates et al..1985; Easterbrooks & Goldberg, 1984) used instructions simi-lar to those used in this study.

The measure developed by Sroufe and his colleagues is notthe only one that has yielded inconsistent findings. Briggs,Schiller, Regan, LaGasse, Rosen, and Lipsitt (1992) used theParental Acceptance measure to code parental behavior andobtained stronger associations with attachment security thanwere obtained in the present study. One possible explanationfor the difference in findings is that the significant associationsobtained by Briggs et al. (1992) were largely attributable to sev-eral dyads with very low Parental Acceptance scores. Prior re-search indicates a very strong relation between parenting that isextremely rejecting or abusive and insecure attachment(Schneider-Rosen et al., 1985). Studies using samples in whichextreme negative forms of parenting are exhibited are likely toobtain stronger associations between parental behavior and at-tachment classifications. Differences between samples in theoccurrence of extreme scores could contribute to the inconsis-tency in findings.

Even if the moderate size effects obtained by Briggs et al.(1992) and others are reflective of parenting-attachment associ-ations in the general population, there is still a substantialamount of as yet unexplained variance in attachment security.Transactional models (e.g., Belsky, 1984; Belsky & Isabella,1988; Crockenberg, 1981; Sameroff & Chandler, 1975) suggestthat there are many variables that are likely to influence thequality of the attachment relationship. Likely sources of influ-ence include child characteristics such as temperament (Bateset al., 1985; Belsky & Isabella, 1988; Goldsmith, Bradshaw, &Rieser-Danner, 1986; Piantaetal., 1989; Rosen &Chmiel, 1992;Vaughn, Lefever, Seifer, & Barglow, 1989; Weber, Levitt, &Clark, 1986) and family/situational factors such as maternalemployment (Easterbrooks & Goldberg, 1985), family func-tioning and the quality of the marital relationship (Goldberg &Easterbrooks, 1984), cultural factors (Grossman et al., 1985),and stressful events (Crockenberg, 1981; Crockenberg &McCluskey, 1986).

The attachment concordance data are relevant to this issue.In the past, investigators have relied on evidence of a lack ofconcordance between maternal and paternal attachment classi-fications to support the claim that Strange Situation behavior isdetermined by prior patterns of interaction within the dyad(Sroufe, 1985). If quality of attachment with one parent is in-deed unrelated to quality of attachment with the other, the ar-gument goes, then it is the unique pattern of interaction ratherthan endogenous (characteristics of the child) or exogenous (sit-

uational) factors that is contributing to the security of the at-tachment relationship. There are data that support this position(e.g., Belsky, Rovine, et al., 1984; Grossman et al., 1981; Lamb etal., 1982; Main & Weston, 1981). However, there is also researchthat yields significant concordance between maternal and pa-ternal attachment classifications (e.g., Fox et al., 1991; Lamb,1978; Owen & Chase-Lansdale, 1982), findings that concurwith those from the present study Careful examination of ex-tant studies indicates that the degree of similarity between clas-sifications is highest when there is a short lag between assess-ments, as in the present study (i.e., 1 week to 1.5 months; Lamb,1978; Owen & Chase-Lansdale, 1982) and lowest when the timelag is greater (2 to 6 months; Grossman et al., 1981; Lamb et al.,1982; Main & Weston, 1981). A significant concordance in-creases the possibility that the attachment relationship reflectsa capacity in the child (e.g., a propensity to become securelyattached, the child's internal working model of attachment re-lationships—Main, Kaplan, & Cassidy, 1985) and environmen-tal factors (including agreement between parents in their care-giving attitudes or beliefs—Belsky, Rovine, et al., 1984; cf. Fox etal., 1991).

This is only the second study to date to examine the associa-tion between parental caregiving and attachment for bothmother-child and father-child dyads. In general, the earlierstudy (Easterbrooks & Goldberg, 1984) also found stronger asso-ciations between parenting and security of attachment formothers than for fathers. The similarity in findings in the twostudies emerged despite important differences in the measuresof parenting and in the types of analyses performed. Futureresearch is needed to further probe differences in the caregiv-ing-attachment security association for mothers and fathers(e.g., by using larger samples and performing separate analysesfor boys and girls).

Also warranting further research are the significant differ-ences between mothers and fathers in Parental Acceptancescores as well as in several reported attitudes and beliefs aboutchild rearing on the CRPR. This is not the only study to reportdifferences in the ways in which mothers and fathers interactwith their children (Belsky, Gilstrap, & Rovine, 1984; Parke,1978; Parke & Swain, 1980). Of particular interest are the dif-ferent correlations between Parental Acceptance scores andCRPR clusters for mothers and fathers: for mothers, ParentalAcceptance is associated with attitudes about expressivity (e.g.,affect and impulsivity), and for fathers, Parental Acceptance isassociated with attitudes about instrumentality (e.g., controland discipline). Further research is needed to determinewhether and how these differences in caregiving contribute todifferences between mothers' and fathers' caregiving-attach-ment associations.

Given the inconsistent and overall modest results regardingthe association between parenting and attachment, togetherwith the attachment concordance data and the mother-fatherdifferences, a reasonable next research strategy is to move fromthe examination of simple main effects—of parental caregiv-ing, child characteristics, or family/situational factors—tomore complex interactions between these and other critical fac-tors as they influence and are influenced by attachment secu-rity.


References

Ainsworth, M. D. S. (1973). The development of infant-mother attach-ment. In B. M. Caldwell & H. N. Ricciuti (Eds.), Review of childdevelopment research (Vol. 3, pp. 1-94). Chicago: University of Chi-cago Press.

Ainsworth, M. D. S., & Bell, S. M. (1969). Some contemporary patternsin the feeding situation. In A. Ambrose (Ed.), Stimulation in earlyinfancy (pp. 133-170). San Diego, CA: Academic Press.

Ainsworth, M. D. S., Bell, S. M., & Stayton, D. J. (1971). Individualdifferences in strange situation behavior of one-year-olds. In H. R.Schaffer (Ed.), The origins of human social relations (pp. 17-57). SanDiego, CA: Academic Press.

Ainsworth, M. D. S., Bell, S. M., & Stayton, D. J. (1974). Infant-motherattachment and social development: "Socialization" as a product ofreciprocal responsiveness to signals. In M. P. M. Richards (Ed.), Theintegration of the child into a social world (pp. 99-135). Cambridge,England: Cambridge University Press.

Ainsworth. M. D. S., Blehar, M. C, Waters, E., & Wall, S. (1978). Pat-terns of attachment: A psychological study of the Strange Situation.Hillsdale, NJ: Erlbaum.

Ainsworth, M. D. S., & Wittig, B. A. (1969). Attachment and the explor-atory behavior of one-year-olds in a strange situation. In B. M. Foss(Ed.), Determinants of infant behavior (\o\. 4, pp. 113-136). London:Methuen.

Bates, J. E., Maslin, C. A., &Frankel, K. A. (1985). Attachment security,mother-child interaction, and temperament as predictors of behav-ior-problem ratings at age three years. In I. Bretherton & E. Waters(Eds.), Growing points of attachment theory and research. Mono-graphs of the Society for Research in Child Development, 50(1-2,Serial No. 209), 167-193.

Bell, R. (1968). A reinterpretation of the direction of effects in studiesof socialization. Psychological Review, 75, 81-95.

Belsky, J. (1984). The determinants of parenting: A process model.Child Development, 55, 83-96.

Belsky, J., Garduque, L., & Hrncir, E. (1984). Assessing performance,competence, and executive capacity in infant play: Relations tohome environment and security of attachment. Developmental Psy-chology, 20, 406-417.

Belsky, J., Gilstrap, B., & Rovine, M. (1984). The Pennsylvania Infantand Family Development Project: I. Stability and change in mother-infant and father-infant interaction in a family setting at one, three,and nine months. Child Development, 55, 692-705.

Belsky, J., & Isabella, R. (1988). Maternal, infant, and social-contextualdeterminants of attachment security. In J. Belsky & T. Nezworski(Eds.), Clinical implications of attachment (pp. 41 -94). Hillsdale, NJ:Erlbaum.

Belsky, J., & Rovine, M. (1987). Temperament and attachment securityin the strange situation: An empirical rapproachment. Child Devel-opment, 58, 787-795.

Belsky, J., Rovine, M., & Taylor, D. G. (1984). The Pennsylvania Infantand Family Development Project: III. The origins of individual dif-ferences in infant-mother attachment: Maternal and infant contri-butions. Child Development, 55, 718-728.

Bender, L., & Yarnell, H. (1941). An observation nursery: A study of250 children on the Psychiatric Division of Bellevue Hospital. Amer-ican Journal of Psychiatry, 97, 1158-1172.

Bishop, S. J., & Rothbaum, F. (in press). Parents' responsiveness tocontrol needs and preschoolers' social behavior: A longitudinalstudy. Canadian Journal of Behavioral Science: Special Issue on thePsychology of Control.

Blehar, M. C, Lieberman, A. E, & Ainsworth, M. D. S. (1977). Earlyface-to-face interaction and its relation to later infant-mother at-tachment. Child Development, 48, 182-194.

Block, J. H. (1981). The child-rearing practices report (CRPR): A set ofQ-items for the description of parental socialization attitudes and val-ues. Unpublished manuscript, University of California, Berkeley, In-stitute of Human Development.

Block, J. H., Block, J., & Morrison, A. (1981). Parental agreement-dis-agreement on child-rearing orientations and gender-related person-ality correlates in children. Child Development, 52, 965-974.

Bohlin, G, Hagekull, B., Germer, M., Andersson, K., & Lindberg, L.(1989). Avoidant and resistant reunion behaviors as predicted bymaternal interactive behavior and temperament. Infant Behaviorand Development, 12,105-117.

Bowlby, J. (1969). Attachment and loss: Vol. 1. Attachment. New York:Basic Books.

Bowlby, J. (1988). A secure base: Parent-child attachment and healthyhuman development. New York: Basic Books.

Bretherton, I. (1985). Attachment theory: Retrospect and prospect. InI. Bretherton & E. Waters (Eds.), Growing points of attachmenttheory and research. Monographs of the Society for Research in ChildDevelopment, 50(1-2, Serial No. 209), 3-35.

Bretherton, I., & Waters, E. (Eds.). (1985). Growing points of attach-ment theory and research. Monographs of the Society for Research inChild Development, 50(1-2, Serial No. 209).

Briggs, M. J., Schiller, M., Regan, C, LaGasse, L., Rosen, K. S., &Lipsitt, L. (1992). The relation of maternal acceptance to the child'sbehavior at 12, 18, and 30 months of age. Unpublished manuscript,Brown University, Department of Psychology.

Carlson, V, Cicchetti, D, Barnett, D, & Braunwald, K. (1989). Disor-ganized/disoriented attachment relationships in maltreated infants.Development Psychology, 25, 525-531.

Chess, S., & Thomas, A. (1982). Infant bonding: Mystique and reality.American Journal of Orthopsychiatry, 52, 213-222.

Cicchetti, D, & Beeghly, M. (Eds.). (1991). The self in transition: Infancyto childhood. Chicago: University of Chicago Press.

Crockenberg, S. B. (1981). Infant irritability, mother responsiveness,and social support influences on the security of mother-infant at-tachment. Child Development, 52, 857-865.

Crockenberg, S. B. (1986). Are temperamental differences in babiesassociated with predictable differences in caregiving? In J. Y Lerner& R. M. Lerner (Eds.), Temperament and social interaction in infantsand child (pp. 53-75). San Francisco: Jossey-Bass.

Crockenberg, S. B., & McCluskey, K. (1986). Change in maternal behav-ior during the baby's first year of life. Child Development, 57, 746-753.

Crowell, J. A., & Feldman, S. (1988). Mothers' internal models of rela-tionships and children's behavioral and developmental status: Astudy of mother-child interaction. Child Development, 59, 1273-1285.

Easterbrooks, M. A., & Goldberg, W A. (1984). Toddler developmentin the family: Impact of father involvement and parenting character-istics. Child Development, 55, 740-752.

Easterbrooks, M. A., & Goldberg, W A. (1985). Effects of maternalemployment on toddlers, mothers, and fathers. Developmental Psy-chology, 21, 774-783.

Egeland, B., & Farber, E. A. (1984). Infant-mother attachment: Factorsrelated to its development and expression over time. Child Develop-ment, 5 5,1 Si-Ill.

Egeland, B., & Sroufe, L. A. (1981). Attachment and early maltreat-ment. Child Development, 52, 44-52.

Erickson, M. F., Sroufe, L. A., & Egeland, B. (1985). The relationshipsbetween quality of attachment and behavior problems in preschoolin a high-risk sample. In I. Bretherton & E. Waters (Eds.), Growingpoints of attachment theory and research. Monographs of the Societyfor Research in Child Development, 50(1-2, Serial No. 209), 147-166.


Fox, N. A., Kimmerly, N. L., & Schafer, W D. (1991). Attachment tomother/attachment to father: A meta-analysis. Child Development,62, 210-225.

Frodi, A., Bridges, L., & Grolnick, W (1985). Correlates of mastery-re-lated behavior: A short-term longitudinal study of infants in theirsecond year. Child Development, 56,1291-1298.

Goldberg, W A., & Easterbrooks, M. A. (1984). The role of maritalquality in toddler development. Developmental Psychology, 20,504-514.

Goldberg, S., Perrotta, M., Minde, K., & Corter, C. (1986). Maternalbehavior and attachment in low birth weight twins and singletons.Child Development, 57, 34-46.

Goldfarb, W (1943). Infant rearing and problem behavior. AmericanJournal ofOrthopsychiatry, 13, 249-256.

Goldsmith, H. H., & Alansky, J. A. (1987). Maternal and infant temper-amental predictors of attachment: A meta-analytic review. JournalofConsulting and Clinical Psychology, 55, 805-816.

Goldsmith, H. H., Bradshaw, D. L., & Rieser-Danner, L. A. (1986).Temperamental dimensions as potential developmental influenceson attachment. In J. V Lerner & R. M. Lerner (Eds.), Temperamentand social interaction during infancy and childhood (pp. 5-34). SanFrancisco: Jossey-Bass.

Greenberg, M. T, Cicchetti, D., & Cummings, E. M. (Eds.). (1991).Attachment in the preschool years: Theory, research, and intervention.Chicago: University of Chicago Press.

Grossman, K. E., Grossman, K., Huber, F., & Wartner, U. (1981). Ger-man children's behavior towards their mothers at 12 months andtheir fathers at 18 months in Ainsworth's Strange Situation. Interna-tional Journal of Behavioral Development, 4, 157-181.

Grossman, K., Grossman, K. E, Spangler, G., Suess, S., & Unzner, L.(1985). Maternal sensitivity and newborns' orientation responses asrelated to quality of attachment in Northern Germany. In I. Brether-ton & E. Waters (Eds.), Growing points of attachment theory andresearch. Monographs of the Society for Research in Child Develop-ment, 50(1-2, Serial No. 209), 233-256.

Hays, W L. (1981). Statistics (3rd. ed.). New York: Holt, Rinehart &Winston.

Holden, G. W, & Edwards, L. A. (1989). Parental attitudes towardchild-rearing: Instruments, issues, and implications. PsychologicalBulletin, 106, 29-58.

Hollingshead, A. B. (1975). Four factor index of social status. Unpub-lished manuscript, Yale University, Department of Sociology.

Isabella, R. A., & Belsky, J. (1991). Interactional synchrony and theorigins of infant-mother attachment: A replication study. Child De-velopment, 62, 373-384.

Isabella, R. A., Belsky, J., & von Eye, A. (1989). Origins of infant-mother attachment: An examination of interactional synchrony dur-ing the infant's first year. Developmental Psychology, 25, 12-21.

Kagan, J. (1984). The nature of the child. New York: Basic Books.Lamb, M. E. (1978). Qualitative aspects of mother- and father-infant

attachments. Infant Behavior and Development, 1, 265-275.Lamb, M. E., Hwang, C. P., Frodi, A., & Frodi, M. (1982). Security of

mother- and father-infant attachment and its relation to sociabilitywith strangers in traditional and nontraditional Swedish families.Infant Behavior and Development. 5, 355-367.

Lamb, M. E., Thompson, R. A., Gardner, W, & Charnov, E. L. (1985).Infant-mother attachment: The origins and developmental signifi-cance of individual differences in strange situation behavior. Hillsdale,NJ: Erlbaum.

Lamb, M. E., Thompson, R. A., Gardner, W, Charnov, E. L., & Estes,D. (1984). Security of infantile attachment as assessed in the "strangesituation": Its study and biological interpretation. Behavioral andBrain Sciences, 7,127-171.

Lewis, M., & Feiring, C. (1989). Infant, mother, and mother-infant

interaction behavior and subsequent attachment. Child Develop-ment, 60, 831 -837.

Londerville, S., & Main, M. (1981). Security of attachment, compli-ance, and maternal training methods in the second year of li fe. Devel-opmental Psychology, 17, 289-299.

Lyons-Ruth, K., Connell, D. B., Zoll, D, & Stahl, J. (1987). Infants atsocial risk: Relations among infant maltreatment, maternal behav-ior, and infant attachment behavior. Developmental Psychology 23223-233.

Main, M., Kaplan, N., & Cassidy, J. (1985). Security in infancy, child-hood, and adulthood: A move to the level of representation. In I.Bretherton & E. Waters (Eds.), Growing points of attachment theoryand research. Monographs of the Society for Research in Child Devel-opment, 50(1-2, Serial No. 209), 66-104.

Main, M., & Solomon, J. (1986). Discovery of a disorganized/disor-iented attachment pattern. In T. B. Brazelton & M. W Yogman(Eds.), Affective development in infancy (pp. 95-124). Norwood, NJ:Ablex.

Main, M., & Weston, D. (1981). The quality of the toddler's relation-ship to mother and father: Related to conflict behavior and the readi-ness to establish new relationships. Child Development, 52,932-940.

Maslin, C. A., & Bates, J. E. (1983). Anxious and secure attachments:Antecedents and consequences in the mother-infant system. Unpub-lished manuscript, Indiana University, Department of Psychology

Matas, L., Arend, R., & Sroufe, L. A. (1978). Continuity of adaptationin the second year: The relationship between quality of attachmentand later competence. Child Development, 49, 547-556.

Miyake, K., Chen, S., & Campos, J. J. (1985). Infant temperament,mother's mode of interaction, and attachment in Japan: An interimreport. In I. Bretherton & E. Waters (Eds.), Growing points of attach-ment theory and research. Monographs of the Society for Research inChild Development, 50(1-2, Serial No. 209), 276-297.

Owen, M. T, & Chase-Lansdale, L. (1982). Similarity between infant-mother and infant-father attachments. Paper presented at the meet-ing of the Southwestern Society for Research in Human Develop-ment, Galveston. TX.

Parke, R. D. (1978). Parent-infant interaction: Progress, paradigms,problems. In G. P. Sackett (Ed.), Observing behavior: Vol. 1. Theoryand application in mental retardation (pp. 69-94). Baltimore: Univer-sity Park Press.

Parke, R. D, & Swain, D. B. (1980). The family in early infancy: Social-interactional and attitudinal analyses. In F. A. Pedersen (Ed.), Thefather-infant relationship (pp. 44-70). New York: Praeger SpecialStudies.

Pettit, G. S., & Bates, J. E. (1989). Family interaction patterns and chil-dren's behavior problems from infancy to four years. DevelopmentalPsychology, 25, 413-420.

Pianta, R. C, Sroufe, L. A.. & Egeland, B. (1989). Continuity and dis-continuity in maternal sensitivity at 6, 24, and 42 months in a high-risk sample. Child Development, 60, 481-487.

Roberts, G. C, Block, J. H., & Block, J. (1984). Continuity and changein parents' child-rearing practices. Child Development. 55,586-597.

Rosen, K. S., & Chmiel, S. (1992). Temperament and attachment duringthe second year. Unpublished manuscript, Boston College, Depart-ment of Psychology.

Rothbaum, F. (1986). Patterns of parental acceptance. Genetic Socialand General Psychology Monographs, 112, 435-458.

Rothbaum, F. (1988). Maternal acceptance and child functioning.Merrill Palmer Quarterly, 34,163-184.

Rothbaum, F., & Crockenberg, S. B. (1992). Maternal acceptance andtwo-year-olds compliance and defiance. Unpublished manuscript.Tufts University, Department of Child Study.

Rothbaum, F, Rosen, K. S., Pott, M., & Beatty, M. (1992). Early care-giving and later problem behavior. Unpublished manuscript, Tufts


University, Department of Child Study, and Boston College, Depart-ment of Psychology.

Rothbaum, F, & Schneider-Rosen, K. (1988). Parental acceptance scor-ing manual: A system for assessing interactions between parents andtheir young children. Unpublished manuscript, Tufts University andBoston College.

Rutter, M. (1979). Maternal deprivation, 1972-1978: New findings,new concepts, new approaches. Child Development, 50, 283-305.

Sameroff, A. J., & Chandler, M. J. (1975). Reproductive risk and thecontinuum of caretaking casualty. In F. D. Horowitz, M. Hethering-ton, S. Scarr-Salapatek, & G. Siegel (Eds.), Review of child develop-ment research (Vol. 4, pp. 187-244). Chicago: University of ChicagoPress.

Schneider-Rosen, K. (1991). The developmental reorganization of at-tachment relationships: Guidelines for classification beyond in-fancy. In M. T. Greenberg, D. Cicchetti, & E. M. Cummings (Eds.),Attachment in the preschool years: Theory, research, and intervention(pp. 185-220). Chicago: University of Chicago Press.

Schneider-Rosen, K., Braunwald, K., Carlson, V, & Cicchetti, D.(1985). Current perspectives in attachment theory: Illustration fromthe study of maltreated infants. In I. Bretherton & E. Waters (Eds.),Growing points of attachment theory and research. Monographs ofthe Society for Research in Child Development, 50(1-2, Serial No.209), 194-210.

Schneider-Rosen, K.., & Cicchetti, D. (1984). The relationship betweenaffect and cognition in maltreated infants: Quality of attachmentand the development of visual self-recognition. Child Development,55, 648-658.

Schneider-Rosen, K., & Wenz-Gross, M. (1990). Patterns of compli-ance from 18 to 30 months of age. Child Development, 61, 104-112.

Skodak, M., & Skeels, H. M. (1949). A final follow-up study of onehundred adopted children. Journal of Genetic Psychology, 75, 85-125.

Smith, P. B., & Pederson, D. R. (1988). Maternal sensitivity and pat-terns of infant-mother attachment. Child Development, 59, 1097-1101.

Sroufe, L. A. (1985). Attachment classification from the perspective ofinfant-caregiver relationships and infant temperament. Child Devel-opment, 56, 1-14.

Sroufe, L. A., Matas, L., & Rosenberg, D. M. (1981). Manual for scoringmother variables in tool-use task applicable for two-year-old children.Unpublished manuscript, University of Minnesota, Institute ofChild Development.

Stayton, D. J., & Ainsworth, M. D. S. (1973). Individual differences ininfant response to brief everyday separations as related to other in-fant and maternal behaviors. Developmental Psychology, 9,226-235.

Susman, E. J., Trickett, P. K., Iannotti, R. J., Hollenbeck, B. E., &Zahn-Waxler, C. (1985). Child rearing patterns in depressed, abusiveand normal. American Journal of Orthopsychiatry, 55, 237-251.

Thompson, R. A. (1986). Temperament, emotionality, and infant cog-nition. In J. VLerner& R. M. Lerner(Eds.), Temperament and socialinteraction in infants and children (pp. 35-53). San Francisco: Jossey-Bass.

Trickett, P. K., & Susman, E. J. (1988). Parental perceptions of child-rearing practices in physically abusive and nonabusive families. De-velopmental Psychology, 24, 270-276.

Vaughn, B. E., Lefever, G. B., Seifer, R., & Barglow, P. (1989). Attach-ment behavior, attachment security, and temperament during in-fancy. Child Development, 60, 728-737.

Wachs, T. D. (1987). Short-term stability of aggregated and nonaggre-gated measures of parental behavior. Child Development, 58, 796-797.

Waters, E. (1978). The reliability and stability of individual differencesin infant-mother attachment. Child Development, 49, 483-494.

Weber, R. A., Levitt, M. J., & Clark, M. C. (1986). Individual variationin attachment security and Strange Situation behavior: The role ofmaternal and infant temperament. Child Development, 57, 56-65.

Received March 25,1991Revision received March 25,1992

Accepted May 1,1992 •

quality of parental caregiving and security of...

Documents