Secure Base Support for Social Competence: Attachment Q-sort Correlates of Security and Social Competence in a Sample of African-American Children Attending Head Start*

Kelly Bost
Assistant Professor, Human and Community Development
University of Illinois at Urbana-Champaign

Brian E. Vaughn

Carrol Heller

The Bowlby/Ainsworth approach to describing and interpreting early mother-infant interactions and emergent relationships was proposed as a broad-band characterization of personality and social development (Ainsworth, 1969, 1973; Bowlby, 1973, 1982, 1980;1988; Sroufe, 1986; Waters & Sroufe, 1983). This perspective is distinguished from most psychodynamic and social learning perspectives on social and personality development by its dual emphasis on the of the child's personality and emotional/social adaptation in the context of ongoing interpersonal relationships and on the organizing influences of a behavioral system shaped by the forces of natural selection to promote proximity, contact, and interaction between the human infant and its caregiver(s) (Bowlby, 1982; Waters, Posada, Kondo-Ikemura, & Richters, 1990). The operation of this behavioral system in the context of ongoing interpersonal relationships is described in terms of the "secure-base phenomenon" (e.g., Ainsworth, 1967; Waters, Vaughn, Posada, & Kondo-Ikemura, 1995). That is, as the relationship is consolidated the child uses the caregiver as a base of exploration, and is confident that assistance, nurturance, and protection will be available if and when needed.

The attachment perspective has proven extremely generative in the sense that it has provided the conceptual tools to advance understandings about personality growth and change across a wide range of ages and populations (e.g., Ainsworth, 1967; Ainsworth, Blehar, Waters, & Wall, 1978; Grossmann & Grossmann, 1993; Kobak, Ferenz-Gillies, Everhart, & Seabrook, 1994; Oppenheim & Waters, 1995; Papini & Roggman, 1993; Posada, Gao, et al., 1995; Urban, Carlson, Egeland, & Sroufe, 1991; Waters et al., 1990). However, most empirical research involving attachment has been focused on intrapersonal growth and adaptation (or maladaptation) and not particularly on the interpersonal adaptations and consequences contingent on the successful negotiation of secure attachment relationships.

When interpersonal adaptation is explored (e.g., Arend, Gove, & Sroufe, 1979; Booth, Rose-Krasnor, MacKinnon, & Rubin, 1994; Kavesh, 1992; LaFreniere & Sroufe, 1985; Suess, Grossmann, & Sroufe, 1992; Youngblade, Park, Belsky, 1993), most researchers have adopted a prospective research design with assessments of attachment security collected prior to assessments of interpersonal adjustment (see Park & Waters, 1989 for an exception to this generalization). The results of these investigations indicate that co-construction of a secure attachment during infancy is predictive of positive interpersonal adaptations with peers and with adults during early and middle childhood. These studies do not, however, consider the contribution of concurrent parent-child relationships vis-á-vis attachment to present interpersonal functioning with peers or adults. Consequently, the patterns of empirical results seem to suggest that features of relationships constructed by infants and their caregivers lead to outcomes for children and adults that are independent of the nature and content of later caregiver-child relationships.

Although this interpretation may seem to follow from the existing data and has led to critiques of the attachment position (see Fox, 1995 for a recent example), attachment theory does not rigidly require early attachments to determine later outcomes. Indeed, Bowlby (e.g., 1973, 1982) argued that the secure-base relationship and its resulting mental models are under construction well into childhood and that the mental representation of the parent-child attachment used by a six- to seven-year-old would reflect both continuities and changes in the caregiving history of the child. When the caregiving environment shows continuity from infancy, then assessments made during infancy should yield strong associations with current adaptations. But, when caregiving environments are discontinuous from infancy, current features of the attachment relationship should be more strongly associated with social adaptation than features of the attachment relationship from infancy.

One reason for the failure of researchers to examine the concurrent associations between parent-child attachment and social adaptations with peers and adults during childhood has been the lack of widely accepted measures of attachment for children beyond infancy. In the past decade, this problem has been identified and attempts to design such measures have been largely successful (e.g., Cassidy, Marvin, et al., 1992; Greenberg, Cicchetti, & Cummings, 1990; Waters, 1987; Waters & Deane, 1985; Waters et al., 1995). It is now generally accepted that individual differences in the organization of secure-base behavior (i.e., attachment security) can be assessed both from "emergency" behavior observed during separations in a laboratory setting and from "ordinary" behavior observed at home with the caregiver (see, however, Posada, Cassidy, Marvin, Silverman, & Waters for a discussion of differences in the interpretations of data from these different assessments). It is now feasible to examine concurrent associations between the organization of secure-base behavior with attachment figures and aspects of social adaptations with non-attachment figures (e.g., peers, adults) in the child's broader network of social relationships. The purpose of this report is to describe just such a study.

The data for this report come from a much larger study of the supports for social competence with peers in the lives of preschool children attending Head Start programs in the State of Alabama. Because the vast majority of the children served in these programs are African-American, a population remarkably understudied by attachment researchers, these data are of special interest. The larger study relied on direct observations of child behavior in the classrooms and on face-to-face sociometric interviews with the children to assess individual differences along a broad "social-competence" dimension. Observed behaviors were summarized using Q-techniques and scores for social competence were derived for each child (see Block & Block, 1980; Waters, Noyes, Vaughn, & Ricks, 1985). Additional observations were used to estimate the frequency and rate at which children received visual attention from peers. This measure of attention structure has been established as a valid indicator of social competence in previous research (Vaughn & Martino, 1988; Vaughn & Waters, 1981; Waters, Garber, Gomal, & Vaughn, 1983). Data from the larger sample in this study indicated that these three sources of information can be characterized in terms of a hierarchically organized factorial construct relating each of the three assessment procedures. For the purposes of this report, we consider scores for both the higher-order "social competence" construct and for the three sub-factors that make up the facets of this construct.

Our larger study relies heavily on naturalistic observations in the classroom and on summaries of these observational data using Q-sort techniques. To parallel this assessment strategy in the evaluations of individual differences along the attachment security dimension, we chose to observe children at home with their mothers and to summarize these observations using the Attachment Q-sort (AQS) (Waters, 1987; Waters, et al., 1995). This instrument has been used widely and has been validated as an index of attachment security against a range of relevant criteria (see Vereijken & van Ijzendoorn). To our knowledge, this is the first report on the organization of secure-base behavior in a sample of African-American children of preschool age using the AQS. Consequently, we examine features of the attachment Q-set at a finer level than might otherwise be the case for a majority sample. That is, in addition to the general hypotheses concerning associations between the attachment security score derived from the Q-sort descriptions of the child's behavior at home and prosocial behaviors with peers in the classroom, we also consider the individual item-level correlates of both attachment security and social competence with peers. These item-correlates provide a more detailed picture of the organization of secure-base behavior for these children and afford opportunities to speculate about which aspects of behavior supporting the child's attachment to mother are also supporting competent interactions with peers.



Sixty-nine, low-income African American mothers and their preschool children (26 girls, 43 boys) participated in this study. Mother-child pairs were recruited from 5 Head Start sites (5 four year-old classrooms and 6 three year-old classrooms) located in central Alabama. Class sizes ranged from 18 to 22 and the child participation rates across classrooms ranged from 75% to 100% (total N for 11 classrooms = 216, of whom 191 had complete classroom data). Children's ages ranged from 36-48 months of age at the beginning of the Head Start year. Maternal ages ranged from 17 to 32 years at the time they were observed with their child (M=21.3). Over 85% of these mothers were single parents with at least one additional child residing in the home.


This study was conducted as part of a larger project designed to examine the social ecologies of Head Start children and their families (Vaughn & Bradbard, 1992). Data obtained from the larger study included classroom-based assessments of social competence with peers derived from direct observation of behavior and from sociometric interviews. Data for the children with complete classroom assessments (N = 191) were used to derive composite "social competence" scores (Bost, 1995; Vaughn, Bost, Cielinski, Newell, & Bradbard, 1993) and it is to this social competence composite that we refer in the primary analyses reported here. Classroom data pertaining to social competence were collected by teams of trained graduate and undergraduate students, each team working independently on a separate aspect of the assessment battery. An additional team of observers independently conducted two-hour home observations designed to assess the quality of the mother-child relationship as indexed by the AQS security score (Vaughn & Waters, 1990; Waters & Deane, 1985; Waters, Posada, Kondo-Ikemura, & Richters, 1990) for the 69 children included in this report.



In order to assess attachment security, the Water's Attachment Behavior Q-Sort (AQS) was used (Vaughn & Waters, 1990; Waters, 1987; Waters & Deane, 1985). This measure consists of 90 items designed to describe children's behaviors observed during periods of interaction with primary caregivers. Items were specifically developed to provide a comprehensive characterization of the use of the parent as a secure base (i.e., the balance between proximity seeking and exploration behaviors). The AQS was completed after a two-hour home visit conducted by two well-trained observers. The observations took place in the natural family context, such that all household residents (e.g., siblings) were typically present during the observation period.

The task of the observers was to arrange the items in nine piles from those that were "least descriptive" to those that were "most descriptive" of the target child using a forced-distribution format (i.e., 4, 6, 10, 15, 20, 15, 10, 6, 4). Interobserver agreement for the total sample averaged .79 with a range of .52 to .89. Item scores for each child were then averaged across the two observers' sorts. Security and dependency composite scores were then derived by computing a Q-correlation between the averaged sorts for each child and a standard criterion sort for the "most secure" child (see Waters, 1987). Security scores for the entire sample averaged .33 ranging from -.38 to .66 (SD = .25).

Classroom Measures of Social Competence

Q-sorting measures and procedure.

Two Q-sorting item sets were used in order to assess classroom-based social competence: The 100-item California Child Q-set (Block & Block, 1980) and Bronson's revision of Baumrind's (1967) CQ-set (72-items). Two observers spent a minimum of 20 hours in each classroom, taking detailed notes of behaviors relevant to the Q-set items. Observers then assigned scores to each item in the Q-sample by sorting the items into nine categories that ranged from those "least descriptive" to those "most descriptive" of the target child. The distribution of items were rectangular such that the CCQ and CQ distributions were 11,11,11,11,12,11,11,11,11 and 8,8,8,8,8,8,8,8,8 respectively. The higher category placement (i.e., pile 1-9) indicated greater salience of the item for the subject.

Scores for each of the two sorts were computed for each child by calculating a Q-correlation between the observed sorts and a standard criterion sort of "social competence" derived from expert judges (Block & Block, 1980; Waters, Noyes, Vaughn & Ricks, 1985). Thus, each child received two scores for social competence: one derived from the 100 item Q-set and one from the 72 item Q-set. Analyses revealed a high degree of association between composite scores of social competence derived from the two sorts ([underlined r] = .59, [underlined p] < .001). The magnitude of this correlation was anticipated since both sorts are presumed to reflect the "social competence" construct.

Sociometric interviews.

Peer acceptance and liking was assessed by using a standard picture nominations procedure (Asher, Singleton, Tinsley, & Hymel, 1979; McCandless & Marshall, 1957) and a paired comparisons procedure (see Koch, 1933; Vaughn & Waters, 1981). For the nomination task, each participating child was shown photographs of all classmates and be asked to identify each child by name. After identifying all children correctly, the child was asked to name three classmates whom he/she especially "liked to play with" and three classmates whom he/she did not especially "like to play with." The photos of each preferred and non-preferred playmate were turned faced down and recorded.

For the paired comparisons procedure, each child was presented with the array of photos for familiarization. In this task, each child made a positive judgment (i.e., "who would you prefer to play with") for all possible pairs of classmates presented. Each pair of children (i.e., classmates) were presented one at a time with photos being placed an equal number of times on the right and left hand sides of the stimulus cards. The judgment from each pair of photos was recorded for all possible pairs of photos. Studies using this procedure have suggested that paired comparison scores are very stable over periods as long as 5 months (Vaughn & Waters, 1981) and, in the present sample, average paired comparison scores were significantly correlated with average positive nomination scores ([underlined r] = .40, [underlined p] < .01).

Visual regard.

All children from participating classrooms were assessed with respect to the amount and rate of visual attention received from peers in the classroom. To evaluate visual regard, a target child was observed for a 6-sec. interval and all recipients of visual attention were recorded. Two to four observers worked independently watching the children during daily classroom routines for a total of 180-210 observations per child. Indices reflecting visual regard for a given classroom were derived from the complete matrix of looks and glances received across all children in the classroom. Two composite scores were created for each child: (1) the total number of looks received from classmates and (2) the rate of visual attention received (calculated by dividing the total visual regard score by the number of rounds of observations actually present). Reliability was assessed by determining whether the rate of visual attention received by a particular child during one set of observations was correlated with the rate of visual attention received during a second set of observations. Correlations ranged from .20 to .86 with a median of .70.

Data Analysis and Reduction

Data analyses proceeded in two waves beginning with a molar presentation of the relations between security and competence and then turning to a more molecular analysis of specific item correlates. In so doing, correlations were first computed between AQS security score and the composite social competence score derived from a principal components analysis of the classroom assessments. Next, we calculated correlations between the AQS security score and the three sets of factors making up the "social competence" score. This was carried out in order to examine the relative strength of the correlations as well as differential relations between security and the different assessments of social competence. To reduce redundancy in the presentation of outcome measures, a composite of the two Q-sorts (i.e., CCQ and CQ) was computed by first making r to z transformations and then summing the two scores. This resulted in a single Q-sort composite score reflecting social competence (QCOMP). Likewise, a composite was created for the two sociometric acceptance scores (paired comparison and three-choice nominations). Similarly, a composite score derived from the standardized values of the visual regard total received and rate received scores was calculated. These three composites serve as indicators for the "first-order" factors revealed in Bost's (1995) analysis of social competence as a hierarchically organized dimension.

In the second wave of data analysis, item correlates of security and a composite score of social competence were computed in order to determine salient behaviors relevant to secure base and competent behavior for these preschool children. The index of social competence was derived by computing a principal components factor analysis using indices from sociometric (i.e., average paired comparison and nomination scores), Q-sort (CQ and CCQ social competence scores) and visual regard (total looks received and average rate of visual attention received) measures. Values for the first (unrotated) factor were used as a single score reflecting social competence for each child (for a detailed discussion of the factor structure of social competence as a higher-order construct and derivation of factor scores, see Vaughn, Bost, Newell, Cielinski, & Bradbard, 1993).


Attachment Security as a Support for Social Competence with Peers

The correlation between children's AQS security score and the Social Competence composite score derived from the principal components analysis (Bost, 1995) was positive and significant ([underlined r] (68) = .33, [underlined p] < .01). Subsequent analyses revealed that each of the social competence indices (i.e., Q-sorts, visual regard, sociometric assessments) was positively correlated with AQS attachment security. Correlations between attachment security and Q-sort descriptions of social competence ([underlined r] = .37, [underlined p] < .01) and visual attention received from peers ([underlined r]= .25, [underlined p] < .05) reached the conventional level of significance. The association between security and a composite created by averaging the standard scores for the paired-comparisons and the nominations sociometric assessments did not reach significance ([underlined r] = .18, ns). These findings are consistent with expectations based on previous studies (e.g., Arend et al., 1979; Wheeler & Seifer) in that secure-base behavior with mother predicted behavioral assessments of competence with peers. It is interesting to note here that the weakest association obtained was between security and sociometric acceptance because sociometric measures are very frequently used as proxies for the broader "social competence" construct. We return to the possible implications of this finding in the discussion.

AQS Item Correlates of Security and Social Competence

The associations between the organization of secure-base behavior with the mother as summarized by the AQS score for security and peer social competence indicators support the hypothesis that the attachment relationship is a key element in interpersonal adjustment. However, the details of these relations remain obscure. That is to say, it is not clear from the correlations of summary scores whether or not specific aspects or features of secure-base behavior account for the obtained relations. In this section, we present the AQS item correlates of the AQS security score and of the social competence composite derived from the principal components analysis of the classroom assessments reported by Bost (1995). These analyses identify those elements of secure-base behavior that are critical to both attachment security and to peer competence (see Table 1). Also presented in Table 1 are items found to be significant correlates of peer social competence that were not also correlates of attachment security.

Table 1: Item Correlates of Security and Composite Social Competence Scores
AQS ITEM NUMBER AND DESCRIPTION Attachment Security Social Competence
21. Child keeps track of mother's location around the house .74*** .35**
36. Child clearly shows a pattern of using mother as secure base .70*** .50***
62. When child is in a happy mood, he/she likely to stay that way .65*** .20
83. When child is bored, he/she goes to mother for something to do .63*** .13
 9. Child is lighthearted and playful most of the time .62*** .25*
87. If mother laughs or approves, child repeats it again and again .58*** .35**
11. Child often hugs or cuddles against mother without her inviting .57*** .10
66. Child easily grows fond of adults who visit and are friendly .57*** .24*
44. Child asks for and enjoys having mother hold, hug, and cuddle .57*** .12
28. Child enjoys relaxing in mothers lap .54*** .18
78. Child enjoys being hugged and held by people other than parent .53*** .25*
90. If mother moves very far, child follows along and plays in area .52*** .21
 7. Child laughs and smiles easily with a lot of different people .52*** .21
89. Child's facial expressions are clear and strong when playing .51*** .24*
15. Child is willing to talk to new people show toys if mother asks .49*** .34**
14. When child finds something new to play with shows to mom .49*** .35**
18. Child follows mother's suggestions readily .48*** .18
70. Child quickly greets mom with a big smile when she enters room .47*** .16
43. Child stays closer to mother or returns more often than keeping track .46*** .22
 1. Child readily shares with mom or lets her hold things .44*** .20
67. When family has visitors, child wants them to pay attention to him .41*** .26*
19. When mother tells child to bring something, he or she obeys .37*** .05
32. When mother says "no" or punishes, child stops misbehaving .36** .21
64. Child enjoys climbing all over mother when they play .31** -.05
12. Child quickly gets used to people or things that made him/her shy .31** .15
48. Child readily lets new adults hold things, if they ask .31** .18
 5. Child is more interested in people than in things .26* .04
27. Child laughs when mother teases him/her .23 * .04
75. At home, child gets upset or cries when mother walks out of room -.23* .13
82. Child spends most of play time with a few favorite toys -.23* -.23*
35. Child is independent of mother, prefers to play on own -.23* -.02
26. Child cries when mother leaves him/her at home with babysitter -.26* -.13
20. Child ignores most bumps, falls or startles -.33** -.04
88. When something upsets the child, he stays where he is and cries -.42*** -.22
76. When given a choice, child would rather play with toys than adults -.43*** -.11
69. Rarely asks mother for help -.43*** -.42***
50. Child's initial reaction when adults visit the home is to ignore them -.44*** -.20
29. At times, child attends so deeply to something doesn't hear what -.50*** -.32**
52. Child has trouble handling small objects or putting them together -.51*** -.43***
61. Plays roughly with mother. Bumps, scratches, or bites during play -.52*** -.03
39. Child is often serious or businesslike when playing away from mom -.53*** -.28*
30. Child easily becomes angry with toys -.54*** -.10
17. Child quickly loses interest in new adults if they annoy him/her -.55*** -.15
58. Child largely ignores adults who visit the home -.55*** -.31**
65. Child is easily upset when mother makes him change activities -.56*** -.28*
56. Child becomes shy or loses interest when activity is difficult -.56*** -.37**
79. Child easily becomes angry at mother -.61*** -.09

As shown in Table 1, 55 AQS items were found to be significant correlates of security (28 signed positively and 27 signed negatively) and 21 of these items were found to be significant correlates (and in the same direction) of peer social competence as well. As should be expected, AQS correlates of security cover the full range of content represented in the AQS, especially items that refer to the organization of secure-base behavior and items describing affective aspects of mother-child interaction (e.g., child keeps track of mother's location around the house; child clearly shows a pattern of using mother as secure base; if mother laughs or approves, child repeats activity over and over again; if mother moves very far, child follows along and plays in area where she has moved; child asks for and enjoys having mother hold, hug, and cuddle; child follows mother's suggestions readily; child quickly greets mom with big smile when she enters the room, etc.). Other item-correlates of security are suggestive of a pervasive positive mood and a positive orientation to non-family adults (the observers). Negative correlates of security included items indicative of negative mood, especially in the context of interactions with mother (e.g., when child returns to mother after playing, is sometimes fussy; child is demanding and impatient with mother) and extreme independence from mother (e.g., child is easy to lose track of when playing out of sight; when child finishes with activity, finds something to do on own, without returning to mother).

The AQS item-correlates of social competence with peers that overlap the correlates of security emphasize secure-base behavior, positive mood, and orientation to new adults (e.g., child shows pattern of using mother as secure base; child is lighthearted and playful most of the time; child enjoys being hugged and held by people other than parent; when family has visitors, child wants their attention). Negative item-correlates emphasized competence and persistence descriptors as well (e.g., child has trouble handling small objects; child becomes shy or loses interest if activity becomes difficult). Five additional item-correlates of social competence were not shared with security. These items seem to emphasize sociability and dependence (e.g., child wants to be center of attention; if visitors laugh or approve, child repeats activities again and again; even before trying things himself, child tries to get someone to help).

As might be expected, the AQS item-correlates of security tended to be those items salient to the criterion definition of the security construct (Waters et al., 1995) (i.e., placements in the criterion sorts above 6 and below 4). Although not a specific focus of this project, it is of interest to identify items salient in the criterion definition of attachment security that were not correlates of security in this sample. These are presented in Table 2.

[ Table 2 ]

For the most part, items in this group have mean placements near the center of the distribution for this sample (e.g., if held in mother's arms, child stops crying; child uses mother's facial expressions as cue; when upset about mother's leaving, sits and cries--does not attempt to follow), with narrow ranges of placements. This indicates that behavioral transactions relevant to these items were not usually observed in the sample; very likely due to the relatively advanced ages of the children. The single exception to this generalization refers to the child's ability to understand requests and statements by the mother. This was universally characteristic for these children and the item was placed relatively highly (mean = 6.75) for all children. Thus, while the item contributed to the level of security for all children in the sample, its placement could not contribute to individual differences among children in the sample. We return to these issues below in the discussion.


The purposes of this report were to (1) examine the relations between preschool children's attachment security as indexed by the AQS and concurrent assessments of their social competencies among peers and (2) identify and discuss specific aspects of behaviors relevant to the preschooler's organization of secure-base behavior and his/her peer competence at the item level of analysis. These goals were carried out in a sample of low-income, African-American children attending Head Start and by assessing multiple dimensions of the social competence construct.

Global associations between attachment security and social competence indicated that the AQS security score was positively and significantly correlated with weighted, higher-order factor scores for social competence derived from the principal components analysis involving the three measurement domains (i.e., sociometrics, Q-sorts, visual regard). Furthermore, correlations between security and the three subfactors of social competence revealed that those children characterized as more secure received higher social competence Q-sort scores and were more frequently targets for peers' visual attention than were their less secure counterparts. Taken together, these findings add to a growing body of literature relating AQS descriptions of security to social competence for preschool-age children (e.g., Wheeler & Seifer), and extend current attachment research by documenting these associations in a low-income minority sample. Consistent with Bowlby's (1980, 1988) arguments regarding the ongoing co-construction of mental models and the broad-band implications of attachment for personality and social development (Waters & Sroufe, 1983), the findings also provide empirical evidence linking current aspects of preschool children's organization of secure-base behavior to their interpersonal functioning and adaptation with peers outside the familial domain.

Although the AQS security score was found to be a significant correlate of Q-sort descriptions of social competence and visual attention received from peers, the association between security and sociometric scores did not reach significance. This finding is of special importance since developmental and clinical researchers have frequently employed of a range of related but not isomorphic measures to study social competence (e.g., competence-relevant behavior observations, sociometric assessments, ratings on attributes and behavioral categories by adults and peers, etc.), and since children's social competence is oftentimes characterized using sociometric peer liking and acceptance measures alone (Parker & Asher, 1987). Our results suggest that peer acceptance measures for preschool children may not overlap significantly with those aspects of social competence indexed by the Q-sorts and visual regard measures that are related to attachment security. Indeed, data from our larger project indicated that sociometric measures were less systematically related to external variables, even when numerous variables (in addition to attachment security) reflecting aspects of these children's social ecologies were examined (e.g., Bost, Washington, & Vaughn, 1995). More systematic and multi-method research is needed to determine whether these differential relations represent important demarcations in the influences of attachment on socioemotional functioning (see Arend et al., 1979; Wheeler & Seifer), or whether the use of sociometric measures with 3 and 4 year-old children to index social competence should be called into question for low income, African-American samples.

In addition to examining global relations between security and measures of social competence, our goal was to also conduct a more micro analysis of these relations by examining AQS item-correlates of security and social competence. Results of these analyses were encouraging in that they yielded coherent relations between a wide-range of AQS items and children's security and competence scores with respect to each item's relative salience in the security criterion sort (see Waters et al., in press). Specifically, 89% of the significant item correlates of security and 69% of social competence were shown to be consistent with AQS item placements for security. Moreover, 21 items were found to support both children's organization of secure-base behavior and their social adaptation in the peer context. Thus, not only did expected associations emerge between the full range of content of AQS items and security scores, many of these items were also found to be significant correlates of social competence as well. These findings, in turn, highlight the continued importance of the attachment relationship as a contributor to interpersonal functioning beyond infancy (Ainsworth, 1990; Cicchetti et al., 1990).

This finer level of analysis also allowed for a more detailed examination of specific behaviors that were found to support children's prosocial interactions with peers but not security as well as of those behaviors that were not frequently observed in this age-group but presumed to reflect (or not reflect) secure base behavior. In regard to the former, five items were shown to be significant correlates of social competence but not security . When the content of these items was examined, the significant associations with social competence were interpretable when the children's ages are taken into consideration. In fact, the two significant positive correlates of social competence tended to be behaviors characterizing both social interactions directed toward observers (e.g., "when visitors laugh or approve of something, child repeats again and again") and attempts by the child to initiate conversation with the caregiver (i.e., "child wants to be center of mother's attention"). In contrast, the significant negative correlates of social competence were behaviors that appeared to direct the child's attention away from the caregiver (i.e., "child is strongly attracted to new activities and toys") or which suggest dependent behavior for preschool children (i.e., "even before trying things him/herself, child tries to get someone to help"). Given this information, it becomes more easily understood as to why these particular behaviors were shown to be associated with children's socially skilled behaviors with peers.

In addition to the above findings, interesting results emerged when we identified those items that proved not to be significant correlates of security but were either especially high or low in their placement in the security criterion sort. Upon examination of the means and distributions of these items, it became apparent that this group of items was placed toward the center of the sort and had narrow ranges indicating that the behaviors were infrequently seen in this sample of children. A single exception was the item "when mother asks child to do something, he/she readily understands." This particular item was found to be placed relatively high for all children and may reflect preschoolers' advancements in cognitive and linguistic abilities. Nevertheless, the restricted variance was shown to hinder its contribution to individual differences in security when the sample as a whole was considered.

Of further interest were items such as "child will accept being bounced around in play" and "if held in mother's arms, child stops crying." Although these items are included in the Q-set as indicators of attachment security, it was rare that mothers in this sample were observed to pick up or hold their children. This may be because the home observations were conducted in the natural family context where younger siblings were often present during observation periods. As such, these types of behaviors were typically directed toward the youngest sibling rather than the older preschool child. Although this could be viewed as a potential confound in that the preferred use of the AQS is with the caregiver and target child alone, the use of the AQS in naturally occurring family situations and its demonstrated associations with children's peer competence in this study supports the flexibility and ecological validity of this measure.

Taken together, results of the item analyses have several important implications for the study of preschool attachment. First, these data suggest that AQS characterizations of preschool attachment security that have been evaluated primarily in non-minority populations are also structured in predictable and anticipated ways for minority preschool children. This highlights the comprehensive scope of the AQS, and adds to the systematic examination of the nature and content of parent-child relationships for young, minority children. Second, the findings emphasize specific behaviors that were not frequently observed in this sample of preschool children, but are especially salient or not salient in the definition of the security construct. Future research conducted with diverse populations will help to determine the extent to which these particular items are relevant for describing preschool attachment behaviors.

The discussion thus far has primarily focused on specific behaviors that were shown to be descriptive or not descriptive of the preschooler's organization of secure base behavior. This detailed examination was felt necessary considering the sample under study and important for identifying key attachment behaviors for this age-group of children. However, the overall coherence of the results at the item level of analysis and the more global relations found between security and social competence should not be de-emphasized or overlooked. Indeed, perhaps the most important implication of these findings is that even when considering the items shown to be irrelevant to secure base behavior in this sample, the comprehensive range of behaviors included in the AQS was sufficient to capture many behaviors along the security continuum for these preschool-age children. More importantly, variability in these security scores was shown to be related to individual differences in children's social competencies among peers. The data reported here can only add to the existing empirical evidence (e.g., Howes & Hamilton, 1992; Waters et al., 1990; Waters et al., in press) highlighting the value of the AQS in characterizing secure base phenomenon across a wide-range of contexts and age-groups.


Ainsworth, M. (1990). Some considerations regarding theory and assessment relevant to attachments beyond infancy. In M. Greenberg, D. Cicchetti, & E. M. Cummings (Eds.), Attachment during the preschool years: Theory, research, and intervention (pp. 463-489). Chicago: University of Chicago Press.

Ainsworth, M. D. S. (1967). Infancy in Uganda: Infant care and the growth of love. Baltimore: Johns Hopkins University Press.

Ainsworth, M. D. S. (1973). The development of infant-mother attachment. In. B. M. Caldwell & H. Ricciuti (Eds.), Review of Child Development Research (Vol. 3). pp. 1-94.

Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment: A psychological study of the strange situation. Hillsdale, NJ: Erlbaum.

Arend, R., Gove., F.L., & Sroufe, L.A. (1979). Continuity of individual adaptation from infancy to kindergarten: A predictive study of eco-resiliency and curiosity in preschoolers. Child Development, 50, 950-959.

Asher, S., Singleton, L.C., Tinsley, B. R., & Hymel, S. (1979). A Reliable Sociometric Measure for Preschool Children. Developmental Psychology, 15, 443-44.

Block, J. H., & Block, J. (1980). The role of ego-control and ego-resiliency in the organization of behavior. In W.A. Collins (Ed.) Development of Cognition, Affect, and Social Relations. Minnesota Symposium on Child Psychology, Vol. 13. (pp. 39-101) Hillsdale, NJ: Erlbaum.

Booth, C., Rose-Krasnor, L., MacKinnon, J.. & Rubin, K.H. (1994). Predicting social adjustment in middle childhood: The role of preschool attachment security and maternal style. Special Issue: From family to peer group: Relations between relationships systems. Social Development, 3, 189-204.

Bost, K.K., (1995). A Contextual Model of Preschool Children's Social Competence: A Call for the Integration of Social Network Perspectives and Attachment Theory. Unpublished Ph.D. dissertation, Auburn University.

Bost, K.K., Washington, W.N., & Vaughn, B.E. (March, 1995). Attachment and social networks as supports for peer social competence. In B.E. Vaughn, Chair, Social networks of children and youth: Outcomes and consequences. Symposium presentation at the Meetings of the Society for Research in Child Development, Indianapolis.

Bowlby, J. (1973). Attachment and loss. Vol. 2: Separation. New York: Basic.

Bowlby, J. (1980). Attachment and loss. Vol. 3: Loss, sadness, and depression. New York: Basic.

Bowlby, J. (1982). Attachment and loss. Vol. 1: Attachment (2nd Edition). New York: Basic.

Bowlby, J. (1988). A secure base: Clinical applications of attachment theory. London: Tavistock.

Cassidy, J. A., Marvin, R. S., with the MacArthur Working Group on Attachment (1992). Attachment organization in three-and four-year-olds: Procedures and coding manual. Seattle, WA.

Cicchetti, D., Cummings, M., Greenberg, M., & Marvin, R. (1990). An organizational perspective on attachment beyond infancy: Implications for theory, measurement, and research. In M. Greenberg, D. Cicchetti, & E. M. Cummings (Eds.), Attachment during the preschool years: Theory, research, and intervention (pp. 3-51). Chicago: University of Chicago Press.

Fox, N. A. (1995). Of the way we were: Adult memories about attachment experiences and their role in determining infant-parent relationships: A commentary on van Ijzendoorn. Psychological Bulletin, 117, 404-410.

Greenberg, M. T., Cicchetti, D., & Cummings, E. M. (1990). Attachment in the preschool years: Theory, research, and intervention. Chicago, University of Chicago Press.

Grossmann, K. E., & Grossmann, K. (1993). Legame d'attaccamento infantile e sviluppo delle dinamichepsichiche individual nel corso della vita. (The attachment relationship of children and the development of individual psychological dynamism in the course of life.) Terapia Familiare, 41, 5-18.

Howes, C., & Hamilton, C. E. (1992). Children's relationships with caregivers: Mothers and child care teachers. Child Development, 63, 859-866.

Kavesh, L. (1992). Antecedents of peer competence in childhood. Dissertation Abstracts International. 53 (3-B) 1623.

Kobak, R., Ferenz-Gillies, R.,. Everhart, E., & Seabrook, L. (1994). Maternal attachment strategies and emotion regulation with adolescent offspring. Special Issue: Affective processes in adolescence. Journal of Research on Adolescence, 4, 553-566.

LaFreniere, P. J., & Sroufe, L.A. (1985). Profiles of peer competence in the preschool: Interrelations between measures, influence of social ecology, and relation to attachment history. Developmental Psychology, 21, 55-69.

McCandless, B. R., & Marshall, H. R. (1957). A picture sociometric technique for preschool children and its relation to teacher judgements of friendship. Child Development, 28, 139-147.

Oppenheim, D. & Waters, H.S. (1995). Narrative Processes and Attachment Representations: Issues of Development and Assessment. In E. Waters, B. E. Vaughn, G. Posada, & K. Kondo-Ikemura (Eds.), Caregiving, cultural and cognitive perspectives on secure-base behavior and working models: New growing points of attachment theory and research. Monographs of the Society for Research in Child Development. (Serial No. 244).

Papini, D.R., & Roggman, L. (1993). Parental attachment to early adolescents and parents' emotional and marital adjustment: A longitudinal study. Journal of Early Adolescence, 13, 311-328.

Park, K. A., & Waters, E. (1989). Security of attachment and preschool friendships. Child Development, 60, 1076-1081.

Parker, J., & Asher, S. (1987). Peer relations and later personal adjustment: Are low-accepted children at risk? Psychological Bulletin, 102, 357-389.

Posada, G., Gao, Y., et al. (1995). The Secure Base Phenomenon Across Cultures: Children's Behavior, Mothers' Preferences, and Experts' Concepts. In E. Waters, B.E. Vaughn, G. Posada, & K. Kondo-Ikemura (Eds.), Caregiving, cultural and cognitive perspectives on secure-base behavior and working models: New growing points of attachment theory and research. Monographs of the Society for Research in Child Development. (Serial No. 244).

Sroufe, L. A. (1986). Appraisal: Bowlby's contribution to psychoanalytic theory and developmental psychology--attachment, separation, loss. Special Issue: 30th anniversary of the Association for Child Psychology and Psychiatry and Allied Disciplines. Journal of Child Psychology & Psychiatry & Allied Disciplines, 27, 841-849

Suess, G.J., Grossmann, K. E., & Sroufe, L. A. (1992). Effects of infant attachment to mother and father on quality of adaptation in preschool: From dyadic to individual organisation of self. International Journal of Behavioral Development, 15, 43-65.

Urban, J., Carlson, E., Egeland, B., & Sroufe, L.A. (1991). Patterns of individual adaptation across childhood. Special Issue: Attachment and developmental psychopathology. Development & Psychopathology, 3, 445-460.

Vaughn, B.E. & Waters, E. (1990). Attachment behavior at home and in the laboratory: Q-sort observations and Strange Situation classifications of one-year-olds. Child Development, 61, 1865-1973.

Vaughn, B.E., & Bradbard, M.R. (1992). Relations between social competence, social behavior and social support networks for children attending Head Start. Proposal submitted to the Administration for Children and Families, Head Start Bureau.

Vaughn, B.E., & Martino, D. (1988). Age related Q-sort correlates of visual regard in groups of preschool children. Developmental Psychology, 24, 589-594

Vaughn, B.E., Bost, K.K., Newell, W.H., Cielinski, K.L., & Bradbard, M. (November, 1993). Relations Between Social Competence, Social Behavior and Social Support Networks for Children Attending Head Start. Poster presented at the Second Head Start Research Conference. Washington, DC.

Waters, E. & Deane, K. E. (1985). Defining and assessing individual differences in attachment relationships: Q-methodology and the organization of behavior in infancy and early childhood. In I. Bretherton & E. Waters (Eds.), Growing points in attachment theory and research (pp. 41-65), Monographs of the Society for Research in Child Development, 50 (1-2, Serial No. 209).

Waters, E. (1987). Attachment behavior Q-set (Revision 3.0). Unpublished manuscript, SUNY, Department of Psychology, Stony Brook.

Waters, E., Garber, J., Gornal, M. & Vaughn, B. (1983). Q-sort correlates of visual regard among, preschool peers: Validation of a behavioral index of social competence. Developmental Psychology, 19, 550-560.

Waters, E., Noyes, D.M., Vaughn, B.E. & Ricks, M. (1985). Q-sort definition of social competence and self-esteem: Discriminant validity of related construct in theory and data. Developmental Psychology, 21, 508-552.

Waters, E., Kondo-Ikemura, K., Posada, G. & Richters, J. E. (1990). Learning to love: Mechanisms and milestones. In M. R. Gunnar and L. A. Sroufe (Eds.) Self Processes and Development. The Minnesota Symposia on Child Psychology. Vol. 23, (pp. 217-255). Hillsdale, NJ: Erlbaum.

Youngblade, L., Park, K., & Belsky, J. (1993). Measurement of young children's close friendship: A comparison of two independent assessment systems and their associations with attachment security. International Journal of Behavioral Development, 16, 563-587.

* Notes

The research reported here was supported in part by Grant # CD90-0956 from the Administration for Children and Families, Head Start Bureau. This paper is based on data and analyses submitted as a part of a Ph.D. dissertation completed by Kelly K. Bost, Department of Family and Child Development, Auburn University. An earlier version of this paper was presented at the Biennial Meetings of the Society for Research in Child Development (1995, Indianapolis, IN). We would like to thank the families, teachers, and Head Start programs who participated in this project for their patience and cooperation during the many months of data collection for this project. We also thank Lisa Wakeen Caya, Kerry Cielinski, Lisa Krzysik, Margaret Machara, and Jennifer Paul for their contributions to this study.

Kelly K. Bost is presently at the University of Illinois at Urbana-Champaign, Department of Human and Community Development, 1105 W. Nevada, Urbana, IL 61801 . Requests for reprints should be directed to Kelly K. Bost or to Brian E. Vaughn, Department of Family and Child Development, 203 Spidle Hall, Auburn University, Auburn, AL 36849-5604.

Copyright © 1998 Kelly Bost, Brian E. Vaughn, and Carrol Heller.

stars icon For technical assistance:
Library Technology Group
General Library System
University of Wisconsin-Madison