PREDICTING THE SUCCESS OF FOSTER PLACEMENTS; TEMPERAMENT AND INTERACTION PATTERNS By JENNY L. DOELLING A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1989
To my parents and late grandparents, whose love gave me the courage to pursue my dreams.
ACKNOWLEDGMENTS First and foremost, I would like to express my sincere gratitude and deep respect for my chairperson, Dr. James H. Johnson, whose guidance and support were invaluable to me in this project. I would also like to thank Dr. Sheila Eyberg, Dr. Patricia Miller, Dr. Nancy Norvell and Dr. Connie Sheehan for their valuable suggestions and encouragement. I would further like to thank the foster care staff of the Florida Department of Health and Rehabilitative Services District Three, including Cheryl Dollar, Foster Care Analyst, Betty Baker, Beverly Dunn, Cathy Ferguson, Joann Fuller, Daisy Lynum, Patricia Morelli, Jan Powell and Ruth Ann Wilson, foster care supervisors, as well as the numerous foster care case workers for their vital assistance. Last but not least, I would like to thank the many generous foster parents and foster children who gave their time and energy to participate in the study. Ill
TABLE OF CONTENTS page ACKNOWLEDGMENTS iii ABSTRACT V INTRODUCTION 1 Problems with Foster Care 1 Evaluation of Foster Placements 5 An Interactive Approach 7 METHOD 14 Sub j ects 14 Measures 15 Procedure 28 RESULTS 31 Demographic Data 31 Relationships between the Dependent Measures 31 Demographic and Temperament Variables Related to Placement Success 35 The Goodness-of-Fit Model and Placement Success.... 38 DISCUSSION 50 APPENDIX A QUESTIONNAIRES 65 APPENDIX B RAW DATA 85 REFERENCES 95 BIOGRAPHICAL SKETCH 99 iv
Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PREDICTING THE SUCCESS OF FOSTER PLACEMENTS: TEMPERAMENT AND INTERACTION PATTERNS By Jenny L. Doelling August 1989 Chairperson: Dr. James H. Johnson Major Department: Clinical and Health Psychology The quality and stability of foster care placements has been shown to relate to the level of adjustment of foster children. Until now, most studies have examined primarily single variables relating to placement outcome. This study examined both parent and child variables, specifically temperament variables, in an interactive "goodness-of-f it" model. Goodness-of-f it was examined 1) in terms of combinations of parent and child temperament characteristics predictive of placement success and 2) in terms of congruence between parental expectations for child temperament and actual child temperament as predictive of placement success. Additionally, the study explored the nature of interaction patterns in foster placements. One demographic variable, number of children in the home, was related to placement success, with conflict increasing
with number of children. No single temperament variable was significantly related to placement success. Several temperament variables were significantly associated with observed parent-child interaction patterns. Children low on task orientation received more commands and children with negative mood received more criticism from their foster mothers. Mothers low on approach and mothers with negative mood were more critical of their foster children. Mothers high on approach and mothers with positive mood elicited more noncompliance from their foster children. In terms of the goodness-of-fit model, the combination of a mother low on approach with a child low on task orientation predicted more conflict in the home. No other combination of parent-child temperament variables was directly predictive of placement outcome. Several combinations of temperament variables were related to interaction patterns. Child's age and task orientation together predicted number of commands. The interaction effect of an inflexible mother and a negative mood child predicted maternal criticism. Child's age and the interaction effect of mother approach-withdrawal with child mood and child's age and the interaction effect of mother mood with child approach-withdrawal both predicted child noncompliance. Regarding expected versus actual child temperament, placements in which the child was more active than the foster mother expected were judged less successful. For all other vi
temperament dimensions, actual child temperament and expected child temperament were highly correlated. Vll
INTRODUCTION Problems with Foster Care Foster care is widely used to provide both short-term and long-term care for children for a variety of reasons. Foster care may be used as an interim arrangement for children who hope to be adopted, may be used for abused children to provide a safe environment until a decision is made about the best permanant placement for them, or may involve a series of placements for difficult-to-adopt children such as minority and handicapped children. While foster care is preferable to an abusive or neglectful home or to institutional care, it is not without its problems. Fanshel and Shinn (1978b) conclude that the foster care literature strongly suggests that children separated from their parents by long-term foster care are at high risk for developing cognitive and personality impairments. Studies in the area have indicated that the degree of trauma experienced by the child in foster care is related to qualities of the foster parents (Cautley and Aldridge, 1975; Fanshel and Shinn, 1978b; Martin and Beezley, 1976; Murray, 1984; Rowe, 1976; Stone and Stone, 1983; Wiehe, 1982), certain child characteristics (Cautley and Aldridge, 1975; Fanshel and
2 Shinn, 1978a; Murray, 1984; Rowe, 1976; Stone and Stone, 1983) , and the number of prior placements experienced by the child (Fanshel and Shinn, 1978a; Murray, 1984; Pardeck, 1983; Stone and Stone, 1983). Abuse and neglect by foster parents does occur, and less than adequate care in foster homes may be common (Martin and Beezley, 1976) . Many researchers have found that the number of replacements a child experiences relates to the emotional stability of the child. Maas and Engler (1959) found that emotional disturbance is associated more with number of replacements than total time in care, and that existing emotional disturbance is exacerbated by multiple replacements. Eisenberg (1962) also found that children experiencing multiple placements displayed more aggressive behavior, were poorly socialized and highly mistrustful. Although children who are already disturbed experience a disproportionately high rate of replacement, their symptoms also become worse with time in care (Fanshel and Shinn, 1978a; Maas and Engler, 1959) . The admission rate for foster children to psychiatric clinics was found by Eisenberg (1962) to be ten times the rate of the general population, and children entering foster care without significant behavior problems were found to show increasing behavior problems with time in care (Fanshel and Shinn, 1978a) . It becomes evident that taking care to place children in foster homes that can accommodate their needs will result in both better care and greater stability for them.
3 Fanshel and Shinn (1978b), who conducted a five-year longitudinal study of 624 children in foster care, state: "Most child welfare experts would agree that children who have already suffered separation from their own parents should be spared unnecessary transfer while living under substitute living arrangements, as such children have a strong need for continuity of relationships with significant others, especially parents" (p. 115) . The Fanshel and Shinn (1978b) study indicates that 58.2% of their sample experienced two or more placements while Eisenberg (1962) reports 50% of his sample to have experienced four or more placements. An unplanned replacement or "breakdown" may be the most traumatic situation for a foster child, yet it also seems to be the least studied or discussed issue in the foster care literature (Stone and Stone, 1983) . Planned removals may occur when a child is returned to his or her natural parents following some remediation in the home situation or when a child is adopted by parents other than the foster parents. Unplanned removals, on the other hand, occur when either the foster parents find themselves unable to continue caring for a child because they find it too difficult, or when the foster care case worker decides the placement is detrimental to the child. Either the foster parents or the case worker may decide to terminate a placement prematurely.
4 Martin and Beezley (1976) estimated that 25% of foster children are shifted from foster home to foster home. However, in one sample where incidence of breakdown was actually calculated, it was found that nearly half of the sample were involved in at least one unplanned replacement (Stone and Stone, 1983) . Interestingly, they also found that the majority of these breakdowns occurred within four weeks after placement. In light of the potential damage repeated placement moves are likely to have on a child and given such a high rate of placement breakdown, it becomes evident that more care in matching children to foster homes should be exercised. Aside from breakdowns, some children may also remain in foster homes for long periods when the placement is not entirely satisfactory or is not adequately meeting the child's needs. There may also be cases in which a child is removed from a home in which he or she nonetheless derived something of value or received better care than in some other provision (George, 1970) . The first situation is likely to occur when there is such a shortage of foster homes that case workers are reluctant to remove a child from a less than satisfactory placement due to the unavailability of alternative homes. The second situation may occur when foster parents are quick to have a difficult child removed from their home despite the fact that they may have been providing excellent care in the face of a stressful situation.
5 Evaluation of Foster Placements It seems that there is a need to judge the success of placements by some criterion besides merely breakdown. Kraus (1971) notes that there is a tendency to judge foster placements subjectively and calls for the development of objective criteria. George (1970) suggests degree of adjustment of the child to be a useful criterion. Some indicators judged by foster care workers to be important in assessing quality of care in the foster home are the ability to deal with child care problems, the amount of affection shown the child, having realistic expectations of children, the ability to handle contact with the child's natural parents (George, 1970; Wolins, 1963) , adequate physical care, good relationships with other children in the home, and the amount of time devoted to the child (Wolins, 1963) . These variables have been compiled into a rating scale for use by foster care workers as a means of formally assessing foster placements. This scale was developed and used in a previous study of foster placements (Doelling, 1987) . Several studies have examined variables related to characteristics of foster parents and characteristics of child behavior thought to be predictive of success of placement. Rowe (1976) found that foster parents who are tolerant of child behavior that opposes their values or wishes such as poor academic performance, early adult behavior (i.e., smoking, drinking, sexual activity) , difficult social behavior
6 from the latency age child or lack of strict religious observance are more successful as foster parents. Similarly, Cautley and Aldridge (1975) found that foster parents more skilled in handling common problem behaviors without harshness or excessive discipline and with an understanding of the reasons for certain behaviors were more successful. Additionally, they found that the ability to handle defiant and withdrawn behavior was predictive of success. Wiehe (1982) examined the differences between personality types of foster mothers and foster fathers in an attempt to explain the lesser degree of involvement found in foster fathers. It was found that foster fathers were more often described as "realistic" types who prefer ordered, systematic activities, whereas foster mothers were more often described as "social" types who prefer activities that inform, train, develop or cure others. As degree of parental involvement with the foster child is seen by foster care workers as one index of quality of care, it may be that the more "flexible" mothers may better fulfill the role of foster parents than more "rigid" foster fathers. Ordering versus informing parenting styles seem to bear some relation to the concepts of tolerance (Rowe, 1976) and understanding (Cautley and Aldridge, 1975) discussed earlier and found to be predictive of success of placement . Characteristics of foster children have also been related to success of placement. Fanshel and Shinn (1978a) found that
7 children whose behavior was characterized as defiant and hostile tended to experience more replacements. Stone and Stone (1983) found that less aggressive, better socialized children were more likely to remain in their assigned placement. An Interactive Approach Most studies have examined foster parent and foster child variables independently. A case can be made, however, for assessing how the characteristics of foster parents and foster children interact. Indeed, the Cautley and Aldridge study (1975) found that no single foster parentor foster childrelated characteristic was predictive of success of placement. It seems to be more informative to examine these variables as they interact with each other. A parent-child interactive approach relevant to the issue of foster placement has been taken by Thomas, Chess and Birch (1968) and Thomas and Chess (1977) in their description of the concept of temperament. Temperament is generally defined as the style of behavior rather than the content (Plomin, 1983) . Individual differences in temperament styles have been observed in children as young as two to three months and are presumed to have some biological basis (Thomas et al., 1968). Thomas et al. have conceived of childhood temperament along an easy-to-difficult continuum. Difficult children are described as having irregular biological functioning (eating, sleeping, toileting) , withdrawing from and adapting slowly to
8 environmental changes, having high-intensity emotional responses, and showing more negative than positive mood. More recent conceptualizations have included high activity level, rigid responding and distractability (Lerner, Palermo, Spiro and Nesselroade, 1982) . Easy children have the opposite pattern on each of these dimensions. It has been shown that difficult children pose more parenting problems and are more at-risk for developing later pathology, particularly when there is a mismatch between child temperament and the demands of the child's environment, including parental behavior (Thomas and Chess, 1977) . Parents with different parenting styles may react differently to children of each temperament type (Thomas et al., 1968). For example, some parents of a difficult child may engage in repeated conflict, while others may be more tolerant. Some parents may ignore difficult behavior, whereas others may be oversol icitous . Thomas and Chess (1977) stress the importance of the interaction between the temperament of the child and the parental response. They found that easy children reassure parents that they are adequate, healthy and loving parents, while difficult children can make parents feel inadequate, unloving and inept. Thomas and Chess (1981) also found that parental pressure on a difficult child may actually enhance the temperament traits leading to the negative reaction which only results in greater parental disapproval and more
9 conflict. Thus, an intolerant parent may, in essence, cause his or her child to become even more difficult. Conversely, the easy child will elicit a positive parental response which reinforces and perpetuates the child's "easy" behavior. Graham, Rutter and George (1973) found that particular temperament characteristics render children more vulnerable to adverse family conditions (i.e., a mentally ill parent). They also found that certain temperament characteristics are harder for parents to deal with and may result in more conflict between parent and child. Children with difficult temperament have been found to be at greater risk for developing behavior disorders (Thomas and Chess, 1981) . However, it has also been suggested that the parental response to child temperament mediates between the temperament and the possible development of behavior problems (Cameron, 1977) . The idea is that certain parental behaviors may put strain on the child's style of responding and actually cause the child to become more difficult (Cameron, 1977) . Cameron (1977) found children with difficult temperament at age one who had parents who reacted with disapproval, intolerance and rejection to have become even more difficult by age five. He, too, emphasizes the importance of the child's temperament as it interacts with parental response. The "goodness-of-f it" concept proposed by Thomas and Chess (1977) refers to the goodness of fit between the demands
10 of the environment and the individual's style of behaving. This goodness-of-fit model has been applied to a variety of contexts. Lerner, Baker and Lerner (1985) review the literature and report that the match between person and context has implications for a person's overall performance, satisfaction and adjustment and for a person's ability to cope with stress. Temperament matches have been shown to influence classroom adjustment and ability, peer relations, behavior problems, parent-child conflict, perceived competence and commitment in dating relationships. Overall, the reviewed research suggests that a person's not meeting the demands of his or her context (having a poor fit) is associated with adverse individual and interpersonal functioning (Lerner, Baker and Lerner, 1985) . While a good fit seems important for all parents and children, it may have particular implications for foster parents and foster children. For example, a difficult foster child placed with an intolerant foster parent might result in conflict within the family, an actual intensification of the child's characteristics, possible behavior problems, and the foster parents' feeling inept and unsuccessful. All of these outcomes could, in turn, influence both a case worker's evaluation of the placement as well as the parents' decision of whether to keep a particular child or even of whether to remain in the foster care system at all. The goodness of fit
11 between foster parent and foster child temperament, then, can be viewed as being related to success of placement. In a study of risk factors in abusive behavior, Johnson, Floyd and Isleib (1986) found that a "mismatch" between parent and child temperament was highly predictive of abusive and neglectful behavior. Children were characterized as easy or difficult and parents were characterized as adaptive or unadaptive (i.e., intolerant, inflexible and rigid). Here it was found that the mismatch of a difficult child paired with an unadaptive parent was present in almost half of the abusive and neglectful families and was not present in any of the nonabusive families. While difficult children and unadaptive parents were more commonly found in abusive and neglectful families, they were present in nonabusive families as well. It was when these two variables were considered together that they were most predictive of abuse and neglect. A previous study (Doelling, 1987) showed a similar relationship to be operating within foster placements. While single temperament variables predicted very little in terms of placement outcome, certain combinations of foster parent and foster child variables were found to be predictive of placement success. In particular, the mismatch of an inflexible mother paired with a child high on the negative mood dimension of temperament was found to be predictive of poorer placement outcome. The inflexible mother/negative mood child mismatch was predictive of poorer placement outcome in
12 terms of there being more conflict in the home, the foster mother being less satisfied with the placement and the case workers rating the placements as less satisfactory. Placements in which children had a more negative mood than expected by mothers also were less successful as rated by case workers. Additionally, the combination of an inflexible mother paired with a low-rhythmic child was found to be predictive of poorer outcome as assessed by ratings of mothers' satisfaction with the placement. The flexibility-rigidity temperament dimension seems to bear a close resemblance to the qualities of tolerance (Rowe, 1976) , harshness of discipline (Cautley and Aldridge, 1975) and realistic versus social personality types (Wiehe, 1982) discussed by previous authors as related to placement success. Likewise children with negative mood may be seen as displaying certain of the child characteristics shown to relate to success of placement such as hostility, defiance and aggressiveness (Fanshel and Shinn, 1978a; Stone and Stone, 1983) . However, the results of the study clearly suggested that it is the interaction between these two that is crucial. While Thomas and Chess (1977) have considered goodness-of-fit as it relates to the development of later child psychopathology, it seems that a similar concept can also be applied to outcomes of child abuse and to failure of foster placements.
13 The mechanism through which temperament mismatches exert their influence is not clear. However, the finding of greater conflict in "mismatched" placements suggests that the problem lies, at least in part, in negative interaction patterns. Direct observation of the interaction patterns between foster parents and foster children would perhaps help reveal through what channels a mismatch exerts its influence. While the initial study (Doelling, 1987) suggests that more conflict occurs in mismatched placements, it would be even more informative to define more precisely the nature of these interactions. The present study proposed first of all to replicate the previous findings documenting poorer outcome in mismatched placements. A second purpose was to explore the nature of the interaction patterns in matched and mismatched placements and the relationship of specific parent-child behaviors to placement outcome.
METHOD Subjects Subjects included 44 foster children between the ages of 2 and 10 and their foster mothers. Subjects were recruited through two local district offices of the Florida Department of Health and Rehabilitative Services (HRS) . Foster families from sixteen north Florida counties were included in the study. Names of foster children in the designated age range and their foster parents were provided by the foster care supervisor in each county after receiving approval from the district offices of HRS. Only one child per foster home was included. In homes with multiple foster children, the child who had been placed most recently was selected. Children with major impairments such as mental retardation and physical handicaps were excluded from the study as it was felt that these children bring a unique set of problems to their placements which might override temperament factors. Children intentionally placed for a short period as an interim placement were also excluded. Placements of longer than one year were excluded as such placements were felt to represent relative successes in and of themselves. 14
15 Measures Demographic Questionnaire A demographic questionnaire was completed by each foster mother. Information sought included child's age, parents' ages, length of the placement, number of natural, adopted and other foster children in the home, number of siblings of the target child in the home, income, educational level of parents, number of hours spent with each parent, number of parents in the home and which parent was the primary caretaker. The questionnaire is presented in Appendix A. Dimensions of Temperament Survey-Revised The Dimensions of Temperament Survey (DOTS) is a measure developed by Lerner, et al. (1982) and revised (DOTS-R) by Windle and Lerner (1986) . Both versions were designed to identify features of temperament continuous in the behavioral repertoire from early childhood through adulthood. The original DOTS was developed from a pool of over 400 items representing the nine categories of temperament discussed by Thomas and Chess (1977, 1981). It was tested on a sample of 1386 subjects from each of three age groups (early childhood, middle to late childhood, and early adulthood) and was factor analyzed revealing five factors labelled Activity Level, Attention Span/Distractability, Rhythmicity, Reactivity and Adaptability/Approach-Withdrawal. Data on the original DOTS suggest internal consistency coefficients (Cronbach's alphas) ranging from .31 to .96, and stability coefficients (test-
16 retest reliability) ranging from .60 to .93 (Lerner et al., 1982) . The revised version was developed to overcome several of the limitations of the original version. The response format was expanded from two choices (true or false) to four choices (usually true, more true than false, more false than true, usually false) to permit greater latitude and finer distinctions by the respondent. Some items from the original 34-item version were retained or modified, and some new items were generated. Items tapping activity level were expanded from just assessing sleep behavior to include items measuring general activity level as well. Expert raters initially categorized 117 items representative of the nine temperament categories proposed by Thomas and Chess (1977) in the New York Longitudinal Study. An overall aggreement rate of 97% was reached for 106 of the 117 items. Then 638 subjects from the three age groups responded to these 106 items. Items with low item-to-total subscale correlations or high item-to-other subscale correlations were deleted. This was followed by factor analyses in which items showing random-loading or nonmeaningful item-factor relations were deleted, leaving the 54 items making up the DOTS-R. Factor analyses of these remaining items revealed a ninefactor model for the child versions and a ten-factor model for the adult version. Eight of the attributes are identical for both versions including Activity Level-General, Activity
17 Level-Sleep, Approach-Withdrawal, Flexibility-Rigidity, Mood, Rhythmicity-Sleep, Rhythmicity-Eating and Rhythmicity-Daily Habits. The ninth attribute on the child versions is Task Orientation which differentiates into two attributes in the adult version: Distractability and Persistence. Internal consistency coefficients (Cronbach's alpha) for these dimensions range from .54 to .91. In an initial study using the DOTS-R (Windle, Hooker, Lernerz , East, Lerner and Lerner, 1986) , significant relationships were found between DOTS-R temperament dimensions and perceived efficacy of psychosocial functioning in samples of early and late adolescents. These correlations were greater than those found with the DOTS and other temperament measures, providing some initial support for the predictive usefulness of the revised version. While additional research with the DOTS-R is needed, the available data support its use as a research measure for studies of child and adult temperament. There are three versions of the DOTS-R. The DOTS-R CHILD is for use with children preschool through middle elementary school and is designed to be completed by a parent or other caregiver. The DOTS-R CHILD (Self) is for use with children from late elementary school through high school and is designed to be completed by the child himself or herself. The DOTS-R ADULT is for adults and is completed by the subject himself or herself. Only the DOTS-R CHILD and DOTS-R ADULT versions were used in the present study. Both versions contain
18 the same 54 items with only minor variations in instructions and changes in pronouns and verbs depending on who is completing the survey. The DOTS-R CHILD is composed of nine attributes while the DOTS-R ADULT is composed of ten attributes as described above. The DOTS-R ADULT was completed by each foster mother. The DOTS-R CHILD was completed on each child by the foster mother. Additionally, a modified version of the DOTS-R CHILD described by Lerner et al. (1985) was used to assess foster mothers' general expectations of child temperament. For example, the DOTS-R item "My child smiles often" was changed to "I expect a child to smile often. " In a study of fourthgrade children, Lerner et al. (1985) found that children whose temperaments met or exceeded teachers' expectations for attention span-distractability and reactivity performed better at school than children whose temperaments fell below teachers' expectations for children on those two attributes. Thus it seems that at least for these two temperament dimensions, goodness-of-f it between child temperament and environmental expectations is predictive of adjustment and degree of success in school. The scoring process is identical for each of the three versions. A score of 1,2,3 or 4 is assigned to each item, with a higher score indicating a higher level of a particular attribute. Item scores are summed to derive a total score for
19 each dimension. Copies of all three forms of the DOTS-R are illustrated in Appendix A. Family Environment Scale The Family Environment Scale (FES) (Moos and Moos, 1981) is a 90-item scale which measures the social environmental characteristics of all types of families. The Real Form (Form R) which measures family members ' perceptions of their existing family environments was used in the present study. The FES consists of ten subscales which assess three underlying domains. The Relationship dimensions are measured by the Cohesion, Expressiveness and Conflict subscales. The Personal Growth dimensions are measured by the Independence, Achievement Orientation, Intellectual-Cultural Orientation, Active-Recreational Orientation and Moral-Religious Emphasis subscales. The System Maintenance dimensions are measured by the Organization and Control subscales. Internal consistencies (Cronbach's alpha) for the subscales range from .61 to .78. Test-retest reliabilities (at a two-month interval) for the subscales range from .68 to .86. Normative data collected from 1125 families included families from all areas of the country, single-parent and multigenerational families, families drawn from ethnic minority groups and families of all age groups. Foster mothers completed the Conflict subscale of the FES in each case. There are nine items on the subscale with a dichotomous (true-false) response choice. Raw scores are
20 calculated for the subscale and then converted to standard scores. The FES was included in the study to assess family environment characteristics of the placements. The Conflict subscale in particular was targeted as a dependent measure relevant to success of placement. Thomas et al. (1968) describe conflict as an outcome of a poor fit between child temperament and parental demands, and this measure also might be viewed as a general index of harmony in the foster home. Previous research with the instrument has shown the Conflict subscale to relate to a variety of family problems. Moos and Moos (1981) found that distressed families (families with a member under psychiatric care) had more conflict. Families with a delinquent child were higher on conflict as were families with an alcoholic member. Larger families scored higher on conflict. Number of family arguments also related to conflict. Parental self-criticism and lack of involvement with their children were also related to conflict. As parental involvement is discussed as an indicator of quality of care, its relationship with conflict seems especially relevant to the study of foster care placements. In a previous study examining success in foster care (Doelling, 1987) , more conflict was found in foster families with a temperament mismatch between the foster mother and foster child in terms of an inflexible mother paired with a negative mood child. A copy of the Conflict subscale is provided in Appendix A.
21 Dyadic Parent-Child Interaction Coding System The Dyadic Parent-Child Interaction Coding System (DPICS) (Robinson and Eyberg, 1981) is a standardized observational system for coding interaction patterns between parents and children. It was included in the present study as a means of assessing mother-child interaction patterns which might reflect goodness-of-f it between mother and child and impact upon placement outcome. The DPICS consists of 22 separate behavior categories covering parent and child behaviors that are coded in three 5-minute interactions. In the first 5minute segment, the child-directed interaction (CDI) , the parent is instructed to allow the child to choose an activity and to play along with the child. In the parent-directed interaction (PDI) , the parent is instructed to select an activity and keep the child playing according to the parents' rules. In the third 5-minute segment (Clean-Up) , the parent is instructed to have the child clean up all the toys by himself or herself. DPICS categories for mother behavior include direct command, indirect command, labeled praise, unlabeled praise, physical positive, physical negative, critical statement, descriptive statement, descriptive question, acknowledgement and irrelevant verbalization. Categories for child deviant behaviors include whine, cry, physical negative, smart talk, yell and destructive. Additionally the parent's response following a child's deviant behavior (i.e., ignores or responds) and the child's response
22 to a parent's command (i.e., complies, noncomplies or no opportunity) are also coded. Observations are continuous and result in the total frequency of each behavior or behavioral sequence per 5-minute interval. The DPICS was validated on 20 families referred for treatment of a conduct problem child (2-7 years of age) and 22 normal families (Robinson and Eyberg, 1981) . The mean inter-rater reliability coefficient for parent behaviors was found to be .91 (range=. 67-1. 0) and for child behaviors was .92 (range=.76-1.0) . Parents of conduct problem children were found to give more critical statements and commands, and conduct problem children demonstrated more noncompliance. Conduct disordered children were found to show significantly higher rates of deviant behavior than their non-conduct disordered siblings. DPICS observations were also found to be highly related to parent report of home behavior problems. In a study of neglected children (Aragona and Eyberg, 1981) , neglectful mothers and mothers of behavior problem children were found to be more critical and give more direct commands. The number of behavior categories was reduced for purposes of statistical analyses. First of all, the categories were totalled across the three play situations. Then from the 22 behavior categories, four separate variables were created for mother behaviors: total praise (labeled + unlabeled praise) , total critical statements, percent criticism (critical statements/critical statments + praise) and total
23 commands. Three variables were created for child behaviors: total child deviant behaviors (whine + cry + physical negative + smart talk + yell + destructive) , total noncompliance (to both direct and indirect commands) and percent noncompliance (noncompliance/compliance + noncompliance) . These are the same categories that have been used in previous studies employing the DPICS (Robinson and Eyberg, 1981; Webster-Stratton, 1984) and found to be relevant. To assess reliability in the present study, two observers were used on 19 of the observations (43%) . Reliability was computed using Pearson product -moment correlations between observers for each behavior category. The mean inter-rater reliability coefficient for mother behaviors was .89 (range=.84-.98) and for child behaviors was also .89 (range=.86-.90) . Correlations for each individual category are shown in Table 1. A copy of the DPICS coding sheet is provided in Appendix A. Foster Placement Evaluation Scale The Foster Placement Evaluation Scale (FPES) is a 14-item scale designed to measure success of foster placements as assessed by foster care case workers. Foster care workers respond to each item on a 5-point scale indicating degree of agreement, with higher scores indicating more successful placements. Item scores are then summed in order to obtain a total score.
24 TABLE 1 Inter-Rater Reliability Coefficients for DPICS Categories Total critical statements .87 Total praise .98 Percent criticism .86 Total commands . 84 Total child deviants .90 Total noncompliance .90 Percent noncompliance .86
25 In the initial development of this measure (Doelling, 1987) , interviews were conducted with nine foster care workers to develop a preliminary item pool. Workers were asked what things they found themselves considering when evaluating foster placements. When necessary, they were asked follow-up questions for elaboration, specification, and clarification and were probed for additional dimensions. From the information obtained in these interviews an initial sixteenitem pool was constructed from the dimensions which were mentioned as important by several workers. There was, in fact, a high degree of consistency and agreement among workers on important evaluational dimensions as mentioned in the interviews. This preliminary scale was then distributed to these same workers as well as to ten clinical psychology graduate students and faculty at the University of Florida. All graduate students surveyed had completed at least a onesemester practicum in clinical child assessment, and many were clinical child psychology minors. Respondents were asked to indicate whether they felt each item was reflective of success of outcome of foster placement and to elaborate on the clarity of the items and identify any pertinent omissions. Overall, foster care workers, graduate students and faculty indicated that they found all but one of the items to be useful, and no omission was mentioned consistently enough to be added to the scale. Some minor rewordings were made on a few of the items where a lack of clarity or explicitness was
26 indicated. Fifteen items were thus retained. However, one item ("The foster parent (s) is involved in and supportive of the child's therapy") was indicated as not applicable in several cases in the original sample and so was omitted from the final version. In the present sample, it was also necessary to omit four additional items. As the sample included many pre-school children, the three items pertaining to school performance were omitted. Additionally, several children in the sample had no visitation with their natural parents, so the item relating to visitation was omitted as well. In cases of single-parent homes, the item "There is ample affection shown between the foster father and the child" was given the same score as the identical item referring to the foster mother so as not to automatically penalize single-parent foster families on the scale. The form of the scale used in the present study thus consisted of ten items. Items retained on the scale are highly similar to and supported by the dimensions discussed by Wolins (1963) as being important in evaluating foster placements. In particular, he discusses physical care, affection, good rolemodeling, parents' ability to get along with the child's natural parents, ability to handle stress, good relationships with other children in the home, tolerance of behavior problems, amount of time spent playing with the child and amount of time spent in general with the child. In the study for which the scale was designed (Doelling, 1987) , the FPES
27 was found to identify placements which were more or less successful in relation to the temperament variables being studied. In particular, it was found that the combination of an inflexible mother and a child with a negative mood frequently resulted in less successful placements as measured by the FPES. Measures of internal consistency in the present study indicate excellent reliability of the scale. The split-half correlation coefficient (Spearman-Brown formula) for odd and even items was .96. The mean item-to-total correlation (Cronbach's alpha) was .91. Reliability as calculated on the twenty cases on which both a primary case worker and supervisor rating were available was also acceptable. The correlation coefficient between the two raters' scores was .56, indicating moderate agreement. This measure of reliability is believed to be lower due both to the reduced number of cases on which two ratings were available and to the fact that case workers and supervisors had different degrees of familiarity and different sources of information about the families. In the original study in which the FPES was used, the correlation between case worker and supervisor ratings was .65. Previous investigators have found inter-rater reliabilities on pairs of foster care workers to be consistently low to moderate (Fanshel, 1961; Wolins, 1959; Wolins, 1963) . Fanshel (1961) reported reliabilities on pairs of case workers evaluating the same families on the Foster
28 Parent Appraisal Form ranging from .25 to .81 (median .51). Given these findings and limitations, the obtained inter-rater reliability coefficient appears quite adequate and higher than that reported by Fanshel (1961) . Overall reliability of the scale, therefore, appears good and supports the use of the measure in the present study. A copy of the FPES is provided in Appendix A. Mother Satisfaction Rating An additional measure of placement success was the foster mothers' satisfaction ratings. Here each foster mother was asked to rate her degree of satisfaction with the placement on a 5-point scale (very dissatisfied, somewhat dissatisfied, neutral, somewhat satisfied, very satisfied) . While perhaps more subjective than case worker evaluations, these ratings were thought to produce a useful source of additional data regarding placement outcome. Procedure Data on the children were obtained from their foster mothers, case workers and home observers. Subjects were recruited through eight local county offices of HRS as described earlier. Foster mothers whose names were provided were contacted by phone and made aware of the nature of the study and the requirements for participation and were assured confidentiality. The large majority (70%) of mothers agreed to participate, and the investigator met them in their home at a mutually agreed upon time. A written description of the
29 study was provided and written consent was obtained. The investigator clarified instructions and read test items whenever necessary. The foster mother completed the demographic questionnaire, a DOTS-R ADULT on themselves, the modified form of the DOTS-R CHILD described earlier assessing parents' expectations of child temperament, a DOTS-R CHILD on the target child and the Conflict subscale of the Family Environment Scale, in that order. The foster mother was then instructed regarding the three DPICS play situations, and she and the target child were provided with a standardized basket of toys. Two basket of toys were available: one for children ages 2-6 and another for children ages 7-10. The toys used were ones which encourage creativity and interaction and did not include toys with rules or games in which there was a winner or loser. Two-to sixyear-olds were given blocks, large plastic stacking pieces, Mr. Potato Head, paper and crayons. Seven-to ten-year-olds were given an Etch-A-Sketch, small plastic stacking pieces. Tinker Toys, paper and crayons. The child was allowed a warmup period with the toys while the mother completed the questionnaires. The two coders used for the DPICS observations had extensive training and experience with the coding system. The coders both had prior experience with DPICS through live coding of parents and children in a clinic setting. In addition, the coders received over 30 hours of formal training which included reading and studying a manual defining the
30 behavioral categories (Eyberg, 1983) , practice coding of videotaped interactions and live coding of clinic families. The coding was done live by the primary investigator and in 43% of cases by a second reliability coder as well. Both coders were blind as to the status of the subjects on the other measures used in the study. Foster care case workers, who were made aware of the nature of the study, were mailed a Foster Placement Evaluation Scale for each placement for which they were primarily responsible. Foster care supervisors were asked to identify those cases with which they had both independent contact and adequate familiarity and were mailed forms for those placements. This was possible in twenty of the cases. The primary case workers' ratings, which were available in all cases and deemed the more accurate ratings, were used in the analyses. The supervisors' ratings were used only in analyses assessing the correlations between the two ratings. Each child was assigned a subject number which appeared on each of the forms rather than the names of the children, the foster parents, the case workers, or the supervisors. This was done to assure confidentiality for all parties and to encourage candor.
RESULTS Demographic Data Twenty-three girls and twenty-one boys were included having an average age of five years, five months. Average length of placement at the time of data collection was five months. Mothers' ages ranged from 29 to 68 (mean age 45) and fathers' ages ranged from 30 to 67 (mean age 46) . Families had on, the average, one natural child, one or no adopted children, and one or two other foster children living in the home, or an average of four total children. Some homes, however, had as many as four natural, two adopted, and four other foster children. Mean annual family income was $25,842. Average educational level for both mothers and fathers was high school. The primary caretaker was the mother in 36 cases, and was both parents in the remaining eight. Thirty-six children had two parents, and eight children had a foster mother only. Relationships between the Dependent Measures In terms of data analyses, it was first necessary to examine the dependent measures to determine whether they measured the same or different outcome attributes. Scores on the Foster Placement Evaluation Scale were not significantly 31
32 correlated with either mother satisfaction ratings or scores on the Conflict subscale of the Family Environment Scale. Neither were mother satisfaction ratings significantly correlated with scores on the Conflict subscale of the Family Environment Scale. These correlations are illustrated in Table 2. For the Dyadic Parent-Child Interaction Coding System variables, percent noncompliance was highly correlated with total noncompliance (r=.71, p<.001); percent criticism was significantly positively correlated with total criticism (r=.54, p<.001) and significantly negatively correlated with total praise (r=-.38, p<.05); and percent noncompliance was highly correlated with total child deviant behaviors (r=.65, p<.001). DPICS variables were not significantly correlated with FPES scores, mother satisfaction ratings, or Conflict scores. These correlations are also illustrated in Table 2. In order to simplify the number of dependent variables and avoid redundancy, only those DPICS variables providing unique information were retained. Due to its high correlation with both total noncompliance and total child deviant behaviors, percent noncompliance was selected as the primary child behavior of interest. Percent criticism was felt to be the best representative measure accounting for both praise and criticism on the mothers' part as this percentage reflects total criticism relative to the sum of criticism and praise, and total commands was also retained as a dependent variable
33 TABLE 2 Correlations between Dependent Measures FPES FPES Conflict Criticism -.027 Praise -.009 Commands -.042 Deviants -.094 Compliance -.102 %Noncompl iance -.151 %Criticism -.178 Conflict Mother Rating .046 -.128 .158 .229 .049 .065 ,176 ,093 ,138 265 .063 .158 .026 -.093 -.078 .051 -.241
34 of interest because this number would seem to reflect the degree of demandingness on the part of the mother toward the child. The six dependent variables used in the analyses, then, were FPES scores. Conflict scores, mother satisfaction scores, percent noncompliance, percent criticism, and total commands. In order to decrease the probability of obtaining spurious results, analyses focused only on those independent variables thought to be relevant to the dependent measures and issues addressed in the study. Three mother temperament dimensions and four child temperament dimensions were chosen. Mother flexibility-rigidity, the mother temperament dimension shown in the previous study to relate to placement success (Doelling, 1987) , was selected along with a related dimension, approach-withdrawal, and mood. These three dimensions were felt to be parent temperament characteristics likely to be most relevant to parenting style and the ability to deal with and adapt to new situations as foster parents must continually do. Child mood, the child temperament dimension shown in the previous study (Doelling, 1987) to relate to placement success, was included along with approach-withdrawal, flexibility-rigidity, and task orientation. Flexibilityrigidity and approach-withdrawal were included because these are measures of adaptability, a trait which is seen as highly relevant for foster children and their adjustment to a home. Finally, task orientation was chosen as a child variable of interest because highly distractible and nonpersistent
35 children would be expected to be especially frustrating for parents to deal with. The two activity level and the three rhythmicity dimensions were felt to be less relevant to the issue of foster placement success as these would seem to be less overtly problematic behaviors. Demographic and Temperament Variables Related to Placement Success Initial analyses were performed to determine which, if any, single variables among those considered were related to the dependent measures. These analyses were aimed at assessing the relationship between demographic variables (e.g., child's age, length of placement, number of children in the home, etc.) and both mother and child temperament variables and indices of placement success. Here the FPES and mother satisfaction scores were considered as direct measures of placement success, while Conflict scores and DPICS variables were viewed as assessing the quality of interactions between the foster mother and child and in the family as a whole. No single demographic or parent-child temperament variable was found to be significantly related to success of placement as measured by the FPES or mother satisfaction ratings. Only one demographic variable was significantly related to the degree of conflict in the home as measured by the Family Environment Scale Conflict subscale. Total number of children in the home (natural + adopted + foster children) was positively correlated with degree of conflict (r=.56,
36 p<.001). The number of foster children alone in the home was also positively correlated with degree of conflict (r=.36, p<.05) (number of total children and number of foster children obviously being correlated) . In this sample then, number of children was the best predictor of conflict in the home. In terms of DPICS variables, several single demographic and temperament variables were significantly associated with observed parent-child interaction patterns. Regarding demographic variables, child's age was related to both total commands by the parent (r=-.65, p<.001) and percent noncompliance by the child (r=-.44, p<.01). That is, younger children received more commands from their foster mothers and were more noncompliant with commands than older children. With regard to child temperament, two variables were significantly associated with DPICS measures. Child task orientation was negatively correlated with total commands (r=-.36, p<.05). That is, children low on task orientation (more distractible and less persistent) received more commands, presumably because such children need more parental direction to stay on task. Child mood was negatively correlated with percent criticism (r=-.51, p<.001); children with more negative mood received more criticism from their foster mothers. Two mother temperament variables were significantly associated with DPICS variables. Here approach-withdrawal was negatively correlated with percent criticism (r=-.33, p<.05).
37 and was positively correlated with percent noncompliance (r=.31, p<.05). That is, mothers high on the withdrawal dimension (who presumably tend to respond negatively to new situations) were more critical of their foster children. Those high on the approach end of this dimension had more noncompliant foster children. While the first finding is comprehensible, the second, at first glance, appears counterintuitive. It may be, however, that high approach mothers are simply more tolerant and perhaps more unintentionally reinforcing of child noncompliance than mothers who are less accepting of novel situations. Maternal mood was found to be negatively correlated with percent criticism (r=-.40, p<.01) and positively correlated with percent noncompliance (r=.39, p<.01). That is, mothers with a negative mood gave more criticism to their foster children, criticism of the child presumably being reflective of their own mood. The finding that mothers high in positive mood elicit more noncompliance in their foster children is also somewhat hard to understand yet is likely related to the previous finding regarding approach-withdrawal. A positive mood mother may also be more likely to be tolerant and perhaps reinforcing of child noncompliance. The fact that these two mother temperament dimensions, approach-withdrawal and mood, are positively correlated (r=.41, p<.01), lends support to this hypothesis. Interestingly, no direct relationships
38 between the parent or child dimension of flexibility-rigidity and any dependent measures were found. The Goodness-of-Fit Model and Placement Success As a major assumption of the study was that a combination of mother-child variables would be most predictive of placement success (as was found in the previous investigation) , analyses designed to consider these variables together were subsequently explored. The analysis of choice was multiple regression with and without interaction terms. Before conducting such analyses, however, it was necessary to examine the dependent measures to determine if they satisfied the normality assumptions as required for multiple regression. Unfortunately, all dependent measures were found to be non-normally distributed. In particular, the FPES (ShapiroWilks W=.83, p<.001) and mother satisfaction scores (ShapiroWilks W=.60, p<.01) were found to be highly skewed. FPES scores were skewed (skewness=-1.08) toward the high end, with most placements reported as being successful. Likewise, the large majority of mothers reported being extremely satisfied with the placements, resulting in a concentration of scores at the high end as well (skewness=-1.7) . Conflict subscale scores were also found to be non-normal (Shapiro-Wilks W=.91, p<.01) although less highly skewed (skewness=. 06) , giving a relatively normal appearing probability plot with a slightly higher proportion of mothers reporting relatively low conflict.
39 Exponential transformations were performed on both FPES scores and mother satisfaction scores; however, due to the concentration of scores on both scales at the highest possible value and the small standard deviation relative to the mean (Afifi and Clark, 1984) , these transformations were not successful in inducing normality. A log transformation was performed on Conflict scores, but due to the already relatively normal distribution and the small standard deviation relative to the mean (Afifi and Clark, 1984) , this transformation was likewise unsuccessful in improving normality. The three DPICS variables were also found to be nonnormally distributed. Percent criticism (Shapiro-Wilks W=.83, p<.001, skewness=.74) , total commands (Shapiro-Wilks W=.87, p<.001, skewness=1.2) , and percent noncompliance (ShapiroWilks W=.90, p<.01, skewness=1.0) were all positively skewed such that low levels of criticism, commands, and noncompliance were more common. Log transformations were performed on these variables which were relatively successful in inducing normality (although the log transformation for percent noncompliance was marginally significant) . Tests for normality on both the untransformed and transformed DPICS variables are shown in Table 3. Due to the non-normal distributions of FPES, Conflict and mother satisfaction scores and the failure of attempts at transformation to induce normality, it appeared most
40 TABLE 3 Tests for Normality of DPICS Variables Variable Untransformed ; Log transformation ; Shapiro-Wilks v p-value Shapiro-Wilks v p-value %Criticism .83 <.001 .93 .05 Commands .87 <.001 .96 .23 %Noncompliance .90 <.01 .93 .02
41 appropriate to categorize subjects into two groups ("successes" and "failures") according to whether they fell above or below the median on each of these measures. Likewise, mothers and children were labeled "difficult" on specific temperament dimensions if they scored above the median on that dimension and were labeled "easy" if they scored below the median on that dimension. "Mismatches" were defined as motherchild pairs in which the foster mother and the foster child were both difficult on any particular pair of mother-child temperament variables. All other pairs were defined as "matches". These definitions are illustrated in Table 4. Matches and mismatches were determined for each of the mother temperament dimensions of interest (mood, approachwithdrawal and flexibility-rigidity) in combination with each of the child temperament dimensions of interest (approachwithdrawal, flexibility-rigidity, mood and task orientation) or for twelve combinations in all. Matches were then compared with mismatches with regard to outcome via chi-square tests. These analyses resulted in a single significant finding. Placements displaying a mismatch in terms of a mother low on approach who had a child low on task orientation had more conflict as measured by the Conflict subscale of the FES (Chisquare=5.4, p<.05, Phi=-.35). This is illustrated in Table 5. A mother low on approach is likely to react negatively when presented with new or challenging situations and thus not provide the degree of structure a highly distractible child
42 TABLE 4 Definition of Match and Mismatch Mother Child T ype of Match easy easy match difficult easy match easy difficult match difficult difficult mismatch
43 TABLE 5 Chi-square Test for Match and Mismatch in Terms of Mother Approach-Withdrawal and Child Task Orientation and Conflict Score High Conflict Low Conflict Match n=10 n=24 Mismatch n=7 n=3 Chi-square=5 .369 df=l p<.05
44 might need. It is easy to see how such a combination might cause frustration on both sides and result in more family conflict. It was possible to conduct multiple regression analyses on the normally distributed log transformations of the DPICS variables. Here there were several significant findings. To assess the usefulness of the transformations, however, these analyses were performed both with and without the log transformations. In every case the conclusions were unaltered by the transformations. To enhance interpretability (Affifi and Clark, 1984) , the results are presented in terms of the untransformed variables. Initial exploratory analyses involved stepwise regressions entering mother and child variables of interest along with the demographic variables shown to be related to the dependent variables (namely, child's age and number of children in the home) . Those variables shown to contribute significantly to the prediction of each dependent variable were then used in multiple regression analyses to arrive at the best prediction equation for each dependent variable. Finally, interaction terms were explored for each of the mother temperament dimensions of interest in combination with each of the child temperament variables of interest (again 12 combinations in all) . Significant findings resulting from these analyses are illustrated in Table 6.
45 TABLE 6 Multiple Regression Results for DPICS Variables DPICS Variable Commands Independent Variable F child's age 34.62 task orientation 6.94 child's age + task orientation 20.78 p< Model R .001 .05 001 50 Percent mother flexibilitycriticism rigidity * child mood 7, .00 .05 Percent child's age 10, .96 .01 noncompl iance mother approachwithdrawal * child approach-withdrawal 4. .36 .05 child's age + mother approach-withdrawal * child approachwithdrawal 7. ,66 .01 Percent child's age 11. ,26 .01 noncompliance mother mood * child approach-withdrawal child's age + mother mood * child approachwithdrawal 5.60 8.43 05 001 14 27 29
46 As found earlier, child's age and child task orientation together accounted for a large portion of the variation in total commands (R^=.50, p<.001). Thus, number of commands given is most dependent on characteristics of the child, indicating that younger, less task-oriented children elicit more commands from their foster mothers. Here none of the interaction terms resulted in a significant prediction equation. For percent criticism, no group of single variables resulted in a significant prediction equation. However, a regression equation employing a mother flexibility-rigidity * child mood interaction term was found to be significant (R^=.14, p<.05). That is, in cases of a rigid mother and a negative mood child, mothers gave a higher percentage of criticism to the child in the play interactions. It should be noted that this was the combination found in the previous study (Doelling, 1987) to predict poorer placement outcome as well. It seems that a high degree of criticism by the foster mother would indicate dissatisfaction with the child, and hence the placement, and be a legitimate index of placement success. This result involving a dependent measure reflective of actual maternal behavior then serves to at least partially replicate the results of the earlier study. For percent noncompliance, two equations employing interaction terms were significant. The first involved child's
47 age and the interaction of mother approach -withdrawal with child approach-withdrawal (R^=.27, p<.01). This equation indicates that a younger child high on approach with a foster mother also high on approach showed more noncompliance. The second equation involved child's age and the interaction of mother mood with child approachwithdrawal (R^=.29, p<.001). This finding suggests that a younger child high on approach with a positive mood mother also showed more noncompliance. Since mother approachwithdrawal and mother mood were positively correlated, these two equations can perhaps be interpreted jointly. An approaching, positive mood mother in combination with an approaching child results in more child noncompliance. Since approaching, positive mood mothers elicit more noncompliance, presumably because they are more tolerant of it, and since noncompliance tends to occur more often in children who are high on approach (r=.25, p<.10) , the combination would clearly result in significantly more child noncompliance. The final way in which the data was examined was by looking at matches and mismatches with regard to mother expectations for child temperament in relation to actual ratings of child temperament. In these analyses, placements were classified as matches on a dimension if the foster child's actual temperament score on that dimension equaled or was easier than that expected by the foster mother. If the child's actual temperament score was more difficult than that
48 expected by the mother, that placement was classified as a mismatch on that temperament dimension. These matches and mismatches were also compared with respect to outcome. Chisquare tests were again used with the FPES, Conflict scores, and mother satisfaction scores, and multiple regression interaction terms of actual temperament with expected temperament were used with the three DPICS variables. A major impediment in these analyses was the fact that for every temperament dimension except one (activity levelgeneral) , actual child temperament and expected child temperament were significantly positively correlated, indicating that most children met mothers Â• expectations except for the dimension of general activity level. That is, there were relatively few cases of mismatch in this sense. Although general activity level was not one of the originally targeted variables, it was examined here since it was the only dimension for which there were any significant differences between expected and actual temperament. In fact, placements in which the child had a higher general activity level than that expected by the foster mother were judged less successful as measured by the FPES (Chi-square=4.5, p<.05, Phi=-.32). This is illustrated in Table 7. No significant relationships were found when employing mother satisfaction scores or Conflict scores as dependent variables.
49 TABLE 7 Chi-square Test for Match and Mismatch in Terms of Mother Expectation for Child Activity Level-General and FPES Score Low FPES Score High PFES Score Child Meets or Exceeds Mother Expectations n=7 n=15 Child Does Not Meet Mother Expectations n=14 n=8 Chi-square=4.464 df=l p<.05
DISCUSSION The goals of this study were twofold. The first aim was to replicate the findings of an earlier study (Doelling, 1987) which found that particular mother-child temperament combinations resulted in less successful foster placements as measured by FPES scores, Conflict scores, and mother satisfaction scores. The second aim was to obtain some information regarding how parent-child mismatches may exert their influence. Thus the DPICS variables were included as a means of exploring what actually occurs between the mother and child, in the case of a mother-child temperament mismatch, that causes conflict and results in the placement being judged less successful. Regarding the first goal, replication of the previous study was only partially achieved. In the earlier study it was found that the mismatch of an inflexible mother paired with a child high on the negative mood dimension of temperament was predictive of poorer placement outcome in terms of higher Conflict scores, lower mother satisfaction scores, and lower FPES scores. These results were not replicated in the present study. It was found, however, that the combination of an inflexible mother and a negative mood 50
51 child was related to more criticism relative to praise given by the foster mother to her child in play interactions. Thus the fact that this inflexible mother/negative mood child combination was found to relate to a negative interaction pattern (a large degree of criticism by the mother toward the child) is noteworthy. In neither the present nor the previous study were scores on the three rating scales (FPES, Conflict subscale, mother satisfaction ratings) significantly correlated with each other. This suggests that while each can be considered a measure of placement success, they appear to measure different outcome attributes. It was somewhat disappointing to find that the DPICS variables were unrelated to the other dependent measures. There are, however, several possible reasons for this. The major difference between the two groups of dependent variables is that the DPICS variables are derived from direct observations while the FPES, mother satisfaction ratings, and the Conflict subscale of the FES are questionnaire measures. It is possible that certain demand characteristics were involved in the self-report measures. In particular. Conflict scores, which were predicted to be most strongly related to interaction patterns, would have also been the measure most likely to have pulled for a socially desirable response style. Although numbers rather than names appeared on all the questionnaires, it is still possible that foster mothers may have been inclined to report minimal family conflict. This is
52 particularly likely since foster parents are accustomed to being evaluated and may already have a set to respond in a socially desirable manner. The fact that both Conflict scores and mother satisfaction scores were skewed in the favorable direction also suggests that such a response pattern may have occurred . Although case workers might be expected to be more objective in their ratings, this may not necessarily be the case. FPES scores were also skewed in the favorable direction, indicating that the majority of placements were judged to be highly successful by case workers. Workers might also be more likely to judge placements as more successful due to a bias noted by George (1970) of foster care workers to present foster parents as ideal people due to the existence of stringent selection criteria. That is, given a shortage of foster homes, workers may have a tendency to present existing or prospective foster parents as better than they are for fear of losing potential placement sources. Although foster mothers may have wanted to give a favorable impression during the DPICS assessment as well, it is more difficult to ascertain what constitutes desirable or undesirable responses in this situation. It is also unlikely that the mother would be able to completely control the behavior of her child unless that was already their normal pattern. Mothers may have some sense that it is desirable for her child to obey and not display deviant behaviors but
53 probably have little idea that commands, criticism, or praise are relevant responses on their part. Thus, the DPICS variables may be less influenced by response biases than the others. It is interesting in light of this that the majority of the significant results were found for DPICS variables rather than the other three dependent measures. As in the previous study, little was found in terms of single demographic or temperament variables being related to placement success as measured by FPES scores, mother satisfaction scores, and Conflict scores. The only significant finding here was that conflict increased with the number of children in the home. This is hardly surprising yet is highly relevant to the issue of foster placements. It would seem desirable in light of this finding not to overcrowd foster homes with large numbers of children, especially if there are other natural or adopted children already in the home. The reality of the situation, however, is that there is a shortage of foster homes, making multiple placements often a necessity. This finding, then, merely underscores the already recognized need for more foster homes. It also indicates that when given the choice, rather than add a child to a full foster home that might nonetheless be willing to take another child, a child might preferably be sent to a less crowded home. One mother-child temperament combination was also significantly related to conflict. Placements with a mother low on approach and a child low on task orientation had more
54 conflict. Highly distractible children are likely to need a high degree of parental direction which a mother low on approach, who responds unfavorably to new situations, might not adequately provide. In such cases, the child may be more likely to exhibit deviant behaviors which the mother may be unsuccessful in controlling, with the result being more conflict in the home. In terms of the three questionnaire measures of placement success, then, Conflict was the only one found to be related to the independent variables. The extreme lack of variability in mother satisfaction scores likely accounts for this measure's lack of association with the other variables. It is disappointing, however, that FPES scores were not predicted by any combination of demographic or temperament variables. This is surprising both because the FPES was designated as the primary measure of placement success and because it was the measure found to be most relevant in the previous study. It is possible, as noted above, that case worker ratings were biased towards presenting foster homes favorably, however, any such bias should have been operating in the first study as well. Although the samples in the two studies were drawn from the same population and same geographic area, there were three notable differences. First, in the present study only placements of less than one year were included as a way of excluding what were thought to be obviously successful long-
55 term placements. This was done in hopes that not only would the original findings be replicated but in fact be stronger through including more new and thus potentially unsuccessful placements. Therefore, if there were a bias in terms of case worker report, it might have been operating more strongly in the present study. That is, if the present sample did in fact include a larger number of less successful placements, case workers might have found it more necessary to bias their ratings than in the first sample, in which more placements truly were successful. The second difference between the two samples was income level. The annual reported income for foster families rose from $18,265 in the first sample to $25,842 in the present sample. It is possible that higher level income homes might be viewed by foster care workers as more desirable, at least in terms of their being able to provide for the material needs of the children in their care. Workers might also be more reluctant to lose higher level income foster homes and thus have reason to want to report placements in these homes as being more successful. Were there more available foster homes than children needing to be placed, such biases as those suggested above would be minimized. However, at least in the area sampled in the study, a shortage of foster homes did exist, making the hypothesized response bias conceivable. The third difference between the two samples was that a younger group of children was included in the present sample.
56 Tha first study included 5 to 10 year olds, while the present study included 2 to 10 year olds. The change in age range could have had two possible effects. First, it is unlikely but possible that the relationships between temperament and the FPES hold only for the 5 to 10 year age range and not the 2 to 5 year age range. The more likely possibility is that the items relating to school performance which were omitted from the FPES in the present study were highly relevant items. If this is the case, it may be that use of the FPES should be restricted to school-age children, or perhaps another version of the scale should be created for placements of younger aged children. The majority of significant findings, then, related to DPICS variables. Child's age was the one significant demographic variable found to relate to interaction patterns. It was found that younger children elicited more maternal commands and were less compliant with commands than older children. It is certainly not surprising that young children require more parental direction and do not obey as well, and it was not the case that child's age by itself predicted placement success. Since child's age did not relate to percent criticism or any of the other measures of placement success, it seems that younger children in general do not have less successful placements. In terms of temperament variables, children with difficult temperament on certain dimensions encountered more
57 negative interactions in the play situations. Specifically, children low on task orientation received more commands, and children with negative mood received more criticism. Thus, a child's temperament does appear to be related to parental response and perhaps impact on the mother-child relationship as evidenced by the observed interaction patterns. Mother temperament was also found to be related to interaction patterns. Mothers low on approach and mothers with negative mood were more critical of their foster children, and mothers high on approach and mothers with positive mood elicited more noncompliance from their foster children. Thus mother temperament seems related both to the way she responds to her child as well as to how the child responds to her. It seems, then, that both child and mother temperament affects the types of interactions that occur between the two and presumably ultimately affects the quality of the relationship. It was also the case that certain combinations of the above variables were predictive of interaction patterns. Child's age and task orientation together were most predictive of number of commands. As already discussed, the interacton effect of an inflexible mother and a negative mood child predicted maternal criticism. Two different combinations of variables predicted child noncompliance. Child's age and the interaction effect of mother approach-withdrawal with child mood and child's age and the interaction effect of mother mood with child approach-withdrawal both predicted child
58 noncompliance. Thus, child approach-withdrawal seems to be a relevant variable only as it interacts with mother approachwithdrawal and mother mood and is not influential by itself. Presumably, approaching, positive mood mothers are more tolerant of child noncompliance, which coupled with the tendency of children high on approach to be more noncompliant, likely explains this interaction effect. Although child noncompliance is typically seen as undesirable, it may not be seen as problematic in every case. That is, some parents may not necessarily expect their child to obey them all the time, and some commands may be seen as more important than others. A mother high on the withdrawal dimension or with negative mood may expect her child to comply with every command, whereas an approaching or positive mood mother may be less demanding. Both types of mothers likely expect certain types of commands, such as those pertaining to the child's safety, to always be obeyed but may be less concerned with other types of commands, such as those given in a play situation. Thus, child noncompliance during play may not be seen as particularly negative and may not result in conflict or further negative interactions between the mother and child. It is also possible that foster parents may be more tolerant of difficult child behavior in general due to their experience with a large number of children, many of whom may have behavioral problems.
59 Thus, in terms of DPICS variables, percent criticism seems to be the variable most indicative of negative interaction patterns and unsatisfactory placements. Negative mood on the part of both the mother and the child and withdrawal on the part of the mother were the single variables that predicted more criticism. Additionally, the combination of an inflexible mother with a negative mood child predicted more criticism. If criticism is viewed as a clearly negative parental response and is also viewed as an index of maternal unacceptance or dissatisfaction with the child, it can be considered one legitimate measure of placement success and quality of the mother-child relationship. When viewed in this way, the results appear more consistent with those of the first study in that they indicate more negative placement outcome in cases of mismatch consisting of an inflexible mother with a negative mood child. Little was found with regard to mismatch in terms of a child not meeting a foster mother's expectations for temperament. Placements in which the child was more active than the foster mother expected were judged less successful by the FPES. For all other temperament dimensions, actual child temperament and expected child temperament were highly correlated, indicating that very few children did not meet mother expectations on these dimensions. One potential limitation of the study is that only mother ratings of child temperament were used. It could be argued
60 that mothers' ratings of child temperament might be biased such that mothers either rate children as uniformly more easy or difficult than they actually are or that mothers' own temperament characteristics might influence the way they rate their children. In the first study, teacher ratings of child temperament were also obtained for as many cases as possible. However, the DOTS-R proved to be a poor measure for use with teachers, as teachers were unable to respond to items on four of the nine dimensions which referred to behaviors not observable in the school setting. For the dimensions on which both teacher and parent ratings were available, correlations were moderately low and consistent with those obtained in other research (Bates, 1980; Field and Greenberg, 1982). As this lack of agreement did exist, however, analyses were done to examine whether there was a consistent bias on the part of the mothers and whether any such bias may have affected the results. Mothers were found not to have rated their children as uniformly more difficult or more easy than teachers. Furthermore, whatever differences existed were shown not to have affected the results. That is, mothers with difficult temperament themselves did not rate their children as more difficult than mothers with easy temperament. Since the DOTS-R was obviously not designed for teacher use and since in the previous study mothers were shown not to have been negatively biased or biased in a way that affected the results, it was felt that mother ratings were the most
61 representative and accurate ratings available and thus were chosen for use in the present study. It would also have been desirable to obtain father ratings both as a second measure of child temperament and as a way of assessing the father-child relationship and its impact on placement outcome. This was also unsuccessfully attempted in the previous study. Although father data was sought in every case in the first study, only 13 of 51 fathers were able or willing to participate. As the sample in the present study was highly similar to the first, any further attempt to collect father data was considered fruitless. This situation does suggest that in the majority of cases foster fathers are less involved as discussed by Wiehe (1982) and is also supported by the fact that 36 mothers reported themselves to be the primary caretaker, while only eight reported both parents to be equal caretakers. In no case was the father reported to be the primary caretaker of the target child. For this reason, it appears that the relationship between foster mother and child was in most cases the crucial one and the one with the greatest consequence in terms of placement outcome. It is possible, therefore, that foster fathers are so uninvolved as to have little impact on the outcome of placements. However, it is also possible that the temperament of foster fathers may interact in some way with that of the child and impact on placement outcome. Any future studies that
62 are able to obtain foster father data would thus be highly desirable. Another limitation of the study was that already existing placements were used due to their availability. Although placements of more than one year were excluded, and every attempt was made to include the most recent placements, there were in the sample a number of placements of several weeks or months duration. As previous research indicates that most placement breakdowns occur within four weeks (Stone and Stone, 1983) , the sample may have been biased in favor of relatively successful placements. This fact helps explain the skewness of the dependent measures (i.e., more placements rated as successful) . However, it should be noted that a broader range of scores on the FPES was obtained in the present study as compared with the previous one in which even more long-term placements were included. That is, by including a larger number of recent placements and excluding extremely long placements, it appears that more relatively unsuccessful placements were present in the sample. Still, it is likely that more powerful results would have been obtained had only new placements been used, however. It is impossible to know how many of the sampled placements will eventually end in breakdown. However, at three to four month follow-up, eleven of the children had been moved to another foster home for reasons of poor outcome as judged by either the foster parents or the case worker. The small
63 number of cases prevented performing a statistical test of significance, but it should be noted that the mean FPES score of these placements was 42 while the mean score for continued placements was 45. Future research examining cases as they are placed, thus maximizing the inclusion of the least successful placements (early breakdowns) , would be desirable as would be a study which includes long-term follow-up of placements. It is clear that while the individual demographic and temperament variables as well as the interactive effects of parent-child temperament variables explained a significant amount of variation in the DPICS variables and Conflict 1 Â• Â• Â• scores , the fact that these variables were not predictive of FPES scores or mother satisfaction scores suggests the presence of other factors as well. Such factors as marital discord, financial problems, amount of case worker involvement, or difficult child behavior other than that addressed by temperament variables might be more relevant than those explored here or might interact with temperament variables. Future research would do well to explore other factors in addition to those addressed by the present study. It is also questionable as to whether potentially problematic placements should be avoided or whether specfic It is interesting to note that while statistical analyses comparing the present sample with normals were not done due to differences in the two populations, on the two measures for which norms were available (the DPICS and the FES) , foster parents and children were not remarkably different from normal families.
64 parent training for "at-risk" placements might be effective. For example, a training program which teaches foster parents to have more positive interactions with their children may succeed in overriding temperamental tendencies. Teaching foster parents effective discipline strategies may also be useful. Research aimed at developing and assessing such training programs is suggested to explore this possibility.
DEMOGRAPHIC INFORMATION Subject number HRS counselor How long has this counselor been involved with this placement? Target child's age Birthdate Date of placement Foster mother's age Foster father's age Number of natural children in home Ages Number of adopted children in home Ages Number of other foster children in home Ages Does the target child have any natural siblings in the home? If yes, how many? Ages Level of education completed by foster mother Level of education completed by foster father Approximate total annual family income How long have you been foster parents? Which parent is the primary caretaker of the target child? Approximately how many hours per day does the target child spend with the mother? with the father? Does the child have any major handicaps or disabilities? If yes, please explain Are there currently any plans for another placement for the child? If yes, what? How satisfied are you with this placement? 1 2 3 4 5 very somewhat neutral somewhat very dissatisfied dissatisfied satisfied satisfied 66
67 REVISED DIMENSIONS OF TEMPERAMENT SURVEY Â— ADULT* Subject number Today's date Sex: Male Female HOW TO ANSWER: On the following pages are some statements about how people like you may behave. Some of the statements may be true of your own behavior and others may not apply to you. For each statement we would like you to indicate if the statement is usually true of you, is more true than false of you, is more false than true of you, or is usually false of you. There are no "right" or "wrong" answers because all people behave in different ways. All you have to do is answer what is true for you . Here is an example of how to fill out this questionnaire. Suppose a statement said: "I eat about the same things for breakfast every day. " If the statement were usually false for you, you would respond: "A," usually FALSE. If the statement were more false than true for you, you would respond: "B," more FALSE than true. If the statement were more true than false for you, you would respond: "C," more TRUE than false. If the statement were usually true for you, you would respond: "D," usually TRUE. On the line to the left of each statement write an A if the statement is usually false for you, write a B if the statement is more false than true for you, write a C if the statement is more true than false for you, or write a D if the statement is usually true for you. PLEASE KEEP THESE FOUR THINGS IN MIND AS YOU ANSWER: 1. Give only answers that are true or false for you. It is best to say what you really think. 2. Don't spend too much time thinking over each question. Give the first, natural answer as it comes to you . Of course, the statements are too short to give all the information you might like, but give the best answer you can under the circumstances. Some statements may seem similar to each other because they ask about the same situation. However, each one looks at a different area of your behavior. Therefore, your answers may be different in each case. 3. Answer every question one way or another. Don't skip any.
68 4. Remember, A = usually FALSE B = more FALSE than true C = more TRUE than false D = usually TRUE THANK YOU FOR YOUR COOPERATION *(Windle and Lerner, 1986)
69 A = usually FALSE C = more TRUE than FALSE B = more FALSE than TRUE D = usually TRUE 1. It takes me a long time to get used to a new thing in the home. 2. I can't stay still for long. 3. I laugh and smile at a lot of things. 4. I wake up at different times. 5. Once I am involved in a task, nothing can distract me from it. 6. I persist at a task until it's finished. 7. I move around a lot. 8. I can make myself at home anywhere. 9. I can always be distracted by something else, no matter what I may be doing. 10. I stay with an activity for a long time. 11. If I have to stay in one place for a long time, I get very restless. 12. I usually move towards new objects shown to me. 13. It takes me a long time to adjust to new schedules. 14. I do not laugh or smile at many things. 15. If I am doing one thing, something else occurring won't get me to stop. 16. I eat about the same amount for dinner whether I am home, visiting someone, or traveling. 17. My first reaction is to reject something new or unfamiliar to me. 18. Changes in plans make me restless. 19. I often stay still for long periods of time. 20. Things going on around me can not take me away from what I am doing.
70 21. I take a nap, rest or break at the same times every day. 22. Once I take something up, I stay with it. 23. Even when I am supposed to be still, I get very fidgety after a few minutes. 24. I am hard to distract. 25. I usually get the same amount of sleep each night. 26. On meeting a new person I tend to move towards him or her. 27. I get hungry about the same time each day. 28. I smile often. 29. I never seem to stop moving. 30. It takes me no time at all to get used to new people. 31. I usually eat the same amount each day. 32. I move a great deal in my sleep. 33. I seem to get sleepy just about the same time every night. 34. I do not find that I laugh often. 35. I move towards new situations. 36. When I am away from home I still wake up at the same time each morning. 37. I eat about the same amount at breakfast from day today. 38. I move a lot in bed. 39. I feel full of pep and energy at the same time each day. 40. I have bowel movements at about the same time each day. 41. No matter when I go to sleep, I wake up at the same time the next morning.
71 42. In the morning, I am still in the same place as I was when I fell asleep. 43. I eat about the same amount at supper from day to day. 44. When things are out of place, it takes me a long time to get used to it. 45. I wake up at the same time on weekends and holidays as on other days of the week. 46. I don't move around much at all in my sleep. 47. My appetite seems to stay the same day after day. 48. My mood is generally cheerful. 49. I resist changes in routine. 50. I laugh several times a day. 51. My first response to anything new is to move my head toward it . 52. Generally I am happy. 53. The number of times I have a bowel movement on any day varies from day to day. 54. I never seem to be in the same place for long.
72 REVISED DIMENSIONS OF TEMPERAMENT SURVEY Â— CHILD* Subject number Today's date Sex: Male Female HOW TO ANSWER: On the following pages are some statements about how children like your own may behave. Some of the statements may be true of your child's behavior, and others may not apply to him or her. For each statement we would like you to indicate if the statement is usually true of your child, is more true than false of your child, is more false than true of your child, or is usually false of your child. There are no "right" or "wrong" answers because all children behave in different ways. All you have to do is answer what is true or false for your child . Here is an example of how to fill out this questionnaire. Suppose a statement said: "My child eats the same things for breakfast every day. " If the statement were usually false for your child, you would respond: "A," usually FALSE. If the statement were more false than true for your child, you would respond: "B," more FALSE than true. If the statement were more true than false for your child, you would respond: "C," more TRUE than false. If the statement were usually true for your child, you would respond: "D," usually TRUE. On the line to the left of each statement write an A if the statement is usually false for your child, write a B if the statement is more false than true for your child, write a C if the statement is more true than false for your child, or write a D if the statement is usually true for your child. PLEASE KEEP THESE FOUR THINGS IN MIND AS YOU ANSWER: 1. Give only answers that are true or false for your child. It is best to say what you really think. 2. Don't spend too much time thinking over each question. Give the first, natural answer as it comes to you . Of course, the statements are too short to give all the information you might like, but give the best answer you can under the circumstances. Some statements may seem similar to each other because they ask about the same situation. However, each one looks at a different area of your behavior. Therefore, your answers may be different in each case.
73 3. Answer every question one way or another. Don't skip any. 4. Remember, A = usually FALSE B = more FALSE than true C = more TRUE than false D = usually TRUE THANK YOU FOR YOUR COOPERATION *(Windle and Lerner, 1986)
74 A = usually FALSE C = more TRUE than FALSE B = more FALSE than TRUE D = usually TRUE It takes my child a long time to get used to a new thing in the home. 2. My child can't stay still for long. 3. My child laughs and smiles at a lot of things. 4. My child wakes up at different times. 5. Once my child is involved in a task, nothing can distract him or her from it. 6. My child persists at a task until it's finished. 7. My child moves around a lot. 8. My child can make him/herself at home anywhere. 9. My child can always be distracted by something else, no matter what he or she may be doing. 10. My child stays with an activity for a long time. 11. If my child has to stay in one place for a long time, he/she gets very restless. 12. My child usually moves towards new objects shown to him/her. 13. It takes my child a long time to adjust to new schedules. 14. My child does not laugh or smile at many things. 15. If my child is doing one thing, something else occurring won't get him/her to stop. 16. My child eats about the same amount for dinner whether he/she is home, visiting someone, or traveling. 17. My child's first reaction is to reject something new or unfamiliar to him/her. 18. Changes in plans make my child restless.
y(h 75 19. My child often stays still for long periods of time. 20. Things going on around my child can not take him/her away from what he/she is doing. 21. My child takes a nap, rest, or break at the same times every day. 22. Once my child takes something up, he/she stays with it. 23. Even when my child is supposed to be still, he/she gets very fidgety after a few minutes. 24. My child is hard to distract. 25. My child usually gets the same amount of sleep each night. 26. On meeting a new person my child tends to move towards him or her. 27. My child gets hungry about the same time each day. 28. My child smiles often. 29. My child never seems to stop moving. 30. It takes my child no time at all to get used to new people. 31. My child usually eats the same amount each day. 32. My child moves a great deal in his/her sleep. 33. My child seems to get sleepy just about the same time every night. 34. I do not find my child laughing often. 35. My child moves towards new situations. 36. When my child is away from home he/she still wakes up at the same time each morning. 37. My child eats about the same amount at breakfast from day to day. 38. My child moves a lot in bed. 39. My child feels full of pep and energy at the same time each day.
76 40. My child has bowel movements at about the same time each day. 41. No matter when my child goes to sleep, he/she wakes up at the same time the next morning. 42. In the morning, my child is still in the same place as he/she was when he/she fell asleep. 43. My child eats about the same amount at supper from day to day. 44. When things are out of place, it takes my child a long time to get used to it. 45. My child wakes up at the same time on weekends and holidays as on other days of the week. 46. My child doesn't move around much at all in his/her sleep. 47. My child's appetite seems to stay the same day after day. 48. My child's mood is generally cheerful. 49. My child resists changes in routine. 50. My child laughs several times a day. 51. My child's first response to anything new is to move his or her head toward it. 52. Generally my child is happy. 53. The number of times my child has a bowel movement on any day varies from day to day. 54. My child never seems to be in the same place for long.
77 REVISED DIMENSIONS OF TEMPERAMENT SURVEY Â— CHILD EXPECTATIONS Subject number Today's date Sex: Male Female HOW TO ANSWER: On the following pages are some statements about how you expect children to behave. For each statement we would like you to indicate if the statement is mostly true of what you expect of children, is more true than false of what you expect, is more false than true of what you expect, or is mostly false of what you expect. There are no "right" or "wrong" answers. Just answer what you expect of children's behavior. Here is an example of how to fill out this questionnaire. Suppose a statement said: "I expect a child to eat the same things for breakfast every day. " If the statement were mostly false of what you expect of children, you would respond: "A," mostly FALSE. If the statement were more false than true of what you expect, you would respond: "B," more FALSE than true. If the statement were more true than false of what you expect, you would respond: "C," more TRUE than false. If the statement were mostly true of what you would expect, you would respond: "D," mostly TRUE. On the line to the left of each statement write an A if the statement is usually false of what you expect, write a B if the statement is more false than true of what you expect, write a C if the statement is more true than false of what you expect, or write a D if the statement is usually true of what you expect. PLEASE KEEP THESE FOUR THINGS IN MIND AS YOU ANSWER: 1. Give only answers that are true or false for you. It is best to say what you really think. 2. Don't spend too much time thinking over each question. Give the first, natural answer as it comes to you . Of course, the statements are too short to give all the information you might like, but give the best answer you can under the circumstances. Some statements may seem similar to each other because they ask about the same situation. However, each one looks at a different area of your behavior. Therefore, your answers may be different in each case.
78 3. Answer every question one way or another. Don't skip any. 4. Remember, A = usually FALSE B = more FALSE than true C = more TRUE than false D = usually TRUE THANK YOU FOR YOUR COOPERATION *(Windle and Lerner, 1986)
Â• ^it Â•i . \ I 79 A = usually FALSE C = more TRUE than FALSE B = more FALSE than TRUE D = usually TRUE 1. I expect that it will take a child a long time to get used to a new thing in the home. 2. I expect that a child won't be able to stay still for long. 3. I expect a child to laugh and smile at a lot of things . 4. I expect a child to wake up at different times. 5. Once a child is involved in a task, I expect nothing will distract him or her from it. 6. I expect a child to persist at a task until it's finished. 7. I expect a child to move around a lot. 8. I expect a child to make him/herself at home anywhere . 9. I expect a child will always be distracted by something else, no matter what he or she may be doing. 10. I expect a child to stay with an activity for a long time. 11. If a child has to stay in one place for a long time, I expect he/she to get very restless. 12. I expect a child to move toward new objects shown to him/her. 13. I expect a child to take a long time to adjust to new schedules. 14. I expect a child will not laugh or smile at many things. 15. If a child is doing one thing, I expect that something else occurring won't get him/her to stop. 16. I expect a child to eat about the same amount of dinner whether he/she is home, visiting someone, or traveling.
80 17. I expect a child's first reaction to be to reject something new or unfamiliar to him/her. 18. I expect that changes in plans will make a child restless. 19. I expect a child to often stay still for long periods of time. 20. I expect that things going on around a child will not take him/her away from what he/she is doing. 21. I expect a child to take a nap, rest, or break at the same times every day. 22. Once a child takes something up, I expect he/she to stay with it. 23. Even when a child is supposed to be still, I expect he/she to get very fidgety after a few minutes. 24. I expect a child to be hard to distract. 25. I expect a child to usually get the same amount of sleep each night. 26. On meeting a new person I expect a child to tend to move toward him or her. 27. I expect a child to get hungry about the same time each day. 28. I expect a child to smile often. 29. I expect a child to never seem to stop moving. 30. I expect it will take a child no time at all to get used to new people. 31. I expect a child to usually eat the same amount each day. 32. I expect a child to move a great deal in his/her sleep. 33. I expect a child to get sleepy just about the same time every night. 34. I expect not to find a child laughing often. 35. I expect a child to move toward new situations.
81 36. When a child is away from home, I expect he/she will still wake up at the same time each morning. 37. I expect a child will eat about the same amount at breakfast from day to day. 38. I expect a child to move a lot in bed. 39. I expect a child to feel full of pep and energy at the same time each day. 40. I expect a child to have bowel movements at about the same time each day. 41. No matter when a child goes to sleep, I expect him/her to wake up at the same time the next morning. 42. In the morning, I expect a child to still be in the same place as he/she was when he/she fell asleep. 43. I expect a child to eat about the same amount at supper from day to day. 44. When things are out of place, I expect it will take a child a long time to get used to it. 45. I expect a child to wake up at the same time on weekends and holidays as on other days of the week. 46. I expect a child to not move around much at all in his/her sleep. 47. I expect a child's appetite to stay the same day after day. 48. I expect a child's mood to be generally cheerful. 49. I expect a child to resist changes in routine. 50. I expect a child to laugh several times a day. 51. I expect a child's first response to anything new to be to move his/her head toward it. 52. I expect a child to be generally happy. 53 . I expect the number of times a child has a bowel movement on any day will vary from day to day. 54. I expect a child will never be in the same place for long.
/ i I : 82 FAMILY ENVIRONMENT SCALE* FORM R CONFLICT SUBSCALE Subject number Today's date Sex: Male Female Instructions: Read each statement and decide if it is true of your family or false of your family. Circle T if it is true of your family, and circle F if it is false of your family. T F 1. We fight a lot in our family. T F 2. Family members rarely become openly angry. T F 3. Family members sometimes get so angry they throw things. T F 4. Family members hardly ever lose their tempers. T F 5. Family members often criticize each other. T F 6. Family members somethimes hit each other. T F 7. If there's disagreement in our family, we try hard to smooth things over and keep the peace. T F 8. Family members often try to one-up or out-do each other. T F 9. In our family, we believe you don't ever get anywhere by raising your voice. *(Moos and Moos, 1981)
I ! / ' 83 DPICS DATA RECORDING SHEET* PARENT BEHAVIORS TOTALS Acknowledge Irrelevant Verbalization Critical Statement Physical Negative Physical Positive Unlabeled Praise Labeled Praise Descriptive/Reflective Question Reflective Statement Descriptive Statement Indirect Command followed by: No Opportunity Compliance Noncompl iance Direct Command followed by: No Opportunity Compliance Noncompl iance CHILD BEHAVIORS Cry Ignored Responded to Yell Ignored Responded to Whine Ignored Responded to Smart Talk Ignored Responded to Destructive Ignored Responded to Physical Ignored Negative Responded to * (Robinson and Eyberg, 1981)
84 FOSTER PLACEMENT EVALUATION SCALE Subject number Today's date The following is a list of statements pertaining to various aspects of foster care placements. Please read each item, decide how descriptive that statement is of this particular placement and circle the appropriate number. Thank you for your help. 1 = strongly disagree 2 = slightly disagree 3 = neither agree nor disagree 4 = slightly agree 5 = strongly agree 1. The foster parent (s) spends an adequate amount of time with the child in fun activities. 12 3 4 5 2. The foster parent (s) treats the child equally well to the other children in the home. 12 3 4 5 3. There is ample affection shown between the foster mother and the child. 12 3 4 5 4. There is ample affection shown between the foster father and the child. 12 3 4 5 5. The child seems to enjoy spending time with the other children in the home. 12 3 4 5 6. The foster parent (s) adequately takes care of the medical and other needs of the child (food, clothing, other appts . , etc . ) 12 3 4 5 7. The foster parent (s) is able to deal effectively with difficult behaviors the child exhibits. 12 3 4 5 8. The foster parent (s) shows an attitude of acceptance toward the child regardless of his/her behavior. 12 3 4 5 9. The child seems to have adapted well to the family structure. 12 3 4 5 10. The foster parent (s) is receptive to and aware of the child's individual needs. 12 3 4 5
RAW DATA ID LOP CAGE MAGE FAGE NONAT NOADOPT NOFOST NOSIB MED FED 1 3 45 65 Â• 10 Â• 2 1 60 35 Â• 2 12 Â• 3 8 79 33 35 2 1 1 10 10 4 7 116 53 56 1 1 16 12 5 2 37 36 Â• 3 2 12 Â• 6 7 27 55 55 1 12 13 7 12 40 42 46 2 2 13 16 8 9 51 56 66 1 2 1 10 10 9 6 37 33 41 1 2 3 16 16 10 1 48 68 60 2 8 6 11 6 33 49 54 2 2 7 7 12 5 76 36 31 3 2 2 13 12 13 6 27 41 43 12 14 14 6 31 48 43 1 4 12 12 15 2 81 29 32 2 1 12 12 16 4 93 47 50 1 11 12 17 2 104 61 Â• 4 1 12 Â• 18 12 120 37 41 2 1 12 11 19 3 36 51 54 2 11 12 20 10 87 61 Â• 4 1 12 Â• 21 4 63 35 37 1 1 14 12 22 11 97 47 48 2 1 16 19 23 2 65 39 36 3 12 7 24 2 85 52 Â• 1 1 1 12 Â• 25 3 27 50 40 1 3 12 12 26 8 104 32 37 2 1 1 12 12 27 1 63 64 Â• 1 2 14 Â• 28 2 35 34 38 1 1 1 12 12 29 7 82 Â• Â• 1 3 1 12 Â• 30 8 29 29 34 3 12 12 31 7 42 34 45 3 2 12 13 32 4 56 51 45 2 1 12 11 33 12 38 35 35 4 1 12 12 34 12 59 41 44 2 14 15 35 2 104 49 44 1 2 12 12 36 7 26 48 56 2 2 12 16 37 3 30 42 46 2 2 1 9 11 38 1 50 60 60 1 1 16 12 39 4 33 31 30 1 1 1 12 12 40 1 45 31 36 2 12 12 41 2 59 43 44 1 3 12 12 42 11 41 51 67 1 1 9 14 43 6 99 50 56 1 2 2 11 12 44 5 98 59 63 1 11 11 86
87 ID INC TIME PRIM MHRS FHRS NOPAR SEX MRACE CRA( 1 6000 144 M 16 Â• 1 F W w 2 10000 36 M 16 Â• 1 M W w 3 48 M 5 5 2 F N N 4 12 B 18 18 2 M N N 5 12000 6 M 16 Â• 1 M N N 6 23000 60 M 13 3 2 F W W 7 35000 48 M 8 5 2 M W W 8 20000 12 M 5 3 2 M W W 9 34000 96 M 5 2 2 M W N 10 6000 9 B 4 4 2 F N N 11 7000 5 M 12 4 2 M N N 12 13000 24 M 6 4 2 F N N 13 41000 24 M 16 4 2 F W W 14 73000 120 M 14 5 2 F W W 15 20000 24 M 12 3 2 M W W 16 24 M 6 2 2 F W W 17 11000 252 M 19 Â• 1 F N N 18 12000 144 M 4 1 2 F W W 19 15000 72 M 8 4 2 F W W 20 15000 96 M 18 Â• 1 F N N 21 37000 4 M 16 6 2 M N N 22 60000 12 M 7 3 2 F N N 23 84 M 6 2 2 F N W 24 17000 Â• M 7 Â• 1 F N N 25 50000 48 M 16 6 2 M N N 26 34000 24 M 15 6 2 F W W 27 5000 36 M 10 Â• 1 F N N 28 35000 3 M 16 4 2 M W W 29 36 M 9 Â• 1 M N N 30 20000 24 M 16 4 2 F W W 31 45000 30 M 16 5 2 M W W 32 11000 24 B 8 8 2 F N N 33 35000 120 M 12 3 2 M W W 34 20000 132 B 16 16 2 M W W 35 44000 178 M 7 3 2 F W W 36 30000 18 M 14 1 2 M W W 37 20000 30 M 5 5 2 F W W 38 30000 60 B 16 16 2 M W W 39 18 M 5 3 2 F W W 40 60000 12 B 12 8 2 M W W 41 28000 9 M 9 3 2 M W W 42 7000 48 B 16 16 2 F W W 43 29000 Â• M 4 4 2 M W W 44 12000 14 B 6 6 2 N W W
88 ID MAI MA2 MA3 MA4 MA5 MA6 MA7 MAS MAIO MAll 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 19 12 13 12 20 21 21 23 20 21 16 17 23 19 14 16 27 20 22 28 20 18 24 23 23 25 23 17 24 19 17 22 23 24 17 17 19 18 16 16 25 25 18 24 9 8 6 9 10 10 8 10 5 15 10 13 16 7 4 8 7 8 6 6 9 14 4 7 5 12 4 10 12 14 4 6 6 6 4 6 12 4 16 4 15 26 19 12 12 21 18 15 20 23 22 27 21 22 20 26 15 26 20 21 18 20 23 23 22 28 25 26 21 16 28 26 24 19 23 19 21 23 17 20 21 24 23 28 19 27 20 21 16 18 18 13 15 19 14 19 18 13 12 19 11 20 13 18 11 19 11 17 18 17 10 18 18 10 20 20 15 10 19 16 12 18 15 12 15 16 18 20 15 17 18 14 18 28 27 24 21 26 21 27 28 26 21 22 28 28 23 24 22 16 23 28 24 13 28 27 26 24 28 20 28 25 28 28 28 25 21 28 22 24 28 28 19 27 27 22 11 16 15 16 12 18 21 15 19 16 22 19 22 18 20 22 17 15 20 21 7 22 20 9 18 17 20 21 9 17 14 19 17 18 17 14 15 21 18 21 21 19 14 16 7 14 6 6 10 19 15 20 14 9 16 8 5 20 18 13 14 10 12 17 6 20 20 7 18 19 20 17 5 8 20 8 16 10 14 17 14 20 18 20 20 16 5 15 8 9 9 10 10 12 9 15 15 11 12 12 14 20 14 13 8 11 12 13 7 20 20 8 18 15 17 20 8 10 9 16 13 11 12 16 15 15 12 11 16 16 11 15 8 10 10 13 13 7 17 18 11 16 11 17 11 8 11 10 15 10 14 11 14 12 12 7 19 11 16 6 14 7 9 5 11 8 13 11 13 12 8 14 5 6 10 12 5 10 27 11 9 9 9 11 9 10 14 11 8 9 11 10 10 8 10 10 9 12 12 9 12 10 12 12 10 11 9 5 7 12 9 11 9 10 5 9 12 6 4 10
89 ID MEl ME2 ME3 ME4 ME5 ME6 ME7 ME8 ME9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 19 18 23 21 19 21 28 15 22 18 27 25 23 29 26 19 26 16 28 23 23 25 21 25 24 23 28 25 27 24 27 21 25 12 21 26 21 28 22 25 22 28 22 20 14 10 12 13 12 15 16 16 12 13 14 15 14 10 11 10 9 15 16 12 13 16 4 12 16 13 14 11 16 15 15 16 8 18 12 12 10 14 14 16 10 15 13 12 16 21 13 9 18 15 23 22 21 17 23 18 21 17 16 15 17 15 25 19 19 21 23 21 17 15 21 14 25 14 14 16 17 10 17 21 19 24 18 16 19 19 16 17 13 14 11 10 13 11 15 14 14 7 10 11 14 14 9 20 16 12 7 14 11 14 9 12 9 10 12 16 14 6 9 17 14 27 11 6 13 14 10 8 14 6 6 10 18 23 11 21 20 25 27 26 25 24 24 28 27 27 19 21 26 18 23 25 15 24 27 21 21 27 27 18 28 20 26 22 22 19 18 28 21 25 22 18 23 26 15 21 11 18 17 13 14 17 18 21 17 10 18 14 19 22 19 19 14 14 13 17 11 23 15 13 14 18 21 18 6 17 11 9 16 13 13 12 17 18 8 9 20 16 13 17 5 15 6 5 13 15 11 18 10 7 18 11 9 20 29 15 17 11 5 6 9 17 9 11 10 14 14 7 5 19 11 5 14 14 11 11 14 14 8 5 13 15 6 15 11 13 12 12 14 15 13 14 12 8 13 11 13 19 17 14 14 13 13 14 12 16 15 14 14 11 16 13 5 16 10 11 15 12 11 14 15 15 7 8 14 17 9 13 14 17 16 13 21 16 12 18 10 19 23 11 10 14 13 18 12 12 17 16 11 16 24 16 11 17 12 10 9 9 12 13 14 28 17 9 18 13 11 9 15 9 14 19
90 ID MCI MC2 MC3 MC4 MC5 MC6 MC7 MC8 MC9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 24 15 25 24 21 25 26 21 16 28 25 16 27 22 26 24 26 16 27 22 23 16 25 28 20 28 15 24 28 26 25 28 25 28 24 28 21 26 28 28 24 28 27 21 16 7 4 13 12 13 16 14 7 16 16 13 16 10 15 9 10 12 16 15 11 16 4 12 15 15 4 9 16 16 10 6 8 11 12 13 12 16 16 16 13 5 15 11 14 25 22 13 21 22 27 21 21 18 23 24 21 28 20 22 23 18 25 24 18 28 28 25 12 21 28 25 24 14 19 27 18 24 23 28 19 27 20 25 24 28 19 17 12 16 15 11 11 15 18 14 15 7 6 14 16 20 15 15 15 15 5 16 12 14 16 16 13 14 20 18 17 11 17 10 13 18 15 8 15 15 19 20 16 14 15 10 18 27 27 24 20 24 28 24 27 15 28 28 23 27 17 23 27 11 22 25 19 26 27 26 15 27 28 18 28 28 22 27 25 28 21 28 21 26 28 28 21 27 21 21 7 21 19 10 13 15 19 23 19 12 21 18 18 15 21 17 13 14 16 21 13 24 24 10 16 22 24 24 24 16 16 23 19 19 14 16 17 22 10 21 20 24 12 17 5 17 10 7 15 16 15 20 15 10 20 11 8 20 20 15 16 13 13 17 12 19 17 16 11 18 20 16 20 17 16 19 14 15 15 29 15 18 5 8 15 17 8 15 10 12 10 7 13 17 13 14 17 10 20 14 12 20 17 21 17 14 13 16 13 16 20 12 13 13 20 18 20 11 12 14 14 12 11 16 15 16 6 14 16 19 8 14 17 17 15 15 16 8 11 17 17 11 20 24 14 17 11 16 9 8 15 19 8 26 19 14 12 12 16 15 16 9 15 11 17 18 15 8 16 15 11 8 17 8 9 21
91 ID CON MS FPESl FPES2 CRIT PRAIS COMM DEV COM? PNON 1 2 5 50 45 19 8 77 2 6 8 2 3 5 50 3 18 1 3 17 3 3 45 46 1 3 40 1 3 4 1 44 45 3 1 6 5 3 5 30 8 85 2 14 16 6 3 3 40 37 2 13 132 2 15 11 7 5 49 47 11 18 119 13 11 8 2 5 46 2 37 1 3 9 5 5 48 5 34 1 3 10 5 42 16 45 3 7 11 2 5 50 3 4 48 7 15 12 4 5 48 1 2 9 13 5 44 6 2 50 5 16 32 14 2 5 50 1 16 40 6 9 21 15 4 5 40 12 16 5 40 1 4 22 4 3 14 17 4 5 38 19 1 5 18 1 3 35 1 26 1 4 19 2 5 50 7 22 38 5 13 20 4 2 50 1 1 19 2 11 21 1 5 50 1 2 21 22 5 30 5 11 2 18 23 3 5 35 1 17 2 7 41 24 2 5 50 3( 3 5 1 33 4 12 25 3 5 46 4 41 8 20 26 3 4 49 5 27 3 5 30 4 31 40 4 10 28 1 5 39 4 66 6 25 38 29 2 5 40 47 6 13 30 5 5 50 4( ) 2 21 117 11 9 31 3 4 50 4S ) 2 31 4 13 32 3 5 44 4: > 5 2 29 3 10 33 2 1 50 5( ) 5 6 42 7 14 33 34 2 5 50 5( ) 6 24 3 13 35 2 50 2 5 36 3 5 46 4^ I 9 84 10 12 37 2 5 43 4J 5 10 15 84 5 17 20 38 2 5 50 3' 7 2 9 18 2 11 39 1 4 41 3' 7 11 25 57 4 8 14 40 1 2 47 4: J 4 93 25 27 41 4 3 42 3i i 9 1 59 3 8 14 42 5 50 4' 7 2 6 27 2 5 19 43 2 5 49 A' 7 1 4 29 4 14 44 2 5 45 4' 7 1 1 16 2 13
92 ID CRIT2 PRAIS2 C0MM2 DEV2 C0MP2 PN0N2 1 14 8 103 2 17 17 2 1 3 109 4 21 3 Â• Â• Â• Â• Â• Â• 4 Â• Â• Â• Â• Â• Â• 5 Â• Â• Â• Â• Â• Â• 6 1 11 140 3 15 11 7 2 17 124 16 13 8 5 34 2 6 9 6 39 3 8 10 Â• Â• Â• Â• Â• Â• 11 3 4 36 3 8 12 Â• a Â• Â• Â• Â• 13 Â• Â• Â• Â• Â• Â• 14 2 15 43 6 10 23 15 15 1 7 16 Â• Â• Â• Â• Â• Â• 17 Â• Â• Â• Â• Â• Â• 18 1 20 19 Â• Â• Â• Â• Â• Â• 20 2 15 21 2 23 22 Â• Â• Â• Â• Â• Â• 23 3 21 10 48 24 5 2 34 1 5 15 25 Â• Â• Â• Â• Â• Â• 26 6 27 Â• Â• Â• Â• Â• Â• 28 Â• Â• Â• Â• Â• Â• 29 40 30 Â• Â• Â• Â• Â• Â• 31 1 2 39 1 3 32 Â• Â• Â• Â• Â• Â• 33 * Â• Â• Â• Â• 34 Â• Â• Â• Â• Â• Â• 35 Â• Â• Â• Â• Â• 36 Â• Â• Â• Â• Â• 37 Â• Â• Â• Â• Â• Â• 38 Â• Â• Â• Â• Â• . 39 Â• Â• Â• Â• Â• Â• 40 6 86 32 37 41 12 1 58 5 5 9 42 Â• Â• Â• Â• Â• 43 Â• Â• Â• Â« Â• Â• 44 Â• Â• Â• Â• Â• Â•
93 RAW DATA CODES ID = identification number LOP = length of placement (in months) CAGE = child's age (in months) MAGE = foster mother's age FAGE = foster father's age NONAT = number of natural children in the home NOADOPT = number of adopted children in the home NOFOST = number of other foster children in the home NOSIB = number of target child's siblings in the home MED = foster mother's educational level (in years) FED = foster father's educational level (in years) INC = annual family income TIME = time as foster parents (in months) PRIM = primary caretaker (M = mother B = both parents) MHRS = number of hours per day child spends with foster mother FHRS = number of hours per day child spends with foster father NOPAR = number of foster parents SEX = sex of child MRACE = foster mother's race (W = white N = nonwhite) GRACE = child's race (W = white N = nonwhite) MAI = mother activity level general MA2 = mother activity level sleep MA3 = mother approach withdrawal MA4 = mother flexibility rigidity MA5 = mother mood MA6 = mother rhythmicity sleep MA7 = mother rhythmicity eating MAS = mother rhythmicity daily habits MAIO = mother distractibity MAll = mother persistence MEl = expectations for activity level general ME2 = expectations for activity level sleep ME3 = expectations for approach withdrawal ME4 = expectations for flexibility rigidity ME5 = expectations for mood ME6 = expectations for rhythmicity sleep ME7 = expectations for rhythmicity eating ME8 = expectations for rhythmicity daily habits ME9 = expectations for task orientation MCI = child activity level general MC2 = child activity level sleep MC3 = child approach withdrawal MC4 = child flexibility rigidity MC5 = child mood MC6 = child rhythmicity sleep MC7 = child rhythmicity eating MC8 = child rhythmicity daily habits MC9 = child task orientation
94 CON = FES Conflict score MS = mother satisfaction score FPESl = case worker rating on FPES FPES2 = supervisor rating on FPES CRIT = number of critical statements PRAIS = number of praises COMM = number of commands DEV = number of child deviant behaviors COMP = number of noncompliant behaviors by child PNON = percent noncompliance by child CRIT2 = number of critical statements (coder 2) PRAIS2 = number of praises (coder 2) C0MM2 = number of commands (coder 2) DEV2 = number of child deviant behaviors (coder 2) C0MP2 = number of noncompliant behaviors by child (coder 2) PN0N2 = percent noncompliance by child (coder 2)
REFERENCES Achenbach, T.M. , & Edelbrock, C. (1983). Manual for the Child Behavior Checklist and Revised Child Behavior Profile . Burlington, VT: Queen City Printers. Afifi, A. A., & Clark. V. (1984). Computer aided multivariate analysis . Belmont, CA: Wadsworth, Inc. Agresti, A., & Agresti, B.F. (1979). Statistical methods for the social sciences . San Francisco: Dellen Publishing Co. Aragona, J.A. , & Eyberg, S.M. (1981). Neglected children: Mothers' report of child behavior problems and observed verbal behavior. Child Development . 52, 596-602. Bates, J.E. (1980) . The concept of difficult temperament. Merrill-Palmer Ouarterlv . 26, 299-319. Cameron, J.R. (1977). Parental treatment, children's temperament, and the risk of childhood behavior problems: 1. Relationship between parental characteristics and changes in children's temperament over time. American Journal of Orthopsychiatry. 47 . 568-576. Cautley, P.W. , & Aldridge, M.J. (1975). Predicting success for new foster parents. Social Work . 20, 48-53. Doelling, J.L. (1987). Predicting the success of foster care placements . Unpublished master's thesis, University of Florida, Gainesville. Eisenberg, L. (1962) . The sins of the fathers: Urban decay and social pathology. American Journal of Orthopsychiatry . 32 . 5-17. Eyberg, S.M., & Robinson, E.A. (1983). Dyadic parent-child interaction coding system: A manual. Psychological Documents . 13, Ms. #2582. Fanshel, D. (1961) . Studying the role performance of foster parents. Social Work . 6, 74-81. 95
96 Fanshel, D. , & Shinn, E.B. (1978a). Child behavior characteristics of foster children. In D. Fanshel & E.B. Shinn (Eds.), Children in foster care: A longitudinal investigation (pp. 325-372) . New York: Columbia University Press. Fanshel, D. , & Shinn, E.B. (1978b). Discharge and other status outcomes. In D. Fanshel & E.B. Shinn (Eds.), Children in foster care: A longitudinal investigation (pp. 112-144) . New York: Columbia University Press. Field, T., & Greenberg, R. (1982). Temperament ratings by parents and teachers of infants, toddlers and preschool children. Child Development . 53 . 160-163. George, V. (1970) . Foster care: Theory and practice . London: Rutledge & Kegan Paul. Graham, P., Rutter, M. , & George, S. (1973). Temperamental characteristics as predictors of behavioral disorders in children. American Journal of Orthopsychiatry . 43., 328-337. Hampson, R.F. (1980). Feedback from the experts: A study of foster mothers. Social Work . 25 . 108-113. Johnson, J.H., Floyd, B.J., & Isleib, R. (1986). Parent, child and social variables as predictors of child abuse: Implications for a temperament mismatch view of abusive behavior . Unpublished manuscript. University of Florida, Gainesville. Kraus, J. (1971) . Predicting success of foster placements for school-age children. Social Work . 16 . 63-72. Lerner, J.V., Baker, N., & Lerner, R.M. (1985). A personcontext goodness fit model of adjustment. In P. Kendall (Ed.), Advances in cognitive-behavioral research and therapy (Vol. 4, pp. 111-136). New York: Adacemic Press. Lerner, J.V. , Lerner, R.M., & Zabiski, S. (1985). Temperament and elementary school children's actual and rated academic performance: A test of a "goodness-of-f it" model. Journal of Child Psychology and Psychiatry . 26, 125-136. Lerner, R.M. , Palermo, M. , Spiro, A., & Nesselroade, J.R. (1982) . Assessing the dimensions of temperamental individuality across the life span: The Dimensions of Temperament Survey (DOTS) . Child Development . 53 . 149-159.
97 Maas, H.S., & Engler, R.E. (1959). Children in need of parents . New York: Columbia University Press. Martin, H.P., & Beezley, P. (1976). Foster placement. In H.P. Martin (Ed.), The abused child (pp. 189-199). Cambridge, MA: Bal linger. Moos, R.H., & Moos, B.S. (1981). Family Environment Scale Manual . Palo Alto, CA: Consulting Psychologists Press. Murray, L. (1984). A review of selected foster care-adoption research from 1978 to mid-1982. Child Welfare . 58, 113-124. Pardeck, J.T. (1983) . Marital status and family source of income: potential predictors for determining the stability of foster family care? Adolescence . 18, 631-635. Plomin, R. (1983) . Childhood temperament. In B. Lahey & A. Kazdin (Eds.), Advances in clinical child psychology (Vol. 6, pp. 45-89). New York: Plenum. Robinson, E.A., & Eyberg, S.M. (1981). The dyadic parent-child interaction coding system: Standardization and validation. Journal of Consulting and Clinical Psychology . 49 . 245-250. Rowe, D.C. (1976). Attitudes, social class, and the quality of foster care. Social Service Review . 50 . 506-514. Stone, N.M., & Stone, S.F. (1983). The prediction of successful foster placement. Social Casework . 64, 11-17. Thomas, A., & Chess, S. (1977). Temperament and Development . New York: Brunner/Mazel . Thomas, A., & Chess, S. (1981). The role of temperament in the contributions of individuals to their development. In R.M. Lerner & N.A. Busch-Rossnagel (Eds.) , Individuals as producers of their development: A life-span perspective (pp. 231-255). New York: Academic Press. Thomas, A., Chess, S., & Birch, H. (1968). Temperament and behavior disorders in children . New York: New York University Press.
98 Webster-Stratton, C. (1984) . Randomized trial of two parenttraining programs for families with conduct-disordered children. Journal of Consulting and Clinical Psychology . 52 . 666-678. Wiehe, V.R. (1982) . Differential personality types of foster parents. Social Work Research & Abstracts . 18 . 16-20. Windle, M. , Hooker, K. , Lernerz, K. , East, P.L., Lerner, J.V., & Lerner, R.M. (1986) . Temperament, perceived competence, and depression in earlyand late-adolescents. Developmental Psychology . 22 . 384-392. Windle, M. , & Lerner, R.M. (1986). Reassessing the dimensions of temperamental individuality across the life span: The Revised Dimensions of Temperament Survey (DOTS-R) . Journal of Adolescent Research . 1, 213-230. Wolins, M. (1959) . The problem of choice in foster home finding. Social Work . 4, 40-48. Wolins, M. (1963). Selecting foster parents; The ideal and the reality . New York: Columbia University Press.
-f \i BIOGRAPHICAL SKETCH Jenny Lynn Doelling was born on October 19, 1959, in Akron, Ohio to Carolyn Hathaway Doelling and Allan Frank Doelling. She graduated from Firestone High School in June 1978. Jenny attended the University of North Carolina from which she received a Bachelor of Arts degree in psychology in May of 1982. After graduation, Jenny was employed for two years as a child care worker at Daniel Memorial, a residential treatment facility for adolescents, in Jacksonville, Florida. She entered the graduate clinical psychology program at the University of Florida in September of 1984 and received her Master of Science degree in May of 1987. She began her clinical internship in the Department of Psychiatry at Kaiser Permanente in Los Angeles, California in September of 1988. She is a clinical child psychology minor and plans to continue working with children and adolescents after completion of her internship. 99
I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree 9^f Doctor of Philosophy. James H. Johrjaon, Chair Professor of vlinical and Health Psychology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Sheila M. Eybe^ f\ Professor of Clinical and Health Psychology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Patricia H/ Miller ' Professor of Psychology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Nancy K. Norvell Assistant/Professor of Clinical and Health Psychology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Constance L. Shehan Associate Professor of Sociology ym
This dissertation was submitted to the Graduate Faculty of the College of Health Related Professions and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. . ^ . ' \\ . August 1989 ^ -:....Â„,. > 'Â•^^rCT.. ... ^ ' Dean, College of Health Related Professions Dean, Graduate School I
UNIVERSITY OF FLORIDA 3 1262 08554 4160
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EGLGC5RUN_YI57SY INGEST_TIME 2011-11-08T19:48:20Z PACKAGE AA00004779_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC