Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Stress, maternal distress, and child adjustment following immigration : exploring the buffering role… Short, Kathryn Helen 1995

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1995-060640.pdf [ 4.65MB ]
JSON: 831-1.0088370.json
JSON-LD: 831-1.0088370-ld.json
RDF/XML (Pretty): 831-1.0088370-rdf.xml
RDF/JSON: 831-1.0088370-rdf.json
Turtle: 831-1.0088370-turtle.txt
N-Triples: 831-1.0088370-rdf-ntriples.txt
Original Record: 831-1.0088370-source.json
Full Text

Full Text

STRESS, MATERNAL DISTRESS, AND CH[LD ADJUSTMENT FOLLOWINGIMMIGRATION:EXPLORING THE BUFFERING ROLE OF SOCIAL SUPPORTbyKATHRYN HELEN SHORTBachelor of Arts, McMaster University, 1986Master of Arts, University of British Columbia, 1990A THESIS SUBMITTED iN PARTIAL FULFILMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESDepartment of PsychologyWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIASeptember, 1995®Kathryn Helen Short, 1995In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of Bntish Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives, It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)_____________________Department of Psyc- logyThe University of British ColumbiaVancouver, CanadaDate ] Se.p4emb’e- MdIDE-6 (2/88)ABSTRACTImmigration is typically deemed a stressful life event. For adults, the experience ofuprooting and settling in a new country has been associated with elevated rates ofpsychological distress. Basic North American parenting models would predict thatimmigrant children are also at risk for developing adjustment problems; both as a directfunction of immigration stress, and indirectly through the influence of parent distress anddisrupted parenting behavior. Although some empirical studies support this contention,many researchers have described lower or equivalent rates of problems in immigrant,relative to nonimmigrant, children. In the present study, in an attempt to understand whyit is that some children develop problems following migration whereas others remainresilient, a model that highlights the role of potential protective variables was empiricallytested. New immigrant mothers from Hong Kong completed a series of questionnairesregarding extrafamilial stress, personal distress, social support, and child behavior.Another adult familiar with the child’s adjustment also completed a child behaviorquestionnaire. Consistent with the Basic Model, results of Moderated MultipleRegression analyses revealed that extrafamilial stress and maternal distress were significantpredictors of child behavior problems. However, no support was found for the ModeratorModel. When the sample was split along gender lines and the analysis was conducted forboys only, findings were consistent with the Moderator Model in that the relationshipbetween extrafamilial stress and child behavior problems was weaker in the presence ofhigher levels of social support. At the same time, however, it was determined that therelationship between maternal distress and boys’ behavior was stronger at higher levels ofIIsupport. There were no significant interaction effects when the analysis was conductedexclusively with families of girls. Support for the Moderator Model was morestraightforward at the level of adult fi.rnctioning. In keeping with findings in the NorthAmerican literature, the relationship between stress and maternal distress was moderatedby social support in this immigrant sample. Cultural explanations for these findings werediscussed. An integrative model that follows from the results of this study was presentedas a heuristic to guide future study in this area.111TABLE OF CONTENTSAbstract iiTable of Contents ivList of Tables ‘1iList of Figures ViiiAcknowledgement ixDedication XIntroduction 1Overview IBasic North American Model of Family Stress and Parent-Child Outcomes 3- Description ofContemporary Models 3- Operationalization of the Basic Model 7- Empirical Support for Links within the Basic Model 11Consideration of the Basic Model Amongst Immigrant Families 17- Empirical Support for the Basic Model from the Immigrant Literature 17Resiliency and Moderator Variables 23- Resilient Children 23Proposed Immigrant Parent-Child Moderator Model 25- Description of the Moderator Model 25- Empirical Support for Links within the Moderator Model 27- Additional and Exploratory Links in the Moderator Model 32Summary of the Proposed Parent-Child Moderator Model and ResearchHypotheses 36- Summary of the Proposed Model 36- Hypotheses 37ivMethod 39Sample 39- Inclusion Criteria 39- Recruitment of Participants 42- Sample Size 44- Description of the Sample 44Procedures 47- Research Assistants 47- Preparation of the Questionnaire Package 47- Data Collection Procedures 51Measurement 53- Mothers’ Questionnaire Package 53- Childrens’ Interview 63- Global Ratings 65Results 66Primary Analyses regarding the Moderator Model 66- Summary Statistics 66- Data Aggregation and Transformation 73- Primary Regression Analyses 76Secondary Analyses regarding the Proposed Model 92- Social Competence 92- Stress-Distress Relationship 96Summary 99Discussion 101Review of Findings 101-Summary Statistics 101- Moderated Multiple Regression Analysis on Child Behavior Problems 102- Summary of Primary Analysis 108- Moderated Multiple Regression Analysis on Child Social Competence 110- Moderated Multiple Regression Analysis on Maternal Distress 112VMethodological Issues 114- Reliance on Maternal Perceptions of Self and Family 114- Reliance on a Correlational Design 118- Underreporting of Psychological Concerns 119- Translation of and Modifications to, Standard Instruments 121- The Challenges of Cross-Cultural Research 122Applications 123Future Directions 124Conclusions 126Footnotes 127Bibliography 129Appendix A. Family Information Questionnaire. 154Appendix B. Immigration Stress Scale. 155Appendix C. Parent Support Scale. 157Appendix D. Stress Thermometer. 158Appendix E. Scale of Children’s Stress. 159Appendix F. Mother Global Ratings. 160Appendix G. Observer-Child Rating Scales. 161viLIST OF TABLESTable 1. Demographic Information for Study Participants. 45Table 2. Means, Standard Deviations, and Ranges for Variables in theModerator Model. 67Table 3. Zero-order Correlations amongst Variables in the Moderator Model. 71Table 4. Intercorrelations amongst Composite Measures. 75Table 5. Summary of Hierarchical Multiple Regression Analysis predictingChild Behavior Problems. 81Table 6. Regressions ofMaternal Distress on Child Behavior Problems atLower, Medium, and Higher Levels of Social Support. 83Table 7. Separate Regressions on Child Behavior Problems; Stress, Support,and Stress x Support, and Distress, Support, and Distress x Support. 86Table 8. Means, Standard Deviations, and Ranges for all Variables in theModerator Model, by Child Gender. 88Table 9. Intercorrelations amongst Variables in the Moderator Model forFamilies of Boys and for Families of Girls. 90Table 10. Hierarchical Multiple Regression Analyses Predicting ChildBehavior Problems, Conducted by Child Gender. 91Table 11. Intercorrelations amongst Variables in the Moderator ModelPredicting Child Social Competence. 94Table 12. Hierarchical Multiple Regression Analysis Predicting Child SocialCompetence. 95Table 13. Summary of Hierarchical Multiple Regression Analyses PredictingChild Social Competence, Conducted by Child Gender. 97Table 14. Hierarchical Multiple Regression Analysis Predicting MaternalDistress. 99viiLIST OF FIGURESFigure 1. Basic Model of Stress and Parent-Child Behavior. 7Figure 2. Parent-Child Moderator Model. 27Figure 3. Scatterplots Depicting the Relationship between Distress and ChildBehavior Problems at Lower, Medium, and Higher Levels of Support. 84Figure 4. Proposed Model Integrating the Moderating Role of Social Support,and the Mediating Role ofMaternal Distress, in Predicting ChildBehavior. 114ViiiACKNOWLEDGMENTThis work has come to fruition only through the support and assistance of manyindividuals. Foremost, I would like to extend my thanks to Charlotte Johnston, mysupervisor and mentor throughout my graduate school career. It is an understatement tosay that this research would not have been possible without the thoughtful guidance andsupport consistently offered. Thanks also to members of my Departmental Committee,Anita DeLongis, Janet Werker, Rod Wong, and Peter Graf for their helpful commentsregarding an earlier version of this manuscript. Respected colleagues Georgia Tiedemann,Susan Cross, Angela Lamensdorf Josie Geller, Wendy Freeman, and Shawn Reynoldsalso made valued contributions. This research was made financially possible through aUBC-HSS grant awarded to Dr. Johnston.I would like to offer a special word of thanks to the research assistants who dedicatedtheir time and talents to this research project. In particular, I would like to acknowledgethe contributions of: Betty Au, Celia Chan, Doris Chu, Cathy Hui, Helen Kwoo, SylviaLam, Laura Lee, Debbie Leung, Mei Kei Leung, Yvette Leung, Linda Sung, Karrie Tam,and Sherlyn Yeap.Individuals in several agencies in the Lower Mainland were extremely generous with theirtime. Aberdeen Centre Mall, B.C. Association for Chinese Language, Chinese CulturalCentre, Ming Pau, Parker Place, SUCCESS, Vancouver Community College, and YaohanCentre were all very supportive of this project and assisted with subject recruitment. Inaddition, individuals at the Chinese Cultural Centre and SUCCESS served as consultantsearly in the research process. Reverend Hui at the Richmond Peace Chinese-MennoniteChurch was also very gracious about allowing us to conduct the pilot study duringSaturday morning Chinese language classes at the Church.I feel particularly indebted to the families who participated in this research. We werewarmly welcomed into their homes and they offered thoughtful, honest descriptions oftheir experiences. I believe that for most of the families involved, the greatest motivationfor participating in this research was a desire to help other new immigrant families.More personally, the achievement that is represented in this work must be credited in largepart to the patience, sacrifice, and unwavering support of my family. I share thisaccomplishment with my parents, Hank and Jacqui Krech, and with my sister and brotherin-law, Sue Krech and Michael Staffieri. And to my husband, Mike, there aren’t enoughwords. Thank you.ixDEDICATIONI would like to dedicate this work to the memories of:Doris & Stanley CoulthardLouise & Andrew KrechImmigrants who exemplified the spirit of resiliency.xINTRODUCTIONOverviewCanada has evolved as a country of immigrants. For over a century, members ofFirst Nations communities have shared this country with individuals who have left theirhomes to settle here. Since “the door was opened” most recently in 1967, an increasingnumber of newcomers, from a diversity of homelands, have contributed to Canada’s richmulticultural landscape. During the first 5 years of this decade, a record number offamilies immigrated to this country (Employment & Immigration Canada, 1991). In 1993,approximately 35% of Vancouver’s 37,132 newcomers were classified as family classimmigrants (Citizenship and Immigration Canada, 1994c). Although recentannouncements have suggested that the federal vision for 1996-2000 favours economicimmigrants over the family class (i.e., it is projected that by 2000, 53% of immigrants willbe economic class and 44% will be family class, compared to 43% and 51% respectively in1994), given the stated committment to maintaining a strong family program, immigrantfamilies will likely continue to represent a significant segment of the Canadian populationinto the 21st century (Citizenship and Immigration Canada, 1994b). Further, because theyalso come to Canada as dependents within the Business and Independent Classes, it isexpected that annual numbers of immigrant children will remain steady (Employment andImmigration Canada, 1989; Citizenship and Immigration Canada, 1994a). In 1993, almost50,000 children migrated to this country (Citizenship and Immigration Canada, 1994c).Immigration is widely regarded as a stressfiil life event (Beiser & Wood, 1988;Berry & Kim, 1988). North American models for understanding stress and well-being1would predict that immigrant children and families are therefore at particular risk fordeveloping adjustment difficulties. Empirical fmdings within the adult immigrant literaturelargely support these predictions. Many immigrants have unique mental health challengesrelated to experiences such as language dysfluency, underemployment, and culturaldistance (Padilla, Cervantes, Maldonado, & Garcia, 1988; Trovato, 1986; Vega, Kolody,& Valle, 1987). Existing studies regarding child behavior problems vary considerably,however, in their findings (Aronowitz, 1984). That is, whereas some studies reveal theexpected finding that immigrant children show more behavior difficulties thannonimmigrant children (e.g., Rutter et al., 1974), others have demonstrated that immigrantchildren show fewer (e.g., Touliatos & Lindholm, 1980) or an equivalent number (e.g.,Monroe-Blum, Boyle, Offord, & Kates, 1989) of problems.While methodological inconsistencies may account for some of the discrepanciesacross studies, an alternative explanation is that some children are protected from thestress of immigration by various family-level and/or intrapersonal factors, whereas othersare not, In the present study, a model of family stress that recognizes the potentialcontribution of moderator variables was tested within a sample of new immigrant familiesfrom Hong Kong. It was hypothesized that this model would adequately portray theexperiences of these families, and that children would show more or fewer adjustmentproblems largely as a fi.rnction of the interplay between stress, parent distress, andbuffering variables such as maternal social support. In summary, rather than focussing onwhether or not immigrant children show more behavior problems than nonimmigrants, thepresent study begins to explore the conditions that either facilitate or prevent thedevelopment of child problems following immigration. This chapter of the manuscript is2divided into five major sections. The first section of the chapter describes a basic NorthAmerican conceptual model related to stress and parent-child adjustment. This modeldraws from the health psychology literature as well as from theoretical developments thatare more specific to parenting stress and child outcomes. Empirical evidence regardingeach variable (e.g., stress) in the model is detailed. Secondly, evidence that supports thisBasic Model within immigrant samples is reviewed. In this way it is possible to considerhow various aspects of the model might function in immigrant families. The literatureregarding behavior problems exhibited by immigrant children, relative to nonimmigrantchildren, is highlighted in this section. Methodological and conceptual explanations forinconsistencies across studies are explored. In the third section, the literature concerningresiliency and moderator variables is examined. Fourth, a more complex, ModeratorModel that may better fit the findings within immigrant samples is presented. Aspects ofthis Moderator Model, and the role of maternal social support in particular, are explored indetail. Finally, the Moderator Model is summarized and a series of hypotheses for theproposed research are presented.Basic North American Model of Family Stress and Parent-Child Outcomes:Description and Empirical SupportDescription of Contemporary ModelsAlthough recent models advanced to explain children’s development and mentalhealth recognize the role of biology in child outcomes, social influences have beenincreasingly highlighted. In his ecological model, Bronfenbrenner (1979) articulated the3layers of social systems that interact with the child to influence developmental processes.He suggested that there exist four interrelated environments affecting the child: themicrosystem (i.e., the most proximal, immediate setting of the child (e.g., home, daycare)),the mesosystem (i.e., the relationships between microsystems (e.g., home and school)), theexosystem (i.e., settings that are removed from the child but have an influence ondevelopment indirectly, usually via the caretakers (e.g., parent’s workplace)), and themacrosystem (i.e., broad institutional and ideological patterns of a culture (e.g., religion)).Each of these social systems is said to have a role in shaping the child’s behavior anddevelopment.At the level of the microsystem, contemporary researchers have demonstrated theimportance of the family context in child outcomes. Developmental psychologists havestressed the importance of sensitive and responsive parenting behavior (Ainsworth & Bell,1974; Skinner, 1985) and warm, supportive care-giving (Beckwith, Chown, Kopp,Parmelee, & Marci, 1976; Bradley, Caldwell, & Elardo, 1979) for normal childdevelopment. Theorists have further articulated that receptive child behavior isreciprocally, or transactionally, related to sensitive parent behavior (Bell & Harper, 1977;Hammen, Burge, & Stansbury, 1990; Patterson, 1982; Samerofl 1987). It has beensuggested that when this positive cycle is interrupted or absent, coercive parent-childinteractions may develop, and ultimately lead to child behavior problems (Christensen,Phillips, Glasgow, & Johnson, 1983; Patterson, 1982). Microsystem variables such asparent psychopathology and parental history of childhood abuse are increasingly beingrecognized as pertinent contextual factors in the parent-child relationship (Downey &4Walker, 1992; Rogosch, Mowbray, & Bogat, 1992; Straus, 1980; Watt, Anthony, Wynne,&Rolf 1984).Researchers have also begun to consider the importance of the exosystem as aninfluence on parent and child functioning (Bronfenbrenner, 1979; Sameroff& Seifer, 1983;Webster-Stratton, 1990). Depending upon life circumstances, the exosystem will haveeither a positive or a negative association with parent-child functioning. For example,everything else being equal, the relationship between parent and child will be more strainedin a family that is experiencing serious economic hardship, relative to a family unburdenedby financial matters (Elder, Caspi, & Downey, 1986). In response to the growing demandfor information regarding the identification of children at risk, researchers have sought todetermine which family-level variables may be associated with disruptions in parentingand, subsequently, with negative child outcomes. Examples of exosystem variables relatedto parent-child difficulties include: sociostructural influences such as crime and violence,poverty, and community disorganization (Kessler & Neighbours, 1986; Wilson, 1987) andmore transient stressful life events and hassles (Belsky, 1984; Krech & Johnston, 1992;Patterson, 1983; Webster-Stratton, 1988). In summary, contemporary models that aim topredict child outcomes go beyond individual child characteristics to appreciate the role ofthe family context, and, further, to consider the exosystem variables that influence parentbehavior.The present study is designed to determine the degree to which such modelscapture the experience of new immigrant families. The model that will be consideredappears in Figure 1 and will be termed the Basic Model. In this model, mother-child5interactions are seen as being associated with two often-cited risk variables, extrafamilialstress and maternal distress. In general, the Basic Model predicts that children of motherswho are subjected to high levels of extrafamilial stress or who experience substantialpersonal distress will show more behavior problems than children of mothers who arefaced with a relatively low degree of stress and distress. It is important to note that theexisting literature upon which this Basic Model is grounded has relied almost exclusivelyon maternal reports and experiences. Although father-child relationships are clearlyrelevant, in order to allow for comparison with the existing literature, the present studywill focus on mother-child associations. In addition, it must be clearly stated that the linksto be described in this model are not presumed to be causal. The correlational nature ofthe supporting empirical evidence must be stressed. For example, although much of theliterature suggests that the relationship between parent and child behavior is causal in thedirection of parent to child, it is also recognized that there exists a reciprocal influence andthat children also affect parent behavior (Bell & Harper, 1977; Hammen et al., 1990). Inthe same way, whereas stress and maternal distress are often thought of as impinging onparent-child behavior, aspects of the parent-child relationship can certainly contribute toincreased stress and/or distress (Ahidin, 1990). Rather than representing putative causalpathways, the model that is presented is meant to serve as a heuristic for contemplatingassociations between stress, distress, and parent and child behavior.6Figure 1. Basic Model of Stress and Parent-Child Behavior.Operationalization of the Basic Model and Related Measurement IssuesChild Behavior ProblemsData from the Ontario Child Health Study suggested that child behavior problemsaffect roughly 15% of the school-aged population (Offord, et al., 1 987a). Overall, moreboys than girls show behavioral difficulties (i.e., 15% versus 11%, respectively) (Zill &Schoenbom, 1990). Children who show behaviors that reflect socially/developmentallyinappropriate functioning often continue to experience mental health and social problemsin adulthood (Loeber, 1990). Two empirically derived broad-band syndromes of childbehavior problems have been identified; externalizing symptoms which reflectundercontrolled, acting-out behaviors and internalizing symptoms which reflectovercontrolled behaviors (Achenbach & Edeibrock, 1978). These two types of childproblem have been recognized in children from several cultures (e.g., Hayashi, Toyama, &Quay, 1976; Lambert, Weisz, & Thesiger, 1989; Quay & Parskeuopoulos, 1972; Weisz,Suwanlert, Chaiyasit, & Walter, 1987) suggesting that they represent an etic, or universal,7way of describing child behavior problems. In the present study, informants will be askedto report on a range of child difficulties, drawn from both internalizing and externalizingdomains.Primarily for practical reasons, researchers who study child behavior have reliedmore on adult ratings than on information provided by the child or on observations ofchild behavior. Most often, maternal perceptions have been targeted. There are bothbenefits and limitations to using mothers to gather data about child behavior (Achenbach,McConaughy, & Howell, 1987). On the positive side, because mothers typically haveexperience with the child in a variety of circumstances over a long period of time, theirreports reflect a comprehensive base of information. Interpersonal, intrapersonal, andenvironmental factors, however, may all function to produce biases in the mother’sperceptions (Griest, Wells, & Forehand, 1979; Schaughency & Lahey, 1985), making itimpossible to determine the degree to which perceived child behavior is equivalent toactual child behavior. Investigators have grappled with this issue throughout the parent-child literature (Achenbach, et al., 1987; Forehand, Wells, McMahon, Griest, & Rogers,1982; Jensen, Xenakis, Davis, & DeGroot, 1988) and it has been generally recommendedthat researchers use multiple informants and weigh equally the information gleaned fromeach rater (Loeber, Green, Lahey, & Stouthamer-Loeber, 1991; Piacentini, Cohen, &Cohen, 1992). In the present study, child behavior ratings will be made by the child’smother, another significant adult, and by the child him/herself.8Extrafarnilial StressConsistent with contemporary definitions (Webster-Stratton, 1990), in this modelextrafamilial stress includes demographic hardships, major life events, and daily hassles.This definition excludes interfamilial stressors such as those involved in the marital orparent-child relationship. In keeping with the understanding of stress proposed by Jessor(1979), immigration is seen as a major life stressor that has both acute and chronicsequelae. That is, the immediate consequences of uprooting and settling elsewhere,although clearly difficult, are not the only factors that contribute to the perception ofstress. Post-migration experiences (e.g., language dysfiuency and communicationdifficulties, status inconsistency and un/underemployment, racial discrimination) may beassociated with daily hassles and frustrations that can persist for many months, and evenyears (Kuo & Tsai, 1986; Padilla, et al., 1988). Immigration stressors that are both distaland proximal to the parent-child relationship will be considered in the present study.Maternal DistressMaternal distress has been characterized and measured in a variety of ways. Someresearchers have used classification systems and cutpoints to determine the status ofresearch participants as either distressed or non-distressed (Williams, Tarnopoisky, &Hand, 1980; Wing, Bebbington, & Robins, 1981). Rather than relying on this categoricalapproach, in the present study, distress will be defined as a continuous variable andmeasurement will reflect this dimensional approach.“Distress” in this model is understood to represent the mother’s experience ofdepressed mood, somatic complaints, and/or anxiety. This relatively broad definition9follows from suggestions in the cross-cultural literature that psychological distress may notbe experienced universally in the narrow way that it is understood in Western culture(Draguns, 1980). That is, although some psychological disorders may be classified as eticand appear in similar forms across cultures (e.g., schizophrenia), culture-bound, emic,syndromes have also been identified (e.g., chat or pan, the culturally identified addictionfor sweets or salty-spicy snacks in Northern India) (Tseng & Hsu, 1980). It is importantto be aware that categories of disorder, as ernie constructs, may not exist in the same wayacross cultures. For instance, it has been argued that depression, as defined by Westernresearchers, is culture-specific (Marsella, 1980). Specifically, some investigators havenoted that there are few conceptual equivalents for words expressing the emotionalcomponents of depression in Asian cultures and instead, when Asian patients complain ofexcessive somatic symptoms, they often appear depressed by Western standards(Kleinman & Kleinman, 1985; Kleinman, 1987). Although support for this contention inthe area of depression is not unanimous (Beiser & Fleming, 1986; Mumford, 1989; Noh,Avison, & Kaspar, 1992), it is important that researchers consider the possibility thatmeasures that tap a given construct in North American culture may not be adequate inmeasuring that construct in another culture. In the present study, measures were selectedwith a sensitivity to differences in symptom expression cross-culturally such that theconstruct of depression, as manifested either emotionally or somatically, would becaptured.10Empirical Support for Links within the Basic ModelMaternal Behavior and Child BehaviorResearchers who study clinical populations have found evidence suggesting thatthere is a strong association between parenting practices and child behavior problems. Forexample, it has been determined that a significant amount of the variance involved in thedevelopment and maintenance of child deviance may be accounted for by poor parentalmonitoring and inept discipline (Frick et al., 1992; Patterson, 1986). More specifically,researchers have utilized sequential analysis of observed mother and child behavior todemonstrate that vague nonspecific maternal commands increase child noncompliance(Christensen et al., 1983; Patterson, 1982). There is also a growing body of literature thatlinks childhood depression and anxiety with family environments (Kaslow, Rehm, &Siegal, 1984; Puig-Antich et at, 1985; Stark, Humphrey, Crook, & Lewis, 1990). Hopsand his colleagues (Hops, Sherman, & Biglan, 1990) showed that parental depression andmarital discord are related to the later development of depression in children andadolescents. Parent behavior in these families has been described as punitive, conflictual,and intolerant (Arieti & Bemporad, 1980; Forehand et al., 1988; Puig-Antich et al., 1985).Patterson and Capaldi (1992) noted that there is an established relationship between self-esteem deficits and depressed mood in children, and suggested that coercive interactionswith parents operate as one risk factor in the development of children’s low self-esteem.There is evidence that boys may be particularly at risk when parenting behavior isdisrupted. For example, Hetherington and her colleagues (1982) determined that parentsare more likely to argue in the presence of their sons than in front of their daughters.11Similarly, Dunn and Kendrick (1982) reported that mothers are consistently more punitivein their interactions with their sons than their daughters. Rutter (1992) suggested thatgirls may be more sheltered than boys from disruptions in parenting.In sum, the link between parent behavior and child behavior is well-established inthe literature. In the present study, this association is presumed, although not explicitlymeasured. It is expected that any relationships between extrafamilial stress and childbehavior, or maternal distress and child behavior, are mediated in large part throughdisruptions in parenting.Stress and Mother-Child BehaviorParent-child researchers have begun to investigate the role of stress in parentingbehavior and related child outcomes. Several correlational studies have indicated that highlevels of maternal stress are associated with disrupted discipline practices (Capaldi &Patterson, 1987; Forgatch, Patterson, & Skinner, 1988; Patterson, 1983). Similarly, intheir study of families during the Great Depression, Elder, Caspi, and Downey (1986)showed that fathers who experienced the stress of major economic losses becameincreasingly explosive in their parenting, with a concurrent elevation in levels of antisocialbehavior in the sons. Several studies have also noted that there is a relationship betweenstress and abusive parenting (Belsky & Vondra, 1989; Egelancl, Breitenbucher, &Rosenberg, 1980; Gaines, Sandgrund, Green, & Power, 1978; Straus, 1980). Forexample, Justice and Justice (1976) found that abusive parents had experiencedsignificantly more life change in the year prior to their abusive episode than did non-abusing parents. Studies using behavioral observations also have found that mothers who12reported high levels of negative life stress issued more commands, communicated lessoptimally, demonstrated less nurturance and used more critical or negative physicalbehaviors when interacting with their child than nonstressed mothers (Longfellow,Zelkowitz & Saunders, 1982; Webster-Stratton, 1988; Weinraub & Ansul, 1984;Weinraub & Wolf 1983). Together these studies of stress and parenting point toassociations between stressful family circumstances, parenting behavior, and child behaviorproblems.Available evidence suggests that stress may influence boys and girls to differentdegrees. For example, in a sample of generally low-risk children, Masten and hercolleagues (Masten, Morison, Pellegrini, & Tellegen, 1992) found that, when stressed,boys showed high levels of disruptiveness relative to their nonstressed counterparts,whereas girls continued to display low levels of problem behavior. Similarly, in familysituations involving marital discord or violence, researchers have found that boys are morelikely to show an immediate behavioral reaction than girls (Rutter & Quinton, 1984;Wolfe, Jaffe, Wilson, & Zak, 1985). Rutter and Quinton (1984) noted, however, that thisgender difference narrows over time as the marital discord becomes more chronic.Findings from the health psychology literature suggest that, although significant,the relationship between major life events and somatic and mental health outcomes ismodest, accounting for only a small portion of the variance in symptoms (Rabkin &Struening, 1976). More of the variance in mental and physical health outcomes has beenexplained using minor daily stressors, or hassles (e.g., arguments, job stress), than bymajor life events (DeLongis, Coyne, Dakof Folkman & Lazarus, 1982; Kanner, Coyne,Schaefer, & Lazarus, 1981; Monroe, 1983; Wagner, Compas, & Howell, 1988). These13daily stressors may be discriminated from major life events in terms of their comparativelyhigh frequency, low severity and close temporal proximity to symptomatology. In keepingwith this focus, researchers in the parent-child area have demonstrated that daily hasslesassert more influence than major life events on parent perceptions of children and onparenting behavior (Cmic & Greenberg, 1990; Krech & Johnston, 1992).A number of explanations have been offered for the link between stress and parentbehavior. For example, in an analogue study, Zussman (1980) demonstrated that whenparents were asked to attend to a task and to monitor their children’s play simultaneously,they exhibited a pattern of ineffective “minimal” parenting. That is, when stressed bycompeting cognitive demands, parents withdrew positive behaviors such asresponsiveness, support, and stimulation and increased negative responses such asinterference, criticism, and punishment. The implication is that environmental stressors inthe real world compete with the child for attention and therefore interfere with the adult’scapacity to parent.There also exists a growing literature that documents the impact of stress onchildren, without reference to mediating parent behavior (Compas, 1987; Johnson, 1986).Cross-sectional investigations, considering a wide range of stressors, have consistentlyrevealed modest associations between children’s experience of life stress and heightenedbehavioral and emotional problems (e.g., Muffins, Siegel, & Hodges, 1985; Sterling,Cowen, Weissberg, Lotyczewski, & Boike, 1985). Similarly, prospective longitudinalstudies have demonstrated that child life stress is a strong predictor of later adjustmentdifficulties (Compas, Howell, Phares, Williams, & Giunta, 1989; Siegel & Brown, 1988).14Further, researchers have found that these problems may be evident for several yearsfollowing a incident (McFarlane, Policansky, & Irwin, 1987; Waflerstein & Kelly,1980; Yule & Williams, 1990) and are often of clinical severity (Yule, 1992).Maternal Distress and Mother-Child behaviorMaternal distress has been associated with mothers’ interactions with their childrenand with child behavior problems. For instance, Lobitz and Johnson (1975) found thatelevations in parent responses on several Minnesota Multiphasic Personality Inventory(MMPI) scales were related to both observed parent negativeness toward the child and toobserved child deviant behavior. The existence of maternal depressed mood, in particular,has been repeatedly identified as a factor associated with poor parent-child relationships(Cohler, Grunebaum, Weiss, Gamer, & Gallant, 1977; Cohn & Tronick, 1983; Weissman,Paykel, & Klerman, 1972). Researchers have found that depressed mothers have difficultycommunicating, express overt hostility, and show emotional detachment in interactionswith their children (Cohler et al., 1977; Weissman & Paykel, 1974). Depressive symptomshave also been shown to negatively influence adult perceptions of child behavior (Johnston& Short, 1993). Behavioral observations have provided additional support, indicating thatmothers experiencing a depressed mood tend to use more critical and aversive parentingstrategies than control group mothers (Biglan, Hops, & Sherman, 1988; Forehand,Lautenschlager, Faust, & Graziano, 1986; Hops et al., 1987; Webster-Stratton &Hammond, 1988). In the same way, parent irritability appears to be related to coerciveinteractions with children (Patterson, 1982).15The relationship between maternal distress and mother-child behavior has also beenemphasized through a focus on the psychological flinctioning of mothers of clinic-referredchildren. For instance, investigators have demonstrated that mothers of disturbed children,particularly socially-aggressive children, display more maladjustment on the M1VIPI thando mothers of normal children (Goodstein & Rowley, 1961; Patterson, 1982).Researchers have also found that mothers of clinic-referred children report significantlyhigher levels of depressed mood than mothers of nonclinic children (Griest, Forehand,Wells, & McMahon, 1980; Rickard, Forehand, Wells, Griest, & McMahon, 1981). Forinstance, in studies comparing parents of hyperactive and nonnal children, investigatorshave found that mothers of hyperactive children are more likely to report depressivesymptomatology (Befera & Barkley, 1985; Cunningham, Benness, & Siegal, 1988). Thesestudies serve to demonstrate the link between parent distress and child problem behaviorwithin families of clinic-referred children.There has been limited study regarding the degree to which parental distressinfluences boys and girls differentially. Existing investigations suggest that boys aresomewhat more vulnerable to parent psychopathology than girls (Rutter & Quinton,1984). At the same time, however, researchers have noted that there is an associationbetween the gender of the affected parent and the gender of children who showdisturbance (Rutter, 1966; Rutter & Quinton, 1984). That is, there is a tendency for boysto exhibit behavior problems when fathers are distressed, and a tendency for girls to showproblems when mothers are distressed. Similarly, Hops and his colleagues (1990) foundthat adolescent girls of depressed mothers displayed lower levels of happy affect than their16male counterparts. Further study regarding the differential influence of parent distress onschool-aged children is clearly warranted.Consideration of the Basic Model among Immigrant FamiliesEmpirical Support for the Basic Model from the Immigration LiteratureAlthough some theorists have framed immigration as an opportunity forexploration and personal growth (Adler, 1975), the experience of migration is generallydeemed to be stressilil (Padilla et al., 1988; Walsh & Walsh, 1987). Social-anthropological theorists suggest that the stress is largely experienced in relation to thecultural changes and conflicts inherent in the migration process (Nann, 1982). Accordingto the Fabrega migration model (Fabrega, 1969), stress is involved in leaving a country oforigin, in difficulties in passage, in the adaptation process in the receiving society, and inunmet expectations for social and economic attainment in the new country. Focussing onthe adaptation process within the host country, Berry and his colleagues (Berry, Kim,Minde, & Mok, 1987) identified five categories of change that immigrants may anticipate.The changes may be physical (e.g., new type of housing, more pollution), biological (e.g.,new diseases, new nutritional status), cultural (e.g. new political/economic circumstances,new language), social (e.g., new set of relationships, ingroup and outgroup experiences)and/or behavioral (e.g., new routines, new mental health risks). Specific examples of thestressors experienced by new immigrants include: language difficulties, homesickness,food differences, climatic changes, underemployment, and discrimination (Kuo & Tsai,1986; Padilla et al., 1988).17Like adults, children are often faced with a variety of life circumstancesfollowing immigration. Findings from the recent Ontario Child Health Studydemonstrated that, compared to non-immigrant children, immigrant children are found tobe 3.7 times more likely to live in an urban setting and to live in overcrowded conditions,3.1 times as likely to live in subsidized housing, 2.4 times as likely to have a mother whohas less than an eighth grade education and 1.5 times as likely to experience some familydysfunction (Monroe-Blum et al., 1989). In addition, like their parents, children oftenmust leave behind a way of life, ffiends, and loved ones. Their entry into the new culturecan be confusing and frightening, particularly as they struggle to learn a new language andestablish a new ethnic identity. Furthermore, during this tumultuous time, the haven thatchildren typically would turn to for support, their family, may be unavailable to provideprotection. In fact, family stability may be challenged as parents cope with the demands offinding employment, securing adequate housing, and learning the host language andcustoms (Hicks, Lalonde, & Pepler, 1993). No empirical data exists regarding therelationship between immigration stress and parenting behavior and/or child problems. Itseems clear, however, that immigrant children experience the types of stressors that,according to the Basic Model, are predictive of subsequent mental health problems.Several British researchers have noted that, relative to nonimmigrants, immigrantadults from a variety of countries show higher mental hospital admission rates(Bebbington, Hurry & Tennant, 1981; Coebrane, 1980; Cochrane & Bal, 1987; Glover,1989). In general, immigrants are reported to experience a variety of psychological18problems, including depression, schizophrenia and anxiety-related difficulties (Amaro &Russo, 1987; Cochrane & Bal, 1987; Comas-Dias, 1988; Flaskerud & Soldevile, 1986).Other researchers have noted subclinical adjustment difficulties among inmiigrants relatedto language, employment, and family issues (Committee on multiculturalism and mentalhealth and education, 1989; Padifla et al., 1988).Overall, given that immigration is generally perceived as a stressful life event forboth parents and children, and that immigrant adults tend to experience elevated rates ofpsychological distress relative to nonimmigrants, the Basic Model would predict thatchildren of these distressed parents should suffer ill effects as a result of their parents’diminished capacity to provide nurturing care-giving. The empirical literature describingthe mental health of immigrant children is, however, not wholly supportive of thiscontention. In fact, some studies have found surprisingly lower rates of behavior problemsin immigrant children relative to nonimmigrant children. For example, comparing thebehavioral adjustment of immigrant children from India to British children, Kaflarackal andHerbert (1976) determined that Indian children showed significantly less maladjustmentthan their British counterparts. Similarly, in a large-scale epidemiological study, Touliatosand Lindholm (1980) demonstrated that children in the United States with parents ofChinese, Japanese or Southeast Asian descent, showed significantly fewer behaviorproblems than children with parents born in the United States. Steinhausen (1985) arrivedat a similar conclusion comparing a sample of Greek immigrant children to their WestGerman nonimniigrant counterparts. In a study specific to Attention Deficit HyperactivityDisorder (ADHD), Yao, Solanto and Wender (1988) found that recently immigrated19Chinese-American children showed fewer cases of ADRD than children assessed in othercultures (e.g., North American, Italian). Contrary to prediction, these researchers haveidentified immigrant populations in which child behavior problems are less evident than inthe host population.This is not to say that all immigrant children are free from adjustment difficulties.Several studies have described children of immigrants as showing elevated rates ofbehavior problems compared to nonimmigrant children. For example, Mmdc and Mmdc(1976) found that 26% of the children in a sample of 51 families who moved from Ugandato Canada exhibited psychological adjustment problems. In Britain, Bagley (1972)demonstrated that West Indian immigrant children in this sample showed more behaviorproblems than nonimmigrant children. Likewise, Rutter and colleagues (1974) indicatedthat teachers rated West Indian immigrant children as showing conduct problems twice asoften as nonimmigrant children. Surprisingly, the West Indian immigrant girls displayedrates that approximated those of boys in this sample. No differences between immigrantand nonimmigrant children were evident, however, when parents were asked to rate theirchildrens’ behavior. Other researchers have indicated that immigrant children experiencelanguage-related social and academic difficulties (Taft, 1977a). Heightened problemsamongst immigrant children have also been found when considering behavior difficultiessuch as elective mutism (Bradley & Sloman, 1975), autism (Harper & Wiffiams, 1976),and parent-child interaction difficulties (Christiansen, Thornley-Brown & Robinson,1982).20A third category of studies indicates no significant differences in behavior problemsbetween inmñgrant and nonimmigrant children. Notably, Monroe-Blum and colleagues(1989), analyzing data from the Ontario Child Health Study, found that despite living indisadvantaged conditions, the immigrant children in their sample experienced rates ofbehavioral disturbance that were similar to those of nonimmigrant children. This studyused a large stratified sample, psychometrically-sound measures of child behavior, bothparent and teacher reports, and multivariate analyses. In summary, although the BasicModel is clear in predicting elevated rates of behavior problems amongst child inmiigrants,the empirical literature fails to yield consistent results.There are a host of methodological and conceptual explanations that may accountfor the divergence of findings across studies. First, there have been inconsistencies acrossstudies regarding the way in which immigrant samples have been defined (e.g., child versusparent as immigrant). Similarly, characteristics of the control or comparison group varyacross studies. Although some researchers appropriately stratif’ or match on demographicvariables such as socioeconomic status (Monroe-Blum et al., 1989; Rutter et al., 1974),others describe control groups in a post hoc way (Kallarackal & Herbert, 1976; Touliatos& Lindholm, 1980) or use no comparison group (Yao, Solanto, & Wender, 1988).Secondly, studies vary in the degree to which samples are homogeneous with respect tovariables such as country of origin, refugee status, and time since immigration.Differences in acculturation, or proportion of refugee respondents in the sample, could inpart explain inconsistencies in findings across studies. Finally, variability in the way that21child behavior problems are defined and measured (e.g., parent versus teacher report)could contribute to differences in findings.While these methodological discrepancies clearly exist in the immigrant childliterature, they are not in and of themselves a sufficient explanation for findings contrary toprediction. That is, it is not the case that studies demonstrating fewer immigrant childbehavior problems are systematically different methodologically from those that showhigher rates of problems amongst immigrant children. Conceptual explanations forinconsistencies in findings must also be considered. For example, within any sample ofimmigrant children, whether or not a given child will exhibit adjustment difficulties is likelyto depend on a multitude of factors that are related to, but separate from, immigration. Afamily’s status with respect to risk variables present at pre-migration (e.g., little time forpreparation before leaving), migration (refugee status), and post-migration (no Englishlanguage fluency) (Fabrega, 1969) may be influential in the development or maintenanceof child behavior problems. Viewed another way, certain protective variables may existwhich serve to decrease the likelihood of child behavior problems, thereby creating asubset of resilient children. In a recent review of the literature, Hicks and her colleagues(1993) identified several possible protective variables that might relate to immigrant childoutcomes. These variables include: language competence, social involvement, goodparent adjustment, comfortable socioeconomic status, favourable host reception, and thepresence of an established ethnocultural community. These researchers emphasized,however, that empirical investigations regarding the role of such variables are sparse. Thepresent study aims to make a contribution in this area by examining a model that considers22variables associated with increased and decreased risk for child behavior problems within asample of immigrant children. It is clear from the conflicting empirical evidence reportedthat the Basic Model provides inadequate complexity for explaining the impact of theimmigration experience in parents and children. A more elaborate model that includesprotective, or moderator, variables will be considered.Resiliency and Moderator VariablesIn the parenting literature, although there is clear support for a model in whichstress and parent distress are related to parent-child behavior outcomes, not all childrenwho are exposed to extrafamilial stress and/or to parent mental health difficulties developbehavior problems (Masten & Garmezy, 1985). It has been posited that there existvariables that moderate; that is, buffer or amplify, the effects of stress and parent distresson child outcomes.Resilient ChildrenEven under the most extreme circumstances, there appear to be a group of childrenwho remain psychologically healthy despite the odds against them. Over the past decade,our understanding of the link between stress and child behavior problems has beenenhanced through a shifting of focus from stress-affected youngsters to children who areraised in environments but appear to suffer few adverse effects. By targetingthese resilient children for study, investigators hope to determine which factors moderatethe stress-child behavior relationship (Garmezy, 1985; Wyman, Cowen, Work, & Parker,1991).The study of resilient children has its roots in epidemiological risk research.Children deemed to be at-risk by virtue of having a schizophrenic parent were followed23longitudinally, and, contrary to expectation, investigators noted that a substantialproportion of these children developed normally (Bleuler, 1978; Mednick & Schulsinger,1968). Early studies of resiliency focussed on personal qualities of the child. Forexample, Anthony (1974) concluded that assertive children who exhibited the ability toemotionally distance themselves from the affected parent did not succumb to mentalillness. Beyond identifiing internal protective factors, such as temperament and gender,Rutter (1970) suggested that environmental variables may also protect children fromdeveloping behavior disorders. For instance, he found that children who had a positiverelationship with one parent were less likely to succumb to disorder than children withoutsuch a relationship. Similarly, in their 30 year study on the island of Kauai, Werner andSmith (1982) concluded that children who were invincible in the face of majorenvironmental stressors possessed both internal protective characteristics (e.g., self-helpskills) and exogenous sources of support (e.g., kin). In a review of this literature,Garmezy (1983), identified a triad of protective factors that consistently seemed tocharacterize resilient children. The triad consists of: personality dispositions of the child, asupportive family environment, and extrafamilial support sources. Despite the recognitionthat enviromnental factors play a role in resiliency, most of the empirical literature hasfocussed on child characteristics. Some researchers have noted that this exclusiveemphasis may be misplaced and that fhrther study regarding the role of extraindividualresources is wananted (Anthony & Cohler, 1987). In keeping with this shift towards aconsideration of exosystem and microsystem variables, the present research will emphasizefamily-level, as opposed to child-focussed, protective variables.24Proposed Immigrant Parent-Child Moderator ModelDescription of the Moderator ModelThe parent-child Moderator Model is an extension of the Basic Model presentedearlier. This more complex model shares with the Basic Model the assumption that ifstressors from a variety of sources accumulate, parenting and child behavior may bedisrupted. Uniquely emphasized in this latter model, however, is the influence of variablesthat serve as moderators between stress or parent distress and parent-child outcomes.Several factors; community support, a supportive family system, and the parent’sexperience of a nurturant childhood, have been identified as potential moderators(Webster-Stratton, 1990). The Moderator Model suggests that, depending upon theirvalence, such variables will either amplify or buffer the negative impacts of stress anddistress. That is, according to the model, a parent who experiences strong family andcommunity support, and/or has had a nurturing childhood, may be protected fromdeveloping parent-child interaction problems in the face of stress. In contrast, the modelpredicts that a socially isolated individual might be more vulnerable to parentingdifficulties in stressfhl circumstances. This model allows for the possibility that childrenfrom families under stress may not show behavior problems, if protective factors areoperating.Although any one of several family-level protective variables could have beenselected for examination, maternal social support (i.e., the mother’s perception of supportfrom family and friends) has attracted substantial research attention. For example, there isan expansive literature describing the role of perceived social support in moderating the25effects of stress on physical and mental health (House, 1981; Kessler, 1982). Further,evidence has accumulated in the parenting literature to show that social support serves asa buffer between stress and distress and parent-child problems. Because social support hasbeen established as a viable construct for examining stress-buffering models, it stands outas the variable of choice when attempting to determine the degree to which the ModeratorModel extends to immigrant families.The proposed Moderator Model is displayed in Figure 2. Immigrant children havenot been targeted extensively in the study of resiliency and/or family-level moderatorvariables. Given their high risk status, in terms of extrafamilial stress and parent distress,and the empirical finding that they do not uniformly demonstrate behavioral difficulties, itseems that immigrant children would provide an excellent sample in which to examinestress and child behavior problems in relation to protective variables. This model predictsthat, in this immigrant sample, the relationships between stress and distress and childbehavior will be moderated by maternal social support. That is, among children ofstressed or distressed mothers, those whose mothers believe that they are receiving a highdegree of support will show fewer child behavior problems than those whose mothers feelrelatively isolated.26Figure 2. Parent-Child Moderator Model.Maternal Social SupportEnthusiasm for the notion that social relationships might promote health andprotect individuals from the negative effects of stress was stimulated by two seminalreview papers published in the mid-1970s (Cassel, 1976; Cobb, 1976). Since then,research has accumulated regarding the relationship between social support and healthoutcomes (Berkman & Syme, 1979; House, Robbins, & Metzner, 1982). The mechanismsand processes linking social relationships to health are, however, still unclear. Whereasseveral researchers have postulated that social support moderates the effects of stress onhealth, evidence for the buffering role of social relationships is inconsistent (Alloway &Bebbington, 1987). A lack of clear support for the hypothesis has been associated withmethodological inconsistencies (Frydman, 1981), which have been partly attributed to theEmpirical Support for Links within the Moderator Model27profusion of existing definitions and measures of social support. Researchers haveidentified several aspects of the construct, some of which appear to have a closerrelationship to outcomes than others. For instance, it appears that perceived socialsupport provides a better measure than other dimensions of support such as network sizeand density, or support-seeking (Kessler & McLeod, 1984; Wethington & Kessler, 1986).The bulk of the evidence suggests that perceived support is related to both stress andpsychological distress, and may well function as a moderator of outcomes (House, Landis,& Umberson, 1988). As a result, maternal perceived social support was the focus of thepresent study.There are several reasons to suspect that social support might be beneficial toparent-child relationships and child outcomes. For example, Cochran and Brassard (1979)suggested that in addition to providing companionship, emotional comfort, and tangibleassistance, support networks often serve as models for appropriate parenting behavior.Researchers in the parent-child area have begun to examine the influence of social supporton parenting and child outcomes. In a recent meta-analysis summarizing the findings of 66studies in this area, Andresen and Telleen (1992) reported that both emotional andtangible support were significantly related to maternal behavior. In addition to anassociation with parenting behavior (Crnic, Greenberg, Ragozin, Robinson, & Basham,1983; Weinraub & Wolf 1983), investigators have demonstrated that social support isrelated to child behavior (Blackwell, 1991; Cmic, Greenberg, & Slough, 1986), and tomaternal attitudes towards child behavior (Crnic et al., 1983). For example, Crockenberg(1981) determined that social support is associated with maternal responsiveness to herinfant. Similarly, in an experimental study in which primiparous women were assigned28either to a social support condition, where volunteer coaches provided support andinformation, or to a control condition, Jacobson and Frye (1991) found that infants withmothers in the group with supplied social supports scored higher on attachment ratingsthan control infants. With slightly older children, Blackwell (1991) showed that çy ofdepressed mothers exhibited more externalizing behavior problems, as rated by daycareworkers and mothers, when mothers reported low levels of social support. Apart from anemerging emphasis on the role of social support in parent-child relationships, there exists aconsistent body of literature that links parenting difficulties with social isolation (Wahier,1980; Dumas & Wahier, 1985). In their description of the meta-analysis conducted,Andresen and Telleen (1992) noted that studies to date have focussed perhaps tooexclusively on relatively homogeneous Caucasian samples.Blackwell’s (1991) finding that maternal social support served as a buffer only forboys is one of several studies that has recognized the importance of child gender in testingmoderator models. Using an inner-city sample of Black families, Myers and his colleagues(Myers, Taylor, Alvy, Arrington, & Richardson, 1992) found that maternal distress andfamily stress were related to heightened child problems, and that these relationships weremoderated for boys in families where mothers actively acquired social support andmobilized the family. Interestingly, these active help-seeking strategies actually seemed toexacerbate the relationship between maternal distress and female child problems. Incontrast, other researchers have found that variables related to acquiring support, such asfamily sociability and maternal social competence, are protective exclusively for girls intheir high-risk samples (Masten, et al., 1992; Pianta, Egeland, & Sroufe, 1992). More29study is required in order to determine the degree to which buffering models are applicablefor boys relative to girls.Recognizing that intimate ties are a particularly important source of support(Berkman & Syme, 1979; Coyne & DeLongis, 1986), investigators are beginning to focuson support received from within the family unit. Several studies have illustrated thatfamily support is correlated with positive child outcomes (Holahan & Moos, 1987).Vami, Wilcox, and Hanson (1988) demonstrated in a sample of children with juvenilerheumatoid arthritis that family support was a significant negative predictor of childinternalizing and externalizing problems, accounting for 22% of the variance in each.Other studies have demonstrated the moderating effects that family support asserts onchild outcomes. For instance, Crrnc and colleagues (Crrnc et al., 1983) found that intimatesupport moderated the effects of stress on maternal interactive behavior with infants.Overall, however, the evidence that has accumulated to date suggests that families providean important source of support for both parents and children and that this support maymoderate the effects of stress on parent-child outcomes.Social support has been identified as an important resource for immigrant families.Many social scientists have recognized that the composition of ethnocultural families tendsto follow an extended rather than a nuclear form (Mindel & Habenstein, 1981). It hasbeen suggested that within ethnocultural groups, extended family members fulfil severalroles including assisting with child-rearing, participating in family decisions, sustaining thefamily’s cultural identity, promoting a sense of belonging, and serving as a support system(Almirol, 1982; Raphael, 1988; Wilson, 1986). Harrison and colleagues (Harrison,30Wilson, Pine, Chan, & Buriel, 1990) suggested that this emphasis on community andfamilial interdependency has its foundation in culturally-based ancestral world views. Forexample, collectivity follows parsimoniously from the Chinese teachings that emphasizefilial piety and reciprocity (Suen & Ng, 1987). As such, immigrant families may beparticularly fluent in the provision and receipt of social support.There is evidence to suggest that an established community of like-ethnic membersin the host country may serve as an important resource for newcomers. Ethnoculturalcommunity members can frequently be relied upon for practical assistance (e.g., assistancewith orientation, securing employment and housing) (Sue & Chin, 1983). In addition, thepresence of a familiar community group may ease the cultural transition by providing thenew immigrant with a sense of belonging and identity (Bhatnagar, 1980; Hitch & Rack,1980). As such, the community may be an important source of validation in the newcountry (Ishiyama, 1989). Individuals who do not have access to this type of “ethnicenclave” tend to exhibit higher levels of psychological distress than immigrants who belongto a larger cultural community (Bland & Orn, 1981; Murphy, 1973). There is evidence tosuggest that these enclaves provide protection from stress for children as well as for theirparents (Way, 1985).In summary, in the health psychology literature, perceived social support hasgarnered substantial attention as a moderator variable. With respect to parenting and childoutcomes, it appears that mothers under stress who have access to sources of support intheir environment, and have the ability to utilize these resources, show fewer difficulties inparenting and have children with fewer behavior problems relative to mothers who lack31support. Similarly, researchers have recognized the direct positive influence of support inparent-child interactions. Although social support is widely regarded as a valued aspect ofimmigrant family life, there is no direct evidence that immigrant child outcomes areassociated with maternal perceived support. The present study will empirically examinethe degree to which this social resource serves to moderate the relationship betweenimmigration stress/maternal distress and child outcomes.Additional and Exploratory Links in the Moderator ModelDimensions of Child Behavior: Social CompetenciesChild behavior may be viewed as occurring along a continuum from healthy todeviant. Generally, healthy behavior is understood to encompass more than simply theabsence of deviant behavior; it also includes physical, mental and social well-being(Kazdin, 1995; Terris, 1973). Increasingly, researchers concerned with children’s mentalhealth have been asking questions not only about behavior problems, but also aboutchildren’s strengths and competencies (Cowen & Work, 1988; Masten, Morison,Pellegrini, & Tellegen, 1992; Waters & Sroufe, 1983). This interest has been sparked, inlarge part, by the growing movement towards primary prevention and the related quest toidentify, and to target for early intervention, precursors of mental health problems(Cowen, 1991; Weissberg, Caplan, & Harwood, 1991). Researchers have detennined thatchildren who lack social competencies are at-risk for later life difficulties. In their reviewof this literature, Parker and Asher (1987) found, for example, evidence for a strongassociation between early peer-relationship adjustment and school drop-out status in32adolescence. Similarly, the bulk of the literature supports a link between socialcompetence in childhood and later criminality (Parker & Asher, 1987; Roff, Sell, &Gorden, 1972; Roff& Wirt, 1984). It is widely believed that, even amongst children whoappear to be free from behavioral difficulties, those who exhibit a disturbed capacity formaintaining social relationships are at-risk for later problems. In response to suchfindings, child mental health professionals have devoted considerable attention to thedevelopment of social competence-enhancing programming (e.g., Conduct ProblemsPrevention Research Group, 1992; Rae Grant & Cr111 Russell, 1989; Selman, et al., 1992;Shure & Spivack, 1982; Weissberg et a!, 1991). The thinking is that if children who areexperiencing social problems can be identified and taught the skills necessary fordeveloping and sustaining positive relationships, then the likelihood that these children willcontinue along a pathway towards mental health problems will be decreased.For several reasons, including language dysfiuency and cultural conflict, newimmigrant children may be at particular risk for showing deficits in social relationships.Moreover, given that parents play a large role in children’s acquisition of prosocial skills,through coaching and exposure to positive peer experiences (Parke & Bhavnagri, 1988;Parke, MacDonald, Beitel, & Bhavnagri, 1988), the Basic Model would predict thatstressed immigrant parents would have a diminished capacity to fulfil this role. To date,however, most researchers in the immigrant child area have focussed on detecting thepresence of behavioral difficulties, to the exclusion of assessing child strengths or the lackthereof There may be a category of immigrant children who show no overt behaviorproblems, but who lack the skills for optimal social functioning. These children are at risk,33but are not detected in studies with an exclusive focus on internalizing and externalizingbehavior. In the present study, child social competence will be measured and examined asa secondary dependent variable in the Moderator Model.The Stress and Maternal Distress RelationshipAlthough stress and maternal distress have been presented as independentconstructs in the preceding discussion, it is recognized that they are related (Brown &Harris, 1978). An association has been demonstrated between chronic life strains andmental health problems (Dumas, 1986; Garbarino, 1976; Herrenkohl, Herrenkohi,Toedtker, & Yanushefski, 1984; Longfellow et al., 1982). It is therefore possible that, inaddition to asserting an independent effect on parent behavior by draining resources, stressmay be related to parenting through its association with distress. That is, the effect ofstress on parenting and child behavior may be mediated in part by maternal distress(Conger et al., 1993; Conger, Patterson, & Ge, 1995; Webster-Stratton, 1990).The link between stress and distress has also been articulated in the immigrationliterature. Researchers have identified general and acculturative stress variables that arerelated to immigrant adult distress. For example, using a sample of immigrant Mexicanwomen, Vega and colleagues (1987) found that the variables that best predict depressionincluded stress-related factors such as parent income and education, perceived economicopportunity, and perceived distance between the country of origin and the host country.Similarly, among refi.igees, post-migration stressors that have been identified as beingsignificantly related to adjustment problems include: separation from family members,painful memories about pre-migration and migration experiences, placement with sponsors34that do not share fundamental spiritual/cultural beliefs, and homesickness (Nicassio &Pate, 1984; Westermeyer, 1988). Thus, the studies conducted to date seem to emphasizethe importance of post-migration stressors as influences on adult mental health.As mentioned previously, it has been further recognized in the health psychologyliterature that not only might stress and distress be related, but that this relationship maybe moderated by social support (House, 1981). Applying this stress-buffering model toparenting, researchers have postulated that under stress mothers with high levels ofsupport will show less distress than mothers who are relatively isolated. For example,D’Arcy and Siddique (1984) found that, among mothers of preschool and school-agedchildren, spousal and community social support served a buffering function in relation tothe impact of stress on maternal distress. More recently, researchers have extended this toconsider child outcomes. For instance, Rogosch and his colleagues (1992) conducted apath analysis through which it was determined that the influence of maternal social supporton parenting was mediated by self-esteem. That is, mothers in this study, all of whomevidenced serious psychopathology, regarded themselves more positively when theyperceived that they were receiving emotional support. This in turn contributed to a morefavourable approach to parenting.There is also evidence to support a link between social support and immigrantparent distress. For example, Nakagawa and colleagues (Nakagawa, Teti, Lamb, &Shigemura, 1991) detennined that recently immigrated Japanese mothers who experiencedlow social support, displayed more depression, anxiety and anger than mothers with highratings of social support. Similarly, using a large sample comprised of Chinese, Filipino,35Japanese and Korean immigrants, Kuo and Tsai (1986) found that a lack of social supportwas predictive of acculturative adjustment difficulties. In the same way, Hurh and Kim(1990) found that marital status and family satisfaction were related to Korean immigrants’mental health, and Padilla and colleagues (1988) found that Central American immigrantsidentified the use of a social support network as the most effective coping response whensettling in a new country. Although a stress-buffering role for social support has not beeninvariably detected in such studies, it has been established that support is positively relatedto success in the acculturation process. In the present study, the role of social support inmoderating the relationship between stress and maternal distress will be tested.Summary of the Proposed Parent-Child Moderator Model and Research HypothesesSummary of the proposed modelThe available evidence suggests that there is support for the application of theNorth American Moderator Model of parent stress and child behavior to immigrantsamples. This model posits relationships among stress, maternal distress, and childbehavior problems, which are moderated by protective variables. Of the many possibleprotective factors that could have been included in the present study, maternal socialsupport was selected because there are theoretical reasons to believe that this variable maybe an important source of strength within immigrant families. Many of the relationships inthe Moderator Model, although possessing theoretical justification, are as yet untested inimmigrant samples. For instance, although it seems likely that immigrants experience preand post-migration stress, this variable has not been examined in relation to child behavior36problems. Similarly, despite evidence that inmigrant parents are likely to experiencepsychological distress, the degree to which this is related to child outcomes is not known.It is the goal of the present study to take the literature a step further by examining theserelationships together in the context of making predictions about child behavior outcomes.HypothesesPrimary Hypotheses regarding the Proposed Moderator ModelIt is hypothesized that extrafamilial stress, maternal distress and social support willeach assert a main effect on child behavior problems. Specifically, it is anticipated thatextrafamilial stress and maternal distress will predict child problems, and higher levels ofsocial support will be predictive of lower levels of child behavior problems. Further, inkeeping with the buffer theory, it is postulated that maternal social support will moderatethe relationships between extrafamilial stress and child behavior problems, and betweenmaternal distress and child behavior problems. That is, stress and maternal distress willeach assert less of an impact on child outcomes when perceived social support is high andmore of an effect when support is low. Based on findings scattered throughout theliterature that the results may vary according to child gender, this model will also be testedfor boys and girls separately. No specific predictions are forwarded for this exploratoryanalysis.Both main and interactive effects are predicted for the social support variable.Researchers have noted that it is possible for a moderator variable to serve a bufferingfunction at high levels of stress and to assert a direct effect on outcomes when stress islow (House, 1981; Finney, Mitchell, Cronkite, & Moos, 1984). Wheaton (1985) agreed37that “there is no necessary tension between the stress-moderator and the distress deterrentrole” (p. 360) and noted that there would be a reduction in mental health problems even atlow levels of stress if support was high, and even larger discrepancies in distress betweenthose with and without support as stress levels increased.Secondary Hypotheses regarding the Proposed Moderator ModelTwo additional relationships in the model will be tested in an exploratory manner.First, child social competence will take the place of behavior problems as the dependentvariable in the model. Because no firm hypotheses can be made regarding the degree towhich risk and protective variables will be associated with social competencies, thisapplication of the model is exploratory.Secondly, the association between stress, perceived support and maternal distresswill be investigated as a test of the general stress-buffering model. Consistent withfindings using other samples, it is anticipated that perceived support will moderate therelationship between extrafamilial stress and maternal distress in this immigrant sample. Inaddition, it is postulated that extrafa.milial stress and social support will each assert a directeffect on maternal distress.38METHODSampleInclusion CriteriaTo be eligible for participation in this research project, families were required tomeet three basic criteria. First, only families that had recently immigrated from HongKong were considered. Hong Kong was selected as the source country of interest largelybecause Vancouver School Board records suggested that, in recent years, roughly twice asmany children immigrated from Hong Kong as from any other source country (VancouverSchool Board, 1990, 1991, 1992). In addition, at a theoretical level, unique aspects of thisimmigrant community make it a particularly relevant sample in which to explore the utilityof North American stress and coping models. For example, because Hong Kong is arelatively Westernized, affluent society, and because there is an established Chinesecommunity in Greater Vancouver, it may be expected that immigrants from this countrywould experience less acculturative stress and fewer adjustment problems. While this maybe the case, there are other aspects of Hong Kong culture that challenge this suggestion.For instance, it has been suggested that the cultural distance between Chinese and NorthAmerican societies, particularly in terms of differences in core values, is significant (Sue &Chin, 1983). Also, the value placed on achievement and family honour in Chinese culturemight place additional pressures on families struggling with the realities of loss and cultureconflict inherent in the migration process (Sue & Chin, 1983; Sue & Sue, 1990). Inaddition, for some families within this immigrant community it is necessary that fatherscontinue to reside and work in Hong Kong for much of the year (Ng, 1993).39Participants were drawn from a single source country for both logistical andconceptual reasons. From a practical point of view, the expense involved in translatingquestionnaire packages into more than one language was prohibitive. By way of example,estimates received regarding the cost of translating the questionnaire package into Chinesehovered around 22 cents per word, plus $50 per page for typesetting. Given that thequestionnaire package was 16 pages long and contained roughly 5000 words, and thatlanguages involving more complex script (e.g., Punjabi) were even more costly, gatheringdata from multiple source countries was not possible. In addition, it was not logisticallyfeasible to recruit and train research assistants representing more than one language group.Conceptually, there are both advantages and disadvantages to recruitingparticipants from a single source country. On the positive side, an immigrant sampledrawn from a single source country is more likely to be homogeneous in terms of; forexample, family values, sociopolitical background, cultural beliefs, and lifestyle. Thispermits data interpretation that is based on knowledge of the culture (Foster & Martinez,1995). The converse of this, however, is that findings may not generalize to immigrantsfrom other countries. Given the large number of immigrant children and families movingfrom Hong Kong to Vancouver and the Lower Mainland (Committee on Multiculturalismand Mental Health Education, 1987; Vancouver School Board, 1992), the results of thisstudy are likely to be of practical significance locally, even in the absence of demonstratedgeneralizability to additional cultural communities.A second criteria for inclusion in the research project was that families must haveimmigrated within the past 4 years. Time since immigration has been identified as an40important variable associated with both the experience of stress (Berry, et al., 1987) andwith mental health outcomes (Beiser, 1988). In general, newcomers are thought to be atgreatest risk for developing psychological adjustment problems in the first few yearsfollowing migration (Grinberg & Grinberg, 1984; Tyhurst, 1982). According totheoretical models that describe different stages of adjustment following migration, aninitial honeymoon period is generally followed by a period of cultural conflict and distress,before a state of bicultural resolution is ultimately reached (Lysgaard, 1955; Oberg, 1960;Sue & Sue, 1990). Most estimates identify the 3-18 month period following migration asthe range of time during which emotional difficulties are most likely to surface amongstadults (Beiser, 1988; Rumbaut, 1985; Sokoloff Carlin, & Pham, 1984). Although mostimmigrants ultimately reach some resolution regarding their new cultural identity andlifestyle, for some, symptoms of anxiety and depressed mood persist (Beiser, 1990). Fongand Peskin (1969) found that students from China continued to exhibit symptoms ofdistress 3.5 years following migration. Krupinski (1967) likewise noted that, whereas formen the incidence of schizophrenia was highest I to 2 years post-migration, for womenthe peak occurred between 7 and 15 years later. Parallel studies focussing on the mentalhealth of immigrant children in relation to time since immigration have yet to beconducted. Research that examines children’s emotional responding to stress moregenerally, however, has suggested that while behavior problems may not surface forseveral weeks following a major life event (Goodyer, Kolvin, & Gatzanis, 1987), thesedifficulties can persist for one or more years (McFarlane, et al., 1987). In order to capturethe time period most germane for the development of behavior problems, the present41study included only recent arrivals; families who had lived in Canada for 4 years or less.This also corresponds with the timing involved in becoming a citizen of Canada, which istypically a 3-4 year process.The final criteria for inclusion in the study was that families had at least oneelementary school-aged child who themselves had immigrated from Hong Kong.Canadian-born children of Hong Kong immigrant parents were not eligible to participate inthe study. Children under age 6 or above age 11 were also excluded from this researchproject. An effort was made to include in the study an equal number of boys and girls, andan equal number of younger children (age 6-8) and older children (age 9-1 1). In caseswhere families had more than one eligible child, a target child was selected with a view tomaintaining this gender and age balance, When the accumulating sample was evenlymatched on these dimensions, a random numbers table was used to select the target child.Recruitment of ParticipantsParticipation in this research project was generated via several methods ofoutreach. First, and perhaps most importantly, community leaders and ethnospeciflcagencies were contacted and informed about the study. Several key agencies took aninterest in the project and assisted in publicizing the research to their members (e.g.,United Chinese Community Enrichment Services Society, Chinese Cultural Center, B.C.Association for Chinese Language). In addition, several organizations that house Englishas a Second Language (ESL) programs allowed announcements of the research projectwithin their classes. Approximately 22% of participants (=28)were recruited throughethnospecific agencies and ESL classes.42The study was also publicized within the community of Hong Kong immigrantsmore widely. For example, booths were set up for 2 to 3 day periods at three Asianshopping centers in Richmond, B.C.. At these booths, posters were distributed tointerested patrons and bilingual research assistants answered questions about the study.Several eligible families (=42, or 33% of recruited sample) signed up to participate in thestudy during these display periods.In addition, posters translated into Chinese were placed in Asian communitiesthroughout the Lower Mainland (primarily Vancouver, Richmond, Burnaby, andCoquitlam). Key locations included: libraries, community centers, Asian markets, doctors’offices, and children’s recreation areas. The Chinese media was also approached topublicize the study. Announcements appeared on Chinese television and were heard on alocal Chinese radio station. The research project was also described in detail in a storythat appeared in a popular Chinese newspaper. Many individuals contacted us havinglearned about the study through the media or via posters (=56, or 44% of the recruitedsample). Through this combination of outreach methods, information regarding this studywas deemed to have reached most segments of the immigrant Hong Kong community.Families could indicate their desire to participate in one of two ways, depending onhow they learned of the research project. First, if they were present at a venue in whichthe research was described (eg., Asian mall display), they could write their first name andtelephone number on a sign-up sheet. 1f on the other hand, the family learned of the studythrough a poster or media publication, then they were asked to call a central telephonenumber. An answering machine was always in operation at this number and families were43instructed, in a translated message, to leave their first name and telephone number. Ineither case, a bilingual research assistant then contacted the interested family, providedadditional information about the study, and scheduled an appointment to visit the motherand target child at their home.Sample SizeThe results of a power analysis, using Cohen’s (1985) tables and estimating smallto medium effect sizes, indicated that 108 study participants would be required for theplanned analyses. A total of 164 families expressed an interest in this research project. Ofthese, 123 were eligible to participate and received questionnaire packages. A total of 122children agreed to be interviewed during the home visit. One hundred and eight motherpackages were returned to UBC (an 86% rate of return) but only 104 (96% of returnedquestionnaires) were usable. Three packages were deemed unusable because theycontained an excessive amount of missing data (i.e., more than 20% of questions leftblank). As an aside, smaller amounts of missing data were handled through prorating ofscores. The fourth package was omitted because it was determined that the family didnot, in fact, meet eligibility criteria.Description of the SampleCharacteristics of mothers and children in the study are displayed in Table 1.Briefly, most mothers in the sample were in their mid- to late thirties, and indicated thatthey had lived in Canada with their children, on average, for 17 months. The mean age forchildren who were involved in the study was 8 years. Of the 104 usable questionnairesreturned, 55 of the mothers reported on girls and 49 reported on boys. Mothers indicatedthat they and their children resided here for roughly 11 months of the year, whereas their44spouses lived in Canada for only 9 months of the year on average. Households includedan average of 2 children and a maximum of 4 extended family members. Most mothers(77%), however, reported having no extended family members in the home. Families hadan average of three additional relatives living in B.C..Table 1. Demographic Information for Study Participants.Variable M SD Range37.4 4.0 27-468.4 1.7 6-1116.6 12.8 1 - 4811.7 1.0 3-129.2 4.0 0-12NMother age (years) 102Child age (years) 104Time since migration (months) 101# Months here - mother 102# Months here - father 101# of children 1.8 0.8 1 - 5 102# extended family in home 0.4 0.9 0 - 4 100#relativesinBC 3.4 5.4 0-40 101SES in Canadaa 31.8 12.8 8 - 60 102SESinHongKong 45.0 9.4 16-64 102a SES = socioeconomic status as calculated using the Hollingshead Index of Social Status,where scores of 8-19 are the lowest social strata (e.g., unskilled labourers) and scores of55 - 66 are the highest social strata (e.g., major professionals).45Families’ socioeconomic status since moving to Canada was significantly lowerthan families’ status in Hong Kong (t(101) = -11.8, p(OO1), as measured by theHollingshead Index of Social Status (Hollingshead, 1975). In qualitative terms, familiesmoved, on average, from an upper middle class status (i.e., minor professionals,technicians) to a middle class status (i.e., skilled workers, clerical, sales). Seventy-fivepercent of mothers stated that they had no plans to return to Hong Kong. Amongst thosewho indicated that they intend to go back, the return was anticipated for almost 3 yearshence, on average.Research assistants provided ratings of the mothers’ openness and acculturativestatus following each home visit (see Global Ratings below). According to theseassistants, mothers were reasonably open and comfortable with the research process, inthat they received a mean score of 20 (SD=4. 1) on an instmment with a maximum scoreof 30 points. Also, on a scale from 0 to 6, where 0 is “not at all” acculturated and 6 is“very” acculturated, the mean rating for mothers in this sample was 3.5 (SD=l.4).Research assistants estimated that they spent about 20 minutes with each mother duringthe home visit.As a part of the Social Skills Rating System (Gresham & Elliott, 1990) (seebelow), mothers were asked to describe their children’s problem behavior both prior to,and since, immigrating to Canada. Scores were converted to standard scores based on ageand gender (scale mean is 100, standard deviation is 15). Mothers remembered theirchildren’s behavior to be significantly less problematic in Hong Kong relative to their postmigration adjustment (t(103) = -4.65, <.0O1). The ratings of behavior both here and in46Hong Kong were within the normal range for children of the same age and gender (i.e.,standard scores of 97 and 94, respectively).ProceduresResearch AssistantsBilingual research assistants were recruited for this project for several reasons.First, it was anticipated that the presence of a bicultural individual would enhance families’comfort and increase their willingness to participate in the study. Second, many aspects ofthis study required facility in both English and Cantonese (e.g., translating writtenmaterials, receiving telephone calls, conducting home visits). Finally, as immigrants fromHong Kong themselves, these women could offer insights that would assist withculturally-sensitive data collection.Bilingual research assistants were initially recruited through notices posted in UBCstudent areas. A total of seven undergraduate women, fluent in both Cantonese andEnglish, joined during the first phase of the project. In addition to completingapproximately 10 hours of orientation and training’, these individuals assisted inconducting a pilot study, and served as bicultural consultants as procedural details werefinalized. For the second phase of the project, data collection, six additional bilingualwomen were recruited (i.e., for a total of 13 research assistants).Preparation of Ouestionnaire PackageTranslationBecause it was assumed that many new immigrants from Hong Kong would beunable to respond confidently to questions written in English, the questionnaire package47was translated into Chinese. Participants were given the choice of completing the packagein either English or Chinese (98% selected the Chinese version). In response to widelyheld concerns that a failure to convey the intended meaning often results if words areliterally translated by one individual, a four-step translation process was used. First,bilingual/bicultural consultants reviewed the proposed questionnaires. These consultantswere mental health professionals who serve the Chinese community and wererecommended by the Multicultural Health coalition of the Affiliation of MulticulturalSocial Service Agencies (AMSSA) in Vancouver. Consultants reviewed the items in eachinstrument with a view to cultural appropriateness and ease of translation. This included,for example, identifring items that might offend respondents or might be considered toopersonal by Chinese individuals. Consultants also identified the words and phrases forwhich literal translation would not be possible.Second, the measures were prepared for translation. Brislin (1980) offered severalguidelines for the preparation of the English version of materials to be translated. Forinstance, he stressed the importance of unambiguous wording and suggested the use ofshort, simple sentences of less than 16 words to assist in achieving clarity of meaning. Healso recommended that vague terms, such as “now and then,” be avoided and thatmetaphors and colloquialisms be replaced with plain language. He encouraged the use ofspecific terms (e.g., dog and cat rather than pets) and redundant language (e.g., repeatnouns instead of using pronouns). Of course, if following such suggestions required majorchanges to an original English language questions, then the reliability and validity of themodified measure would have to be determined. In general, researchers have found that48careflully translated instruments retain their psychometric properties (e.g., Noh, Avison, &Kaspar, 1992; Roberts, Vernon & Rhoades, 1989; Solis & Abidin, 1991). In the presentstudy, an attempt was made to select English language measures that had been used cross-culturally in previous investigations, and, without modification, met Brislin’s (1980)criteria. In addition, a central consideration in choosing instruments was that items afreadyconsist of objective terms, rather than vague, global labels that might be misinterpreted.Very few wording changes were suggested by cultural consultants; instead, most of theirconcerns centered on the total length of the questionnaire package and on wording of thenonstandardized demographic and immigration stress questions. These consultants did,however, recommend specific structural changes designed to minimize the chance ofconfusion by enhancing consistency across measures (e.g., use of 4-point response scalesfor all instruments).The third step in the translation process was to have the questionnairesprofessionally translated. This service was provided by a Chinese counsellor at Surrey-Delta Immigrant Services. As part of the service, the package was back-translated byanother Chinese staff member. There were five or six phrases that produced some degreeof discrepancy between translators. These were discussed and the term that best conveyedthe intended meaning was chosen in each case. The questionnaire package was thenindependently checked for accuracy by the Editor of the UBC Chinese Studentsnewspaper, and then checked again by three bilingual research assistants. The assistantsdetected 8 to 10 phrases that could have been better worded and the resulting changeswere made.49Fourth, after the translation was finalized, a pilot study was conducted.Questionnaire packages were presented to mothers who had recently immigrated fromHong Kong with children aged 6 to 11 years. In addition to completing the questionnairesas instructed, mothers were asked to indicate any terms that were difficult to understand,grammatically incorrect, or culturally insensitive.Pilot StudyThe pilot study was conducted at a Chinese Language School in Richmond, B.C..Mothers who consented to participate received the package of questionnaires in astamped, addressed envelope. They were asked to return the package to UBC aftercompleting the questions at home at a quiet time. Of 30 packages distributed, 18 weremailed back (a 60% rate of return).On the basis of results from the pilot study, one major change in instrumentationwas made. It became clear from the responses that mothers were confused by the formatof the social competence measure proposed. In addition, they endorsed relatively fewitems on the measure tapping child behavior problems. Based on these findings, twochanges were made. A simpler, more sensitive measure of social competence wasemployed, and when presenting the measure of child behavior problems research assistantsmade a statement intended to normalize the existence of these child problems so thatmothers would feel more comfortable endorsing problems.Ivlinor changes to the package were also made on the basis of suggestions made bypilot study mothers. For example, one item on the child behavior problem measure wassupposed to read, “allergies,” but it became clear by mother’s examples (e.g., “afraid that I50talk about his faults in front of others”) that the translation could also be interpreted tomean “sensitivities.” In addition to wording changes, the layout of the package wasaltered somewhat to facilitate easy completion. Also, a sample item was provided for thesocial support measure because it had a slightly different format from other instruments inthe questionnaire package. Overall, there appeared to be no concern regarding the contentof questionnaire items. Most mothers completed even those items that might beconsidered sensitive.Data Collection ProceduresThe main source of data for this research initiative was the questionnaire packageto be completed by mothers. Because this package was lengthy, and contained items of apersonal nature, written questions were favoured over an interview format. At the sametime, a personal approach to questionnaire distribution was deemed necessary forincreasing interest in the study and for ensuring comprehension of items. As a result,questionnaire packages were hand-delivered to participant& homes at a pre-arranged time.Counterbalanced packages were prepared and placed in stamped, addressed envelopeslabelled with subject numbers. During the visit, details of the study were explained (e.g.,means for maintaining confidentiality, etc.), and mothers agreeing to participate signed atranslated consent form. General instructions (e.g., complete the package in one sitting, ata quiet time) were then followed by specific guidelines for completing each of thequestionnaires. Packages were left with mothers and they were asked to return them bymail to UBC as quickly as possible. Immediately after the interview, the research assistantmade global ratings of the mother (see “Global Ratings’ below). Names and telephone51numbers of several research assistants were provided in case questions arose as motherswere completing the questionnaires. When envelopes were received, packages thatincluded payment of $10, a list of local community resources, and a note of thanks, weresent to families.A second source of data was the child’s own perceptions of immigration stress andhis/her adjustment in Canada. During the home visit, the child was asked to complete abrief structured interview that included a standardized measure of child behavior andrationally-derived questions regarding immigration stress. In order to increase thelikelihood of unbiased responding, this interview took place in a separate location relativeto the meeting with the child’s mother. The rationale for the study was explained to thechild and the interview proceeded with his/her assent (only one child refused toparticipate). If they asked, children were informed that their responses would be sharedwith their parents in a general way if they requested feedback. Afterwards, a series ofglobal ratings of the child’s adjustment were made (see “Global Ratings” below).In addition to data gathered through the mother and target child, an adult familiarwith the child’s daily behavior was asked to complete the rating of child adjustment. Themother was asked to select this person (e.g., child’s father, a family friend, a teacher, etc.)and to give him/her the questionnaire to be completed in private. Eighty-nine percent ofmothers found an appropriate individual to complete this form. Of forms received, 57%were completed by the child’s father, 33% were completed by another family member, and10% were completed by a family friend or professional in the community.52MeasurementMothers’ Questionnaire Package2Family and Immigration InformationFamily Information Ouestionnaire. This brief measure (Appendix A), created forthe present study, was designed to gather information pertinent in defining the sample.Mothers were asked to record, for example, their date of birth and date of arrival inCanada, their educational level, and their occupation here and in Hong Kong. They alsoprovided basic information about their children (e.g., genders, ages, date of arrival inCanada). Finally, one item from the Mother Immigration Stress Scale (below) appearedon this measure. This item pertains to changes in family income subsequent to migrationand seemed most appropriately placed among other questions regarding employment.Extrafamilial StressThe Hassles and Uplifts Scale- Revised (Hassles Scale). Mothers were asked tocomplete a modified version of the 53-item revised Hassles and Uplifts Scale (DeLongis,Folkman, & Lazarus, 1988). Because Uplifts were not a focus of interest for the presentstudy, only the Hassles portion of the scale was presented to mothers. In order tocomplete the modified Hassles Scale, mothers indicated how much of a hassle, from 0(none or not applicable) to 3 (extremely), each item was on the day of questionnairecompletion. In addition to a total score, it is possible to calculate both the frequency andseverity of hassles experienced that day, as well as factor scores. In the present study,only the total score was used in the main analyses. According to DeLongis (1985), thetotal hassles score of this scale is characterized by high autocorrelations (.77 from day to53day) indicating a stability of stress ratings over time. Internal consistencies ranging from.80 to .93 were reported for these subscales. Validity of the scale has been demonstratedin terms of predicting health outcomes (DeLongis, et al., 1982) and psychologicalsymptoms (Kanner et al,, 1981).The Hassles Scale was chosen over other stress measures for several reasons.First, hassles appear to be a better predictor of parent perceptions of child behavior thanare major life events (Crnic & Greenberg, 1990; Krech & Johnston, 1992). Similarly,cross-cultural researchers have suggested that, although the migration experience in itselfmay be stressfiil, it is the more proximal daily stressors associated with post-migrationadjustment that affect mental health (Berry & Kim, 1988; Berry et al., 1987). Second, therevised version of the Hassles Scale has eliminated items potentially confounded withhealth or psychological outcomes. Third, the items in the Hassles Scale are clear, simple,short and objective, allowing for straightforward translation. In addition, the HasslesScale retains its good psychometric properties when it is applied cross-culturally (Lepore,Palsane, & Evans, 1991).Minor changes were made to this scale on the basis of recommendations fromcultural consultants. First, rather than asking participants to circle a number from 0 to 3 torepresent their response, each option (i.e., “none/not applicable, somewhat, quite a bit,extremely) was written out in fi.ill for every item. This simpler, repetitive format wasdeemed more appropriate for this immigrant sample. Secondly, the options for respondingdiffered slightly from the Hassles Scale traditionally used. In the revised version(DeLongis, et al., 1988) the four options were: “none or not applicable,” “somewhat,”54“quite a bit,” and “a great deal.” In the present version, the option, “a great deal” wasreplaced with “extremely” because the original term posed difficulty for translation.Finally, mothers were asked to reflect on hassles over the past week, rather than merelyduring the past day, as is the procedure on the revised Hassles Scale. This modificationwas made on the recommendation of cultural consultants to be consistent with othermeasures included in the package, thereby simplifying the demands placed on respondents.Five items from the Mother Immigration Stress Scale (Appendix B) appeared atthe end of the Hassles Scale because they were well-suited to the format of this scale.These items contributed to the score of the Mother Immigration Stress Scale and not tothe Hassles Scale. The overall reliability (alpha) of the version of the Hassles Scale used inthis study was .92.Immigration Stress Scale. Mothers were asked to respond to 15 questionsdesigned to assess the degree of immigration stress that the family had experienced(Appendix B). Ten of these questions refer to the mother’s experience of stress followingimmigration (termed Mother Immigration Stress Scale). The dimensions included are:pre-migration stress, language difficulties, status inconsistency (i.e., a decline insocioeconomic status upon immigration due to inability to obtain employment in theoccupation in which the individual was trained in the country of origin), homesickness,perceived incompatibility between own and Canadian culture, and racial discrimination.Of the many possible stress-eliciting aspects of immigration to consider, these variableswere chosen because they have received empirical attention (Padilla et al., 1988; Vega etal., 1987). Questions regarding language issues, homesickness, culture incompatibility,55and racial discrimination appeared on the Hassles Scale (see above). These items wereselected from a measure of daily social stress that was developed for immigrantrespondents (Kuo & Tsai, 1986). The status inconsistency question appeared on theFamily Information Questionnaire (above) and remaining items regarding generalmigration stress and language fluency were included on the questionnaire labelledImmigration Stress Scale.The questions were brief and straightforward and required a response on a 4-pointscale that used construct-specific anchors. For instance, mothers were asked to indicatehow much stress they experienced during various stages of the immigration process byendorsing either “no stress”(O), “a little stress”(l), “quite a bit of stress”(2), or “extremestress”(3). Similarly, in reporting their perceived English language fluency, mothers wereto endorse either “fluent”(O), “good”(l), “fair”(2), or “poor”(3). Scores were totalled andhigher scores always reflected more stress. The internal consistency coefficient for thisexploratory scale was .66.In addition to self-reported immigration stress, mothers were asked to comment ontheir child’s experience of stress. Five child-related items, collectively termed the ChildImmigration Stress Scale in subsequent analyses, also appeared on the Immigration StressScale. Mothers were asked to indicate the level of stress that their child experiencedgenerally during immigration, and then to rate his/her English language fluency andexperience of racial discrimination and homesickness. A 4-point scale was also used forrecording these aspects of child immigration stress. Scores were totalled and high scores56reflected higher levels of stress. The internal consistency coefficient for this scale was only.53.Parent DistressCenter for Epidemiological Studies - Depression Scale (CES-D). The CES-D(Radlofl 1977) is a 20-item scale that measures the major dimensions of depression.Radloff (1977) reported that the scale is composed of four symptom clusters: negativeaffect, positive affect, interpersonal problems and somatic or retarded activity. Only thetotal score was used, however, in the present study. Originally created for community use,the CES-D has been widely used with both clinical and community samples (Aneshensel &Frerichs, 1983; Myers & Weissman, 1980; Roberts & Vernon, 1983). According toRadloff (1977), this dimensional scale has demonstrated good reliability and validity. Forinstance, in terms of internal consistency, she reports an alpha coefficient of .85 for thetotal score in community respondents and .90 in patients, and test-retest reliabilities of .51to .59 over periods of 2 to 8 weeks. With respect to validity, Radloff (1977) found thatCES-D scores significantly discriminated psychiatric inpatients from general populationsamples and were sensitive to improvements following treatment. In addition to goodpsychometric properties, this measure was chosen to reflect adult depressed mood in thepresent study because it is comprised of objective items and has been shown to retain itspsychometric qualities when translated for use in other cultures (Nob, et a!., 1992; Robertseta!., 1989).Mothers were asked to indicate their responses on a 4-point scale. For clarity, aswith the Hassles Scale, rather than using a number code short form, each of the four57possible responses was written in fill after each item. The range of responses was: “noneof the time or rarely (less than 1 day a week),” “a little of the time (1-2 days),” “amoderate amount of time (3-4 days)” and “most or all of the time (5-7 days)”. Theinternal consistency of this scale in the present study was .89.Symptom Checklist-90-Revised (SCL-90-R): Somatization and Anxiety subscales.The SCL-90-R (Derogatis, 1983) is a 90-item self-report inventory that provides ameasure of psychological distress in both community and clinic samples. It consists ofnine empirically-derived primary symptom dimensions, including Somatization and Anxietysubscales. For each of the symptom dimensions, respondents are asked to indicate, on a5-point scale ranging from 0 (not at all) to 4 (extremely), the level of discomfortexperienced as a result of each item over the past week.The Somatization scale was selected for use in the present study so that somaticexpressions of depressed mood, reportedly common in Asian cultures, could be measured.The scale consists of 12 items and reflects distress arising from perceptions of bodilydysfhnction. Items on the Somatization scale are simple and straightforward. Derogatis(1983) indicated that this scale is reliable in that it demonstrates high internal consistency(.86) and test-retest reliability (.86). The Anxiety scale was included in the present studyin order to capture this domain of thoughts and feelings that has been found to beassociated with the immigration experience (Berry & Kim, 1988). The Anxiety dimensionon the SCL-90-R consists of behavioral (e.g., trembling) and cognitive (e.g., the feelingthat something bad is going to happen to you) items. Again, the wording is simple andunambiguous. The internal consistency of the scale is .85 and the test-retest reliability is58.80 (Derogatis, 1983). With respect to validity of the entire scale, studies have indicatedthat the SCL-90-R demonstrates concurrent validity with other measures (Derogatis,Rickels & Rock, 1976) and sensitivity to changes due to intervention or experience (Egan,Kogan, Garber, & Jarrett, 1983). In addition to its strong psychometric properties, theSCL-90-R was deemed appropriate for the present study because an effort was made inthe original item development to utilize a basic level vocabulary to convey meanings.Further, the SCL-90-R has been translated into 20 languages and used extensively incross-cultural research (Kim, Kim & Won, 1983; Simoes & Binder, 1980).The SCL-90-R was modified slightly for use in the present study. Specifically, inresponse to suggestions made by cultural consultants, a 4-point scale replaced the original5-point scale so that study participants would not be conli.ased by scale variations acrossmeasures. Response options included: “not at all, l “a little bit,” “moderately,” and“extremely.” As with the other stress/distress measures, these response options werewritten in full after each item rather than using a number code short-form. A total scorewas created by summing scores on both Somatization and Anxiety subscales. The internalconsistency coefficient for these combined subscales was .85 in the present study.Social SupportMultidimensional Scale of Perceived Social Support. The Multidimensional ScaleOf Perceived Social Support (Zimet, Dahlem, Zimet, & Farley, 1988) is a 12-item scalethat measures the perception of support available from family members, friends, and asignificant other. Tn keeping with the evidence that perceptions of support are critical(Barrera, 1981; Sarason, Levine, Basham, & Sarason, 1983), Zimet and his colleagues59(1988) have developed a measure that taps an individual’s perception of the adequacy ofthe support received from important others. To complete the Multidimensional Scale OfPerceived Social Support, respondents circle a number along a 7-point rating scale thatranges from “very strongly disagree”(l) to “very strongly agree”(7) in response to eachstatement. Confirmatory factor analysis revealed three subscale groupings: perceivedsupport from family, from friends, and from a significant other. In the present study, onlythe total score was used in the analysis. Zimet and colleagues (1988) reported goodinternal consistency (alpha=.88) for the scale as a whole, and for each subscale (i.e., forthe Family, Friends, and Significant Other subscales, alpha values were .87, .85, and .91,respectively). The test-retest reliability for the whole scale was .85, indicating adequatestability. Moderate construct validity was also demonstrated as scores on theMultidimensional Scale Of Perceived Social Support were significantly negatively relatedto depressive symptomatology as measured by the Hopkins Symptom Checklist(Derogatis, Lipman, Rickels, Uhienhuth, & Covi, 1974). Relative to many other socialsupport measures, this scale offers clear simple instructions, brevity, and readilytranslatable items.Minor changes were made to the Multidimensional Scale Of Perceived SocialSupport for its use in the present study. First, a sample question was included so thatparticipants could see how to complete this type of scale, which was different from othersin the questionnaire package. Secondly, in the present version, following items tapping theadequacy of support received from a special person, participants were asked to60indicate their relationship to this individual (e.g., spouse, friend, sister, etc). The internalconsistency for this scale in the present study was .94.Parent Support Scale. This scale, designed for use in the present study, focusseson perceptions of the adequacy of support received specifically within the parentingdomain (Appendix C). The scale taps three potential sources of support: family, friends,and spouse. Participants were asked how often they receive parenting support (e.g., childcare, discipline advise) from these sources. Responses were made on a 5-point scale thatincludes options, “never”(O), “rarely”(l), “sometimes”(2), “quite a bit”(3), and“regularly”(4). Participants were also asked how helpflil they deem the support receivedfrom each of these sources to be. Again, a 5-point scale was used, ranging from “not at allhelpfiul”(O), “rarely helpflui”(l), “sometimes”(2), “usually helpfiul”(3), to “extremelyhelpflul”(4). Only the total score was used in the present study. The internal consistencycoefficient for this scale was .70.Child BehaviorChild Behavior Checklist (CBCL). The CBCL (Achenbach, 1991) is a frequentlyused, standardized measure of the behavior problems and competencies of children. Bothinternalizing and externalizing syndromes are represented on the CBCL and are measuredin the form of T-scores. Respondents rate the target child on a 3-point scale, ranging from0 (not true (as far as you know)) to 2 (very true or often true), for each of 113 problemitems. From these ratings, it is possible to determine the degree to which the child exhibitsproblem behaviors relative to an age- and gender-appropriate normative sample.The CBCL possesses good psychometric properties. With respect to internalconsistency, Achenbach (1991) reports that the total problem behavior score has a61Cronbach’s alpha of .96. The Internalizing scale alpha coefficient for girls is .90 and forboys is .89. The alpha coefficient for the externalizing scale is .93 for both girls and boys.Using a 1 week interval, the test-retest reliability of the internalizing scale is .89 and theexternalizing scale is .93. After 1 year, this reliability drops to .82 for the internalizingscale and .86 for the externalizing scale. In terms of validity, Achenbach (1991) providesevidence that the CBCL discriminates significantly between clinic-referred and nonreferredchildren and is associated with other scales designed to tap child behavior problems. Inaddition to these psychometric strengths, the CBCL is a desirable measure for the currentstudy because items are short and objective. The measure has been applied cross-culturally with success (Bird et al., 1987). The total behavior problem I-score, based onage and gender norms, was used in the present study. The internal consistency coefficientfor the CBCL in this sample was .95.Social Skills Rating System. The Social Skills Rating System (Gresham & Elliott,1990) focusses primarily on positive child behaviors deemed important for initiating andmaintaining successfiul relationships with others. The social skills section of the Parentversion has 38 items and includes four subscales: Cooperation, Assertion, Responsibility,and Self-Control. Only the total score was used in the present study. Parents were askedto indicate how often their child shows each positive behavior listed on a 3-point scale(i.e., “never,” “sometimes,” “very often”). Total social competence scores were created bysumming items and determining respective standard scores based on age and gender norms(i.e., Social Skills Rating System mean is 100 and standard deviation is 15).62According to Gresham and Elliott (1990), the social skills section of the SocialSkills Rating System-Parent form for Elementary students has a coefficient alpha of .87and a test-retest reliability correlation of .87. In terms of validity, there is a moderatecorrelation between the Social Skills Rating System total score and the CBCL socialcompetence score. In addition, the Social Skills Rating System reportedly differentiatesgroups of handicapped from nonhandicapped children (Gresham & Elliott, 1990).The Social Skills Rating System was modified slightly for use in the present study.For example, parents were not required to make an importance rating for each social skillas they are in the original Social Skills Rating System. This request was omitted in orderto minimize the complexity of the scale. The internal consistency for this scale in thepresent study was .91.Children’s InterviewStressStress Thermometer. This scale was designed for the present study to providesome assessment of the child’s own perception of the stress involved in immigration(Appendix D). The child completed these ratings with the help of the research assistantduring the home visit. To complete the Stress Thermometer, the child was asked to drawa line on a picture of a thermometer to indicate the amount of stress s/he experienced incoming to Canada from Hong Kong. The thermometer was 20 cm high and the child’sscore was the number of centimeters from the bottom that s/he placed the line. Visualanalog scales have been shown to be a usefhl way of gathering data from children (Jay,Ozolins, Elliott, & Caldwell, 1983).63Scale of Children’s Stress. This series of questions was meant to parallel thoseappearing on the Immigration Stress Scale (Appendix E). The four questions tappedaspects of immigration such as: homesickness, English language fluency, anddiscrimination. The children’s responses were scored as indicating either high stress, lowstress, or medium stress. For example, in response to the question, “Do you like living inCanada or in Hong Kong better?” the response “Canada” received a score of 0, “HongKong” received a score of 2, and “both the same” was scored a 1. The internalconsistency of this exploratory scale was only .24.Child BehaviorChild Rating Scale. The Child Rating Scale (Hightower, Spinell, & Lotyczewski,1990) was developed for use in the Primary Mental Health Project at the University ofRochester. This 24-item instrument assesses children’s perceptions of their ownfunctioning, particularly within the classroom setting. Children rate on a 3-point scale(i.e., usually no,” “sometimes,” “usually yes”) the degree to which each statement appliesto him/her. Children’s responses may be grouped into four empirically-derived subscales:Rule compliance, Anxiety, Peer Social Skills, and School Interest. For the purposes of thepresent study, item scores were summed to produce a total score. Hightower andcolleagues (Hightower et al., 1987) reported that the Child Rating Scale has moderatelyhigh internal consistency (alpha=.78) and satisfactory test-retest reliability (L=.74 over a10-week interval). In terms of validity, the authors indicated that the scale is capable ofdiscriminating groups known to differ in adjustment and shows convergent and divergentvalidity with other measures of child adjustment (Hightower, et al., 1987). The Child64Rating Scale is also a valued measure for the present study because it is brief and is easilytranslated. The order of presentation, between Child Rating Scale and the stress ratings,alternated across children. The reliability (alpha) for this scale in this sample was .75.Global RatingsMother Global RatingsAfter the home visit, research assistants provided global ratings of the mother’sinterview behavior (Appendix F). This rating system was developed so that, in theabsence of a social desirability measure, some assessment of each mother’s openness andmotivation to participate in the study was available. In addition to noting the duration ofthe meeting, research assistants rated, on five 7-point likert scales, each mother’s comfortwith the research project/home visitors. Included on these scales was an item regardingthe mother’s acculturative level. Ratings were reverse-scored as required and weresummed to produce a total openness score. The internal consistency of this measure wasonly .43.Observer-Child Rating ScalesFollowing the home visit, research assistants made ratings of the child’s behaviorduring the meeting. Global ratings were made on five 7-point likert scales (Appendix G).Ratings were summed to produce a total score. During the orientation phase, using asimple coding manual developed for this task3, research assistants practiced making suchratings on videotapes of Clinic children and child actors. This measure was designed toprovide a global impression of the child’s adjustment. No attempt was made to establishreliability across raters.65RESULTSPrimary Analyses regarding the Moderator ModelSummary StatisticsMeans. Standard Deviations, and RangesThe means, standard deviations, and ranges for the variables in the proposedModerator Model are presented in Table 2. For several of the measures used, the meanscores recorded in this Asian immigrant sample can be compared to the normative mean ascited for non-immigrant community samples. For example, Radloff (1977) reported amean overall score of 9.25 (SD 8.58) for the standardization sample on the Center forEpidemiological Studies- Depression scale (CES-D). In a community sample of singlemothers (N = 66), a presumably stressed group, the mean score was 13.89 (SD = 10.67)(Krech & Johnston, 1992). Noh and his colleagues (1992) determined that, despitesuggestions that Asians underreport psychological distress, the mean score for theircommunity sample of immigrant Korean-Canadians (N = 86) was 14.71. Using thesemeans as points of comparison, it appears that the present sample reported a moderatelevel of distress (i.e., M 15.94). Further, contrary to the predictions of several cross-cultural researchers (Kleinman & Kleinman, 1985), somatic items on the CES-D were notdifferentially endorsed over items reflecting psychological symptomatology. In fact, theitems most frequently endorsed in a depressed fashion by mothers in this sample werethose reflecting a lack of positive affect (i.e., there was a tendency among participants toanswer none of the time or rarely or a little of the time to items such as “I felt that I wasjust as good as other people,” “I felt hopeful about the future,” and “I was happy.”). Noh66Table 2. Means. Standard Deviations, and Ranges for Variables in the Moderator Model.Variable M SD RangeAdult RatingsaHassles Scale 20.62 13.66 1 - 62Mother Immigration Stress Scale 11.10 3.83 2 - 20Child Immigration Stress Scale 4.77 2.09 0 - 11CES-D 15.94 9.66 0-41SCL-90-R Subscales 7.45 5.78 0- 26MSPSS 60.98 15.63 18-84Parent Support Scale 13.51 3.73 5 -23CBCL - Mother T Score 53.03 11.27 24 - 75CBCL - Other Adult 1 Score 51.09 11.34 24 - 73Observer-Child Rating Scale 23.46 4.40 2 - 30Child RatingsbStress Thermometer 6.74 5.19 0 - 20Scale of Children’s Stress 3.76 1.45 0- 8Child Rating Scale 62.40 5.54 46- 71Note. CES-D = Center for Epidemiological Studies - Depression Scale, SCL-90-RSubscales = Somatization + Anxiety Subscales of the SCL-90-R, MSPSS =Multidimensional Scale of Perceived Social Support, CBCL = Child Behavior Checklist.a N104, except for CBCL - Other Adult where N=97 and for Observer Child RatingScale whereI 122. ‘ N=121, except for Child Rating Scale where N=1 19.67and his colleagues (1992) found a similar pattern of responding in their research involvingthe Korean community. These researchers suggested that this tendency to responddysphorically to positively framed items may reflect an aspect of Asian culture that existsindependently of depressed mood.With respect to the Multidimensional Scale of Perceived Social Support, Zimet andhis associates (1988) provided normative information that was based on the responses of across-section of college students ( = 275). A mean total score of 69.56 was reported.These researchers acknowledged that this score indicates a relatively high level of socialsupport and noted that, for any given item, few of the respondents rated the support thatthey received as less than 3.5 on a 7-point scale. Although the mean score for the presentsample was also rather high (i.e., M = 60.98), there was a range of scores reported,suggesting that some of the mothers felt truly unsupported. In the college sample,students reported receiving approximately equal levels of support from friends, family, anda significant other. Mothers in the present study were similarly balanced with respect tosource of support (i.e., family (M = 20.91), significant other (M = 20.87), friends (M19.25)).Several of the child ratings may also be compared to a normative sample. Forexample, the Child Behavior Checklist (CBCL) scores describing children in the presentstudy were converted to I-scores based on child age and gender (Achenbach, 1991).Scores of 53.03 for mother ratings of child behavior and 51.09 for other adult ratingstherefore fell about the normative mean. Like the scale total, scores for both theinternalizing and externalizing syndromes also fell about the normative mean for ratings by68both the mother (i.e., 51.29 and 50.29, respectively) and another adult (i.e., 50.45 and49.91, respectively). The five problems most frequently endorsed on the CBCL for thissample were: “argues a lot,” “can’t concentrate, can’t pay attention for long,” “can’t sitstill, restless, or hyperactive,” “clings to adults or too dependent,” and “complains ofloneliness.” For the Child Rating Scale, although no total score comparisons are available(because scores were not totalled in the normative study), subscale scores can becontrasted to an age and gender matched normative sample (Hightower, et al., 1990). Inthe present study, the mean Anxiety/Withdrawal score fell at the 57th percentile, RuleCompliance at the 60th percentile, Social Skills at the 61St percentile, and School Interestat the 71St percentile, relative to the normative sample.North American comparison samples are also available for the Hassles Scale andfor subscales of the Symptom Checklist-90-Revised (SCL-90-R). Modifications made tothese scales for the present study, however, preclude a comparison between these normsand the current sample. For example, the instructions on the version of the Hassles Scaleused in this study asked mothers to think about stressors that had occurred over the pastweek, rather than over the past day. This modification was made on the advice of culturalconsultants who indicated that participants might have difficulty with different time framesacross measures. They suggested that to be consistent mothers should always be askedabout stress/distress over the past week. With respect to the frequency and severity ofhassles reported on the Hassles Scale, mothers indicated that they experienced, onaverage, 17 items as stressful in the I week period surveyed. They rated these hassles asonly mildly severe (i.e., 1.19 on a 4-point scale ranging from 0 to 3). The five items most69frequently endorsed as stressfiul were: “enough money for necessities,” “your spouse,”“eating (at home),” “your children,” and “time spent with family.”The subscales of the SCL-90-R were also altered too much to allow for cross-cultural comparison. Consultants recommended that a 4-point scale be used, rather thanthe standard 5-point scale, in order to be consistent with other measures of stress anddistress. On the subscale tapping somatic concerns, the mean was 4.97 (SD = 3.83), witha range of 0 to 16. A maximum possible score on this subscale would have been 36. Inresponse to the prompt, “How much were you distressed by:”, the three most frequentlyselected items were: “soreness of your muscles,” “headaches,” and “feeling weak in partsof your body.” The mean score for the anxiety subscale was 2.48 (SD = 2.87), with arange of 0 to 11. The highest possible score on this subscale would have been 30. Thethree most frequently selected items were: “nervousness or shakiness inside,” “feelingtense or keyed up,” and “feeling fear.thl.”Finally, it is obvious that no North American comparison group exists for measuresdeveloped for use in this study (i.e., Mother Immigration Stress Scale, Child ImmigrationStress Scale, Parent Support Scale, Stress Thermometer, Scale of Children’s Stress,Mother Global Ratings, Observer-Child Rating Scale).IntercorrelationsTable 3 shows the intercorrelations amongst the variables in the proposed model.Note that each of the mother-rated independent variables was significantly related to childbehavior problems in the hypothesized direction. That is, measures of extrafamilial stress(i.e., Mother Immigration Stress Scale, Child Immigration Stress Scale, and Hassles Scale)70Table3.Zero-orderCorrelationsamongstVariablesintheModeratorModel.—II-Note.ScalenameshavebeenabbreviatedasinTable2.aN=W4exceptforCBCL-OtherAdultwhereN=97andfor Observer-ChildRatingScalewhereN=121.bN=121,exceptfor ChildRatingScalewhereN=119.Variable12345678910111213AdultRatingsa1.HasslesScale--2.MotherImmigrationStressScale.54--...3.ChildImmigrationStressScale.12.42--4.CES-D.32***.5434***--5.SCL-90-RSubscales.38***.43.32***.56***--6.MSPSS-.26’-.37-.29.51*39*--7.ParentSupportScales-.25-.09-.17-.23-.29.44--8.CBCL-Mother TScore.27**.41***37***.38***.48***--9.CBCL-OtherAdultScore•33**.29**35***.34.30.64***--10.ObserverChildRatingScale.02.01-.12.07.05-.10-.16.11-.04--ChildRatingsb11. StressThermometer.22*. ScaleofChildren’sStress.13.27**.30**.20**.14-.07-.01.22*.16.06.18*13. ChildRatingScale-.09-.13-.09.05-.06.04-.04-.15-.18.08.32***33***--*R<05**<y[***<fl[and maternal distress (i.e., CES-D and SCL-90-R subscales) were significantly related tochild behavior ratings (i.e., CBCL) in a positive direction, whereas measures of maternalsocial support (i.e., Multidimensional Scale of Perceived Social Support and ParentSupport Scale) were negatively correlated with child behavior problems. Significantcorrelations that parallel these findings were found when child ratings were made by theother adult who knows the child well.Child ratings of their own stress and behavior were largely consistent with thispattern, but the effects were not as strong. For example, the two measures of children’sown stress, the Stress Thermometer and the Scale of Children’s Stress, were negativelycorrelated with the child’s rating of his/her behavioral strengths (i.e., Child Rating Scale)to a significant degree. In addition, the Scale of Children’s Stress was significantly relatedto mother ratings of child problems in the expected direction, and there was a trendtowards significance in the relationship between the Scale of Children’s Stress and otheradult ratings of the child. There was also a trend towards significance regarding therelationship between the Stress Thermometer scores and mother ratings on the CBCL.Scores on the Stress Thermometer were not associated with other adult ratings, however.Research Assistant ratings of child behavior (Observer-Child Rating Scale) were notrelated to any of the other variables considered. This global rating likely suffered fromproblems related to ceiling effects (i.e., although one child received a very low score, mostof the other ratings clustered in the 20-30 range).72Data Aggregation and TransformationComposite MeasuresEach independent variable (e.g., extrafamilial stress) was measured using anestablished, primary, instrument (e.g., Hassles Scale) and an experimental, secondary,measure (e.g., Mother Immigration Stress Scale). The secondary measures were includedin order to capture aspects of each construct that were deemed to be particularly relevantin this sample of immigrant families. In the case of the stress construct, additionalmeasures pertaining to the child’s experience of stress following immigration were alsoincluded (i.e., Child Immigration Stress Scale, Stress Thermometer, Scale of Children’sStress).With respect to correlations between primary and secondary measures of eachconstruct, Table 3 reveals moderate correlations between each of the two major measuresof extrafamilial stress, maternal distress, and social support. In order to decrease thenumber of measures entered into subsequent analyses, and to minimize the problem ofmulticoffinearity in regression analyses, composite measures were formed. To create theStress Composite, scores were computed for the Hassles Scale and for the MotherImmigration Stress Scale, and then were summed. Ratings of child stress were notincluded in this Composite because of their low internal consistency and because they werenot related to the maternal measures of extrafamilial stress in a strong and consistentmanner. As an aside, it is possible that the low reliability of the Scale of Children’s Stressreflects the grouping together of disparate aspects of immigration stress. A morehomogeneous scale that tapped children’s appraisals of stress may have been more73internally consistent. Because this was not the primary focus of this study, the child stressratings were simply excluded from subsequent analyses. The Distress Composite wasobtained by adding together scores from the CES-D and from the somatic and anxietyscales of the SCL-90-R. Finally, scores from the Multidimensional Scale of PerceivedSocial Support were added to scores from the Parent Support Scale to create theSupport Composite. Reliabilities (alpha coefficients) for these three Composites were .92,.91, and .92, respectively.In addition to describing relationships amongst independent variables, Table 3reveals that mother ratings of child behavior were moderately correlated with theperceptions of another adult rater. In keeping with the above rationale, a Child BehaviorProblem Composite was formed by summing the scores of mother ratings and otheradult ratings. Had the child ratings of their own behavior been correlated with these adultratings, they too would have been included in the Composite. Likewise, the researchassistant ratings would have been added to the Composite if moderate correlations wereobtained with other perceptions of child behavior. As it stands, however, only mother andadult ratings were used in subsequent analyses.Intercorrelations amongst all of the Composite measures appear in Table 4. Asexpected, the Stress, Distress, and Support Composite scores were significantly related tothe Child Behavior Problem Composite in the anticipated directions. It is also apparentfrom Table 4 that the Stress, Distress, and Support Composites were themselvesmoderately correlated.74Table 4. Intercorrelations amongst Composite Measures.Variablea 1 2 3 4Stress Composite--Distress Composite--Support Composite 33*** 47**s --. .Child Problem Composite .40 .48 -.35Note. Stress Composite = Hassles Scale + Mother Immigration Stress Scale; DistressComposite = Center for Epidemiological Studies- Depression Scale + Somatization andAnxiety Subscales of the SCL-90-R; Support Composite Multidimensional Scale ofPerceived Social Support + Parent Support Scale; Child Problem Composite = ChildBehavior Checklist - Mother score + Child Behavior Checklist - Other Adult T score.a N=104, except for Child Problem Composite where N=97.<.001.Linear TransformationIn Moderated Multiple Regression, the influence of a putative moderator variableis reflected in the inclusion of an interaction term in the regression equation. Thisinteraction term, which is the product of two independent variables (e.g., Stress multipliedby Support), carries the influence of the moderated relationship. In the present study, twointeraction terms were included in the equation predicting child problems; Stress x75Support and Distress x Support. Prior to the creation of these interaction terms, inanticipation of problems arising from multicollinearity of independent variables in theregression analyses, a transformation was conducted on these predictor variables. Thislinear transformation, commonly called centering, has been increasingly recommended inModerated Multiple Regression as a way to decrease the association between the productterms that carry the interaction effect and the independent variables upon which they arebased (Coulton & Chow, 1992; Cronbach, 1987; Dunlap & Kemery, 1987; Finney et a!,,1984; Jaccard, Wan, & Turrisi, 1990; Koeske, 1992). Two similar forms of thistransformation have been recommended, and both yield essentially the same results(Jaccard, Wan, & Turrisi, 1990). Cronbach (1987) suggested that mean scores besubtracted from the raw scores of all of the independent variables before product terms areformed (i.e., deviation scores are used instead of raw scores). The second method that hasbeen employed involves the use of standard scores. Each independent variable isconverted to a standard score prior to the formation of product terms (Dunlap & Kemery,1987). In the present research, because the Composite measures were created usingstandard scores, the latter method of controlling multicollinearity was utilized.Prmarv Regression AnalysesRegression DiagnosticsPrior to conducting the Linear Moderated Multiple Regression analyses, severalsteps were taken to ensure that the regression assumptions were not violated. Theseassumptions include: linearity, homogeneity of variance, and normality. In order to test76these assumptions, standardized residuals were first plotted against predicted values. Theassumption of linearity would be deemed violated if the plotted residuals formed asystematic pattern (e.g., curvilinear or u-shaped designs). No discernible patterns could bedetected on these plots, rather, the residuals appeared to be randomly clustered about thehorizontal line through zero. This result suggests that the relationship between theindependent variables and child behavior problems is best represented by a straight line4.The assumption of homogeneity of variance may also be evaluated by examining the plotof residuals and predicted values. Because the spread of the residuals appeared to beconstant across levels of the predicted values, this assumption also seems to have beenmet. In order to assess the normality of the distribution, a histogram of standardizedresiduals was created. The residuals appeared to be roughly normally distributed. Use ofthe nonparametric one-sample Kolmogorov-Smirnov Test confirmed that there was anormal distribution.In addition, several procedures were followed to detect the possible presence ofoutliers or influential points. An examination of standardized residuals determined thatthere were no outliers in this data set. One influential point did surface, however, using acombination of procedures designed to detect unusual cases. That is, this case had a largeMahalanobois Distance (i.e., a large distance (20) between this point and the mean valuesfor the independent variables), a high leverage value (i.e., this point has a significantinfluence on the fit of the model (.21)), and a high covariance ratio (i.e., this point has animpact on the variance-covanance matrix (1.32)) (Statistical Package for the Social77Sciences, Inc., 1993). As a result, this case was deleted from fi.irther analyses regardingthis model.Moderated Multiple RegressionRegression Analysis. Total Sample. The Moderated Multiple Regressionprocedure was first presented by Saunders (1956) as an alternative to analysis of variancemethodology in detecting interaction effects. Statistical methods and procedures forinterpretation have since been formalized (e.g., Aiken & West, 1991; Baron & Kenny,1986; Cohen & Cohen, 1983; Jaccard, Turrisi, & Wan, 1990). A central principle that hasbeen established is that Moderated Multiple Regression is to be conducted hierarchically(Aiken & West, 1991; Cronbach, 1987; Koeske, 1992). Specifically, interaction terms areto be entered into the equation only after the influence of control variables and main effectindependent variables have been considered. In the present study, a block containing theStress, Distress, and Support Composites was entered first. A block containing theinteraction terms Stress x Support and Distress x Support was then entered. Thedependent variable in the analysis was the Child Behavior Problem Composite.With respect to control variables, only 4 of 39 correlations between potentiallyinfluential variables (e.g., socioeconomic status, child age) and the independent variableswere significant (i.e., maternal English language dysfluency was related to Distress, E( 103)= .22, p<.O5, and Stress, r(103) = .37, <.001; the families’ socioeconomic status in HongKong was related to Stress, r(101) = -.22, p<.O5; and time since immigration was relatedto Support, r(100) = -.20, p<.O5). In spite of the limited number and strength of theserelationships, regression analyses were initially conducted controlling for these variables.78Analyses were then re-run in the absence of control variables. It was noted that theoutcome of the analysis was the same, regardless of whether or not control variables wereentered into the equation. As a result, all subsequent analyses were conducted withoutcontrolling for these variables, so as to avoid sacrificing power.The thU regression model, then, was of the form,Child Behavior Problems = bo + b1 Stress + b2 Distress + b3 Support +b4 (Stress x Support) + b5 (Distress x Support).In keeping with guidelines suggested by House (1981), when scores for allvariables are “scaled positively with a real zero point” (p. 133), b1 estimates the maineffect of Stress on Child Behavior Problems, b2 estimates the main effect of Distress onChild Behavior Problems, and b3 estimates the main effect of Support on Child BehaviorProblems. Theory predicts that b1 and b2 should be positive, whereas b3 should benegative. Evidence for stress buffering by Support exists when b1 is significant andpositive, and b4 is significant and negative. This reflects the negative change in slope thatresults from an increase in Support (i.e., the relationship between Stress and ChildBehavior Problems decreases when Support increases). Likewise, evidence for thebuffering of Distress by Support exists when b2 is significantly greater than zero and b5 issignificantly less than zero.Some researchers contend that, in the presence of a significant interaction effect,main effects should not be interpreted because the significant interaction implies thateffects are conditional, rather than constant, across levels of the moderator variable (e.g.,Cleary & Kessler, 1982). Others have suggested that the effect of constituent variables79should be considered (Finney, et al., 1984; House, 1981), although there is some disputeas to how this is best accomplished. The debate appears to lie in the way in which a maineffect is defined. Whereas some researchers suggest that a main effect is to be understoodas the influence of a variable in the absence of any interaction terms (House, 1981), othershave posited that it should be seen as the average effect of a variable across all levels of amoderator variable (Finney, et al., 1984). Finney and his colleagues (1984) make acompelling argument for this latter position and indicate that centered scores will allow forthe meaningful interpretation of main effects. In the present analysis, given that scoreshave already been centered, it is considered appropriate to examine main effects as well asinteractions.The results of the regression analysis appear in Table 5. As is consistent withprevious regression modeling in health psychology, all variables, even those that werenonsignificant, were considered in the interpretation of the full model. Analysesconducted in this area typically do not proceed to re-run regressions, omittingnonsignificant terms, until the best predictive statement is obtained. Rather, the emphasisis on examining the degree to which each variable in the model is important. Two blocksof variables were entered into the equation. First, the main effect independent variableswere entered, producing a significant model that accounted for 27% of the variance inchild behavior problems. Each of the independent variables operated in the anticipateddirection (i.e., b1 and b2 were positive and b3 was negative). Interaction terms wereentered in the next block. The resulting full model was significant overall, accounting for32% of the variance in child problems within this immigrant sample. The change in R280attributable the the addition of the interaction terms was also significant. In keeping withthe suggestions made by Finney and his colleagues (1984), main effects were interpreted atthis point. Both Distress and Stress were significant predictors of Child BehaviorProblems. The Support Composite, however, was not significant in this model. Withrespect to interaction effects, the Stress x Support variable operated in the expecteddirection but was nonsignificant. In contrast, the Distress x Support interaction term wassignificant, but appeared to flrnction in an unanticipated direction. That is, rather thanbeing significantly negative as would be predicted by the buffering model, b5 wassignificant and positive.Table 5. Summary of Hierarchical Multiple Regression Analysis predicting ChildBehavior Problemsa.Predictors B SE B b T Sig TStress Composite 0.24 0.19 0.23 2.00 .048Distress Composite 0.34 0.12 0.32 2.80 .005Support Composite -0.16 0.11 -0.15 -1.50 .126Stressx Support -0.06 0.06 -0.10 -0.90 .388Distress x Support 0.16 0.06 0.27 2.40 .016R2ciiange = .05 (due to interaction terms) E change 12.14 p<.OO1R2 .32 (full model) E(5,90) = 8.40 <.001Note. Tabled values are for the full model. B = regression coefficient, SE B = standarderror of regression coefficient, b = beta.8N=96.81The interaction finding in this study is particularly striking in that, contrary to thestated buffering hypothesis, the relationship between maternal distress and child behaviorproblems appeared to be facilitated by the perceived presence of social support. That is,as social support increased, the slope of the relationship between maternal distress andchild behavior problems became steeper (i.e., the relationship between Distress and ChildBehavior Problems became more pronounced with increasing Support). In an effort tofurther understand this result, the sample was divided into three groups; ordered accordingto higher (i.e., z scores from +.84 to +3.75), medium (i.e., z scores between -.52 and+.83), and lower (i.e., scores from -4.2 to -.51) Support. Separate regression analyseswere conducted for each of these groups (see Table 6). It is apparent from these analysesthat, indeed, as the level of social support increased, the slope of the line relating Distressto Child Behavior Problems became steeper. A series of scatterplots depicting theserelationships may be found in Figure 3.82Table 6. Regressions of Maternal Distress on Child Behavior Problems at Lower.Medium, and Higher Levels of Perceived Support.Predictors B SE B b T Sig TLower Support ( = -4.23 to -0.5 1)Distress Composite 0.24 0.18 0.24 1.33 .193R2=.06 (1,29)= 1.77 p>.O5Medium Support ( = -0.52 to 0.83)”Distress Composite 0.31 0.14 0.36 2.20 .035R2=.13 E(1,32)=4.84 <.O5Higher Support ( = 0.84 to 3•75)CDistress Composite 0.97 0.20 0.67 4.90 .001R2=.45 E(1,29)=24.20 p<001Note. B = regression coefficient, SE B = standard error of regression coefficient, b =beta.a b=34.83Relationship between Distress and Child Behavior ProblemsMedium Maternal Social Support210-24-3 ÔRelationship between Distress and Child Behavior ProblemsHigh Maternal Social SupportFigure 3. Scatterplots of the relationship between Distress and Child Behavior Problemsat Lower, Medium, and Higher Levels of Support.Relationship between Distress and Child Behavior ProblemsLow Maternal Social SupportIDistress ComposfleDistress CompositeDistress Composite84Before concluding that the data truly reflects this counterintuitive finding, itseemed important to rule out alternative computational explanations for this significantinteraction. Two competing explanations were deemed plausible. First, because theguidelines for interpreting Moderated Multiple Regression forwarded by House (1981)presuppose that variables “are scaled positively with a real zero point” (p. 133), there wassome question as to whether or not centered data, with its positive and negative values,yielded interpretable results. In order to ensure that the unexpected finding was not theresult of the multiplication of positive and negative integers, a linear transformation wasapplied that made all data points positive (i.e., added the most negative number for eachvariable to every score), The regression analysis that resulted from the use of thesetransformed Composites produced results that were identical to those obtained in theprimary analyses. The unexpected direction of the interaction effect cannot, therefore, beattributed to the centering procedure.Secondly, it seemed feasible that this unexpected finding might be the result ofcoffinearity among main effect independent variables in that, whereas the centeringprocedure provides protection from collinearity between the interaction terms and thesingle independent measures upon which they are comprised, it does not correct for therelationships amongst single independent variables. As indicated in Table 3, Stress,Distress, and Support were moderately intercorrelated. Therefore, two separateregression analyses were conducted; one testing Stress, Support, and Stress x Support aspredictors of Child Behavior Problems, and the other evaluating the role of Distress,Support, and Distress x Support. These analyses are displayed in Table 7. In the firstmodel, both Stress and Support were significant predictors of Child Behavior Problems85Table 7. Separate Regressions on Child Behavior Problems Stress. Support, andStress x Support, and Distress. Support, and Distress x Support.Regression Model including Distress, Support, and Distress x SupportbDistress Composite 0.46 0.10 0.44 4.40Support Composite -0.19 0.11 -0.18 -1.83Distress x Support 0.12 0.05 0.21 2.32= .04 (due to interaction term) Echge = 2.43 p<O5R2 .29 (full model) F(3, 92) = 12.41 < Tabled values are for the full model. B = regression coefficient, SE B = standarderror of regression coefficient, b = beta.aN96 bN96Predictors B SE B b T Sig TRegression Model including Stress, Support, and Stress x SupportaStress Composite 0.33 0.10 0.32 3.17 .002Support Composite -0.26 0.11 -0.24 -2.39 .019Stress x Support -.001 0.05 -.002 -0.02 .986R2ciiange =0 (due to interaction term) = 0 >.05R2 = .21 (full model) E(3, 92) = 8.14 p<00186and the interaction term (Stress x Support) continued to be nonsignificant. The secondmodel paralleled the findings of the full model, including the significant positive b5.Evidently, the unexpected result cannot be attributed to collinearity between Stress andDistress variables.In summary, the counterintuitive result does not appear to be easily explained bythe computational or procedural methodology employed. Rather, it seems to be a tmereflection of the data. In this sample, the relationship between maternal adjustment andchildren’s behavior was stronger in families where mothers perceived that they weresupported than in families in which mothers felt isolated. It is important to note, however,that mothers who reported lower levels of social support showed higher mean levels ofmaternal distress ( score = .73) and child problems (I score = 55.5) than mothers whoperceived that they were well-supported who reported relatively low rates of bothpersonal distress ( score = -.94) and child problems (I score = 47.4). It is not the case,therefore, that the presence of higher levels of social support facilitated the developmentof child problems under conditions of high maternal distress. Rather, it was the strength ofthe relationship between maternal distress and child problems that increased across levelsof social support.Regression Analysis, by Child Gender. The extent to which the findings obtainedwith the entire sample held across child genders was explored next. The sample wasdichotomized based on the gender of the target child. Although mothers were asked tothink about this target child while completing the questionnaire package, in interpretingthese findings it should be recognized that several families also had children of theopposite gender in their home. Means and standard deviations for the standardized87composite measures in the proposed model are displayed by child gender in Table 8. I-tests were conducted comparing scores in families with male versus female children for allvariables in the Table. Results show that only the difference in Support is significant((102) = -2.41, p<.O5). That is, mothers of girls reported that they receive significantlymore support than mothers of boys. Other scores were not significantly different by childgender.Table 8. Means. Standard Deviations, and Ranges for all Variables in the ModeratorModel, by Child Gender.Variable M SD RangeFamilies of BoySaStress Composite 0.21 1.82 -3.48 to 5.35Distress Composite 0.04 1.82 -2.94 to 4.18Support Composite -0.41 1.73 -4.23 to 2.68Child Problem Composite 0.32 1.70 -3.20 to 3.70Families of GirlsbStress Composite -0.22 1.68 -3.81 to 3.36Distress Composite -0.13 1.61 -2.73 to 3.07Support Composite 0.44 1.54 -3.56 to 3.75Child Problem Composite -0.31 1.86 -4.97 to 3.53Note. Because Composite scores are reported, tabled values are in the form of scores.a=49, except for Child Problem Composite where =42. b =54.88In Table 9, mtercorrelations amongst Composite variables, Stress, Distress,Support, and Child Behavior Problems are displayed; first for families of boys and then forfamilies of girls. Note that, for girls, the independent variables were moderatelyintercorrelated and were all related to child behavior problems to approximately the samedegree. This suggests that multicollinearity may be problematic when using thesemeasures in regression analyses. In contrast, for boys, although independent variableswere related to each other, they were associated differentially to child problems. It seemsthat there was something unique about each of these measures in their relationship withchild adjustment.Hierachical Moderated Multiple Regression analyses were conducted separatelyfor boys (ir49) and for girls (p54). As in the main analysis, independent variablesincluded Stress, Distress, and Support, and the two interaction terms, Stress x Supportand Distress x Support. The dependent variable was Child Behavior Problems. Theresults are presented in Table 10. For families ofboys, there was a main effect for Distressand a significant Distress x Support interaction term that operated in the now familiar,counterintuitive direction. In addition, the Stress x Support interaction was significant,and in the hypothesized direction. That is, for mothers of boys, as Support increased, therelationship between stress and child behavior appeared to weaken. On the other hand, forgirls, there were no significant interaction terms, and there was only a main effect forSupport5.89Table 9. Intercorrelations amongst Variables in the Moderator Model for Families ofBoys and for Families of Girls.Variable 1 2 3 4Families of BOysa1. Stress Composite—2. Distress Composite--3. Support Composite -.20 .52*** --4. Child Problem Composite 34* .52***-.12--Families of Girlsb1. Stress Composite--2. Distress Composite--3. Support Composite .41** 37**..4* 4*. *4*4. Child Problem Composite .42 .43 -.47--a p49, except for Child Problem Composite where =42. b*<05 44<.01.90Table 10. Hierarchical Multiple Regression Analyses Predicting Child Behavior Problems.Conducted by Child Gender.Predictors B SE B b T Sig TFamilies of boySaStress Composite 0.23 0.14 0.25 1.67 .103Distress Composite 0.58 0.15 0.64 3.75 .001Support Composite 0.18 0.15 0.19 1.27 .211Stress x Support -0.16 0.08 -0.32 -2.09 .044Distress x Support 0.24 0.07 0.50 3.35 .002= .17 (due to interaction terms) 5.88 R<01R2= .48 (full model) (5,36) = 6.8 p<.OO1Families ofgirls1’Stress Composite 0.10 0.20 0.09 0.50 .621Distress Composite 0.29 0.18 0.25 1.58 .119Support Composite -0.41 0.16 -0.34 -2.46 .017Stressx Support 0.10 0.12 0.17 0.82 .417Distress x Support -0.04 0.14 -0.06 -0.31 .756R2change = .01 (due to interaction terms) change 0.36 p>.05R2= .33 (full model) E(5,48) = 4.70 p<.OOlNote. Tabled values are for the full model. B = regression coefficient, SE B = standarderror of regression coefficient, b = beta.afl4291Secondary Analyses regarding the Proposed ModelSocial CompetenceDescriptive InformationThe mean child social competence level for this sample (=103), as measured bymother ratings on the Social Skills Rating System (Gresham & Elliott, 1990), was astandard score of 91.7, with a standard deviation of 17.9, and a range of 49 to 130. Themean score for boys was 89.02 (SD = 14.97) and for girls was 94.15 (SD = 19.97). Thisdifference by child gender was not statistically significant.Intercorrelations amongst Stress, Distress, Support and Child Social Competenceare displayed in Table 11. As expected, whereas Child Social Competence was negativelyrelated to Stress and Distress, it had a positive significant relationship with Support.Similarly, ratings of Child Behavior Problems were negatively related to Child SocialCompetence (r (94) = -.4 1, p<OO).Secondary Regression AnalysesRegression Diagnostics. The Main Moderated Multiple Regression Analysis wasrepeated using Child Social Competence as the dependent variable, in place of ChildBehavior Problems. Assumptions of linearity, homogeneity of variance, and normalitywere tested prior to conducting this analysis. Again, plotted residuals showed no distinctpattern (reflecting linearity) and a constant distribution of residuals across levels of thepredicted values (suggesting homogeneity of variance). The histogram of standardizedresiduals showed an approximately normal distribution. A search for possible outliersand/or influential points yielded no outliers, but one very dominant influential point. Thiscase had a very large Mahalanobois Distance (45), and a large Leverage value (.44) and92Covariance Ratio (1.9). This case was deleted from further social competence analyses.It was not the same case that was reported as an influential point in the main analysis.Regression Analysis. Total Sample. The full model tested in this analysis was:Child Social Competence = b0 + b1 Stress + b2 Distress + b3 Support+ b4 (Stress x Support) + b5 (Distress x Support)Following from House (1981), theory would suggest that b1 and b2 should be negative,whereas b3 should be positive. Stress-buffering would be confirmed if b1 was significantand negative and b4 was significant and positive. Similarly, b2 should be significantlynegative, while b5 should be significantly positive.The results of this regression analysis appear in Table 12. The main effectvariables were entered in the first block. All of the main effect independent variablesasserted their influence in the expected directions. When the interaction terms were addedin the next block, the full model accounted for 21 % of the variance in child competencies.However, the change in R2 was not significant, nor were the contributions of either of theproduct terms. Although it is recognized that multicollinearity between interaction termsmay have contributed to a lack of significant findings for one or the other of these terms, itwas deemed inappropriate to explore this further in that, overall, interaction terms werenonsignificant. With respect to main effects, only Support was a significant predictor ofchild social competence. There was a trend towards significance for the main effectDistress.93Table 11. Intercorrelations amongst Variables in the Moderator Model PredictingChild Social Competencea.Variables 1 2 3 41. Stress Composite--2. Distress Composite--3. Support Composite .4O*** --. . **4. Child Social Competence -.27 -.32 .36Note. Independent variable Composites are as described previously. Child SocialCompetence = Total score on Social Skills Rating Scale (completed by mothers).a N=103, except for Child Social Competence where N=102.**p<.ol. p<.ool.94Table 12. Hierarchical Multiple Regression Analysis Predicting Child SocialCompetencea.Predictors B SE B b T Sig TStress Composite -0.64 1.26 -0.06 -0.51 .612Distress Composite -2.34 1.23 -0.23 -1.90 .060Support Composite 2.95 1.11 0.28 2.67 .009Stress x Support -0.30 0.78 -0.05 -0.39 .698Distress x Support -1.10 0.69 -0.20 -1.59 .116R2ciiange .05 (due to interaction terms) Ecige = 3.07 >.05R2= .21 (full model) F(5,96) = 5.22 12<.001Note. Tabled values are for the full model. B = regression coefficient, SE B = standarderror of regression coefficient, b = beta.a N=102.Regression Analysis by Child Gender. Moderated Multiple Regression Analyseswere conducted separately for families of boys and families of girls to determine if thegender differences observed when predicting child behavior problems were evident whenconsidering child social competence. The results of these analyses are displayed in Table13. For families of boys, the full model approached significance (E(47) 2.08, p.O9).Although in a strict statistical sense this precludes interpretation of component variables, itis interesting to note that the main effect Distress variable was the only significant95predictor of child competencies, and there was a trend towards significance, in theunanticipated direction, for Distress x Support. For families of girls, in contrast, the fullmodel was significant and accounted for 26% of the variance in Child Social Competence.However, the interaction terms did not add significantly to the model. Only Support was asignificant predictor of child competencies for girls. It seems that the results by gender arenot dissimilar from those obtained when considering child behavior problems6.Stress-Distress RelationshipSecondary Regression AnalysisAlthough the main focus of the present study involved examining the ModeratorModel as it pertains to immigrant mothers and their children, it was deemed relevant toexplore the degree to which the adult stress and coping model (e.g., Lazarus & Folkman,1984; Kessler, 1982) applied to this immigrant sample of mothers. A Moderated MultipleRegression analysis was conducted to determine the extent to which the relationshipbetween extrafamilial stress and maternal distress was buffered by perceived support.96Predictors BTable 13. Summary of Hierarchical Multiple Regression Analyses Predicting ChildSocial Competence. Conducted by Child Gender.SEBFamilies ofBoysaStress Composite -0.26 1.54 -0.03 -0.17 .865Distress Composite -3.52 1.65 -0.42 -2.13 .039Support Composite 0.97 1.49 0.11 0.65 .517Stress x Support -0.36 1.18 -0.06 -0.30 .766Distress x Support -1.60 0.92 -0.35 -1.74 .090R2c11.ange .11 (due to interaction terms) Echange = 2.89 p>.05R2 0.20 (full model) E(5,42) = 2.08 p>.O5Families of Girls1’Stress Composite .768Distress Composite .287Support Composite .014Stress x Support .837Distress x Support .5892R changeNote. Tabled values are for the full model. B regression coefficient, SE B = standarderror of regression coefficient, b = beta.afl4S b54b T SigT-0.61 2.07 -0.05 -0.30-2.00 1.86 -0.17 -1.084.53 1.77 0.36 2.56-0.24 1.14 -0.04 -0.21-0.58 1.07 -0.10 -0.54.02 (due to interaction terms) Ecinge = 0.65 P>.05R2= 0.26 (full model) E(5,48) 3.33 p<.Ol97Regression Diagnostics. Regression diagnostics confirmed that the sample metassumptions for linearity, normality, and homogeneity of variance. No outliers weredetected. The same point that emerged as influential in the Social Competence analysiswas also omitted from this analysis (i.e., Mahalanobois Distance 30, leverage = .29,covariance ratio = 1.39).Regression Analysis. Centered Composite measures were used, and the equationto be tested took the form of:Distress = b0 + b1 Stress + b2 Support + b3 (Stress x Support)Again, following House’s (1981) guidelines, evidence for stress-buffering would beobtained if b1 were significant and positive, b2 were significant and negative, and b3 weresignificant and negative (i.e., suggesting that the strength of the relationship betweenStress and Distress decreases as Support increases). It may be seen in Table 14 that all ofthese conditions were satisfied in the analysis. Note in particular that, unlike the results ofthe primary analysis predicting Child Behavior Problems, the Stress x Support interactionoperated in the anticipated direction (i.e., b3 was negative and significant). To confirmthis, the sample was again subjected to a three-way split of low, medium, and highSupport scores. Regression analyses across these three groups indicated that, as expected,the slope of the line depicting the relationship between Stress and Distress decreased withhigher levels of Support (at low support =.90, at medium support .43, and at highsupport, . 19). Therefore, the stress-buffering relationship observed in the NorthAmerican health psychology literature was replicated in this Asian sample.98Table 14. Hierarchical Multiple Regression Analysis Predicting Maternal Distressa.Predictors B SE B b T Sig TStress Composite 0.53 0.09 0.50 6.05 .001Support Composite -0.27 0.09 -0.25 -3.05 .003Stressx Support -0.09 0.05 -0.15 -1.99 .049R2ciiange = .02 (due to interaction term) = 3.96 p<.O5.44 (fI.ill model) E(3, 99) = 25.46 p<.OOlNote. Tabled values are for the 11.111 model. B = regression coefficient, SE B = standarderror of regression coefficient, b = beta.a N=1 03.SummaryIn summary, there appears to be clear support for the Basic Model in predictingchild behavior problems within this sample of new immigrant families. Mothers whoreported higher levels of social support and lower extrafamilial stress and personal distressappeared to have children who showed fewer child problems than mothers with lowersocial support and higher stress and distress. Evidence in support of the Moderator Modelwas mixed, and seemed to vary, in part, with child gender. For mothers of boys, theinfluence of extrafamilial stress on child behavior was moderated by maternal perceivedsupport. The counterintuitive significant Distress x Support also emerged only for boys.99For girls, there was no evidence that social support served a moderating role betweenstress or distress and female child behavior. With respect to the Basic Model, althoughonly social support predicted child outcomes in the regression analysis for girls, bivariatecorrelations indicated that extrafamilial stress and maternal distress were also significantlyrelated to child behavior problems.100DISCUSSIONReview of FindingsAlthough researchers have been examining adult and child adjustment followingimmigration for decades, theoretically-driven empirical study that integrates theexperiences of parents and children is lacking. The present study contributes to this areaby exploring, in the context of family-level influences, why it is that some childrenexperience adjustment problems while others seem resilient to the stress of immigration.Based on predictions from the Basic North American Model of family stress and parent-child outcomes, it was hypothesized that in immigrant families in which parents reportedhigher levels of extrafamilial stress, distress, and/or social isolation, children would be atparticular risk for showing behavior problems. Further, the parent-child Moderator Modelsuggests that immigrant children from stressed families, and/or with distressed mothers,should show fewer behavior problems if they are protected by family-level sources ofstrength, such as maternal social support. The present study provides an empirical test ofthese hypotheses.Summary StatisticsLike the findings drawn from the Ontario Child Health Study (Monroe-Blum Ct a!.,1989), the immigrant children in this sample were rated as exhibiting an average level ofbehavior problems, relative to published norms for age and gender (Achenbach, 1991). Arange of scores were reported, however, indicating that some children displayed aclinically-significant level of problems, whereas others did not. Similarly, the process ofimmigration was not uniformly perceived by mothers in this sample to be stressful. Thatis, although some mothers indicated by their responses that they had experienced101significant stress and personal distress since immigrating, others provided scores thatreflect a relatively easy transition to life in Canada. This is in contrast to assumptionsfrequently made in the literature that immigration is, by definition, a stressful life event(Furnham & Bochner, 1986; Oberg, 1960). The results of this study highlight the fact thatthere are individual differences in the way in which the immigration process isexperienced7.The Basic Model of stress and parent-child outcomes contends that exosystemvariables such as stress and microsystem variables such as maternal distress disrupt parentbehavior, and, in turn, influence child adjustment. Within the present sample of immigrantfamilies, correlational analyses were conducted in order to determine the degree to whichmothers’ perceptions of extrafamilial stress and personal distress coincided with behaviorproblems in their children. As predicted, child problems were related to higher levels ofextrafamilial stress and maternal distress, and lower levels of social support for mothers.It appears then, on the basis of these correlational findings, that this North Americanmodel may have utility for understanding parent and child functioning in Asian immigrantfamilies.Moderated Multiple Regression Analysis on Child Behavior ProblemsPrimary FindingsIn order to examine the degree to which the proposed Moderator Model was anappropriate representation of the relationships observed within this immigrant sample ofparents and children, a Moderated Multiple Regression analysis was conducted. Maineffect findings from this analysis support the fundamental contentions of the Basic Model.As expected, within this immigrant sample, extrafamilial stress and maternal distress were102significant predictors of child behavior. That is, mothers who reported higher extrafamillalstress levels had children with more behavior problems than mothers with lower stress, andmothers who reported higher personal distress had children with more behavior problemsthan mothers with lower distress. A significant main effect for maternal social support wasnot detected in this full regression model.It was fI.irther hypothesized that, even amongst families experiencing high levels ofstress and distress, some children would remain resilient due to the presence of family-level protective factors, such as maternal social support. Contrary to expectation, this testof the parent-child Moderator Model failed to find support within the full Asian immigrantsample. The results indicated that social support was not instrumental in minimizing theinfluence of either extrafamilial stress or maternal distress on child outcomes. Further, asignificant Distress x Support interaction was observed operating in a direction opposite towhat would be predicted by the buffering hypothesis (House, 1981).Findings for families ofboysBefore drawing conclusions on the basis of the above results, it is important toconsider the pattern of findings by child gender. For families of boys, not unlike theresults using the full sample, the Moderated Multiple Regression analysis revealed a maineffect for maternal distress and a non-buffering Distress x Support interaction. In contrastto the findings of the full sample, however, a significant Stress x Support interaction termoperating in a direction consistent with the buffering hypothesis was detected when theanalysis was conducted exclusively for families of boys. Caution must be used, however,in interpreting this finding as evidence in favour of the buffering hypothesis. House (1981)asserted that buffering exists if two conditions are satisfied; ifb1 is significant and positive103and if b is significant and negative. While the latter was true in this regression analysis,the main effect for Stress only approached significance in this full model (p=. 103) (seeTable 10). Although it is anticipated that all of these effects would emerge as significant inthe full model with a larger sample size, for now the support for the buffering hypothesiswithin families of boys must be considered tentative.In both the full sample and for families of boys, the Moderated Multiple Regressionanalysis yielded a significant Distress x Support interaction term operating in acounterintuitive direction. Specifically, maternal distress and child behavior problemswere more closely related at higher levels of social support than at lower levels. This doesnot imply, necessarily, that support facilitated the presentation of child problems within thedistressed subset of mothers in this sample. In fact, mean levels suggest that at higherlevels of social support, fewer symptoms of maternal distress çi lower rates of childproblems were generally reported. The interaction term emerged as significant becausethe relationship between these variables was stronger for mothers who reported arelatively high degree of social support and weaker for mothers indicating that theyreceived lower levels of support.Because major computational and statistical explanations for this unexpectedsignificant interaction have been ruled out, it is appropriate to consider theoretical reasonsfor this finding. It is possible, for example, that supported mothers were moreacculturated than mothers lacking social support in this sample. For instance, in traditionalChinese culture, responsibilities regarding child-rearing and household managementgenerally fall to the mother, and she can expect little support or assistance104from her husband whose role it is to work outside the home to provide materially for thefamily (Lai & Yue, 1990; Lee, 1982). It seems plausible therefore that mothers in thissample who reported higher levels of support were from families with less traditionalviews regarding family role structures than were mothers who indicated lower levels ofsocial and parenting support. Further, in this sample, mothers who indicated less supportreported significantly lower English language proficiency than their more supportedcounterparts ((59)=2.42, p<.O2) Language dysfiuency in these low support mothers mayindex a restriction of social contacts to individuals within the Chinese community. Inaddition, lower support mothers were perceived by research assistants using MotherGlobal Ratings to be slightly less acculturated (mean value of 3 on the scale from 0 (not atall acculturated) to 6 (very acculturated)) than mothers with higher levels of support(mean value of 4 on the scale from 0 to 6). Lower support mothers were also rated byresearch assistants as less “open” in their interpersonal style during the home visit.Although neither the acculturation nor the openness ratings resulted in a statisticallysignificant difference between groups, it should be noted that research assistant ratingswere introduced to the procedure only after much of the data was already collected whichresulted in a small sample of ratings (=27).Perhaps then, relative to those indicating a low level of social support, supportedmothers were more acculturated overall and had families that more closely resembled theWestern families upon which the Basic Model was founded. If so, it follows thatpredictions of this model, like the presence of an association between parent and childadjustment, would be more accurate in describing the experiences of these more105acculturated immigrant families. In contrast, the Basic North American Model may be lessapplicable for mothers whose reports of lower support indicate a lesser degree ofacculturation to the Canadian lifestyle. Further, if children are acculturating more quicklythan their mothers, as is likely given children’s mandatory school attendance (Hong, 1989;Lee, 1982), there may be several variables other than parent functioning that assert aninfluence on child adjustment in these low support families. In contrast to children whoare a part of families in which parents are acculturated, perhaps these children of lowersupport mothers more often serve as the liaison between the family and the host societyand are, therefore, more affected by factors in the new community environment than byintrafamilial, parent-child, influences. Thus, behavior problems, or the lack thereof inthese children might be better predicted by variables related to bicultural competence (e.g.,presence of a bicukural mentor, awareness of social nuances and popular culture,membership in a schoolyard Chinese enclave) than by family or mother-level variables.Such a set of circumstances would serve to explain the weaker association betweenmaternal and child adjustment when social support is lower, relative to when support, andacculturative level, is higher.This explanation is also consistent with the finding that the counterintuitivematernal distress by social support interaction term emerged as significant only for familiesof boys. There is evidence in the literature that males tend to acculturate faster thanfemales (Rogler, 1994), suggesting that one of the roles assumed by boys may be to serveas the intercultural liaison person for the family. In the present study, perhaps the boys inmore isolated, and less acculturated, families were taking on this liaison role more often106than the girls in these families. As such, the adjustment of these boys would have beenmore reliant upon a multitude of outside influences than would the behavior of girls.Findings for families of girlsThe results of the Moderated Multiple Regression analysis for girls look differentfrom those depicted in the full model, and from findings for families of boys. In theregression model for girls, the only significant main effect was for Support, whichoperated in the anticipated direction (i.e., the more social support perceived, the fewerchild problems reported, and vice versa). Correlational analyses demonstrated, however,that extrafamilial stress and maternal distress were almost as highly correlated with childbehavior problems as was perceived support, and that all three independent variables weremoderately intercorrelated (see Table 9). This is in contrast to findings for boys for whomeach of the independent variables, although themselves moderately interrelated, wereassociated with child behavior to different degrees. Multicollinearity therefore may havebeen more of an influence in predicting child outcomes within the regression model forfamilies of girls. Two independent regression analyses were conducted for families of girlsin an attempt to examine the separate influences of stress and distress (i.e., to examineeffects without the collinearity between stress and distress). In the regression model thatincluded extrafamilial stress, maternal perceived support, and the interaction betweenstress and support, there were significant main effects for both Stress (h(55) = .27, p<.O5)and Support (h(55) -.37, <.Ol). Similarly, in the model that included maternal distress,perceived support, and the Distress x Support interaction, both Distress ((55) .28,p<.O5) and Support (k(55) = -.36, <Ol) were significant predictors of child outcome.107When considered in the context of the correlational findings, these analyses for families ofgirls largely support the Basic Model.With respect to the buffering role of maternal social support, the ModeratedMultiple Regression analyses revealed no moderator effects for families of girls (i.e., ineither the lull model or in the separate analyses for Stress and Distress). Unlike boys, forwhom the influence of stress on child behavior appeared to be diminished by higher levelsof social support, girls were not afforded protection from extrafamilial stress by thismaternal variable. One might be tempted to explain this difference by suggesting thatbecause boys tend to be more sensitive to family disruptions and therefore more likely toshow behavior problems than girls (Reid & Crisafulli, 1990), mothers of boys sought outsources of support in response to their son’s difficulties, whereas mothers of girls did nothave the need to do so. A comparison of means, however, indicated that mothers of girlshad significantly more social support than mothers of boys. A second possible explanationfor the lack of buffering findings for girls is that there may be some variable other thanmaternal perceived support that better serves to lessen the influence of stress and distresson immigrant child adjustment. It may be that maternal social support is helpful for girlsas a main effect variable, but that some other variable that is more proximal to parent-childinteractions (e.g., marital satisfaction, maintenance of family rituals/routines) may be morepotent in buffering child adjustment when parents are stressed or distressed.Summary of Primary AnalysesTaken together, the findings from this study suggest that the Basic Model was agood fit for the Asian immigrant mothers and their children in this sample. Extrafamilial108stress, maternal distress, and maternal social support, were all correlated with childbehavior outcomes in the expected manner. Although for families of boys, Distress wasthe only significant main effect in the regression model, and for girls only Support wassignificant, in general variables were correlated with outcomes as expected. With respectto the buffering hypothesis, for boys, there was evidence that part of the Moderator Modelalso predicted child outcomes as social support appeared to serve as a buffer forextrafamilial stress. Such buffering did not occur for girls, and maternal distress was notbuffered by support for either gender. In fact, for boys, a significant interaction term thatoperated in an unanticipated direction emerged, such that the presence of high levels ofsocial support facilitated the relationship between maternal distress and child behavior. Itwas suggested that this counterintuitive result may have been a flinction of differentialacculturation between parents and children at lower levels of support, making boyssusceptible to influences from outside of the family and less susceptible to circumstanceswithin the family8.Given that only tentative support for the Moderator Model was found in this study,and even then only for boys, it is worth considering how variability in immigrant childadjustment might be accounted for by the Basic Model of stress and parent-childoutcomes. The results of the present study suggest that this Basic Model, in and of itselfcould explain why some children exhibit behavior problems following migration whereasothers do not. Specifically, y if one assumes that immigration is inherently anduniformly stressful, and always leads to parental distress, does this model predict that allimmigrant children should show behavior problems. 1f on the other hand, it is deemed109possible that immigration may be more or less stressful or distressing depending on premigration and post-migration circumstances, then this model predicts that some childrenwill show problems and others will not, strictly as a function of stress and parent distresslevels. In the present study, it was clear that respondents did not all perceive immigrationto be a stressful life event, and that mothers were not all experiencing post-migrationdistress. Moreover, child problems varied with the degree of stress, distress, and supportreported. Therefore, although it would be premature to abandon the Moderator Model asa tool for understanding parent and child adjustment following immigration, main effectsfindings go a long way to explaining why some children experience difficulties while othersremain invulnerable.Moderated Multiple Regression Analysis on Child Social CompetenceIn response to increasing interest in children’s prosocial capabilities (Parker &Asher, 1987; Selman et al., 1992), and the understanding that social competence is morethan the inverse of behavior problems (Kazdin, 1992), in the present study ModeratedMultiple Regression analyses were also conducted using a measure of child competence asthe dependent variable. Although no specific predictions were made for this exploratoryanalysis, it was anticipated that the relationships suggested by the Basic and ModeratorModels would be reflected in the findings. A second reason for measuring this aspect ofchild behavior evolved from a concern that behavior problems might be underreported inthis sample due to the cultural stigma attached to psychological problems. Child socialcompetence was assessed so as to provide a more benign, and perhaps more culturallyacceptable, forum for describing child adjustment.110At the level of the fi.ill sample, the regression model predicting Child SocialCompetence looked quite different than the results obtained for problematic childbehavior. Although correlational analyses demonstrated significant relationships betweenall of the main effect independent variables and child competencies in the predicteddirections, in the regression analysis, only Support was a significant predictor of ChildSocial Competence. There was a trend towards significance for Distress, but extrafamilialstress did not predict child competencies. With respect to the Moderator Model, neitherthe Stress x Support interaction, nor the Distress x Support interaction, was significant inthe thU model, suggesting that Support did not serve a buffering fi.rnction in relation tochildren’s prosocial competence.Like the findings from the primary analyses, however, gender differences wereapparent when predicting Child Social Competence. When the sample was divided bychild gender, although the results of the regression analyses for families of boys onlyapproached significance, if the contributions of constituent variables were tentativelyinterpreted, the pattern of results more closely mirrored the findings obtained whenconsidering child problems. That is, for families of boys, the main effect for Distress was asignificant predictor of child competencies, and the counterintuitive Distress x Supportinteraction approached significance. Contrary to primary findings, Stress x Support wasnonsignificant. For families of girls, Support was the only significant predictor of ChildSocial Competence, as was the case when considering child problems. It appears thenthat the gender differences may be robust, in that no matter which aspect of child111fl.inctioning is measured, relationships with exosystem and microsystem variables seem tobe present.Ratings of Child Social Competence were only moderately associated with theChild Behavior Problem Composite (r(95) = -.41, 12<.OO1). It seems clear that theseconstructs are tapping unique aspect of child functioning. It is not surprising, therefore,that the results for this dependent variable might vary somewhat from those obtained inthe primary analysis. Also of note is that underreporting of child problems did not appearto be an issue in this study, as mothers reported a range of adjustment difficulties in theirchildren.Moderated Multiple Regression Analysis On Maternal DistressIn the health psychology literature, the role of social support as a moderatorbetween stress and psychological adjustment has been well documented (House, 1981;Kessler, 1982). The bulk of this literature suggests that adults who perceive that they aresupported show fewer mental health problems following a stressor than adults who feelthat social support is lacking. Based on this literature, a secondary hypothesis tested inthis study was that support would behave as a moderator of extrafamilial stress at the levelof maternal distress, irrespective of child outcomes. The results of a Moderated MultipleRegression analysis suggest that maternal social support does serve a buffering function inthe relationship between extrafamilial stress and maternal distress. Immigrant mothers inthis sample who reported higher degrees of stress and lower levels of support had a higherlevel of overall distress than stressed mothers who perceived that they were wellsupported. This is an important finding in that it serves as a cross-cultural replication of112the stress-buffering findings in the North American health psychology literature. Therewere also main effects for extrafainilial stress and social support on maternal distress, inthe anticipated directions (i.e., mothers with higher stress reported higher distress thanmothers with lower stress and mothers with higher support reported lower distress thanmothers with lower support).It is conceivable, given the moderator role of support at the parental level, thatstress not only exerts an independent effect on child behavior, but that it has effects thatare mediated through relationships with support and maternal distress. That is, for familieslow in support, the impact of migration stress on children may be felt both directly, andthrough resultant maternal distress and disruptive parenting abilities. On the other hand,stressed mothers who perceive that they are supported are less likely to develop symptomsof distress and to experience disruptions in parenting. These proposed relationships aredepicted in Figure 4. More sophisticated, large sample, statistical investigation is requiredto test these suggested pathways, but they offer an intuitively appealing, theoreticallygrounded, direction for further study.113Figure 4. Proposed Model Integrating the Moderating Role of Social Support, and theMediating Role of Maternal Distress, in Predicting Child Behavior.Methodological IssuesReliance on Maternal Perceptions of Self and FamilyFor both cultural and theoretical reasons, it was necessary to rely upon maternalreports for much of the information gleaned through this study. At a cultural level, it hasbeen widely reported that privacy is highly valued in Chinese culture (Hong, 1989; Lee,1982). Respecting this, confidential self- and family-report measures were selected as theywere deemed to be the least intmsive means of gathering information within thiscommunity. In terms of reporting child problems, there were also theoretical reasons forasking parents to make ratings. First, clinical research suggests that children come to theattention of mental health professionals largely in response to adult perceptions of childbehavior problems (Griest et al., 1980; Zahner, Pawelkiewicz, DeFrancesco, & Adnopoz,1141992). Thus, if the concern is with perceived child functioning, familiar adults, who cancomment on the child’s behavior across a variety of settings, are probably in the bestposition to rate behavior. It has been recommended that the assessments of severalindividuals be used to provide the most comprehensive picture of the child’s level ofbehavioral/emotional functioning (Achenbach et al., 1987; Reich & Earls, 1987). In thepresent study, mother reports of child adjustment were supplemented by the ratings ofanother adult familiar with the child, child self-reports, and global ratings by researchassistants. Only maternal and other adult perceptions were, however, correlated stronglyenough to be combined into a composite score. It is perhaps not surprising that the adultand child ratings were only mildly associated, as previous research has shown that adultsand children are bothered by different types of behavior (Edelbrock, Costello, Dulcan,Conover, & Kalas, 1986; Tarullo, Richardson, Radke-Yarrow, & Martinez, 1995), andthat they routinely show less agreement in their reports of problems than do mothers andfathers or parents and teachers (Achenbach et al., 1987). The fact that research assistantratings of child behavior were not significantly correlated with the perceptions of otherinformants was likely the result of a limited opportunity to observe child behavior and ofceiling effects in these ratings.The decision to utilize self-report instruments to tap constructs such as stress,distress and support was guided by theoretical models that highlight the importance of thepersonal appraisal of a stressor in predicting subsequent psychological health. Specifically,these models suggest that health outcomes are more dependent on the individual’sperception of a life event, and an appraisal of his/her coping resources, than on any115“actual” threat associated with the event (Lazarus & Folkman, 1984). In this study, aswith child reports of behavior, child perceptions of extrafamilial stress were too differentfrom maternal impressions to permit the formation of a combined stress score. It isrecognized that maternal ratings may not have reflected a true objective reality or theexperiences of others in the family. At the same time, theory predicts that these stressappraisals would be most pertinent in relation to the mother’s own mental health status(Kanner et al., 1984; Lazarus & Folkman, 1984), and perhaps most related to herparenting behavior, attributions for child behavior, and subsequent child outcomes (Crnic& Greenberg, 1990; Krech & Johnston, 1991; Middlebrook & Forehand, 1985).Similarly, perceptions of support have been shown to be better predictors of outcomesthan support-seeking or size of support network (Wethington & Kessler, 1986).Perhaps more at issue than the fact that ratings are subjective, is the associatedconcern that extrafamilial stress, maternal distress, and perceived support are interrelatedconstructs. To the degree that these variables overlap, the unique contribution of each inpredicting child oUtcomes is difficult to ascertain. For instance, it is possible thatdistressed mothers in this sample selectively recalled negative life events and perceivedfailures in support networks, such that little else besides distress is being measured bythese other instruments. This problem of conceptual and operational overlap acrossvariables is present throughout much of the stress and coping literature (e.g.,Dohrenwend, Dohrenwend, Dodson, & Shrout, 1984; Lazarus, DeLongis, Folkman, &Gruen, 1985).116Similarly, with respect to the interplay between independent and dependentvariables, there has been extensive discussion in the parenting literature regarding thedegree to which mothers who are experiencing a depressed mood can accurately provideratings of their children’s behavior (Richters, 1992). The suggestion has been thatdistressed mothers distort the frequency and intensity of child problems as a result of alowered tolerance for behavioral difficulties (Schaughency & Lacey, 1985). The results ofseveral recent studies have, however, begun to refbte this assumption (see review byRichters, 1992). Most recently, comparing mother and father agreement in mood-disordered and well families, Tarullo and her colleagues (1995) determined that there washigher cross-informant concordance in families with one depressed parent than in wellfamilies. Such research calls into question the hypothesis that maternal distress is adistorting variable in reporting about child behavior.Several precautions were taken in the present study to minimize the influence ofconceptual overlap across independent and/or dependent variables. For example,whenever possible, an attempt was made to select measures that were comprised ofobjective items. This was a priority, not only for ease of translation and cross-culturalunderstanding, but also to inhibit overly interpretive responding amongst participants (e.g.,a vague item, such as “I haven’t been feeling good lately” might inspire a range ofresponses depending upon whether the phrase suggested to the individual physical orpsychological well-being, how the respondent defined “lately,” etc.). Also, response biaspotentially induced by context was controlled by counterbalancing measures so thatparticipants were not consistently primed by one of the instruments to respond in one way117or another. For instance, if measures of distress consistently preceded measures of stress,then mothers’ recall of stressful events might have been unduly influenced by attention totheir mood state. Similarly, different response formats were used for measures of support,relative to measures of stress or distress, and several items on each scale were reverse-scored, so as to avoid problems with systematically biased responding (Pauthus, 1991). Inaddition, the composite child behavior score included not only mother ratings, but also theperceptions of another independent adult rater.Finally, study data suggests that variables assessed in this study were sufficientlyindependent to permit meaningful interpretation of unique contributions. A correlationalanalysis revealed that, although Stress, Distress, and Support Composites were related, theoverlap was only moderate. In addition, for boys, these independent variables seemed tooperate differently in predicting child problems, suggesting that they were distinctconstructs.Reliance on a Correlational DesignThe present research was correlational in nature. None of the variables weremanipulated, rather, they were assessed as they occurred naturally in each family’sexperience. As a result, no conclusions regarding causal relationships are possible.Although knowledge of the literature might lead one to favour explanations that suggestthat stress and distress disrupted parenting and, in turn, child behavior (e.g., Conger et al.,1993), it is also conceivable that disturbed child behavior might have caused parents tofeel more stressed and distressed (Bell & Harper, 1977). Alternatively an additional,unmeasured variable, such as marital satisfaction, might have accounted for variations118among exosystem and microsystem variables and child problems. It would be clearlyinappropriate to advance causal explanations for the results of this correlational study withany degree of certainty.Except in analogue studies (e.g., Krecb & Johnston, 1992; Zekoski, O’Hara, &Wills, 1987), variables such as parent stress and distress are not amenable to manipulation.It is therefore difficult to design externally valid studies in this area that speak to the issueof causality. Some investigators have suggested that the findings from cross-sectionalstudies would be complemented by the use of longitudinal designs in which naturalvariations in stress and distress could be examined (Black & Holden, 1995). A program ofresearch that includes longitudinal study of the immigration and acculturation process, andits association with child adjustment, would be a welcome contribution to this area.Underreporting ofPsychological ConcernsAs mentioned previously, symptom underreporting has been identified as an issuewithin Asian cultures (Sue & Morishima, 1982). Two major explanations for this havebeen postulated; social desirability as a fi.inction of cultural stigmatization associated withmental illness and cross-cultural differences in symptom expression (i.e., emphasis onsomatic versus psychological/affective symptomatology) (Kleinman & Kleinman, 1985;Sue & Sue, 1987; Yamamoto & Acosta, 1982; Zhang, 1994). Regardless of its source, inthe present study, problems with underreporting of symptoms were anticipated.Social DesirabilitySeveral steps were taken to minimize the impact of social desirability. Forexample, as outlined in the Methods chapter, care was taken to ensure that questionnaire119items were culturally-sensitive and nonthreatening. In addition, face-to-face contact wasmade with every participant and mothers were ensured that their responses would beconfidential, and that they could leave blank any items that left them feelinguncomfortable. Participants in this study appeared willing to acknowledge feelings ofstress, distress, and isolation. There were very few examples of mothers who provideduniformly low scores across all constructs (i.e., less than 10% scored below one standarddeviation on all measures). Thus, although the mothers in this sample may still have beenunderreporting problems, they were willing to acknowledge that they experienced at leastsome degree of stress, emotional upset, loneliness, and/or difficulties with their child.Cross-Cultural Differences in Symptom ExpressionCognizant of the possibility that study participants might not experience, or attendto, the emotional aspects of their distress, measures were chosen so as to assess mentalhealth problems, broadly defined. Specifically, a somatic scale of distress was used tosupplement the CES-D so as to capture this aspect of psychological upset if it was thepreferred mode of presentation within this sample. The results suggest that despite thereported preference amongst Asian cultures for endorsing somatic expressions of distress,study participants displayed no difficulty in describing emotional/psychologicalexperiences. Mothers described a range of symptoms, covering affective, interpersonal,and health concerns. It seems then, that if asked to disclose personal psychologicalinformation in privacy to a trusted individual, the mothers in this sample were willing andable to do so.120Translation of: and Modifications to. Standard InstrumentsIn order to increase the probability of recruiting a representative sample of HongKong immigrants, it was deemed necessary to translate questionnaires into Chinese. Acareful process was followed in selecting and translating measures for this study. Despiteconcerns that measures would not translate effectively, all instruments used in the formalanalyses were internally consistent and showed anticipated intercorrelations acrossindependent and dependent variables. This is consistent with previous findings thatsuggest that carefully translated measures retain their psychometric properties (e.g., Nohet al., 1992).This sample of immigrant families from Hong Kong seemed reasonablycomfortable with the process of research. For the most part, participants required littleguidance in the task of questionnaire completion and often hurried research assistantsthrough their script of instructions. In retrospect then, it probably was not necessary tomodify standard measures to the extent that was recommended by cultural consultants. Itseems likely that this sample could have easily managed the “confusion” of scales withdiffering ranges of response, particularly given the careful explanations offered by researchassistants during the home visit. For instance, it was probably unnecessary to change thescale of the SCL-90-R from the standard 5-point version to a 4-point scale (as was doneto be consistent with other measures of distress and stress). Similarly, on the HasslesScale, participants could have been asked to think about stressors that they hadexperienced earlier that day, rather than over the past week. Had such formatting changes121been avoided, interpretation of scores relative to published norms would have beenpossible.The Challenges of Cross-Cultural ResearchConducting research within a culture that is different from one’s own can be achallenging endeavor. In its early stages, the present research involved a gradual processof becoming acquainted with cultural values, mores, and nuances, while at the same timebuilding a relationship of trust with community members. In general, Chinese communityagencies were receptive to the proposed research and assisted with promotion of theproject. Accessing the community, however, was only part of the difficulty. Moredaunting still was the language and cultural barrier with potential study participants.Every detail had to be considered from a cultural and logistical perspective. For instance,written and oral translations were required for a number of administrative requirementsthat surfaced as the study progressed (e.g., providing a translated recording for interestedmothers to call for more information about the study, coordinating advertisement of thestudy at exclusively Chinese-speaking media centers, describing the study to interestedChinese mall patrons, etc.). It was not feasible to request assistance from communityagencies for every translation or interpretation need. Bicultural research assistantsprovided a necessary cultural and language bridge to the community.Although the participation of bilingual, bicultural research assistants was critical tothe success of this project, reliance on these undergraduates posed additional challenges.Their involvement required a substantial investment of time and energy. For example, anextensive training protocol was followed in order to be sure that each assistant could122convey the correct information to study participants in a professional manner, couldanticipate questions and problems, and could establish rapport with families in a shortperiod of time. Further, regular meetings were conducted to ensure that procedurescontinued to be followed as agreed upon, to handle administrative details, and to discussissues that arose during home visits. Most demanding logistically was the need tocoordinate recruiting sessions, home visits, and organizational meetings around the busyschedules of these 13 undergraduates.ApplicationsThe results of this study offer several suggestions for meeting the mental healthneeds of immigrant families. First, with respect to assessment, these research findingshighlight the importance of including measures of parental stress, distress, and supportsystems in clinical protocols designed to detect immigrant children at risk. The resultssuggest that these family-level influences are predictive of child problems. In terms ofinstrumentation, at least for this sample of immigrants from Hong Kong, translated NorthAmerican measures appeared culturally-appropriate and were sensitive to detect a range ofsymptomatology.Secondly, and in a related manner, parent well-being appears to be predictive ofchild outcomes, particularly for boys. From an intervention perspective, it follows thatnew immigrant parents who are experiencing significant stress associated with the culturaltransition may be well-served by taking advantage of community services designed toalleviate settlement concerns and to assist with stress management more generally.123Further, the study suggests that even if mothers are experiencing high levels ofstress, if they feel supported, then distress is minimized, relative to mothers who perceivethemselves to be isolated. On the basis of these findings, mental health agencies thatprovide services to new immigrants, particularly newcomers from Hong Kong, might playa role in helping isolated families to build supportive ties in Canada. This might beaccomplished through formal social support groups for adult immigrants, or, moreinformally, through accessible community programming such as cooking classes. Thepresent study suggests that in addition to fl.irthering sources of self-validation for themother (Ishiyama, 1989), the acquistion of support will benefit children indirectly throughmaternal well-being and positive parenting behavior.Future DirectionsAs one of a handfiul of investigations in an emerging field of study, the presentresearch was exploratory in nature and, as such, possibly raised more questions than itanswered. It is hoped that the findings will stimulate interest in this area that is much inneed of fi.irther study. Future research could take any one of several possible directions.For example, an attempt to replicate the findings of this study using a design that goesbeyond self-report data would be useful. In particular, a study in which maternal distresswas assessed independently (e.g., through a clinical interview, behavioral observations,spousal ratings) would be a helpful addition to the literature. Experience with the presentsample would suggest that members of the local Chinese community would be willing totolerate this level of involvement if interviews were handled sensitively and wereconducted in Chinese. A second methodological advance would be to design a study inwhich family adjustment could be followed longitudinally through the immigration and124acculturation process. In this way, problems that lie donnant following the acute stress ofthe move, may be observed as they develop over time. In addition, unlike many stressors,immigration is a predictable life event. In most cases, there is a period of time prior to themove in which families can anticipate their new lifestyle. As such, the phenomenon ofimmigration affords researchers the opportunity to study families both before and after thestressor. This would be an exciting future direction in that longitudinal work of this kindwould better permit an understanding of causal relationships. Similarly, the role ofacculturation and bicultural identity development in both parent and child adjustmentcould be also be measured and more systematically explored. Findings from the presentstudy suggest that this might be a particularly fruitful research avenue. Specifically, use ofmeasures of acculturation designed specifically for adults and for children, in a similarlydesigned study might shed light on the counterintuitive finding regarding the Distress xSupport interaction term.Future study might also be focussed on other variables that could potentially servea buffering function between immigration stress, distress, and child adjustment. Family-level variables such as differential acculturation, maintenance of family/cultural routines,and marital satisfaction are likely candidates. In addition, a more complete model wouldalso include measurement of parent behavior, either through self-report or structuredobservational methods. Similarly, child-level factors such as role flexibility,temperament/personality (e.g., adaptability, emotionality, reactivity, ability to self-regulate) or personal support systems (school peers or teachers) might be considered.Because findings from the present study suggest that many of these variables will protect125boys and girls differentially, it is recommended that gender differences be considered infuture research. Finally, the degree to which the current findings generalize; to otherimmigrant samples, to children of different ages (under 5 years and/or adolescents), and/orto fathers, would be a pertinent avenue for further study.ConclusionsThis study builds on the existing literature by providing a theoretical framework forunderstanding children’s behavioral adjustment following immigration. In so doing, itmoves the field beyond immigrant versus nonimmigrant comparisons to a place ofconsidering differences in the experience of migration at the level of the family or theindividual. Further, given that this theoretical framework was developed through amerging of several contemporary North American literatures, this study demonstrates thatWestern constructs and relationships have cross-cultural utility.126FOOTNOTES1 Training manual is available for review upon request.2 In addition to the measures described in this section, a few instruments wereincluded in the questionnaire package but were not analyzed because they were not thefocus of the present research effort. These include: qualitative questions on the FamilyInformation Questionnaire regarding the family’s strategies for coping with the stress ofmigration, a measure of maternal optimism (Life Orientation Test; Scheier & Carver,1985), and questions regarding past and anticipated mental health service utilization.3Guidelines for child ratings are available for review upon request.To further confirm this statistically, quadratic terms were systematically enteredin the regression model (e.g., Stress2)and residual plots were examined. There was noevidence that the addition of these items improved the quality of the plots.Although the model appears to fit differently depending upon child gender, astatistical comparison revealed no significant difference in the contribution made by eachvariable for families ofboys relative to families of girls.There was no significant difference across child gender in the contribution madeby Distress, or by either of the interaction terms. The slope for Support, however, wassignificantly steeper for families of girls relative to boys, suggesting that Support is astronger predictor for girls.127In response to concerns that newcomers who expressed intent to stay in Canada(75% of sample) may have had different migration experiences than those who planned toreturn to Hong Kong (or who were unsure of their plans), the primary analyses were againconducted, this time excluding those who might be better classified as sojourners. Theoverall results were similar to those obtained using the full sample. The full model wassignificant (R2=.34, E=lO.3, p<.OO1), and there was a significant main effect for Distressand a trend towards significance for the Distress by Support interaction term in thecounterintuitive direction. Contrary to findings with the full sample, however, there was asignificant main effect for Support, but not for Stress. These results suggest that it may beimportant to control for long-term settlement plans in future studies.8 In an attempt to determine if the source of support played a role in thiscounterintuitive finding, the degree of support received from family, friends, and a specialperson, was examined at high and low levels of support. No significant differences weredetected across levels of support for any of these sources, suggesting that the source ofsupport is less important than the overall perceived presence of support in understandingthis finding.128BIBLIOGRAPHYAbidin, R.R. (1990). Introduction to the special issue: The stresses of parenting. Journalof Clinical Child Psychology, i, 298-301.Achenbach, T.M. (1991). Manual for the Child Behavior ChecklistJ4-18 and 1991 Profile.Burlington, VT: University of Vermont, Department ofPsychiatry.Achenbach, T.M. & Edelbrock, C.S. (1978). The classification of child psychopathology:A review and analysis of empirical efforts. Psychological Bulletin, , 1275-1301.Achenbach, T.M., McConaughy, S.H., & Howell, CT. (1987). Child/adolescentbehavioral and emotional problems: Implications of cross-informant correlationsfor situational specificity. Psychological Bulletin, J..Q!, 2 13-232.Adler, P.S. (1975). The transitional experience: An alternative view of culture shock.Journal ofHumanistic Psychology, li(4), 13-23.Aiken, L.S. & West, S.G. (1991). Multiple regression: Testing and interpretinginteractions. Newbury Park, CA: Sage.Ainsworth, M.D. & Bell, S.M. (1974). Mother-infant interaction and the development ofcompetence. In K. Connolly & J. Bruner (Eds.), The growth of competence.London: Academic Press.Alloway, R. & Bebbington, P. (1987). The buffer theory of social support: A review ofthe literature. Psychological Medicine, 17, 91-108.Almirol, E.B. (1982). Rights and obligations in Filipino-American families. Journal ofComparative Family Studies, j, 291-306.Amaro, H. & Russo, N.F. (1987). Hispanic women and mental health: An overview ofcontemporary issues in research and practice. Psychology of Women Ouarterly,U, 393-407.Andresen, P.A. & Telleen, S.L. (1992). The relationship between social support andmaternal behaviors and attitudes: A meta-analytic review. American Journal ofCommunity Psychology, Q(6), 75 3-774.Aneshensel, C.S. & Frerichs, R.R. (1983). Stress, support and depression: A longitudinalcausal model. Journal of Community Psychology, 10, 363-376.129Anthony, E.J. (1974). The syndrome of the psychologically invulnerable chid. In E.J.Anthony & C. Koupernik (Eds.), The child in his family: Children at psychiatricijk (pp. 529-544). New York: Wiley.Anthony, E.J. & Cohier, B.J. (1987). The invulnerable child. New York: Guilford Press.Arieti, S. & Bemporad, JR. (1980). The psychological organization of depression.American Journal of Psychiatry, j.7, 1360-1365.Aronowitz, M. (1984). The social and emotional adjustment of immigrant children: Areview of the literature. International Migration Review, 11(2), 237-257.Bagley, C. (1972). Deviant behaviour in English and West Indian school children.Research in Education,, $, 45- 55.Baron, R.M. & Kenny, D.A. (1986). The moderator-mediator variable distinction insocial psychological research: Conceptual, strategic, and statistical considerations.Journal ofPersonality and Social Psychology, 1173-1182.Barrera, M. (1981). Social support’s role in the adjustment of pregnant adolescents:Assessment issues and findings. In BH. Gottlieb (Ed.), Social networks andsocial support in community mental health (pp. 69-96). Beverly Hills: SagePublications.Bebbington, P.E., Hurry, J., & Tennant, C. (1981). Psychiatric disorders in selectedimmigrant groups in Camberwell. Social Psychiatry, i, 43-51.Beckwith, L., Cohen, S., Kopp, C., Parmelee, A., & Marci, T. (1976). Caregiver-infantinteraction and early cognitive development in preterm infants. ChildDevelopment, 47, 579-587.Befera, M.S. & Barkley, R.A. (1985). Hyperactive and normal girls and boys: Mother-child interaction, parent psychiatric status and child psychopathology. Journal ofChild Psychology and Psychiatry, , 439-452.Beiser, M. (1988). Influences of time, ethnicity, and attachment on depression inSoutheast Asian refugees. American Journal ofPsychiatry, j4, 46-51.Beiser, M. (1990). Migration: Opportunity or mental health risk. Triangle, , 83-89.Beiser, M. & Fleming, l.A. (1986). Measuring psychiatric disorder among SoutheastAsian refugees. Psychological Medicine, 16, 627-639.130Beiser, M. & Wood, M. (1986). Canadian task force on mental health issues affectingimmigrants and refugees: Review of the literature on migrant mental health.Ottawa: Health & Welfare, Canada.Bell, R.Q. & Harper, L.V. (1977). Child effects on adults. Hillsdale, N.J.: Erlbaum.Beisky, 3. (1984). The determinants of parenting: A process model. Child Development,j 83-96.Beisky, J. & Vondra, J. (1989). Lessons from child abuse: The determinants of parenting.In D. Cicchetti & V. Carlson (Eds.), Child maltreatment: Theory and research onthe causes and consequences of child abuse and neglect (pp. 153-202). New York:Cambridge University Press.Berkman, L.F. & Syme, S.L. (1979). Social networks, host resistance and mortality: Anine-year follow-up study of Alameda County residents. American Journal ofEpidemiology, iQ. 186-204.Berry, 3. W. & Kim, U. (1988). Acculturation and mental health. In P. Dasen, J.W.Berry, & N. Sartorius (Eds.), Cross-Cultural Psychology and Health: TowardsApplications (pp. 207-336). London: Sage.Berry, 3., Kim, U., Minde, T. & Mok, D. (1987). Comparative studies of acculturativestress. International Migration Review, i, 491-511.Bhatnagar, J.K. (1980). Linguistic behavior and adjustment of immigrant children inFrench and English schools in Montreal. International Review of AppliedPsychology, 29(1-2), 141-158.Biglan, A., Hops, H., & Sherman, L. (1988). Coercive family processes and maternaldepression. In R.D. Peters Y R.J. McMahon (Eds.), Social learning and systemsapproaches to marriage and the family (pp. 72-103). New York: BrunnerfMa.zel.Bird, H.R. & Canmo, G. (1987). Use of the Child Behavior Checklist as a screeninginstrument for epidemiological research in child psychiatry: Results of a pilotstudy. Journal of the American Academy of Child and Adolescent Psychiatry,(2), 207-2 13.Black, M.M. & Holden, E.W. (1995). Longitudinal intervention research in children’shealth and development. Journal of Clinical Child Psychology, 4(2), 163-172.Blackwell, P. (April, 1991). Maternal perceptions of child behavior as a function ofmaternal affect, social support and child behavior. Paper presented at the biannualmeeting of the Society for Research in Child Development, Seattle, WA.131Bland, R.C. & Orn, H. (1981). Schizophrenia: Sociocultural factors. Canadian Journal ofPsychiatry, (3), 186-188.Bleuler, M. (1978). The schizophrenic disorders: Long-term patient and family studies.New Haven, Yale University Press.Bradley, S. & Sloman, L. (1975). Elective mutism in immigrant families. Journal of theAmerican Academy of Child Psychiatry, 14, 510-514.Bradley, R.H., Caldwell, B.M., & Elardo, R. (1979). Home environment and cognitivedevelopment in the first two years: A cross-lagged panel analysis. DevelopmentalPsychology, j, 246-250.Brislin, R.W. (1980). Translation and content analysis of oral and written material. InH.C. Triandis & J.D. Draguns (Eds.), Handbook of cross-cultural psychology, vol.Z (pp. 3 89-444). Boston: Allyn and Bacon, Inc.Bronfenbrenner, U. (1979). The ecology of human development: Experiments by natureand design. Cambridge, Mass.: Harvard University Press.Brown, G.W. & Harris, T. (1978). Social origins of depression: A study of psychiatricdisorders in women. London: Tavistock Publications Limited.Capaldi, D. & Patterson, G.R. (1987). An approach to the problem of recruitment andretention rates for longitudinal research. Behavioral Assessment, , 169-177.Cassel, J. (1976). The contribution of the social environment to host resistance.American Journal ofEpidemiology, 1Q4, 107-123.Christensen, A., Phillips, S., Glasgow, R.E., & Johnson, S.M. (1983). Parentalcharacteristics and interactional dysfunction in families with child behaviorproblems: A preliminary investigation. Journal of Abnormal Child Psychology,11(1), 153-166.Christiansen, J.M., Thornley-Brown, A., & Robinson, J.A. (1982). West Indians inToronto: Implications for helping professionals. Toronto: Family Services ofMetropolitan Toronto.Citizenship and Immigration Canada (1994a). A broader vision: Immigration andcitizenship plan 1995-2000. Minister of Supply and Services Canada.Citizenship and Immigration Canada (1994b). Into the 21st century: A strategy forimmigration and citizenship. Minister of Supply and Services Canada.132Citizenship and Immigration Canada (1994c). Facts and figures: Overview ofimmigration. Minister of Supply and Services Canada.Cleary, P.D. & Kessler, R.C. (1982). The estimation and interpretation of modifiereffects. Journal of Health and Social Behavior, , 159-169.Cobb, S. (1976). Social support as a moderator of life stress. Psychosomatic Medicine,(5), 300-314.Cochran, M.M., & Brassard, J.A. (1979). Child development and personal socialnetworks. Child Development, Q, 601-616.Cochrane, R. & Bal, S. (1987). Migration and schizophrenia: an examination of fivehypotheses. Social Psychiatry, , 181-191.Cochrane, R. (1980). Mental illness in England, in Scotland and in Scots living inEngland. Social Psychiatry, j, 9-15.Cohen, J. (1985). Statistical power analysis for the behavioral sciences. Hilisdale, NJ:Lawrence Eribaum Associates.Cohen, 3. & Cohen, P. (1983). Applied multiple regressioncorrelation analyses for thebehavioral sciences, second edition. Hilisdale, NJ: Eribaum.Cohier, B.J., Grunebaum, H.V., Weiss, 3. L., Garner, C.J., & Gallant, D.H. (1977).Disturbances of attention among schizophrenic, depressed, and well mothers andtheir young children. Journal of Child Psychology and Psychiatry, J..$., 115-135.Cohn, J.F. & Tronick, E.Z. (1983). Three-month-old infants’ reaction to simulatedmaternal depression. Child Development, 4, 178-184.Comas-Dias, L. (1988). Mainland Puerto Rican women: A sociocultural approach.Journal of Community Psychology, j, 21-31.Committee on Multiculturalism & Mental Health & Education (1987). Task force onChinese Canadians. Unpublished manuscript.Committee on Multiculturalism & Mental Health & Education (1989). Task force onLatin American Canadians. Unpublished manuscript.Compas, B.E. (1987). Stress and life events during childhood and adolescence. ClinicalPsychology Review, 7, 275-302.133Compas, B.E., Howell, D.C., Phares, V., Williams, R.A., & Giunta, C.T. (1989). Riskfactors for emotional behavioral problems in young adolescents: A prospectiveanalysis of adolescent and parental stress and symptoms. Journal of Consultingand Clinical Psychology, 7, 732-740.Conduct Problems Prevention Research Group (1992). A developmental and clinicalmodel for the prevention of conduct disorder: The FAST Track program.Development and Psychopathology, 4, 509-527.Conger, R.D., Conger, K.J., Elder, G.H., Lorenz, F.O., Simons, R.L., & Whitbeck, L.B.(1993). Family economic stress and adjustment of early adolescent girls.Developmental Psychology, (2), 206-219.Conger, R.D., Patterson, G.R., & Ge, X. (1995). It takes two to replicate: A mediationalmodel for the impact of parents’ stress on adolescent adjustment. ChildDevelopment, , 80-97.Coulton, C. & Chow, J. (1992). Interaction effects in multiple regression. Journal ofSocial Service Research, J, 179-199.Cowen, E. (1991). In pursuit ofweilness. American Psychologist, 4(4), 404-408.Cowen, E. & Work, W.C. (1988). Resilient children, psychological wellness, and primaryprevention. American Journal of Community Psychology, j(4), 59 1-607.Coyne, J.C. & DeLongis, A.M. (1986). Going beyond social support: The role of socialrelationships in adaptation. Journal of Consulting and Clinical Psychology,.4,454-460.Crnic, K.A. & Greenberg, M.T. (1990). Minor parenting stressors in young children.Child Development, j, 1628-1637.Crnic, K.A., Greenberg, M.T., & Slough, N.M. (1986). Early stress and social supportinfluences on mothers’ and high-risk infants’ fi.inctioning in late infancy. InfantMental Health Journal, 7, 19-32.Crnic, K.A., Greenberg, M.T., Ragozin, A.S., Robinson, N.M., & Basham, R. (1983).Effects of stress and social support on mother and premature and fill-term infants.Child Development, 4, 209-2 17.Crockenberg, S.B. (1981). Infant irritability, mother responsiveness, and social supportinfluence on the security of mother-infant attachment. Child Development, 52,857-865.134Cronbach, L.J. (1987). Statistical tests for moderator variables: Flaws in analyses recentlyproposed. Psychological Bulletin, iQ2(3), 414-417.Cunningham, C.E., Benness, B.B., & Siegal, L.S. (1988). Family functioning, timeallocation, and parental depression in the families of normal and ADD-H children.Journal of Clinical Child Psychology, !Z(2), 169-177.D’Arcy, C. & Siddique, C.M. (1984). Social support and mental health among mothers ofpreschool and school age children. Social Psychiatry, W, 155-162.DeLongis, A. (1985). The relationship of everyday stress to health and well-being: Inter-and intraindividual approaches. Unpublished doctoral dissertation. University ofCalifornia, Berkeley.DeLongis, A., Coyne, J.C., Dakof 0., Folkman, S., & Lazarus, R.S. (1982). Therelationship of hassles, uplifts, and major life events to health status. HealthPsychology,!, 119-136.DeLongis, A., Folkman, S., & Lazarus, R.S. (1988). The impact of daily stress on healthand mood: Psychological and social resources as mediators. Journal of Personalityand Social Psychology, 4(3), 486-495.Derogatis, L.R. (1983). SCL-90-R Administration., scoring and procedures manual-IT.Towson, MD: Clinical Psychometric Research.Derogatis, L.R., Lipman, R.S., Rickels, K., Uhienhuth, E.H., & Covi, L. (1974). TheHopkins Symptom Checklist (HSCL): A self-report symptom inventory.Behavioral Science, j.., 1-15.Derogatis, L.R., Rickels, K., & Rock, A.F. (1976). The SCL-90 and the IvIMPI: A step inthe validation of a new self-report scale. British Journal of Psychiatry, j, 280-289.Dohrenwend, B.S., Dohrenwend, B.P., Dodson, M., & Shrout, P.E. (1984). Symptoms,hassles, social supports and life events: The problem of confounded measures.Journal of Abnormal Psychology, , 222-230.Downey, 0. & Walker, E. (1992). Distinguishing family-level and child-level influenceson the development of depression and aggression in children at risk. Developmentand Psychopathology, 4, 8 1-95.Draguns, J.D. (1980). Psychological disorders of clinical severity. In H.C. Triandis &J.D. Draguns (Eds.), Handbook of cross-cultural psychology. vol. 6 (pp. 99-174).Boston: Allyn and Bacon.135Dumas, J.E. (1986). Indirect influence of maternal social contacts on mother-childinteractions: A setting event analysis. Journal of Abnormal Child Psychology, 14,205-2 16.Dumas, J. E. & Wahier, R.G. (1985). Indiscriminate mothering as a contextual factor inaggressive-oppositional child behavior: “Damned if you do and damned if youdon’t.” Journal of Abnormal Child Psychology, U(1), 1-17.Dunlap, W.P. & Kemery, E.R. (1987). Failure to detect moderating effects: Ismulticollinearity the problem? Psychological Bulletin, iQ(3), 418-420.Dunn, 3. & Kendrick, C. (1982). Siblings: Love. envy and understanding. London: GrantMcIntyre.Edelbrock, C., Costello, A.J., Dulcan, M.K., Conover, N.C. & Kalas, R. (1986). Parent-child agreement on child psychiatric symptoms assessed via structured interview.Journal of Child Psychology and Psychiatry, Z(2), 181-190.Egan, K.J., Kogan, H.N., Garber, A., & Jarrett, M. (1983). The impact of psychologicaldistress on the control of hypertension. Journal ofHuman Stress, , 4-10.Egeland, B., Breitenbucher, M., & Rosenberg, D. (1980). Prospective study of thesignificance of life stress in the etiology of child abuse. Journal of Consulting andClinical Psychology, 4, 195-205.Elder, G.H., Jr., Caspi, A., & Downey, G. (1986). Problem behavior and familyrelationships: Life course and interpersonal themes. In A.B. Sorensen, F.E.Weinert, & L.R. Sherrod (Eds.), Human development and the life course:Multidisciplinary perspectives (pp. 293 -340). Hillsdale, NJ: Lawrence ErlbaumAssociates, Inc.Employment & Immigration Canada (1989). Immigration to Canada: A statisticaloverview.Employment & Immigration Canada (1991). Annual report to Parliament. ImmigrationPlan for 1991-1995.Fabrega, H. (1969). Social psychiatric aspects of acculturation and migration: A generalstatement. Comprehensive Psychiatry, iQ, 314-329.Finney, J.W., Mitchell, R.E., Cronkite, R.C., & Moos, RH. (1984). Methodologicalissues in estimating main and interactive effects: Examples from the copingsocialsupport and stress field. Journal ofHealth and Social Behavior, , 85-98.136Flaskerud, J.H. & Soldeville, E.Q. (1986). Filipino and Vietnamese clients: Utilizing anAsian mental health center. Journal ofPsychosocial Nursing, 4, 32-3 6.Fong, S.L.M. & Peskin, H. (1969). Sex-role strain and personality adjustment of China-born students in America: A pilot study. Journal of Abnormal Psychology, Z4,563-567.Forehand, R., Brody, G., Slotkin, J., Fauber, R.., McCombs, A., & Long, N. (1988).Young adolescent and maternal depression: Assessment, interrelations, andpredictors. Journal of Consulting and Clinical Psychology, , 422-426.Forehand, R., Lautenschlager, G.J., Faust, J., & Graziano, W.G. (1986). Parentperceptions and parent-child interactions in clinic-referred children: A preliminaryinvestigation of the effects of maternal depressive moods. Behavior Research andTherapy, 4, 73-75.Forehand, R., Wells, K.C., McMahon, R.J., Griest, D., & Rogers, T. (1982). Maternalperception of maladjustment in clinic-referred children: An extension of earlierresearch. Journal of Behavioral Assessment, 4, 145-151.Forgatch, M.S., Patterson, G.R., & Skinner, M. (1988). A mediational model for theeffect of divorce in antisocial behavior in boys. In E.M. Hetherington & J.D.Arasteh (Eds.), Impact of divorce, single parenting, and step-parenting on children(pp. 135-154). Hlilsdale, NJ: Lawrence Eribaum Associates, Inc.Foster, S.L. & Martinez, C.R. (1995). Ethnicity: Conceptual and methodological issues inchild clinical research. Journal of Clinical Child Psychology, 24(2), 2 14-226.Frick, P.1, Lahey, B.B., Loeber, R., Stouthamer-Loeber, M., Christ, M.G., & Hanson, K.(1992). Familial risk factors to Oppositional Defiant Disorder and ConductDisorder: Parent psychopathology and maternal parenting. Journal of Consultingand Clinical Psychology, Q(1), 49-55.Frydman, M.I. (1981). Social support, life events and psychiatric symptoms: A study ofdirect, conditional and interaction effects. Social Psychiatry, ], 69-78.Furnham, A. & Bochner, S. (1986). Culture shock: Psychological reactions to unfamiliarenvironments. London: Methuen.Gaines, R., Sandgrund, A., Green, A. H., & Power, E. (1978). Etiological factors in childmaltreatment; A multivariate study of abusing, neglecting and normal mothers.Journal of Abnormal Psychology, 7, 531-540.Garbarino, J. (1976). A preliminary study of some ecological correlates of child abuse:The impact of socioeconomic stress on others. Child Development, 47, 178-185.137Garmezy, N. (1983). Stressors of childhood. In N. Garmezy & M. Rutter (Eds.), Stress,coping. and development in children (pp. 43-84). New York: McGraw-Hill.Garmezy, N. (1985). Stress-resistent children: The search for protective factors. In J.E.Stevenson (Ed.), Recent research in developmental psychopathology. Journal ofChild Psychology and Psychiatry Book Supplement. No. 4 (pp. 2 13-233). Oxford:Pergamon Press.Glover, G.R. (1989). The pattern of psychiatric admissions of Caribbean-born immigrantsin London. Social Psychiatry, 4, 49-56.Goodstein, L.D. & Rowley, V.N. (1961) A further study of MMPI differences betweenparents of disturbed and non-disturbed children. Journal of ConsultingPsychology, , 460-464.Goodyer, I.M., Kolvin, I., Gatzanis, S. (1987). The impact of recent undesirable lifeevents on psychiatric disorders in childhood and adolescence. British Journal ofPsychiatry, 151, 179-184.Gresham, F.M. & Effiott, S.N. (1990). Social Skills Rating System Manual. Circle Pines,MN: American Guidance Service.Griest, D.L., Forehand, R., Wells, K.C., & McMahon, R.J. (1980). An examination ofdifferences between non-clinic and behavior problem clinic referred children.Journal of Abnormal Psychology, , 277-281.Guest, D.L., Wells, K.C., & Forehand, R. (1979). An examination of predictors ofmaternal perceptions of maladjustment in clinic-referred children. Journal ofAbnormal Psychology, , 277-281.Grinberg, L. & Grinberg, R. (1984). A psychoanalytic study of migration: Its normal andpathological aspects. Journal of the American Psychoanalytic Association, (1),13-38.Hammen, C., Burge, D., & Stansbury, K. (1990). Relationship of mother and childvariables to child outcomes in a high-risk sample: A causal modeling analysis.Developmental Psychology, (l), 24-30.Harper, J. & Williams, S. (1976). Infantile autism: The incidence of national groups in aNew South Wales survey. Medical Journal ofAustralia,!, 299-301.Harrison, A.O., Wilson, M.N., Pine, C.J., Chan, S.Q., & Buriel, R. (1990). Familyecologies of ethnic minority children. Child Developmçfl, j, 347-364.138Hayashi, K., Toyama, B., & Quay, B.C. (1976). A cross-cultural study concerned withdifferential behavioral classification. I. The Behavior Checklist. Japanese Journalof Criminal Psychology, , 2 1-28.Herrenkohi, E.C., Herenkohi, R.C., Toedtker, L. & Yanushefski, A.M. (1984). Parentchid interactions in abusive and nonabusive families. Journal of the AmericanAcademy of Child Psychiatry, , 64 1-648.Hetherington, E.M., Cox, M., & Cox, R. (1982). Effects of divorce on parents andchildren. In M. Lamb (Ed.), Nontraditional families (pp.133-178). HIlisdale, NJ:Erlbaum.Hicks, R., Lalonde, R.N., & Pepler, D. (1993). Psychosocial considerations in the mentalhealth of immigrant and refugee children. Canadian Journal of Community MentalHealth, .j(2), 71-87.Hightower, A.D., Cowen, E.L., Spinell, A.P., Lotyczewski, B.S., Guare, J.C., Rohrbeck,C.A., & Brown, L.P. (1987). The Child Rating Scale: The development of asocioemotional self-rating scale for elementary school children. SchoolPsychology Review, i(2), 239-255.Hightower, A.D., Spinell, A., & Lotyczewski, B.S. (1990). Primary Mental HealthProject: Child Rating Scale (CRS) Guidelines. Rochester, NY: Primary MentalHealth Project, Inc.Hitch, P.J. & Rack, P.H. (1980). Mental illness among Polish and Russian refugees inBradford. British Journal of Psychiatry, 206-211,Holahan, C.J. & Moos, R.H. (1987). Risk, resistance and psychological distress: Alongitudinal analysis with adults and children. Journal of Abnormal Psychology,, 3-13.Hollingshead, A.B. (1975). Four factor index of social status. Unpublished manuscript,Yale University, Department of Sociology, Connecticut.Hong, G.K. (1989). Application of cultural and environmental issues in family therapywith immigrant Chinese Americans. Journal of Strategic and Systemic Therapies,, 14-21.Hops, H., Biglan, A., Sherman, L., Arthur, J., Friedman, L., & Osteen, V. (1987). Homeobservations of family interactions of depressed women. Journal of Consulting andClinical Psychology, (3), 34 1-346.139Hops, H., Sherman, L., & Biglan, A. (1990). Maternal depression, marital discord, andchildren’s behavior: A developmental perspective. In G.R. Patterson (Ed.),Depression and aggression in family interaction (pp. 185-208). Hifisdale, NJ:Eribaum.House, J.S. (1981). Work stress and social support. Reading, MA: Addison-Wesley.House, J.S., Landis, K.R., & Umberson, D. (1988). Social relationships and health.Science, 4j, 540-545.House, J.S., Robbins, C., & Metzner, H.L. (1982). The association of social relationshipsand activities with mortality: Prospective evidence from the Tecumseh CommunityHealth Study. American Journal ofEpidemiology, 116, 123-140.Hurh, W.M. & Kim, K.C. (1990). Correlates of Korean immigrants’ mental health.Journal ofNervous and Mental Disease, jJ, 703-711.Ishiyama, F.I. (1989). Understanding foreign adolescents’ difficulties in cross-culturaladjustment: A self-validation model. Canadian Journal of School Psychology, ,4 1-56.Jaccard, J., Wan, C.K., & Turrisi, R. (1990). The detection and interpretation ofinteraction effects between continuous variables in multiple regression.Multivariate Behavioral Research, (4), 467-478.Jacobson, S.W. & Frye, K.F. (1991). Effect of maternal social support on attachment:Experimental evidence. Child Development, (3), 572-582.Jay, S.M., Ozolins, M., Elliott, C.H., & Caidwell, S. (1983). Assessment of children’sdistress during painfhl medical procedures. Health Psychology, , 13 3-147.Jensen, P.S., Xenakis, S.N., Davis, H., & DeGroot, J. (1988). Child psychopathologyrating scales and interrater agreement: II. Child and family characteristics. Journalof the American Academy of Child Psychiatry, 7, 451-461.Jessor, R. (June, 1979). The perceived environment and the study of adolescent problembehavior. Paper presented at the Symposium on the Situation in PsychologicalTheory and Research at LOVIK, Stolkholm, Sweden.Johnson, J.H. (1986). Life events as stressors in childhood and adolescence. BeverlyHills, CA: Sage.Johnston, C. & Short, K.H. (1993). Depressive symptoms and perceptions of childbehavior. Journal of Social and Clinical Psychology, j, 164-181.140Justice, B. & Justice, R. (1976). The abusing family. New York: Human Sciences Press.Kallarackal, AM. & Herbert, M. (1976). The happiness of Indian immigrant children.New Society, , 422-424.Kanner, A.D., Coyne, J.C., Schaefer, C., & Lazarus, R.S. (1981). Comparison of twomodes of stress measurement: Daily hassles and uplifts versus major life events.Journal of Behavioral Medicine, 4, 1-3 9.Kaslow, N.J., Rehm, L.P., & Siegel, A.W. (1984). Social-cognitive and cognitivecorrelates of depression in children. Journal of Abnormal Child Psychology, j,605-620.Kazdin, A.E. (1992). Child and adolescent dysfunction and paths toward maladjustment:Targets for intervention. Clinical Psychology Review, j, 795-817.Kazdin, A.E. (1995). Scope of child and adolescent psychotherapy research: Limitedsampling of dysfunctions, treatments, and client characteristics. Journal of ClinicalChild Psychology, 4(2), 125-140.Kessler, R.C. (1982). Life events, social support, and mental health. In W.R. Gove (Ed.),Deviance and mental illness (pp.247-271). Beverly Hills, CA: Sage.Kessler, R.C. & McLeod, J. (1984). Social support and mental health in communitysamples. In S. Cohen & S.L. Syme (Eds.), Social support and health (pp. 3-22).New York: Academic Press.Kessler, R.C. & Neighbours, H. (1986). A new perspective on the relationships amongrace, social class, and psychological distress. Journal of Health and SocialBehavior, 7, 107-115.Kim, J.H., Kim K., & Won, H.T. (1983). Symptom Checklist-90-Revision (SCL-90-R) inpsychiatric outpatients. Mental Health Research, 150-168.Kleinman, A. (1987). Anthropology and psychiatry: The role of culture in cross-culturalresearch on illness. British Journal ofPsychiatry, J5i, 447-44.Kleinman, A. & Kleinman, J. (1985). Somatization: The interconnections in Chinesesociety among culture, depressive experiences, and the meanings of pain. In A.Kleinman & B. Good (Eds.), Culture and depression: Studies in the anthropologyand cross-cultural psychiatry of affect and disorder (pp. 429-490). Berkeley, CA:University of California Press, Ltd.Koeske, G.F. (1992). Moderator variables in social work research. Journal of SocialService Research, j., 159-178.141Krech, K.H. & Johnston, C. (1992). The relationship of depressed mood and life stress tomaternal perceptions of child behavior. Journal of Clinical Child Psychology.(2), 115-122.Krupinski, J. (1967). Sociological aspects of mental ill-health in migrants. Social Scienceand Medicine, 1, 267-281.Kuo, W.H. & Tsai, Y.M. (1986). Social networking, hardiness and immigrant’s mentalhealth. Journal of Health and Social Behavior, 7, 133-149.Lai, M.C. & Yue, K.M.K. (1990). The Chinese. In N. Waxier-Morrison, J.M. Anderson,& E. Richardson (Eds.), Cross-cultural caring: A handbook for healthprofessionals in Western Canada (pp.73-89). Vancouver: University of BritishColumbia Press.Lambert, M.C., Weisz, J.R., & Thesiger, C. (1989). Principal components analyses ofbehavior problems in Jamaican clinic-referred children: Teacher reports for ages 6-17. Journal of Abnormal Child Psychology, j.7(5), 553-562.Lazarus, R.S., DeLongis, A., Folkman, S., & Gruen, R. (1985). Stress and adaptationaloutcomes: The problem of confounded measures. American Psychologist, 4Q(7),770-779.Lazarus, R.S. & Folkman, S. (1984). Stress, appraisal. and coping. New York: Springer.Lee, E. (1982). A social systems approach to assessment and treatment for ChineseAmerican families. In M. McGoldrick, J.K. Pearce, & J. Giordano (Eds.),Ethnicity and Family Therapy (pp. 527-55 1). New York: Guilford Press.Lepore, S. J., Palsane, M.N., & Evans, G.W. (1991). Daily hassles and chronic strains: Ahierarchy of stressors? Social Science and Medicine, (9), 1029-1036.Lobitz, O.K. & Johnson, S.M. (1975). Normal versus deviant children: A multimethodcomparison. Journal ofAbnormal Child Psychology, (4), 353-375.Loeber, R. (1990). Development and risk factors of juvenile antisocial behavior anddelinquency. Clinical Psychology Review, ]LQ, 1-41.Loeber, R., Green, S.M., Lahey, B.B., & Stouthamer-Loeber, M. (1991). Differences andsimilarities between children, mothers, and teachers as informants on disruptivebehavior disorders. Journal of Abnormal Child Psychology, j, 75-95.Longfellow, C., Zelkowitz, P., & Saunders, E. (1982). The quality of mother-childrelationships. In D. Belle (Ed.), Lives in stress: Women and depression (pp.163-176). Beverly Hills, CA: Sage.142Lysgaard, S. (1955). Adjustment in a foreign society: Norwegian Fuibright granteesvisiting the United States. International Social Science Bulletin, , 45-51.Marsella, A.J. (1980). Depressive experience and disorder across cultures. In H.C.Triandis & J.D. Draguns (Eds.), Handbook of cross-cultural psychology. vol. 6(pp. 237-290). Boston: Allyn and Bacon.Masten, A.S. & Garmezy, N. (1985). Risk, vulnerability, and protective factors indevelopmental psychopathology. In B. B. Lahey & A.E. Kazdin (Eds.), Advancesin clinical child psychology. vol. 8 (pp. 1-52). New York: Plenum Press.Masten, A.S., Morison, P., Pellegrini, D., & Tellegen, A. (1992). Competence understress: Risk and protective factors. In 3. Rolf A.S. Masten, D. Cicchetti, K.H.Nuechterlein, & S. Weintraub (Eds.), Risk and protective factors in thedevelopment of psychopathology (pp. 236-256). Cambridge: CambridgeUniversity Press.McFarlane, A.C., Policansky, S.K., & Irwin, C. (1987). A longitudinal study of thepsychological morbidity in children due to a natural disaster. PsychologicalMedicine, , 727-73 8.Mednick, S.A. & Schulsinger, F. (1968). Some premorbid characteristics related tobreakdown in children with schizophrenic mothers. In D. Rosenthal & S.S. Kety(Eds.), The transmission of schizophrenia (pp. 267-29 1). Oxford: PergamonPress.Middlebrook, J. & Forehand, R. (1985). Maternal perceptions of deviance in childbehavior as a function of stress behavior and clinic vs. non-clinic status of thechild: An analogue study. Behavior Therapy, j, 494-502.Minde, K., & Minde, R. (1976). Children of immigrants: The adjustment of UgandanAsian primary-school children in Canada. Canadian Psychiatric AssociationJournal, j, 371-381.Mindel, C.H. & Habenstein, R.W. (1981). Family lifestyles of America’s ethnic minorities:An introduction. In C.H. Mmdcl & R.W. Habenstein (Eds.), Ethnic families inAmerica: Patterns and variations (pp. 1-13). New York: Elsevier, North Holland.Monroe, S.M. (1983). Major and minor life events as predictors of psychological distress:Further issues and findings. Journal of Behavioral Medicine, , 189-205.Monroe-Blum, H., Boyle, M.H., Offord, D.R., & Kates, N. (1989). Immigrant children:Psychiatric disorder, school performance and service utilization. American Journalof Orthopsychiatry, , 510-519.143Mullins, L.L., Siegel, L.J., & Hodges, K. (1985). Cognitive problem solving and life eventcorrelates of depressive symptoms in children. Journal of Abnormal ChildPsychology,.i, 305-314.Mumford, D.B. (1989). Somatic sensations and psychological distress among students inBritain and Pakistan. Social psychiatry, 4, 32 1-326.Murphy, H.B.M. (1973). Uprooting and after. New York: Springer-Verlag.Myers, H.F., Taylor, S., Alvy, K.T., Arrington, A., & Richardson, M.A. (1992). Parentaland family predictors of behavior problems in inner-city Black children. AmericanJournal of Community Psychology, (5), 557-576.Myers, J.K. & Weissman, M.M. (1980). Use of a self-report symptom scale to detectdepression in a community sample. American Journal of Psychiatry, j, 108 1-1084.Nakagawa, M., Teti, D.M., Lamb, M.E., & Shigemura, T. (April, 1991). Japanesemothers and children in the U.S.: Stress, support. parenting. and attachment.Paper presented at the Biennial Conference of the Society for Research in ChildDevelopment, Seattle, WA.Nann, (1982). Settlement programs for immigrant women and families. In R.C. Nann(Ed.). Uprooting and Surviving. Dordrecht: Holland: D. Reidel PublishingCompany.Ng, K.K.Y. (1993). Volunteer work and settlement: A study of Chinese immigrantwomen. Canadian Journal of Community Mental Health, IZ(2), 31-45.Nicassio, P.M. & Pate, J.K. (1984). An analysis of problems of resettlement of theIndochinese refugees in the United States. Social Psychiatry,.l, 135-141.Noh, S., Avison, W.R., & Kaspar, V. (1992). Depressive symptoms among Koreanimmigrants: Assessment of a translation of the Center for Epidemiologic Studies-Depression scale. Psychological Assessment, 4(1), 84-91.Oberg, K. (1960). Culture shock: Adjustment to new cultural environments. PracticalAnthropology, 2, 177-182.Offord, D., Rae-Grant, T.L., Links, G.S., Cadman, R.W., Byles, S.R., Crawford, D.M.,Monroe-Blum, C.R., Byrne, E.D., Thomas, W.S., Woodward, T.R., Boyle, I.L., &Szatmari, P. (1987). Ontario Child Health Study: II. 6-month prevalence ofdisorder and rates of service utilization. Archives of General Psychiatry, 44, 832-836.144Padila, A.M., Cervantes, R.C., Maldonado, M. & Garcia, RE. (1988). Coping responsesto psychosocial stressors among Mexican and Central American immigrants.Journal of Community Psychology, !, 418-427.Parke, R.D. & Bhavnagri, N.P. (1988). Parents as managers of children’s peerrelationships. In D. Belle (Ed.), Children’s social networks and social supports(pp. 24 1-259). New York: Wiley.Parke, R.D., MacDonald, K.B., Beitel, A., & Bhavnagri, N. P.(1988). The role of thefamily in the development of peer relationships. In R. Peters & R.J. MeMahon(Eds.), Social learning systems: Approaches to marriage and the family (pp. 17-44). New York: Brunner-Mazel.Parker, J.G. & Asher, S.R. (1987). Peer relations and later personal adjustment: Are low-accepted children at risk? Psychological Bulletin, ].Q (3), 357-389.Patterson, G.R. (1982). Coercive Family Process. Eugene, Oregon: Castalia.Patterson, G.R. (1983). Stress: A change agent for family process. In N. Garmezy andM. Rutter (Eds.), Stress, coping, and development in children (pp. 235-264). NewYork: McGraw-Hill.Patterson, G.R. (1986). Perfonnance models for antisocial boys. American Psychologist,41(4), 432-444.Patterson, G.R., & Capaldi, D.M. (1992). A mediational model for boy& depressed mood.In J.Rolf A.S. Masten, D. Cicchetti, K.H. Nuechterlein, & S. Weintraub (Eds,),Risk and protective factors in the development of psychopathology (pp. 141-163).Cambridge: Cambridge University Press.Paulhus, D.L. (1991). Measurement and control of response bias. In J.P. Robinson, P.R.Shaver, & L.S. Wrightman (Eds.), Measures of Personality and SocialPsychological Attitudes. Vol. 1. (pp. 17-58). San Diego, CA: Academic Press.Piacentini, J.C., Cohen, P. & Cohen, J. (1992). Combining discrepant diagnosticinformation from multiple sources: Are complex algorithms better than simpleones? Journal 9fAbnormal Child Psychology, Q(1), 65-82.Pianta, R.C., Egeland, B., & Sroufe, L.A. (1992). Maternal stress and children’sdevelopment: Prediction of school outcomes and identification of protectivefactors. In J. Rolf A. S. Masten, D. Cicchetti, K.H. Nuechterlein, & S. Weintraub(Eds.), Risk and protective factors in the development of psychopathology (pp.215-235). Cambridge: Cambridge University Press.145Puig-Antich, J., Lukens, E., Davies, M., Goetz, D., Brennan-Quattrock, J., & Todak, G.(1985). Psychosocial functioning in prepubertal major depressive disorders:Interpersonal relationships during the depressive episode. Archives of GeneralPsychiatry, 4, 500-507.Quay, H.C. & Parskeuopoulos, I.N. (1972, August). Dimensions of problem behavior inelementary school children in Greece. Iran and Finland. Paper presented at theXXth International Congress of Psychology, Tokyo, Japan.Rabkin, J.G. & Struening, EL. (1976). Life events, stress, and illness. Science, 124,10 13-1020.Radloff L.S. (1977). The CES-D scale: A self-report depression scale for research in thegeneral population. Applied Psychological Measurement, 1, 385-401.Rae Grant, N. & Crill Russell, C. (1989). Better Beginnings, Better Futures: Anintegrated model of primary prevention of emotional and behavioral problems.Kingston: Queen’s Printer for Ontario.Raphael, E.I. (1988). Grandparents: A study of their role in Hispanic families. Physicaland Occupational Therapy in Geriatrics, , 31-62.Reich, W. & Earls, F. (1987). Rules for making psychiatric diagnoses in children on thebasis of multiple sources of information: Preliminary strategies. Journal ofAbnormal Child Psychology, j., 601-616.Reid, W.J. & Crisafulli, A. (1990). Marital discord and child behavior problems: A metaanalysis. Journal of Abnormal Child Psychology, 1, 105-117.Rickard, K., Forehand, R., Wells, K., Griest, D., & MeMahon, R. (1981). Factors in thereferral of children for behavioral treatment: A comparison of mothers of clinic-referred deviant, clinic-referred non-deviant and nonclinic children. BehaviorResearch & Therapy, 12, 20 1-205.Richters, J.E. (1992). Depressed mothers as informants about their children: A criticalreview of the evidence for distortion. Psychological Bulletin, ll, 485-499.Roberts, R.E. & Vernon, S.W. (1983). The Center for Epidemiologic Studies-Depressionscale: Its use in a community sample. American Journal ofPsychiatry, j4Q, 4 1-46.Roberts, R.E., Vernon, S.W., & Rhoades, H.M. (1989). Effects of language and ethnicstatus on reliability and validity of the Center for Epidemiologic Studies-Depression scale with psychiatric patients. Journal of Nervous and MentalDisease, 177, 581-592.146Roff M., Sells, S.B., & Golden, M.M. (1972). Social adjustment and personalitydevelopment in children. Minneapolis: University of Minnesota Press.Roff J.D. & Wirt, D. (1984). Childhood aggression and social adjustment as antecedentsof delinquency. Journal of Abnormal Child Psychology, j, 111-126.Rogler, L.H. (1994). International migrations: A framework for directing research.American Psychologist, 4(8),701-708.Rogosch, F.A., Mowbray, C.T., & Bogat, GA. (1992). Determinants of parentingattitudes in mothers with severe psychopathology. Development andPsychopathology, 4, 469-487.Rumbaut, RD. (1985). Mental health and the refugee experience: A comparative study ofSoutheast Asian refugees. In T. Owan, B. Bliatout and K.M. Lin (Eds.), SoutheastAsian mental health: Treatment, prevention, services, and research (DHHSPublication ADM-1399) (pp. 433-486). Rockville, MD: National Institute ofMental Health.Rutter, M. (1966). Children of sick parents: An environmental and psychiatric study.Institute of Psychiatry Maudsley Monographs No. 16. London: Oxford UniversityPress.Rutter, M. (1970). Sex differences in children’s responses to family stress. In E.J.Anthony & C. Koupernik (Eds.), The child in his family (pp. 165-196). NewYork: Wiley.Rutter, M. (1992). Psychosocial resilience and protective mechanisms. In J. Rolf A.S.Masten, D. Cicchetti, K.H. Nuechterlein, & S. Weintraub (Eds.), Risk andprotective factors in the development of psychopathology (pp. 2 15-235).Cambridge: Cambridge University Press.Rutter, M. & Quinton, D. (1984). Parental psychiatric disorder: Effects on children.Psychological Medicine, 14, 853-880.Rutter, M., Yule, M., Berger, M., Yule, B., Morton, 3., & Bagley, C. (1974). Children ofWest Indian immigrants: 1. Rates of behavioral deviance and of psychiatricdisorder. Journal of Child Psychology and Psychiatry, 3, 24 1-262.Sameroff A.J. (1987). The social context of development. In N. Eisenberg (Ed.),Contemporary topics in developmental psychology (pp. 273-291). New York:Wiley.Sameroff A.J. & Seifer, R. (1983). Familial risk and child competence. ChildDevelopment, 54, 1254-1268.147Sarason, LG., Levine, H.M., Basham, R.B., & Sarason, B.R. (1983). Assessing socialsupport: The Social Support Questionnaire. Journal of Personality and SocialPsychology, 44, 127-139.Saunders, D.R. (1956). Moderator variables in prediction. Educational and PsychologicalMeasurement, j, 209-222.Schaughency, E.A. & Lahey, B.B. (1985). Mothers’ and fathers’ perceptions of childdeviance: Roles of child behavior, parental depression, and marital satisfaction.Journal of Consulting and Clinical Psychology, , 7 18-723.Scheier, M.F., & Carver, C.S. (1985). Optimism, coping, and health: Assessment andimplications of generalized outcome expectancies. Health Psychology, 4, 2 19-247.Siegel, J.M. & Brown, J.D. (1988). A prospective study of stressful circumstances, illnesssymptoms, and depressed mood among adolescents. Developmental Psychology,4, 715-721.Selman, R.L., Hickey Shultz, L., Nakkula, M., Barr, D., Watts, C., & Richmond, J.B.(1992). Friendship and fighting: A developmental approach to the study of riskand prevention ofviolence. Development and Psychopathology, 4, 529-558.Shure, M. & Spivack, G. (1982). Interpersonal problem-solving in young children: Acognitive approach to prevention. American Journal of Community Psychology,jQ, 341-357.Simoes, M. & Binder, J. (1980). A socio-psychiatnc field study among Portugueseemigrants in Switzerland. Social Psychiatry, j, 1-7.Skinner, E. (1985). Determinants of mother sensitive and contingent-responsive behavior:The role of child-rearing beliefs and SES. In I.E. Sigel (Ed.), Parental beliefsystems: The psychological consequences for children. New Jersey: LawrenceEribaum Associates.Sokolofl B., Carlin, J., & Pham, H. (1984). Five-year follow-up of Vietnamese refugeechildren in the United States. Clinical Pediatrics, (10), 565-570.Solis, M.L. & Abidm, R.R. (1991). The Spanish version Parenting Stress Index: Apsychometric study. Journal of Clinical Child Psychology, Q(4), 372-378.Stark, K.D., Humphrey, L.L., Crook, K., & Lewis, K. (1990). Perceived familyenvironments of depressed and anxious children: Childs and maternal figure’sperspective. Journal of Abnormal Child Psychology, j(5), 527-547.148Statistical Package for the Social Sciences, Inc. (1992). SPSS for Windows: Base systemuser’s guide, release 6.0. Chicago, Illinois: SPSS, Inc.Steinhausen, H.C. (1985). Psychiatric disorders in children and family dysfunction. SocialPsychiatry, , 11-16.Sterling, S., Cowen, E.L., Weissberg, R.P., Lotyczewski, B.C., & Boike, M. (1985).Recent stressful life events and young children’s school adjustment. AmericanJournal of Community Psychology, i, 87-99.Straus, M.A. (1980). Stress and child abuse. In C.H. Kempe & R.E. Heifer (Eds.), Thbattered child. 3rd ed. (pp. 86-103). Chicago: University of Chicago Press.Sue, D. & Sue, S. (1987). Cultural factors in the clinical assessment of Asian Americans.Journal of Consulting and Clinical Psychology, (4), 479-487.Sue, D.W. & Sue, D. (1990). Counseling the culturally different: Theory and practice,2nd edition. New York: John Wiley & Sons.Sue, S. & Chin, R. (1983). The mental health of Chinese-American children: Stressorsand resources. In G.J. Powell (Ed.), The psychosocial development of minoritygroup children (pp. 3 85-397). New York: Brunner / Maze!.Sue, S. & Monshima, J.K. (1982). The mental health of Asian Americans. San Francisco:Jossey-Bass.Suen, P. & Ng, S. (1987). Background paper on aspects of Chinese-Canadian culture.Burnaby, B.C.: Burnaby Multicultural Council.Taft, R. (1977). A comparative study of the initial adjustment of immigrant schoolchildrenin Australia. International Migration Review, 13(1), 71-81.Tarullo, L.B., Richardson, D.T., Radke-Yarrow, M., & Martinez, P.E. (1995). Multiplesources in child diagnosis: Parent-child concordance in affectively ill and wellfamilies. Journal of Clinical Child Psychology, 4(2), 173-183.Ten-is, M. (1973). Approaches to the epidemiology of health. American Journal of PublicHealth, , 1037-1045.Touliatos, J. & Lindholm, B.W. (1980). Behavioral disturbance in children of native-bornand immigrant parents. Journal of Community Psychology, , 28-33.Trovato, F. (1986). A time series analysis of international immigration and suicidemortality in Canada. International Journal of Social Psychiatry, (2), 3 8-46.149Tseng, W.S. & Hsu, J. (1980). Minor psychological disturbances of everyday life. InH.C. Triandis & J.D. Draguns (Eds.), Handbook of cross-cultural psychology. vol.6 (pp. 54-98). Boston: Allyn and Bacon.Tyhurst, L. (1982). Coping with refugees. A Canadian experience: 1948-1981.International Journal of Social Psychiatry, 28(2), 105-109.Vancouver School Board (1990). Oakridge Reception and Orientation Centreregistrations: 1989 July to 1990 June. Unpublished raw data.Vancouver School Board (1991). Oakridge Reception and Orientation Centreregistrations: 1990 July to 1991 June. Unpublished raw data.Vancouver School Board (1992). Oakridge Reception and Orientation Centreregistrations: 1991 July to 1992 June. Unpublished raw data.Varni, J.W., Wilcox, K.T., & Hanson, V. (1988). Mediating effects of family socialsupport on child psychological adjustment in juvenile rheumatoid arthritis. HealthPsychology, 7(5), 421-431.Vega, W.A., Kolody, B., & Valle, J.R. (1987). Migration and mental health: An empiricaltest of depression risk factors among immigrant Mexican women. InternationalMigration Review, j(3), 512-530.Wagner, B.M., Compas, B.E., & Howell, D.C. (1988). Daily and major life events: A testof an integrative model of psychosocial stress. American Journal of CommunityPsychology, j., 189-205.Wahier, R.G. (1980). The insular mother: Her problems in parent-child treatment.Journal of Applied Behavior Analysis, j, 207-2 19.Wallerstein, J.S., & Kelly, J. (1980). Surviving the breakup: How children and parentscope with divorce. New York: Basic Books.Walsh, A. & Walsh, P.A. (1987). Social support, assimilation, and biological effectiveblood pressure levels. International Migration Review, i(3), 577-591.Waters, E. & Sroufe, L.A. (1983). A developmental perspective on competence.Developmental Review, j, 5 9-79.Watt, N.F., Anthony, E.J., Wynne, L.C., & Rolf J.E. (Eds.). (1984). Children at risk forschizophrenia. New York: Cambridge University Press.Way, R.T. (1985). Burmese culture, personality, and mental health. Australian and NewZealand Journal of Psychiatry, j(3), 275-282.150Webster-Stratton, C. (1988). Mothers’ and fathers’ perceptions of child deviance: Roles ofparent and child behaviors and parent adjustment. Journal of Consulting andClinical Psychology, (6), 909-9 15.Webster-Stratton, C. (1990). Stress: A potential disrnpter of parent perceptions andfamily interactions. Journal of Clinical Child Psychology, i(4), 302-312.Webster-Stratton, C. & Hammond, M. (1988). Maternal depression and its relationship tolife stress, perceptions of child behavior problems, parenting behaviors and childconduct problems. Journal of Abnormal Child Psychology, j, 299-3 15.Weinraub, M. & Ansul, S. (1984, April). Children’s responses to strangers: Effects offamily status, stress and mother-child interaction. Paper presented at the FourthBiennial conference on Infant Studies, New York.Weinraub, M. & Wo1f B. M. (1983). Effects of stress and social supports on mother-child interactions in single and two-parent families. Child Development, 4, 1297-1311.Weissberg, R.P., Caplan, M., & Harwood, R.L. (1991). Promoting competent youngpeople in competence-enhancing environments: A systems-based perspective onprimary prevention. Journal of Consulting and Clinical Psychology, (6), 830-841.Weissman, M.M. & Paykel, E.S. (1974). The depressed woman. Chicago: University ofChicago Press.Weissman, M.M., Paykel, E.S., & Kierman, G.L. (1972). The depressed woman asmother. Social Psychiatry, 7, 98-108.Weisz, J.R., Suwanlert, S., Chaiyasit, W., & Walter, B.R. (1987). Over- and under-controlled referral problems among Thai and American children and adolescents:The wat and wai of cultural differences. Journal of Consulting and ClinicalPsychology, 5, 7 19-726.Werner, E.E. & Smith, R.S. (1982). Vulnerable but invincible: A longitudinal study ofresilient children and youth. New York: McGraw-Hill.Westermeyer, J. (1988). A matched pairs study of depression among Hmong refugeeswith particular reference to predisposing factors and treatment outcome. SocialPsychiatry, , 64-87.Wethington, E. & Kessler, R.C. (1986). Perceived support, received support, andadjustment of stressful life events. Journal of Health and Social Behavior, 7, 78-89.151Wheaton, B. (1985). Models for the stress-buffering fi.inctions of coping resources.Journal ofHealth and Social Behavior, , 352-364.Williams, P., Tarnopoisky, A., & Hand, D. (1980). Case definition and case identificationin psychiatric epidemiology: Review and assessment. Psychological Medicine, IQ,101-114.Wilson, M.N. (1986). The black extended family: An analytical consideration.Developmental Psychology, , 246-258.Wilson, W.J. (1987). The truly disadvantaged: The inner city, the underclass, and publicpolicy. Chicago: University of Chicago Press.Wing, J.K., Bebbington, P., & Robins, L. N. (Eds.) (1981). What is a case? London:Grant McIntyre.Wolfe, D.A., Jaffe, P., Wilson, S.K., & Zak, L. (1985). Children of battered women: Therelation of child behavior to family violence and maternal stress. Journal ofConsulting and Clinical Psychology, (5), 657-665.Wyman, P.A., Cowen, E.L., Work, W.C., & Parker, OR. (1991). Developmental andfamily milieu correlates of resilience in urban children who have experienced majorlife stress. American Journal of Community Psychology, W(3), 405-426.Yamamoto, J. & Acosta, F.X. (1982). Treatment of Asian Americans and HispanicAmericans: Similarities and differences. American Academy of Psychoanalysis, j.Q,585-607.Yao, K., Solanto, M.V., & Wender, E.H. (1988). Prevalence of hyperactivity amongnewly immigrated Chinese-American children. Developmental and BehavioralPediatrics, 9, 367-373.Yule, W. (1992). Resilience and vulnerability in child survivors of disasters. In B. Tizard& V. Varma (Eds.), Vulnerability and resilience in human development. London:Jessica Kingsley Publishers.Yule, W., & Williams, R. (1990). Post-traumatic stress reactions in children. Journal ofTraumatic Stress, (2), 279-295.Zahner, G.E.P., Pawelkiewicz, W., DeFrancesco, J.J., & Adnopoz, T. (1992). Children’smental health service needs and utilization patterns in an urban community: Anepidemiological assessment. Journal of the American Academy of Child andAdolescent Psychiatry, j, 95 1-960.152Zekoski, E.M., O’Hara, M.W., & Wills, K.E. (1987). The effects of maternal mood onmother-infant interaction. Journal of Abnormal Child Psychology, j(3), 361-378.Zhang, D. (1994). Depression and culture: A Chinese perspective. Manuscript submittedfor publication.Zill, N. & Schoenborn, C.A. (1990, November). Developmental, learning, and emotionalproblems: Health of our nation’s children, United States 1988. Advance Data:National Center for Health Statistics, Number 190.Zimet, G.D., Dahiem, N.W., Zimet, S.G., & Farley, G.K. (1988). The MultidimensionalScale of Perceived Social Support. Journal of Personality Assessment, Z(1), 30-41.Zussman, J.V. (1980). Situational determinants of parental behavior: Effects ofcompeting cognitive activity. Child Development, 51, 792-800.153Appendix AFamily Information QuestionnaireTodays date (day/month/year):_____________________________________Your date ofbirth (day/month/year):_______________________Date ofyour arrival in Canada (month/year):____Your highest (completed) education level:__ __(e.g., grade 9, B.A., one year of nursing college, etc.)Spouse’s highest (completed) education level:_Your present occupation (be specific):___ _Spouse’s present occupation (be specific):Your occupation in home country (be specific):Spouse’s occupation in home country (be specific):_Generally, has your family income increased or decreased since coming to Canada? (checkone)* *increased or a little a moderate a largeno change_ decrease decrease_ decrease_Age (s), and gender (s) of child (ren) (if you have more than one child, place a * besidethe child that you have been asked to think about for this study):When did this child (*) arrive in Canada (month/year)?__How many months in a year do you and this child live in B.C.?How many months in a year does your spouse live in B.C.?Does anyone else live in your home with you (e.g., your mother your spouse’s parents, afemale friend, a housekeeper)? Please list these members ofyour household.How many members ofyour extended family (e.g., parents, sisters, aunts, in-laws) live inthe Lower Mainland?___________Do you plan to return to Hong Kong to live?If yes, approximately when do you plan to go back?[**Item contributed to the total score of the Mother Immigration Stress Scale.]154Appendix BImmigration Stress Scale1. How much stress did you experience in the year before you came to Canada?no stress_ a little_ quite a bit_ extreme stress_2. How much stress did you experience during the move itself?no stress_ a little_ quite a bit_ extreme stress3. How much stress have you experienced since settling in Canada?no stress_ a little quite a bit_ extreme stress_4. How much stress did your child experience in the year before coming to Canada?*no stress_ a little_ quite a bit_ extreme stress_5. How much stress did your child experience during the move itself?*no stress_ a little quite a bit extreme stress_6. How much stress has your child experienced since settling in Canada?*no stress_ a little_ quite a bit_ extreme stress_7. Please rate your English language fluency: (check one)SPOKEN: poor_ fair_ good_ fluent_WRITTEN: poor_ fair good_ fluent_8. Please rate your child’s English language fluency: (check one)*SPOKEN: poor_ fair_ good_ fluentWRITTEN: poor_ fair_ good fluent_9. How bothered do you think that your child is by things like racial discrimination andhomesickness (for people and things back in Hong Kong)?*not at all_ a little_ quite a bit_ extremely_155Appendix B, continuedThis week, how much of a hassle was (were):54. Language issues?**none/not applicable_ somewhat quite a bit_ extremely_55. Thinking about people/things in home country?**none/not applicable_ somewhat_ quite a bit_ extremely56. Availability of cultural food preferences?* *none/not applicable_ somewhat_ quite a bit_ extremely_57. Canadian ways of doing things?**none/not applicable_ somewhat_ quite a bit extremely_58. Racial discrimination?**none/not applicable_ somewhat_ quite a bit_ extremely_[*These items together comprised the Child Immigration Stress Scale.][**These items appeared on the Hassles Scale but contributed to the total score on theMother Immigration Stress Scale.]156Appendix CParent Support ScaleSometimes others can be helpful to us as parents by offering to babysit, giving gooddiscipline advice, and so on. At other times, people actually interfere with the way wewant to parent our children.1. Do any of your extended family members (living here or in Hong Kong) give youparenting support (e.g., babysitting, discipline advice)?never_ rarely_ sometimes quite a bit_ regularly_2. How helpful is the parenting “support” offered by extended family members? [if nevergive support or no extended family, check hereJnot at all rarely sometimes usually extremelyhelpful helpful_ helpful_ helpful helpful_3. Do any ofyour friends help you with parenting?never_ rarely_ sometimes_ quite a bit regularly4. How helpful is the parenting “support” offered by ffiends? [if never get help or nofriends, check herejnot at all rarely sometimes usually extremelyhelpfiil_ helpful_ helpful_ helpful helpful_5. Does your husband help you with parenting?never_ rarely_ sometimes quite a bit regularly6. How helpful is the parenting “support” offered by your husband? [if never get help orno husband, check herejnot at all rarely sometimes usually extremelyhelpful_ helpful helpful_ helpful_ helpful_157Appendix DWHEN YOU MOVED TOCANADA. HOW MUCHSTRESS DII) YOU FEEL?(HOW HARD WAS IT?)DRAW A LiNE TO SHOWHO\V MUCH STRESSYOU FELT.*Drawing not eNactlv to scale.A LITTLENONEStress ThermometerLOTS!QUITE A BIT158Appendix BScale of Children’s Stress1. Are you happier in Vancouver or Hong Kong?Vancouver —Hong Kong —Both the same —2. Do you miss your friends and family and things back inHong Kong?noyes_ alittle_a lot3. How would you rate your English?poor —fairgood —excellent4. Do you feel that you are not included in things becauseyou are Chinese?noyes — a littlea lot159Appendix FMother Global Ratings1. Approximately how long did you spend talking with this mother?____minutes2. How comfortable did this mother seem to be with you? (think about how easy/hard itwas to establish rapport)-3 -2 -1 0 +1 +2 +3not at neutral veryall3. How comfortable did this mother seem to be with participating in this research project?(did she express enthusiasm? reservations? did she ask questions that showed she wasinterested? wants copy of results?)-3 -2 0 +1 +2 +3not at neutral veryall4. How concerned did this mother seem to be with appearances? (i.e., presenting herselfand her child in a positive light (e.g., tells child to speak English, etc))-3 -2 -1 0 +1 +2 +3not at neutral veryall5. How confident do you feel that this mother will complete and return the questionnairepackage?-3 -2 -l 0 +1 +2 +3not at neutral veryall6. In your opinion, how “acculturatecP’ does this mother seem to be? (consider Englishlanguage fluency, spontaneous comments about Canadian culture/educational system,etc.). You may also probe with the following questions:- How long have you been in Canada?- How do you feel about living in Canada?3 -2 -l 0 +1 +2 +3not at neutral veryall160Appendix GObserver-Child Rating Scales0 1 2 3 4 5 6not at all somewhat verycooperative cooperative cooperative0 1 2 3 4 5 6not at all somewhat veryrelaxed relaxed relaxed0 1 2 3 4 5 6not at all somewhat veryattentive attentive attentive0 1 2 3 4 5 6not at all somewhat veryfriendly friendly friendly0not at allhappy3somewhathappyComments (Note odd behaviors, salient examples of behavior, appearance issues, etc.):1 2 4 5 6veryhappy161


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items