Studying Individual and Family Development: Linking Theory and Research

MARION O’BRIEN University of North Carolina at Greensboro

Studying Individual and Family Development:

Linking Theory and Research

The predominant theoretical orientation of themajority of scientists studying human behav-ioral development and family functioning isderived from systems models. These theories aredynamic in their conceptualization and em-phasize process and context. To a large extent,the analytic strategies we employ are not consis-tent with our theoretical assumptions. Analyticmethods that focus on moderating and mediat-ing processes rather than main effects, onwithin-group variability and the sources of thatvariability, and on the identification of inter-related patterns of change in individuals, families,and environments are needed for our researchresults to mirror the complexity of our theories.

The goal of research into individual and familydevelopment is to understand the processes bywhich individuals and families adapt success-fully to the challenges that confront them or,alternatively, become dysfunctional. In the 30-plus years that have passed since the publicationof Bell’s (1968) descriptions of child effects onparents and Sameroff and Chandler’s (1975)examination of developmental processes, devel-opmental scientists have become sensitized tothe complexity involved in studying these pro-cesses. Contextualism, dynamic systems, andtransactional analyses are now the catchwords

used to indicate that the authors of a researchproposal or report recognize that bidirectionaland reciprocal relations exist among the vari-ables they study. Often, however, these catch-words are emphasized in our introductions andignored in our research designs and analyses(Tudge, 2000). A gap exists between our theo-retical models of development and our empiri-cal approaches to studying change.

This brief essay addresses that gap. First, Idiscuss common characteristics of modern theo-retical approaches to understanding processesof individual and family development thathave become widely accepted in the fields ofdevelopmental psychology, family studies, andhuman development. Following from this base,I examine challenges facing us in the applica-tion of our theoretical approaches to the analysisof developmental processes.

The ideas presented here are not new. Thecomplexity of transactional developmental pro-cesses has been widely discussed by theoreti-cians (e.g., Bertalanffy, 1968; Bronfenbrenner,1979; Gottlieb, 1996, to name just a few), andthe difficulty of turning these complex pro-cesses into analyzable data has been recog-nized (Crouter & Booth, 2003; Sameroff &Mackenzie, 2003). Some observers have, infact, proposed that essentially all the statisticalprocedures currently in use to study develop-ment need to be replaced with new or at leastdifferent methods (Bergman, Magnusson, &El-Khouri, 2003; Tudge, 2003). My position isnot as strong as that. I believe there have beenand will continue to be useful contributions tothe study of developmental process using exist-ing empirical and statistical approaches as well

Family Research Center, Department of Human Develop-ment & Family Studies, University of North Carolina atGreensboro, P.O. Box 26170, Greensboro, NC 27401([email protected]).

Key Words: analytic methods, family systems, human devel-opment, theory.

880 Journal of Marriage and Family 67 (November 2005): 880–890

as new ones. My goal is to identify some incon-sistencies between our theoretical assumptionsand both the design of our studies and the ana-lytic methods we use and to suggest some waysto bring the two into closer concordance.

CURRENT THEORETICAL APPROACHES

TO THE STUDY OF DEVELOPMENT

Living systems are defined by change. Thus,research into individual development or the func-tioning of a family involves an examination ofchange. Our interest may be in the rate of change(i.e., vocabulary acquisition by toddlers), thedirection of change (i.e., positive vs. negativefamily accommodation to chronic illness), oridentifying ways to promote change (i.e., devel-oping interventions to reduce risk of abuse). Thechange we study may be quantitative (i.e., fre-quency of marital conflict) or qualitative (i.e.,divorce). Even when our research designs are notintended to capture change, the rationale for ourquestions assumes that the phenomena we docapture are important as determinants or out-comes of a process of change. For example, fam-ily interaction may be observed at a single timepoint, but the underlying assumption behind thestudy design is that supportive parent-child inter-action at age 12 is important for the child’s suc-cessful negotiation of the adolescent transition.

This emphasis on process and change rests ontheoretical approaches that are dynamic in theirconceptualization. Not so long ago, develop-mental psychologists could be divided into twocamps: those who saw developmental change asmechanistic, imposed by the environment ona largely passive individual, and those who sawchange arising from within an active individualwho selectively sought out experience (Cairns,1998). Today, these positions seem almostquaint in their simplicity. Similarly, within thefamily arena, the structural functionalism of themid-20th century gave way to contextual ap-proaches from such areas as feminist theory,family therapy, and the life course perspec-tive (Boss, Doherty, LaRossa, Schumm, &Steinmetz, 1993). Advances in neurobiology,ethology, and genetics have laid to rest any rem-nants of the nature-nurture dichotomy in favor ofcomplex, multilevel, reciprocal, interconnected,and altogether too complicated heuristics repre-senting the processes of developmental changeand family functioning. These theoretical ap-proaches are subsumed under the general label

of contextual models or systems approaches.They include, among others, ecological systemtheory (Bronfenbrenner, 1979; Bronfenbrenner& Morris, 1998), family systems theory(Whitchurch & Constantine, 1993), epigenetictheory (Gottlieb, 1992), dynamic systems the-ory (Thelen & Smith, 1998), developmentalcontextualism (Lerner & Simi, 2000), andthe holistic-interactionistic paradigm (Bergman,Cairns, Nilsson, & Nystedt, 2000). All thesetheoretical approaches have as their goal thedescription of interactions among biological,behavioral, and environmental processes overdevelopmental time.

Some of these theoretical models appear atfirst glance to be intimidating in their com-plexity and so global as to not readily generateresearchable questions. In many ways, thesetheories more closely resemble metatheories orworldviews, overarching frameworks for think-ing about individual and family development,than classical theories used in experimentalscience. Although differing from one another insome respects, these theoretical approachesshare two core principles that can serve asbenchmarks against which to evaluate the extentto which our research approaches are consistentwith our models of change. These principles,simply stated, are as follows:

d Processes of change operate at multiple lev-els, from the cellular to the cultural

d Processes of change, across all levels, aretransactional and reciprocal

Acceptance of these principles leads us to a con-sideration of analytic methods used in the fieldof human development and family studies.

STUDYING PROCESSES OF CHANGE

AT MULTIPLE LEVELS

Systems theories recognize that development isa complex process involving biology, individualbeliefs and behavior, and interpersonal transac-tions, all embedded within cultural context. Thisprinciple suggests research into child and familydevelopment that is multidisciplinary, broad inscope, and linked to the contexts in which peo-ple live.

The Need for Multidisciplinary Research

The field of human development and familystudies is inherently multidisciplinary. At

Linking Theory and Research 881

many universities, faculty trained in separatedisciplines—psychology, sociology, gerontol-ogy, and early childhood education, amongothers—identify themselves as members ofa single department of human development andfamily studies and work together routinely onacademic business. Interdisciplinary collabora-tion is valued by our professional organizations;both the Society for Research in Child Devel-opment and the National Council on FamilyRelations emphasize and encourage participa-tion by professionals from many disciplines.There is a general recognition of the importanceof integrating theoretical approaches, models,and methods from multiple disciplines into ourresearch approaches.

Much recent attention has been devoted tothe incorporation of biological processes in thestudy of human behavior (e.g., Magnusson &Stattin, 1998). Advances in molecular genetics,brain imaging, and neurochemistry have clari-fied the extensive role biology plays in indi-vidual behavioral development. Unlike earlierswings of the pendulum toward the biologicalside, that have often resulted in simplistic anddeterministic models (‘‘biology is destiny,’’ forexample), the current focus on biological under-pinnings of behavior is multifaceted and ac-knowledges the complex transactions betweennature and nurture. Just as brain structures in-fluence behavior, behavior changes the brain.Just as genes are involved in the display of com-plex human traits, life experience contributes tothe activation of genes. Perhaps the most fullydeveloped current conceptualization of theinterplay between biology, behavior, and cul-ture has been set forth by Gottlieb (1992, 1996),whose epigenetic theory postulates multiple in-terconnected and reciprocal pathways across alllevels of development. This model emphasizesthat the question of interest is not how muchbiology contributes to behavior and develop-ment as compared with the environment, butinstead how biology and the environment inter-act, over developmental time, to produce indi-viduals or families that are more or lesssuccessful in adapting to the context in whichthey live.

Biology is not the only discipline whose con-tributions are shaping ideas about developmentand family processes. With an increased interestin cultural contexts comes a recognition of thepotential role that anthropologists can play inconceptualizing and interpreting aspects of the

broader milieu in which families are embedded(e.g., Weisner, 2002). Demographic and eco-nomic approaches are increasingly being in-tegrated into studies of child and familydevelopment (Foster, 2002). Pediatricians andpublic health specialists emphasize the impor-tance of healthy functioning both in studies ofnormative development and in the goals of pre-vention and intervention programs (Olds et al.,1998).

Yet, the reality is that most of our research isnot interdisciplinary. There are many structuralreasons for this. Within the academic rewardstructure, collaborating with faculty in otherdepartments is not high on the list. Grant pro-posals tend to be reviewed by study groupsmade up largely of members of a single disci-pline. Collaboration takes time.

There are conceptual reasons as well. Theintellectual histories of each of our disciplinesexert an influence on us that is not unlike therole of culture in human behavior. It is not easyto merge these histories. Rules of evidence dif-fer and paradigms clash. The anthropologist’sreliance on qualitative methods and participantinvolvement, although these methods are con-sistent with contextual theories, run counter tothe psychological and sociological tradition ofquantitative statistical analyses (Kidd, 2002). Atthe other end of the spectrum, although the ap-proaches used by economists have offered newways of thinking about individual developmentand especially family functioning, economistshave less interest in the kinds of process analy-ses that are at the heart of systems models andfocus instead on outcomes (Foster, 2002).

There is no easy way to make research moreinterdisciplinary. At some universities, interdis-ciplinary research centers have been establishedto promote such efforts. Another trend in thefield—toward research projects of larger scopeand longer duration—has the potential to en-courage the inclusion of multiple perspectives.

Linking Large-Scale andSmaller Scale Research

‘‘Big’’ science is popular these days. Physicistsseeking to understand the nature of the atomand molecular biologists mapping the humangenome have contributed to the image of sci-ence as progressing through the collaborativeefforts of large groups of individuals workingtogether on a common problem. In behavioral

882 Journal of Marriage and Family

science, this approach has been implemented inthe form of large survey studies using nationallyrepresentative samples and smaller, but stillextensive, multisite studies focusing on parti-cular issues considered to be of widespread im-portance (see Table 1 for a listing of some ofthese studies).

Large-scale studies have many advantagesfrom a systems or contextual viewpoint. Theyusually include samples that are diverse in termsof social class, ethnicity, and cultural back-ground and therefore allow for meaningful anal-ysis of contextual factors. Several of theseinitiatives have been fortunate to receive fund-ing over a long enough period of time that pro-cesses of developmental change can begin to bestudied more comprehensively than has beenpossible in the past. The large number of partici-pants in many of these studies affords adequatestatistical power to address questions aboutdevelopmental processes as they may operatedifferently for girls versus boys, single female-headed households versus two-adult households,and rural areas versus urban neighborhoods.Most of the large-scale research initiativeslaunched in the past 20 years have also includedprovisions for making their databases publiclyavailable to researchers not directly involvedwith the collection of the data. These data re-sources have great potential for promoting inter-disciplinary work as scientists from variousresearch traditions examine questions of shared

interest. When multidisciplinary teams of scien-tists are involved in the design of such studies,or serve on their advisory boards, the opportu-nities for integration across research areas areconsiderable.

These advantages are offset by a number oflimitations common to many large-scale re-search efforts. The broad scope and wide-ranging goals of some large studies have beenaccompanied by a lack of theoretical and meth-odological rigor. Many constructs may be mea-sured, but the theoretical linkages acrossconstructs may be weak or even questionable.Further, it is not uncommon for complex indi-vidual or family level constructs (satisfactionwith parenting is an excellent example citedby Sabatelli & Waldron, 1995) to be measuredby a single item that may or may not captureimportant aspects of the respondent’s experi-ence (‘‘How often would you rather be child-less?’’). Unless a strategy of embedding smaller,more focused studies within the larger frame-work is adopted, projects involving many hun-dreds or thousands of participants also do notlend themselves to in-depth, possibly qualita-tive, analyses of the thought processes con-tributing to decisions that precede an outwardbehavior of interest. Thus, we may be able todescribe a child’s history of child care but knownothing of the underlying reasons the childreceived those types of care for that number ofhours. Finally, although large-scale studies

TABLE 1. EXAMPLES OF LARGE-SCALE AND LONG-TERM LONGITUDINAL OR MULTISITE STUDIES

OF CHILDREN AND FAMILIES

Family and Community Health Study, Institute for Social and Behavioral Research, Iowa State University

(http://www.isbr.iastate.edu/FACHS/)

Fragile Families & Child Wellbeing Study, Center for Research on Child Wellbeing, Princeton University

(http://crcw.princeton.edu/fragilefamilies)

Iowa Youth and Families Project, Institute for Social and Behavioral Research, Iowa State University

(http://www.isbr.iastate.edu/projects/)

Longitudinal Study of Child Neglect, University of Maryland at Baltimore

(http://www.acf.dhhs.gov/programs/cb/publications/compendium/ocan/ocana4.htm)

Longitudinal Study of Generations, University of Southern California (http://www.usc.edu/dept/gero/research/4gen/)

National Longitudinal Study of Adolescent Health, NICHD, Carolina Population Center, University of North Carolina

(http://www.cpc.unc.edu/addhealth)

National Longitudinal Study of Youth, U.S. Department of Labor Bureau of Labor Statistics (http://stats.bls.gov/nls)

NICHD Study of Early Child Care & Youth Development, NICHD Early Child Care Research Network (http://secc.rti.org)

Data archives (access to many large-scale datasets)

Inter-University Consortium for Political and Social Research, University of Michigan (http://www.icpsr.umich.edu)

Murray Research Center Data Archive, Radcliffe Institute for Advanced Study, Harvard University

(http://www.murray.harvard.edu/mra/index/jsp)


could encompass measurement at the biological,individual, interpersonal, and cultural levels,they are too often designed and directed by sci-entists from only one or two disciplines whoseinterests and expertise do not span this fullrange.

Large-scale nationally representative or mul-tisite studies are here to stay, and they havegreat potential to enhance research into devel-opmental processes. For that potential to befully realized, however, these efforts require thecontribution of continued and expanded smallerscale research that can look more closely atreciprocal processes of change within individu-als and families. No one approach to research isthe ‘‘right’’ one. Perhaps a productive strategyis to merge them, providing funding streamsthat link smaller scale research initiatives withlarger studies. In other words, a large-scale coreproject could serve as an overarching structurewithin which would be embedded smaller scaleprojects examining related developmental pro-cesses at a more microlevel. Similarly, projectsdesigned as multisite studies would be en-hanced by the use of dual research strategies:a common protocol across all sites to addresskey theory-driven questions combined withsmaller individual projects conducted by sub-sets of researchers or affiliated researchers thatextend the reach of the larger project andaddress in-depth issues of developmental andfamily process. Linking research at the macro-and microlevels could help to improve the qual-ity and consistency of analytic strategies andbring greater clarity to findings from researchersworking at different levels of analysis. Partici-pation at both levels by scientists from differentdisciplines could also add breadth and depth tothe picture of individual and family develop-ment resulting from the set of linked studies.

Studying the Contexts of Individualand Family Development

The diversity of human experience is recog-nized in current theories about the developmentof individuals and families. The microsystems,exosystems, macrosystems, and chronosystemsof ecological systems theory (Bronfenbrenner,1979) have become an integral part of ourthinking. We know that individual and inter-personal behavior is a function of overlappinginfluences of such factors as economics, familystructure, school or workplace characteristics,

and social networks. The importance of examin-ing behavior in natural situations has beenaccepted, if not always followed. Researchdesigns are much more likely now than in thepast to include children and families of variedethnic backgrounds. Cross-cultural research inthe pursuit of the universals of development isnow amplified by cultural research focusing onthe uniqueness of cultures.

Within much of our published literature,however, diversity is acknowledged but notanalyzed. The standard approach to data analy-sis involves entering factors representing diver-sity as controls, in order that the key variablesof interest can be examined without fully con-sidering the potential importance or meaning ofthe variability associated with that diversity.Thus, we may control family income in studiesof children’s cognitive ability and achievementas if income did not bring with it varying oppor-tunities to learn. We may control ethnicity instudies of parent-adolescent conflict as if cul-tures did not espouse different ‘‘rules’’ for theconduct of arguments. Of course, not everystudy can examine every possible contextualfactor in detail. If we follow the logic of ourcontextual and systems approaches, however,some aspect of context or process has to beclose to the heart of every theoretically drivenquestion we ask. For example, once we identifya trend toward declining marital satisfaction fol-lowing the birth of a child, we begin to ask why.What differentiates couples who become lesssatisfied from those who do not? Is it theirdemographic and economic status? their priorhistories? their expectations regarding parent-hood? Similarly, if we find links between thequality of child care and children’s cognitiveand language skills, the next step is an examina-tion of specific aspects of child-care environ-ments that contribute to learning in childrenwho come to those environments with a specificset of characteristics. Understanding how devel-opmental and family processes vary by contextis, by definition, what we seek to know. Of par-ticular interest to many who study child andfamily development are the intersections ofrace, class, and gender (e.g., Hill & Sprague,1999). Marginalizing these contextual variablesby using them as statistical controls rather thanplacing them at the center of our analyses con-tradicts our theoretical framework and limitsour ability to understand individual and familydevelopment (Newcombe, 2003).


In much behavioral science research, compar-ison of contexts has tended to involve somedegree of criticism of one or the other group.Adolescent girls’ visual-spatial abilities are notas good as those of boys; single mothers pro-vide less of the supervision developing childrenneed than do two-parent families. Value-ladencomparisons have contributed to the call ex-pressed by many minority researchers in theUnited States for more research focusing onwithin-group variation rather than between-groupdifferences (Doucette-Gates, Brooks-Gunn,& Chase-Lansdale, 1998; McLoyd, Cauce,Takeuchi, & Wilson, 2000). Within-group vari-ation can help us to understand the processes ofsuccessful adaptation to environmental de-mands that are unique to a particular context.Between-group comparisons do not have toimply superiority of one group over another,however. Identifying the ways in which devel-opmental and family processes are the sameacross contextual groups and the ways they dif-fer are both important to furthering our appreci-ation for the contribution of culture and context(Dilworth-Anderson, Williams, & Gibson, 2002).In-depth reviews of areas of literature with anemphasis on identifying where main effects andbetween-group analyses have suggested processquestions that have yet to be addressed wouldhelp the field in recognizing the need for morecomplex analyses.

STUDYING TRANSACTIONAL AND

RECIPROCAL PROCESSES OF CHANGE

Systems approaches to the study of individualand family development emphasize the bidirec-tional nature of influence—between geneticsand environment, parent and child, and cogni-tion and behavior. No person, event, or contextis static or passive. Further, development occursin a reciprocal process: One person’s behavioraffects another’s, whose response alters thebehavior of the first, which then in turn againaffects the other’s behavior. The baby criesa lot, the father reliably soothes the crying, thechild shows positive emotion with the dad, thedad becomes more committed to the parentingrole. If the crying baby has a father who be-comes irritable and tense, a different and lesspositive pattern is likely to follow. Concepts oftransaction and reciprocity suggest research intoindividual and family development that focuseson the study of moderating and mediating varia-

bles and processes rather than main effects andthat uses holistic approaches to measurementand analysis.

Studying Moderation and Mediation

Much of developmental science has been con-structed on analyses of main effects. Childrenliving in poverty are less successful in schoolthan other children. Maternal depression ina child’s infancy is linked to emotional dysregu-lation at preschool age. Within families, maineffects models assume that we can learn aboutthe whole by measuring each individual sepa-rately and combining these measurements.Gottlieb and Halpern (2002) have referred tothis kind of linear thinking as ‘‘analysis of vari-ance mentality’’ (p. 421). This description re-minds us that our thinking about the systems ofindividual and family development are oftentied more closely to the analytic methods weknow and are comfortable with than they are toour theoretical assumptions.

Given conceptualizations of individual devel-opment and family functioning as transactionaland reciprocal, main effects and linear modelsare no longer adequate to capture the processesof interest. It is not enough to measure a charac-teristic of an individual, a family, or a context atone time point and then predict an ‘‘outcome’’for that individual or family at a later time pointwithout an examination of the processes thatlink the two. Identifying the link between earlierevents and later functioning can be a helpful,and even necessary, first step in understandingprocess, but it is only a first step. The study ofprocess requires a focus on process, not onprediction. The questions posed by systems the-ories are difficult, and they require more fine-grained analysis than studies of main effects.

One example of a line of research wheremain effect results have been followed up withprocess analyses is the area of divorce. Earlywaves of research into divorce reported thatchildren from families in which parents weredivorced had a number of continuing academicand behavioral difficulties (see Hetherington,1979, for an early review). Replicated findingsof a main effect for divorce served to focus theattention of researchers on divorce and encour-aged them to ask questions of process. Morecomplex studies examined such issues as:How were individuals and families functioningprior to divorce, and how is earlier functioning


linked to child adjustment? What events andperceptions of events led to the decisions ofcouples to divorce? How do parent-child rela-tionships change when parents live apart? Whatis happening within families following divorcethat makes a difference in children’s behavior?These process questions yielded much greaterunderstanding of the dynamics occurring withinfamilies and the processes and contexts that dif-ferentiate children who are able to function suc-cessfully after their parents’ divorce from thosewho face continuing challenges.

For the most part, these process questionsinvolve analysis of moderation and of mediation(for further discussion of analytic approaches tomoderation and mediation, see Baron & Kenny,1986; MacKinnon, Lockwood, Hoffman, West,& Sheets, 2002; Rose, Holmbeck, Coakley, &Franks, 2004). Moderators influence the direc-tion or the strength of the relation between twoother variables. Thus, a family’s ethnicity maymoderate the link between parental disciplinestrategies and child outcomes (e.g., Lansford,Deater-Deckard, Dodge, Bates, & Pettit, 2004).Questions of moderation use contextual factorsto help us understand, for example, why riskfactors are not linked to negative outcomes forall families. Is marital conflict more likely toresult in divorce when wives are employed?Does the drop in household income that typi-cally accompanies divorce lead to more nega-tive outcomes for teenagers than for youngerchildren? Is there a difference in the extent towhich the active involvement of a nonresidentfather in the child’s life ameliorates negativeeffects for boys versus girls?

Mediators identify reasons why one variableis related to another. Thus, the quality of parent-ing may mediate the relation between povertyand its associated stressors and child social andacademic outcomes (McLoyd, 1998). Questionsof mediation use our understanding of the com-plexity of family context to help us understandhow risk factors operate. Do high levels of mari-tal conflict result in divorce because the conflictreduces intimacy? Do children in the first yearafter their parents’ divorce experience higherlevels of anxiety and depression because theirparents are less supportive and involved withthem? The area of divorce is one in whichdevelopmental and family researchers haveactively followed up main effects with processanalyses that address a wide range of complexquestions (for a recent review, see Amato &

Sobolewski, 2001). Too often, we have beensatisfied with main effects and ready to moveon to the next analysis linking early experiencewith later difficulties, rather than asking thehard questions about moderating and mediatingprocesses linked to individual and family char-acteristics and contexts.

Holistic Approaches toMeasurement and Analysis

Within contextual theories, individual develop-ment and family functioning are seen as pro-cesses of adaptation to external challenge(Luthar, 2003). Adaptation takes place on alllevels—biological, perceptual, cognitive, be-havioral, and interpersonal—in an integrated,not random, manner. Similarly, contextual chal-lenges are typically multivariate, in that dis-advantages tend to co-occur just as positivecharacteristics of environments do. Familieswho are poor are usually also undereducated,live in crowded and chaotic homes and danger-ous neighborhoods, have poor health and inade-quate health care, and work at physicallydemanding jobs on the least desirable shifts. Itis not clear that separating out the effects of oneof these factors from the package of disadvan-tage, and then controlling the others, is morehelpful to our understanding of family processor child development than identifying co-occurring patterns of risk. Holistic constructs—sets of related characteristics of individuals,families, and environments—are beginning tobe operationalized as quantitative variablesthrough structural equation modeling approachesand in person-centered analytic approaches,such as cluster analysis and latent class analysis.To qualitative researchers, the search for encom-passing and meaningful themes and patterns haslong been a primary goal. The adoption of ho-listic approaches by quantitative researchersmay help to bridge a methodological divide andcontribute to more comprehensive understand-ing of developmental contexts.

A classic example of a holistic conceptualiza-tion is the idea of ‘‘goodness of fit’’ used byThomas and Chess (1984) to describe the extentto which a young child’s temperamental charac-teristics were adequate to meet the demands foradaptation placed by a family environment. Fromthis viewpoint, an infant’s temperament—atleast in part a biologically based characteristic—can be adaptive or not depending upon the


characteristics of the caregiver and the nature ofthe caregiving environment. A highly activechild whose parents enjoy physical play andoutdoor activities is likely to grow into a skilledathlete, whereas the same child in a sedentary oran anxious family may receive criticism ratherthan encouragement and may be seen as havingattentional or behavioral problems. Studyingtemperament as a predictor of later outcome,without studying the environment, is not likelyto yield results that help to explain the processesof development.

Current approaches to the study of risk provideanother example of the movement toward holisticanalyses (Kraemer, Stice, Kazdin, Offord, &Kupfer, 2001). Because risk factors—even thoseat different levels of analysis—coexist, an impor-tant consideration in identifying risk and inter-vening to reduce risk has to do with defining theunderlying latent structure through which multi-ple factors act to affect outcomes. As one exam-ple, when parents separate and a teenager movesto a new community, the adolescent’s access toa positive peer group may be more important indetermining the course of later development thanthe simple fact of family disruption. The avail-ability of peer groups, however, may dependheavily on complex and interrelated child andfamily factors. Did the custodial parent have torelocate so quickly that neighborhood characteris-tics were ignored? Have the custodial parent’spaid work hours increased, providing moreincome to the family but allowing less time forparental involvement with the child? Is the teensocially skilled or talented in an area of achieve-ment valued by other adolescents? The develop-ment of effective intervention approaches tominimize negative outcomes for children andfamilies requires an understanding of the mecha-nisms by which biological, behavioral, and con-textual risk factors operate in conjunction withone another.

The multivariate clustering methods referredto as person-centered approaches are seen by someinvestigators as capturing the complexity of sys-tems models in ways that traditional variable-centered analyses cannot (Bergman et al., 2000).Whereas variable-centered approaches focus onthe examination of individual differences, theyalso are based on an assumption that develop-mental processes are similar for all individuals.The goal of person-centered approaches, bycontrast, is to identify subgroups of individualsthat differ from one another in important ways.

Once categories or clusters are identified, theways in which they differ can provide insightinto variations in the processes of growth andchange for particular sets of children or fami-lies. When both approaches are used to analyzethe same sets of data, they do not compete forthe right answer but instead may offer increasedclarity in our efforts to understand developmentin children and families. Within the field ofchild development, these approaches have beenapplied primarily in clinical studies focusing ondiscrete subtypes of child behavior problemswhere the outcomes of interest are categorical(von Eye & Bergman, 2003). There is currentlyconsiderable interest, however, in the potentialof these analytic approaches to allow examina-tion of processes of change within individualsand families as they affect and are affected bythe contexts in which they live.

One recent example of the use of person-centered analyses to examine family processessuggests how this approach fits a contextualviewpoint. Mueller and Elder (2003) clusteredgrandparents participating in the Iowa Youthand Families Project into five types on the basisof six aspects of their relationship with theirgrandchild. These grandparent clusters werefound to differ with regard to intergenerationalfamily relationships, indicating that family con-text was more important than demographicstatus factors in contributing to grandparent-grandchild relationships. Supportive and in-volved grandparents were more likely to havehad a positive relationship with a grandparentof their own when they were growing up and tohave children who actively involved them in thegrandchild’s life. Thus, family contexts andprocesses extending across generations areidentified through these analyses as affectingchildren’s relationships in the present.

Because our goal is to understand how de-velopment happens, or the processes throughwhich developmental change occurs, longitudi-nal applications of person-centered analysesare of particular interest. An extension of thegrandparent-grandchild relationship study, forexample, in which the grandparent clusters wererepeated at a later time point, would allowidentification of relationships that changedand an analysis of factors associated with thatchange. Clearly, person-centered approachesoffer new tools for our use and provide a per-spective different from traditional regression-based analyses of individual difference factors.


Person-centered approaches are not without limi-tations, in that they are heavily data and computerdriven rather than relying on theory to detect pat-terns that constitute meaningful categories or sub-groups. Further, person-centered techniques arestill relatively untested, with new applications ap-pearing frequently in the literature. These techni-ques are not the only possible approach tocontextual analyses. But in conjunction withmore traditional and familiar types of analysis,including qualitative analyses, these methodsoffer another opportunity for us to link our theo-retical propositions with our analytic models.

SUMMARY AND CONCLUSIONS

The worldview that guides research into humandevelopment and family studies at the begin-ning of the 21st century is a contextual anddynamic one. Whether presented as ecologicalsystems theory, epigenetic theory, dynamic sys-tems theory, family systems theory, or holisticinteractionism, the underlying and widely sharedconceptualization of development within indi-viduals and families is one in which both natureand nurture are active processes operatingwithin a complex system of reciprocal influ-ence. Much of the actual research published inour journals, in contrast, examines direct or lin-ear effects of individual or contextual character-istics on some outcome that is described as if itwere static and immutable. The reciprocity thatis at the heart of living systems has not yet beenfully incorporated into our analytic methods.

To move the field forward, to make ourresearch findings relevant to the lives of realpeople, and to improve the design of servicesaimed at the promotion of positive individualand family adjustment, we need to bring ouranalytic approaches in line with our theory.Structurally, universities and funding agenciescan assist the movement toward such consis-tency by supporting multidisciplinary researchand process-oriented research using qualitativeas well as quantitative methods, longitudinalstudies across theoretically meaningful periodsof time, and programmatic efforts to link large-scale and smaller scale projects. Individually,researchers can examine their own assumptionsand determine the extent to which their ques-tions and analyses are consistent with the theorythey espouse. Are contextual hypotheses pro-posed without a plan for measuring importantfeatures of contexts? Are main effects of one

variable on another reported with no follow-upto help us understand why or for whom theserelations hold? Are quantitative variable-centered approaches used even when particularquestions could be better addressed by qualita-tive or person-centered methods? No singleresearch design is adequate to capture the com-plexity of the phenomena we seek to under-stand. By expanding the range of analyticpossibilities and by using multiple methods,measures, and informants in all studies, we havethe best opportunity to meet our goal of under-standing the processes of individual and familyadaptation to their changing environments.

NOTE

The writing of this paper was supported in part by a coopera-tive agreement (U10 HD25430) from the National Instituteof Child Health and Human Development to the Universityof Kansas and the University of North Carolina at Greens-boro. I thank Jonathan Tudge for his many contributions tothese ideas and Beth Manke and David H. Demo for helpfulsuggestions on the manuscript.

REFERENCES

Amato, P., & Sobolewski, J. (2001). The effects of

divorce and marital discord on adult children’s

psychological well-being. American SociologicalReview, 66, 900–921.

Baron, R., & Kenny, D. (1986). The moderator-

mediator variable distinction in social psychologi-

cal research: Conceptual, strategic, and statistical

considerations. Journal of Personality and SocialPsychology, 51, 1173–1182.

Bell, R. (1968). A reinterpretation of the direction of

effects in studies of socialization. PsychologicalReview, 75, 81–95.

Bergman, L., Cairns, R., Nilsson, L.-G., & Nystedt,

L. (Eds.). (2000). Developmental science and theholistic approach. Mahwah, NJ: Erlbaum.

Bergman, L., Magnusson, D., & El-Khouri, B.

(2003). Studying individual development in aninterindividual context: A person-oriented approach.

Mahwah, NJ: Erlbaum.

Bertalanffy, L. von. (1968). General system theory.

New York: George Braziller.

Boss, P., Doherty, W., LaRossa, R., Schumm, W., &

Steinmetz, S. (1993). Sourcebook of family theo-ries and methods: A contextual approach. New

York: Plenum.

Bronfenbrenner, U. (1979). The ecology of humandevelopment: Experiments by nature and design.

Cambridge, MA: Harvard University Press.


Bronfenbrenner, U., & Morris, P. (1998). The ecol-

ogy of developmental processes. In W. Damon

(Ed.) & R. Lerner (Vol. Ed.), Handbook of childpsychology (5th ed., Vol. 1, pp. 993–1028). New

York: Wiley.

Cairns, R. (1998). The making of developmental psy-

chology. In W. Damon (Ed.) & R. Lerner (Vol.

Ed.), Handbook of child psychology (5th ed.,

Vol. 1, pp. 25–105). New York: Wiley.

Crouter, A., & Booth, A. (2003). Children’s influenceon family dynamics: The neglected side of familyrelationships. Mahwah, NJ: Erlbaum.

Dilworth-Anderson, P., Williams, I., & Gibson, B.

(2002). Issues of race, ethnicity, and culture in

caregiving research: A 20-year review. Gerontolo-gist, 42, 237–272.

Doucette-Gates, A., Brooks-Gunn, J., & Chase-

Lansdale, P. L. (1998). The role of bias and equiv-

alence in the study of race, class, and ethnicity. In

V. McLoyd & L. Steinberg (Eds.), Studying minor-ity adolescents: Conceptual, methodological, andtheoretical issues (pp. 211–236). Mahwah, NJ:

Erlbaum.

Foster, E. M. (2002). How economists think about

family resources and child development. ChildDevelopment, 73, 1904–1914.

Gottlieb, G. (1992). Individual development and evo-lution: The genesis of novel behavior. London:

Oxford University Press.

Gottlieb, G. (1996). Developmental psychobiological

theory. In R. Cairns, G. Elder, & J. Costello (Eds.),

Developmental science (pp. 63–77). New York:

Cambridge University Press.

Gottlieb, G., & Halpern, C. (2002). A relational view

of causality in normal and abnormal development.

Development and Psychopathology, 14, 421–435.

Hetherington, E. M. (1979). Divorce: A child’s per-

spective. American Psychologist, 34, 851–858.

Hill, S., & Sprague, J. (1999). Parenting in Black and

White families: The interaction of gender with race

and class. Gender and Society, 13, 480–502.

Kidd, S. (2002). The role of qualitative research in

psychological journals. Psychological Methods, 7,

126–138.

Kraemer, H., Stice, E., Kazdin, A., Offord, D., &

Kupfer, D. (2001). How do risk factors work

together? Mediators, moderators, and independent,

overlapping, and proxy risk factors. AmericanJournal of Psychiatry, 158, 848–856.

Lansford, J., Deater-Deckard, K., Dodge, K., Bates,

J., & Pettit, G. (2004). Ethnic differences in the

link between physical discipline and later adoles-

cent externalizing behaviors. Journal of Child Psy-chology and Psychiatry, 45, 801–812.

Lerner, R. M., & Simi, N. (2000). A holistic, inte-

grated model of risk and protection in adolescence:

A developmental contextual perspective about

research, programs, and policies. In R. B. Cairns &

L. R. Bergman (Eds.), Developmental science andthe holistic approach (pp. 421–443). Mahwah, NJ:

Erlbaum.

Luthar, S. (2003). Resilience and vulnerability:Adaptation in the context of childhood adversities.New York: Cambridge University Press.

MacKinnon, D., Lockwood, C., Hoffman, J., West,

S., & Sheets, V. (2002). A comparison of methods

to test mediation and other intervening variable

effects. Psychological Methods, 7, 83–104.

Magnusson, D., & Stattin, H. (1998). Person-

context interaction theories. In W. Damon (Ed.) &

R. Lerner (Vol. Ed.), Handbook of child psychol-ogy (5th ed., Vol. 1, pp. 685–760). New York:

Wiley.

McLoyd, V. (1998). Socioeconomic disadvantage

and child development. American Psychologist,53, 185–204.

McLoyd, V., Cauce, A.-M., Takeuchi, D., & Wilson,

L. (2000). Marital processes and parental socializa-

tion in families of color: A decade review of

research. Journal of Marriage and the Family, 62,

1070–1093.

Mueller, M., & Elder, G. H., Jr. (2003). Family con-

tingencies across the generations: Grandparent-

grandchild relationships in holistic perspective.

Journal of Marriage and Family, 65, 404–417.

Newcombe, N. (2003). Some controls control too

much. Child Development, 74, 1050–1052.

Olds, D., Henderson, C., Kitzman, H., Eckenrode, J.,

Cole, R., & Tatelbaum, R. (1998). The promise of

home visitation: Results of two randomized trials.

Journal of Community Psychology, 26, 5–21.

Rose, B., Holmbeck, G., Coakley, R., & Franks, E.

(2004). Mediator and moderator effects in develop-

mental and behavioral pediatric research. Journalof Developmental and Behavioral Pediatrics, 25,

58–67.

Sabatelli, R., & Waldron, R. (1995). Measurement

issues in the assessment of the experiences of par-

enthood. Journal of Marriage and the Family, 57,

969–980.

Sameroff, A., & Chandler, M. (1975). Reproductive

risk and the continuum of caretaking casualty. In

F. Horowitz, M. Hetherington, S. Scarr-Salapatek,

& G. Siegal (Eds.), Review of child developmentresearch (Vol. 4, pp. 187–244). Chicago: Univer-

sity of Chicago Press.

Sameroff, A., & Mackenzie, M. (2003). Research

strategies for capturing transactional models of


development: The limits of the possible. Develop-ment and Psychopathology, 15, 613–640.

Thelen, E., & Smith, L. (1998). Dynamic systems

theories. In W. Damon (Ed.) & R. Lerner (Vol.

Ed.), Handbook of child psychology (5th ed.,

Vol. 1, pp. 563–634). New York: Wiley.

Thomas, A., & Chess, S. (1984). Genesis and evolu-

tion of behavioral disorders: From infancy to adult

life. American Journal of Psychiatry, 141, 1–9.

Tudge, J. (2000). Theory, method, and analysis in

research on the relations between peer collabora-

tion and cognitive development. Journal of Experi-mental Education, 69, 98–112.

Tudge, J. (2003). Trying to apply an ecologicaltheory: Metatheoretical, methodological, and

statistical issues. Unpublished manuscript, Univer-

sity of North Carolina at Greensboro, Greensboro.

von Eye, A., & Bergman, L. (2003). Research

strategies in developmental psychopathology:

Dimensional identity and the person-oriented

approach. Development and Psychopathology, 15,

553–580.

Weisner, T. (2002). Ecocultural understanding of

children’s developmental pathways. Human Devel-opment, 45, 275–281.

Whitchurch, G., & Constantine, L. (1993). Systems

theory. In P. Boss, W. Doherty, R. LaRossa, W.

Schumm, & S. Steinmetz (Eds.), Sourcebook offamily theories and methods: A contextualapproach (pp. 325–352). New York: Plenum.


Documents

Studying Individual and Family Development: Linking Theory and Research