From Sibling Differences to Internal Representations

May 1, 2006

Differences between siblings are affected by internal representations and non-shared environmental influences such as gender, birth rank order, different peer influences, illnesses, and accidents.

They were as different as can be imagined. The olderbrother, only 18 months senior, was much quieter andmore structured and disciplined, had fewer friends, andseemed content to be alone. The younger brother wasfar more spontaneous, expressive of his feelings, affectionate,and had a larger circle of friends. These differences were obviousfrom early childhood and seemed to account for their differentdevelopmental trajectories. The older brother was by far thebetter student and went on to graduate school. In contrast, theyounger brother spent only 2 years in college, leaving it to takea job in sales. At midlife, the older brother had a stable professionalcareer and marriage while the younger brother had gonefrom one sales job to another and was in his third marriage.

Not surprisingly, their parents andextended family believed that thesedifferences in personality were entirelyinherited. The differences were apparentso early in life that it was widelyheld that their subsequent developmentaloutcomes flowed directly from theirdifferent genetic endowments. After all,what difference could there be in theirenvironment? They were raised by thesame parents, lived in the same neighborhoods,went to the same schools, andhad the same opportunities.

This formulation—sibling differencesas a result of only genetic differences—has been shown to be a simplisticand misleading reductionism. What issurprising, perhaps, is that the challengeto this reductionism has come, to aconsiderable extent, from researchaimed at studying how differences ingenetic makeup account for differencesin siblings' personalities.

Quantitative behavioral geneticsrelies on a methodology in which thepersonalities (and development ofpsychiatric disorders) are compared insiblings of different degrees of geneticrelatedness (eg, twins or adoptedsiblings) raised together or apart. Alarge number of studies have demonstratedthat sibling differences arisefrom 3 categories of variables: genetics,shared environmental influences,and non-shared environmental factors.In addition to documenting the role ofgenetic factors in personality, thesestudies have consistently indicated thatnon-shared environmental factors playa major role in producing personalitydifferences in siblings.

Non-shared environmental factors

Non-shared environmental factors aremultiple and include differential parentaltreatment, gender, birth rank order, different experiences by siblings of thesibling relationship itself, different peerinfluences, and the occurrence of suchnonsystematic events as illnesses, accidents,and the like. Some of these nonsharedenvironmental factors interact with each other in complex ways. Ourown research, for example, suggests thatin more functional families, sibling rankorder is very influential, whereas in moredysfunctional families, gender is moredecisive. We concluded that the influencesof sibling rank order and genderdiffered as a function of family competence.1 Emphasis must be placed,however, on the fact that these nonsharedenvironmental factors areassumed to be interwoven with geneticdifferences; siblings, to a significantextent, do create some of the environmentalresponses through the influenceof partially genetically determinedpersonality characteristics.

The attempt to document the exactnature of non-shared environmentalfactors has proved to be most difficult.For example, David Reiss, the notedfamily researcher, and his colleagueshave reported the results of a monumentallongitudinal study of adolescentdevelopment that by its sheer scope,surpasses most other developmentalstudies.2 More than 700 familiescontaining 2 same-sex siblings whowere 4 or less years apart were studiedon 2 occasions separated by a 3-year interval. With use of the proceduresof quantitative behavioral genetics, the 2 adolescent siblings were categorizedaccording to their different degrees ofgenetic relatedness (from monozygotictwins to adopted siblings who have nogenetic relatedness). What is mostimpressive about this study, however,is that environmental influences—atleast those operating within thefamily—were directly observed andmeasured, using state-of-the-art interactionaltechniques. Thus, this researchproject was carefully designed to elicitinformation about non-shared environmentalinfluences (as well as geneticand shared environmental influences).

The study's results add much to ourunderstanding of adolescent development:the important role of geneticinfluences was documented and theearlier results from quantitative behavioralgenetic research that suggested amodest role, if any, for shared environmentalinfluences were contradicted.Shared environmental influences(eg, stable family characteristics) doaccount for a significant part of adolescentpersonality in crucial areas suchas autonomy and sociability.

Internal representations

Even in this remarkable study, however,the nature of the non-shared environmentalinfluences—a very significantcontributor to personality differences—remains elusive. The data supplied donot inform about what these factors are.Reiss discusses this surprising turn ofevents in a careful manner and makesa cogent case that the answer may liein the psychoanalytic concept of internalrepresentations of self and others.

In other words, the manner in whichexperiences are taken into the self,stored, recalled, and narrated is believedto be a highly individualistic process.Two siblings could internalize the sameexperiences within the family or in theirsocial world in very different ways, andthese differences in representationscould play important roles in personalitydevelopment and consequentsibling differences. Reiss also discussesthe fact that although techniques havebeen developed to identify internalrepresentations, his study's methodology(a sample of more than 700 families and 1400 adolescents)did notinclude the opportunityto conduct theintensive interviewsor carry out the othermeasures necessaryto identify internalrepresentations.

This step, however, has recently beendescribed in another remarkable longitudinalstudy. Sroufe and colleagues3report on persons from adverse socioeconomiccircumstances who were studiedfrom birth into young adulthood. Theinvestigators describe the developmentand use of age-appropriate techniquesto get at the nature of the individual'sinternal representations. What is veryexciting about their findings is that thesemeasures of representations substantiallypredict behavior at the next developmentalstage, with those experiencesat that developmental stage predictingthe nature of the internal representationsat the following developmental stage.Thus, the inability of Reiss's study toidentify internal representations as partof non-shared environmental influencesis addressed in the longitudinal study bySroufe and colleagues.

All of this is (or should be) mostexciting to clinicians of diverse theoreticalorientations. These 2 remarkableresearch projects have begun to documentthe central importance of the waysin which relationships are experienced,stored, recalled, and narrated. Whetherwe use the language of internal representations,internal working models,cognitive schemata, or core assumptionsand expectations, a fundamental clinicalpremise derived from variouspsychotherapies is being documentedby systematic longitudinal research.

How sweet it is.

Dr Lewis is chairman emeritus of theTimberlawn Psychiatric Research Foundationand clinical professor of psychiatry at theUniversity of Texas Southwestern MedicalSchool in Dallas. He also has a private practicein individual, marital, and family therapyin Dallas.




Lewis JM, Beavers WR, Gossett JT, Phillips VA.

No Single Thread: Psychological Health in FamilySystems

. New York: Brunner/Mazel; 1976.


Reiss D, Plomin R, Neiderhiser JM, HetheringtonEM.

The Relationship Code: Deciphering Genetic andSocial Influences on Adolescent Development

.Cambridge, Mass: The Harvard University Press; 2000.


Sroufe LA, Carlson EA, Collins WA, Egeland B.

The Development of the Person: The MinnesotaStudy of Risk and Adaptation From Birth toAdulthood

. New York: The Guilford Press; 2005.