Family assessment: a review

Journal of Family Therapy (1987)g: 367-380

Family assessment: a review

Ian Wilkinson* This article is a review of developments in the assessment of the family in recent years. The review is selective, in order to draw out certain themes and principles from the literature.

Introduction and background to this work At the inaugural European conference on developmental psychology (Groningen,Netherlands,August 1984) arathercuriousstatistic emerged. O u t of exactly I O O papers presented, only four mentioned the family and one more discussed the rcile of the parents. Clearly, this is entirely at variance with the generally accepted importance of the family in child development. It seems likely that this anomaly reflects the problems encountered by empirical workers in assessing a phenomenon as complex as the family. A family of five consists of a network of five individuals, or ten dyads, or tenpossible triads, or five possible foursomes, or one complete network. I n practice,thecomplexitiesmultiplyduetothevastarray of agencies and professions involved in child and family work. There is a clear need for assessment methods which employ widelyrecognized and validated concepts and which refer to them in everyday terms. Assessment methods which use such a ‘common language’ will enable clarity and comparison of practice, encourage multi-disciplinary work and facilitate communication. The methods should be broad enough to assess variety a of complex problems yet also be readily communicable to beginners.As a member of a small, newly developing unit I am ideally placed to appreciate these needs. Over a period of Accepted version received March 1987. * Principal Clinical Psychologist, Department of Psychology, Darlington Memorial Hospital, Darlington DL3 6HX. 367 0163-4445/87/040367+ 14 $03.00/0

0 1987 The Association for Family Therapy

368 I. Wilkinson three to four years, we have tried out various ideas in this unit, out of which has emerged a n assessment system which at face value seems to fulfil some of these needs. O u r particular system will be described in another paper. Thefocus of this paper will be to outline developments in family assessment which are of particular relevance to a clinical context.

Family assessment in context Nearly two decades ago, Lickorish (1968) commented that few well tried methods were available for the assessment of the family, partly because the study of the family was a relatively recent psychological undertaking. Eleven years later, Litman and Venters (1979) gave a methodological overview of research on health care and the family. They concluded that despite interest from behavioural scientists in the rble of the family inhealthand illness, empiricalresearchhad remained rather limited. They emphasized a basic need for measuring instruments to help with research. In fact, Strauss (1969) had summed up the problem in a comprehensivereview of existing techniques when he stated. it is onlyslightly stretching the pointtosay that the conceptual status of measurement is not more primitive in the social sciences than in the physical sciences. The key differencelies in the vastly more primitive state of measurement technology in the social sciences.

Clearly, assessment methods vary according to the characteristics of the assessor. It is, therefore, useful to state our main characteristics in relation to assessment. We have a view of assessment and therapy as a problem-solving process. We believe that it is important to understand both the subjective and objectiveviews of a family’s problems, and then toform a contract which reconciles thetwo views. O u r theoretical position is a developmental and normalizing one which seeks to avoid stigmatizingindividualsor families. Hence,thedistinctionwhich is often made between clinical and other forms of assessment seems to us misleading. It seems more fruitful to consider the context, purpose and function of the assessment. O u r context is certainly a clinical oneindeed, it is a very typical small team in a district general hospital. As stated, we view the ultimate purposeof the assessment as clarifying and enhancingproblem-solving activities. T h e majorfunction of the assessment is to establish whether a contractualarrangement for

Familyassessment: a review

369

problem solving can be established with the family (and if it can, what kind of contract is agreed). In order to do this,it is importantto examine the family view(s) of the problems and also how the family members view our context, purpose and function. Finally, let us considerthemeanings of suchterms as diagnosis, assessment, description andmeasurement. I n thisarticle, we shall define assessment as the process of understanding problems in order to make a n intervention which is designed to help solve them. We shall avoid using the term diagnosis, except when reporting others’ work, since it has too many confusing medical connotations. If pressed, we would define diagnosis as referring more to the final, decision-making, aspects of assessment. The assessment of families is oftenpurely descriptive, and cleqrJv the qualities of that description are extremely important. To makea crud: distinction, science observes andart creates; but clearly, a good assessment relies on good observationand is creative. Hence, assessment and therapy should involve the strengths of bothartand science-they should be creativeand rigorous. Although assessment is often purely descriptive, it can include some kind of measurement. I n its crudest sense, measurement refers to the use of some hypothetical ordefined standard of comparison. It might, therefore, be reasonable to assert that a more scientific approach to assessment should incorporate some elementof measurement. Clearly,the differences betweenexistingmethods of assessment depend on several factors, but one of the most important is the broad purpose of the assessment (e.g. between clinical work and theoretical research). Different tasks require different technologies. We shall now considerthedevelopment of methods of assessment for particular purposes, considering each ina roughly chronological sequence.

Methods for obtaining information I n commonpractice,theinterview is used toobtain a purely descriptive form of assessment. However, as early as 1966 Watzlawick had developed a semi-structured interview for families. The imposition of structure within the context of the interview is a critical step which incorporatesmeasurementintowhat wouldotherwisebeapurely descriptive process. By having a standard procedure, the interviewer comes to define his own standards of comparison after he has used the as interviewonseveral occasions.(Responses arethenunderstood either normal or unusual.) I n fact, the imposition of structure within a n interview situation can

370 I. Wilkinson be taken so far as to include a directive or request for the family to performastructuredtask.Gilbert and Christensen(1985) reviewed and classified task approaches in marital and family work. One of the most relevant for clinical purposes has been developed by Kinston and Loader (1986~) which contains a number of different tasks to evaluate different aspects of family functioning. Family tasks are an interesting variant of the structured interview because they reveal a different type of information.Family tasks are designedtorevealtheongoing patterns of behaviouror ‘family process’ ratherthanthe family’s verbal understanding of the problems. I n fact, the family may not be willing to give an accuratedescription of the problems that areseen by a n observer. In family therapy terms, the family task is a standardized ‘enactment’.Note,however,that a n interviewmayrevealverbal understanding and process, whereas a task in itself will usually only yield process information. Towards the mid-1g70s, a number of developments began to take place in rapid succession. I n the field of adult psychiatry, a number of very elegant structured interviews began to appear which were rapidly acknowledged as extremely useful assessment techniques(e.g.the ‘Present State Examination’). These interviews have observer ratings integratedintotheprocedure. Inthe field of childpsychiatry,a growing awareness of the importance of psychosocial factors led to the abandonment of exclusively ‘categorical’diagnosis infavour of a multi-axial or dimensional system (e.g. Rutter et al., 1975). Theuse of categorical systems can be challenged on twomaingrounds. On a theoretical level, it can be argued that there is a finitelimit to the descriptive meanings that can be packed into one label or descriptor. Hence, assessment usingcategories is likely to result in stereotyping and other confusing misconceptions (e.g. as can occur with depression and its subtypes). There are also empirical arguments as to whether such ‘clusters’ or ‘syndromes’ exist (e.g. see Fisher, I 977). However, structured interviews do not yet seem to have had any wide impact upon family therapy. Hodges et al. (1982) have produced a structuredinterview for use withchildren.KinstonandLoader ( 1984) recently reported a standardized.clinica1 interview for families, based upon a conceptual framework used with other instruments the group have produced, and providedsome psychometric data in a later paper(KinstonandLoader, 19866). They mentioned some other instrumentsbutnotedthatdetails of thesepsychometric data had rarely been provided. It is certainly difficult to establish validity and reliability for such a n instrument in a clinicalsetting, and they list

Family assessment: a review

37 I

many of the difficulties. Perhaps for this reason, itis certainly clear that structured interviews are currently not widely used in familywork, either clinically or in research. Another of the common information-gatheringtools is the self-report questionnaire. I t developed partly out of a desiretoreduce routine repetitive tasks but also out of an explicitdesire for measurement. Locke and Wallace published one of the earliest marital questionnaires to be used by clinicians in 1959. However, it was not until the 1970s that questionnaires which examined whole family functioning began to be used byclinicians(Pless and Satterwhite, 1973; Moos and Moos, 1976). These earlyfamily questionnaires were designed by researchers ratherthan clinicians.However, in1983Epstein et al. published a questionnaire based upon the McMaster modelof family therapy. This would appear to be a significant development, since it is one of the first ‘family’ self-report forms which has been purpose-designed by clinical workers around an explicit model of family therapy (see later).

Methods for recording and organizing information Clinicians working with families tended initially to opt for a check-list of areas of family functioning-which might be considered to be a crudeandpurelydescriptive form of observer rating scale. For example, in 197 I Kadushin produced an instrumentof this type which attemptedto systematically integrateconcepts from a variety of theories and studies. As such, it was a n interesting instrument which gave a wide vFiriety of perspectives upon family life but lacked a central theme. Observer rating scales have tended to be popular with clinicians because they usually provide a brief and efficient record form which summarizes the important aspects of the ‘thing observed’ (in this case, thefamily).Onceone is familiarwith a rating scale, they can be completed extremely quickly. Provided that the content of the scale matches the theoretical perspective of the observer, they are genuinely labour-saving. They also have the attraction of introducing a crude type of measurement into the situation. More sophisticated scales may also provide in-built guidelines for summarizing the information and guidingtheclinicianthroughthedecision-making process. Their disadvantages are that,because the judgements required only relate to a small range of specific issues, rating scales tend to provide a ‘narrow focus’. They also rely heavily upon the skill of the observer as well as the clarity of definition of the terms used. In essence, although they can

372 I. Wilkinson claimed be tve’ method of assessment they are liable to profound subjective distortions by the user. Reliability, which may be poor initially, drops when it is not assessed (Reid, 1970). I n Britain, theGreatOrmondStreetgroupproduced a succession of clinical instruments which can also be categorized as descriptive rating scales. The Summary Format of Family Functioning (Loader et al., 1983) used foursections,some of whichhavefurthersubdimensions. The instrument is clear,comprehensiveand uses conceptswhichare commonly acknowledged to be of importance in clinical work, but it lacks a central theme. By contrast, some members of the same group also produced the Focal Hypothesis Assessment Form (Kinston and Bentovim, I 982). Based on a model of therapy which explicitly seeks to develop a central hypothesis (Bentovim and Kinston, 1978), the form consists of a set of guidelines for doing so. T h e method forces the clinician to find a central theme with respect to a particular family. Those interested in ‘transgenerational’ or historical issues will find this an extremely useful, albeitpurelydescriptive, assessment method. However, it assumes a high level of clinical knowledge and skills if it is to be used effectively. Despite its inherent problems, the rating scale has continued to be one of the most popular types of instruments among clinical workers. I n America,Epsteinand his co-workersdevelopedthe ‘McMaster model’ of family functioning and therapy, based upon six dimensions of family functioning (Epstein and Bishop, 1981). Thedimensions used areproblem solving, communication, rbles, affective responsiveness, affective involvement and behaviour control.T o accompany the model of therapy,therearevarious assessment forms,some of which are descriptiverating scales, as well as the self-reportformmentioned earlier. I n the field of social psychology, researchers began to use coding systems to examine interpersonal behaviour in groups long before the techniques would be applied clinically (e.g. Borgatta, 1962). Coding systems rely on the precise linguistic definition of behaviours so that theiroccurrencecan be ‘objectively’scored.Provided the linguistic distinctions and definitions are clear and meaningful, they do provide more reliable methods of measurement in social science since they are derived from more objective (process) typeof information. Because the data can be collected temporally, they also enable a detailed sequential analysis of what is happeninghereand now. The clarityand meaningfulness of thecoding definitions determinesthemeasuring power and clinical usefulness of the system. However, these methods

Family assessment: a review

373

can alsobeverytime-consuming and it is probably this limitation which prevented them from becoming more popular with clinicians, untilrecentapplicationshavebeendeveloped by behaviourally oriented workers. A number of behavioural coding systems had been developed following on from the work of Patterson et al. (1973) and others. A particularly useful clinical technique for the assessment of parent-childrelationshipswasdevelopedbyEyberg andRobinson (1981). Unlike many coding systems, it is relatively brief and easy to administer and relies on coding a standard task. It is also supported by a number of empirical studies. Other attempts to apply coding systems to marital and family problems have been reviewed by Filsinger ( I 983) andGrotevantandCarlson(1987).Additionally,themethods of assessing expressed emotiondevelopedby Vaughn and Leff (1976) have proven clinical utility but have not so far been widely applied in familywork; and the attributional ,coding system of Stratton et al. (1986) may be useful to analyse families’ belief systems. However, it doesseem that for wholefamily transactions attempts to develop a coding systemwhich is brief but also meaningful and thus efficient havenot yet bornfruit. It seems thatinordertobeclearand unambiguousit is difficult to retain a n ability to analyse the complexity of family functioning.

General trends and theoretical considerations T h e trends in all of these methods have to be set against the general background of familywork. T h e development of thestructural approach(Minuchin,1974)andstrategicapproaches(e.g.Haley, 1976) can both be placed within the context of a n increased understanding of families as total systems rather than as networks of individuals. The application of systems typetheorieshascreated a quiet but massive revolution in family work by giving a conceptual framework with which to understand the family as a whole, and the relationships of the partsof the family to the whole, rather than simply understanding parts the ‘individuals’. as Since nearly all psychotherapyhad previously beenconceptualizedinterms of its impact upon ‘the individual’,this was a massive conceptual leap. Anindication of thechangeinthinkingthat this provided is indicated by a review of family assessment methods by Fisher in 1976. Fisher reviewed the existing literature at that time and extracted five key ‘dimensions of family assessment’. I n simple terms, these were the ‘commonthreads’thatranthroughallthevariousmethods of

ds

374 I. Wilkinson assessment hereviewed, in terms of the types of phenomena which were thought to beimportantin clinicalworkwith families. In a further important review in 1977, Fisher looked at previous attempts to ‘classify’ families and concluded thatthenotion of ‘types’ or ‘categories’ was simplynotsupportedbyempiricalevidence. He suggested that it wouldbe farbetterto use ‘dimensional’analyses which were purpose-designed for broad types of problems (e.g. childfocused problems). Meanwhile in 1976, Cromwell et al. had reviewed existing marital and family assessment methods. They pointed out, among otherthings, that differentmethods of assessment providequalitatively different 1977 by information (see Table I ) . These ideaswere addedtoin KeeneyandCromwellin a n article entitled ‘Towards systemic diagnosis’ (which they defined as, Laway of evaluating various system levels andtheirinterplay’). Inother words,they arguedthatone T A B L E I . (From Cromwell et al., 1.976)

Type of d a t a Ob-jective Subjective Behaviour Self-report Insider

Reporter frame of Observer reference methods reports jective Outsider

should not only evaluate the system as a whole, but also the interplay of the various individuals andsubsystems within the family, in order to obtain a more holistic picture of the situation. Cromwell and Keeney further elaborated their ideas in 1979. They outlined a training model based upon their earlier notion of ‘systemic diagnosis’ in which students were trained to use a variety of different types of assessment techniques on the same family. In this way, they were encouraged to examine a number of different perspectives on the families and to attempt to integratethese perspectives and thus achieve a holistic view of the family. T h e conclusion suggests that the students were rather confused by this process. This is not really surprising, since the ‘package’of instruments lacked a unifying theme. Olson produced a typology of the family using two major dimensions of cohesion and adaptability in the same year (Olson etal., 1979). Although this has generated much research work, it does not

Family assessment: a review

375

seem to have interested clinical workers. The typology is neat and tidy but there is no real evidence that it is of any pragmatic value. Reiss (1980)delineatedsome of the choices that face researchers and clinicians in assessing the family. He also described his own work the same year (Reiss and Oliveri, 1980). In a series of experimental studies using family tasks, he suggested that the major factors which affected problem-solving performance were shared family paradigms to do with attitudes to the social environment. This is a n interesting finding as it underscores the need to examine the family belief systems. I n 1982, Fisher once again weighed in with the critical observation that despite having ‘transactional’ theories of family functioning the methods of assessment used in research were still predominantly using ‘individual’ concepts and perspectives. He also argued strongly for the use of direct observational procedures suchas coding methods. Perhaps the most important notion has been ‘multi-system multimethod’ (MSMM) assessment proposed by Cromwell and Peterson in 1983. Thiselaborated even furtheruponthenotion of ‘systemic diagnosis’. T o summarize their arguments, they proposed that it was not only important to evaluate the parts and their relationship to the whole (‘multi-system’), but it was also important to utilizedifferent kinds of assessment since they yield very different types of information. This latter argument for ‘multi-method’ assessment can be illustrated by Tables 2 and 3. Table 2 shows a classification of the various types of assessment methods which clarifies the distinctions made in Table I in moreconcrete terms.If we rely ononlyone of these assessment methods, we are likely to have problems, since an observer’s view of a family often differs from their own view of themselves, and what they saythey do may not beconfirmedbyobservations of theiractual behaviour. It is not within thescope of this review to examine this issue in detail, since it has proved to be a major source of conflict between TABLE 2 . Assessment methods ‘Perspective’ Subjective methods Cognitive Assessment Questionnaire levels verbal Ratingscale topic Behavioural Interview Behaviour level process

Objective methods

coding tasks

376 I. Wilkinson TABLE 3. Type of change required Subjective problems,i.e. acknowledged by family Yes

Objective problems, behaviour i.e. observed by outsider

Yes

Behaviour change

No

Perceptual and change

No

Perceptual change

No change

proponents of varioustheories. The relativeemphasisplacedupon subjective and objective views is amajorparadigm whichshapes theories of assessment and therapy. As we have suggested, we believe both views are necessary. Liddle (1982) has listed otherimportant therapeutic paradigms in the contextof giving useful advice about how to be eclectic without being confused. There is also a difficult ethical question involved in this issue. This hinges upon the choice of whether one views either the subjective or objective perspectives as ‘more correct’ than the other. When does a clinical worker have a right or duty to impose his or her view on the family? Inour view, theconsiderationsrequiredare, firstly, the potential risksof not challenging the family view, and secondly, the contractual issues involved in the relationship (i.e. seeking permission to confront). As a final illustration of the need for MSMM assessment, let us take the observer’s standpointand assume that his perspective is more valid. Table 3 then shows how the changes required in therapy vary according to whether thefamily and the observer agree about whether the problems exist. Without considering both views, the nature of these changescannot be defined in thisway. The MSMM notionhas aroused some controversy and the debate about it has been rich (see Reiss, I 983; Peterson and Cromwell, 1983).At the end of the day, it seems likely that the objections to MSMM assessment will be pragmatic rather than theoretical. How can such anassessment device remain brief and efficient? T o finish the review on a pragmatic level, there are two particular books which will be of interest to the beginning practitioner of family assessment. The work of the Brief Therapy Group (Fisch et al., I 982) is assessment. knownmainly as a method of therapyratherthan However,theirwork is veryimportantin thiscontextsinceit (a)

Family assessment: a review

3 77

underscores the importance of clearly understanding the client’s view of the problems; (b) provides a highly detailed account of the manner inwhich assessment shouldbecarriedout. O n a moreconceptual level, Karpeland Strauss (1984) haveproducedprobablythe best overview of the subject in recent years. However, to return to our centralquestion-how can we reconcile a reasonably comprehensive assessment method with brevity and clinical efficiency? Looking over the development of assessment technology in recent years, we can perhaps conclude that the following principles will be useful when assessing families. Having a central theme such as a coherent theory of family functioning and therapy. Adopting a broadly based systems theory approach, i.e. one which allows holistic analysis but does not neglect examination of the parts and their relationship to the whole system. Using a common language, i.e. widely recognized and validated concepts referred to in everyday terms. Using the best principles from various theoretical perspectives (ifit is possible without destroying the central theme). Using both subjectiveand objective methods of assessment. Integrating these methods of assessment to form a n efficient ‘package’. Incorporating a structured interview format witha n integrated rating scale. Incorporating someelement of task enactment to clarify important family processes. Incorporating a method of codingfamilytransactions(if possible). ( 1 0 ) identifying family strengths as well as problems. ( I I ) Using a dimensional rather than a categorical analysis. ( I 2 ) Using operational definitions and behavioural descriptions to clarify concepts. ( 1 3 ) Usinggenuinely‘transactional’perspectivesonthewhole family. There is, in fact, some empirical support for many of these notions. I n a major study of normal and clinical families utilizing a variety of assessment measures, theTimberlawngroup(Lewis et al., 1976) advocatedmany of these ideas. Theirstudy looked mainly at the normal family, but they draw a number of conclusions for clinicians that aresimilar to many of these principles. It has to be admitted that the application intotal of these principles v

378 I. Wilkinson presents a daunting task. Efforts to apply them in our clinical team in Darlington will be described in a further paper.

Acknowledgements Thanks are due to a great many people (too numerous to mention names) who contributeda wealth of ideas.

References BENI'OVIM,A. and KINSTON,W. (197%)Brief focal family therapy when the child is the referred patient. I. Clinical. 11. Methodology and results. Journal of' Child PsychologyandPsychialry, rg: 1 - 1 2 , I 19-144. BORGATTA, E. F. (1962)A systematic study of interaction process scores, peer and self assessments, personality andother variables. Genetic P.ychologyMonographs, 65: 2IQ2g'. CROMWELL, R. E., OLSON, D. H. L. andFOURNIER,D. G . (1976) Tools and techniques for diagnosis and evaluation in marital and family therapy. Family Process, 15: 1-49, CROMWELL, R. E. andPETERSON,G. W . (1983) Multisystem-multimethod family assessment in clinical contexts. Family Process, 22: I 47-163. DSTEIN,I%. B. and BISHOP, D. S. ( 1 9 8 1 )Problemcentred systems therapy of the family.Journa1 of Marital and Family Therapy,7: 23-31. D U E I N , N. B., BLDWIN. L. M. andBISHOP,D. S. (1983) TheMcMaster family assessment device. Journal of Marital and Family Thera,?), g: I 7 I - I 80. EYBERG,S. M. andROBINSOX, E. A. ( 1 9 8 1 )DyadicParent-Child Interaction Coding Systems: A Manual. Washington, D.C. University of Washington. FILSINGER, E. (Ed.) ( I 983) Marriage and Family Assessment. London. Sage Publications. FISCH,R., WEAKLAND,H. J. and SEGAI.,L. (1982) The Tactics of Change: Doing Therapy Briefly. London. Jossey Bass. FISHER, L. (1976) Dimensions of family assessment. Journal of MarriageandFamily Counselling, 4: 367-382. FISHER,L. (1977) O n the classification of families. Archives of GeneralPgchiatr-y, 34: 424-433. FISHER,L. (1982) Transactional theories butindividual assessment: afrequent discrepancy in family research. Family Process, 21:3 I 3-320. GILBERT,R. and CHRISTENSEX, A. (1985)Observational assessment of marital and T h e Handbook of Family Psychology and family interaction. In L. L'Abate (Ed.), Therapy. Illinois. Dorsey Press. GROTEVANT, H. and CARISON, C. (1987) Family interaction coding systems: a descriptive review. Family Process, 26: 49-74. HALEY,J. (1976) Problem Solving Therapy. San Francisco. Jossey Bass. € B n G E & K. MCKNEW, D., CYTRYN,L., STERN,L. and KLINE, J. (1982)The child assessment schedule (CAS) diagnosticinterview: a reporton reliability and validity. Journal of the American Academy of Child Psychiatry, 21:468-473. KADUSHIN,P. (1971) Towardsa family diagnostic system. FamilyCoordinator, 20: 279289. KARPEL, M. andSTRAUSS, E. (1984) Family Evaluation. London. Gardiner Press.

Family assessment: a review 379 B. P. and CROMWELL, R.E. (1977) Towards systemic diagnosis. Family Therapy,4: 225-236. KINSTON, W. and BENTOVIM, A. (1982) Constructing a focal formulationand hypothesis in familytherapy. Australian Journal of Family Therapy, 4:37-50. KINSTOW W. and LOADER,P. (1984) Elicitinga whole-family interaction witha standardized clinical interview. Journal of Family Therapy, 6 347-363. KINSTON, W . and LOADER, P. (1986a) The Family Task Interview. Available from the Family Research Program, B I O S S , Brunel University. K I N S T O ~W. and LOADER,P. (19866) Preliminary psychometric evaluation of a standardized clinical family interview. Journal of Family Therapy, 8: 351-369. IXWIS, J., BREAVERSW . , GOSSETT, J. and FHILLIPS, V. (1976) No SingleThreadPsychological Health in Family Systems. New York. Brunner/Mazel. U C K O R I S H , J. R . (1968) The psychometric assessment of the family. In: J. G. Howells (Ed.), Theory and Practice of Psychiatry. New York. Brunner/Mazel. L I D D L ~H. (1982)O n the problems of eclecticism: a call for epistemologic clarification and human-scale theories. Family Process, 21 (2): 243-250. LITMAN, T. J. and VENTERS, M. (1979) Research on health care and the family. A methodological overview. Social Science and Medicine, 13: 379387. IDADER, P.,BURCK, C., B I N G L E Y ,L . , KINSTON,W. and BENTOVIM, A. (1983) T h e SummaryFormat o f FamilyFunctioning: A Guide to itsUse. Great Ormond Street, London. Family Studies Group. L O C K 6 H. and WALLACE, K. (1959) Short marital adjustment and prediction tests: their reliability and validity. Marriage and Family Living, 21:251-255. M I N U C H I N , S. (1974)Families and Family Therapy. London. Tavistock. MOOS,R. H. andMOOS,B. S. (1976) A typology of family social environments. Family Process, 15:357-371. O!!SON, D. H., SPRENKLE, D. H. and RUSSELL, C. S. (1979) Circumflexmodel of maritaland family systems. I . Cohesion andadaptability dimensions, family types and clinical applications. Family Process, 18 (I): 3-28. PATTERSON, G . R . , COBB, J. A. and RAY,R. S. (1973) A social engineering technology for retraining the families of aggressive boys. In: H. E. Adams and J. P. Unikel (Eds),Issues and Trends in Behaviour Therapy. Springfield, Illinois. Thomas. PETERSON, G. W. and C ~ O M W E L L , E. R. (1983) clarification A of multisystem-multimethod assessment: reductionism versus wholism. FamilyProcess, 22: 173-1 77. R E S & B. and QTTERWHITE, B. (1973) A measure of family functioning and its application. Social Science Medicine, 7: 6 I 3-62 I . REID. J . B. (1970)Reliability of observation data: a possible methodological problem. ChildDeveloprnent, 41: I 143-1 150. R E I % , D. (1980) Pathways to assessing the family: some choice points and a sample route. In: C. K. Hofling and I. M. Lewis (Eds), TheFamily:Evaluationand Treatment. New York. BrunnerIMazel. REISS, D. and O L I V E R I , M. (1980) Family paradigmand family coping. Family Realtions, 29: 43 1-444. R E I % , D. (1963)Sensory extenders versus meters and predictors: clarifying strategies for the use of objective tests in family therapy. FamilyProcess, 22: 165-171. RUTTER,M,, SHAFFER, D. and SHEPHERD, M. (1975) A multiaxial classification of child psychiatric disorders. Geneva. World Health Organization. KEENEY,

380 I. Wilkinson P., HEAR4 D., HANKS, H.,M U N T O N , A., BREWIN, c. and DAVIDSON, c. (1986) Codingcausal beliefs innatural discourse. British Journal of Social Psychology, 25:2993 I 3 . SI‘RAUSS, M. A. (1969) FamilyMeasurementTechniques. Minneapolis.University of Minnesota Press. V A U C HC. ~ E. and LEFF,J . (1976) T h e measurement of expressed emotion in the families of psychiatric patients. British Journal o f Sociul and Clinical Psychology, 15: 157-165. P. (1966) A structured family interview. FumibProcess, 5: 256-271. WATZLAWICK, SrRATTON,