of 21/21
Psychological Bulletin 1979, Vol. 86, No. 4, 831-851 Cognitive Development in Retarded and Nonretarded Persons: Piagetian Tests of the Similar Sequence Hypothesis John R. Weisz University of North Carolina at Chapel Hill Edward Zigler Yale University From the debate over developmental "universals" in Piagetian theory and the controversy between developmental and difference theories of mental retarda- tion, an important hypothesis emerges—one that is testable via cognitive- developmental comparisons between retarded and nonretarded persons. This similar sequence hypothesis holds that retarded and nonretarded persons tra- verse the same stages of cognitive development in the same order, differing only in the rate at which they progress and in the ultimate developmental ceiling they attain. Current evidence relevant to this hypothesis is drawn from 3 longi- tudinal and 28 cross-sectional studies of developmental phenomena described by Piaget. The great preponderance of this evidence supports the hypothesis with respect to every subject group, with the possible exception of individuals suffering from pronounced electroencephalogram abnormalities. The quality of current evidence is critically evaluated, and procedures by which more precise tests of the hypothesis might be fashioned are proposed. Overall, the review illustrates that developmental research with atypical populations can be a potent tool in testing general developmental theory. Conversely, it illustrates the power of general developmental theory to enrich our understanding of atypical de- velopment. In recent years two important theoretical issues have stimulated interest in Piagetian research with retarded and nonretarded pop- ulations. One is the question of whether developmental "universals" exist. Many psy- chologists regard the sequence of develop- mental stages described by Piaget (e.g., 1970) and elaborated by other cognitive-develop- mental theorists (e.g., Kohlberg, 1969) as one of psychology's few current candidates for universality (see Weisz, 1978). Piaget (1956, 1966) took a psychological universalist position, with qualifications, and Kohlberg (1969,1971) argued strongly for the invariance of what he regarded as a cognitive-develop- This review was supported by a New York State Human Ecology research grant to the first author and by National Institute of Child Health and Human Development Grant 5 POI HD 03008 to the second author. The authors are grateful to Thomas Achenbach and Sally Styfco for their thoughtful comments on a draft of the article. Requests for reprints should be sent to John R. Weisz, Department of Psychology, University of North Carolina, Chapel Hill, North Carolina 27514. mental sequence rooted in an inherent logic and in universal characteristics of both the nervous system and the environment. Of course it is impossible to know that any given developmental phenomenon occurs everywhere without exception, since one can never test all possible exceptions (Popper, 1959). However, if one is not to let the claims of cognitive-developmental theorists go un- challenged, it is important to evaluate the extent to which transcontextual validity (see Weisz, 1978) has been demonstrated for the Piagetian account of development. One approach to assessing such validity across experimental contexts is to examine developmental sequences across various cul- tures (Buck-Morss, 1975; Simpson, 1974). Another approach, of particular interest because of the cognitive emphasis of Piagetian theory, is to compare groups of children who differ markedly in measured intelligence, that is, groups of mentally retarded and non- retarded children. If children at very different IQ levels were to show identical Piagetian developmental sequences, then the transcon- Copyright 1979by the American Psychological Association, Inc. 0033-2909/79/8604-0831$00.75 831

Cognitive Development in Retarded and Nonretarded …weiszlab.fas.harvard.edu/files/jweisz/files/weisz_zigler_1979_psyc... · Cognitive Development in Retarded and Nonretarded Persons:

  • View

  • Download

Embed Size (px)

Text of Cognitive Development in Retarded and Nonretarded...

  • Psychological Bulletin1979, Vol. 86, No. 4, 831-851

    Cognitive Development in Retarded and Nonretarded Persons:Piagetian Tests of the Similar Sequence Hypothesis

    John R. WeiszUniversity of North Carolina at Chapel Hill

    Edward ZiglerYale University

    From the debate over developmental "universals" in Piagetian theory and thecontroversy between developmental and difference theories of mental retarda-tion, an important hypothesis emergesone that is testable via cognitive-developmental comparisons between retarded and nonretarded persons. Thissimilar sequence hypothesis holds that retarded and nonretarded persons tra-verse the same stages of cognitive development in the same order, differing onlyin the rate at which they progress and in the ultimate developmental ceilingthey attain. Current evidence relevant to this hypothesis is drawn from 3 longi-tudinal and 28 cross-sectional studies of developmental phenomena describedby Piaget. The great preponderance of this evidence supports the hypothesiswith respect to every subject group, with the possible exception of individualssuffering from pronounced electroencephalogram abnormalities. The quality ofcurrent evidence is critically evaluated, and procedures by which more precisetests of the hypothesis might be fashioned are proposed. Overall, the reviewillustrates that developmental research with atypical populations can be a potenttool in testing general developmental theory. Conversely, it illustrates the powerof general developmental theory to enrich our understanding of atypical de-velopment.

    In recent years two important theoreticalissues have stimulated interest in Piagetianresearch with retarded and nonretarded pop-ulations. One is the question of whetherdevelopmental "universals" exist. Many psy-chologists regard the sequence of develop-mental stages described by Piaget (e.g., 1970)and elaborated by other cognitive-develop-mental theorists (e.g., Kohlberg, 1969) asone of psychology's few current candidatesfor universality (see Weisz, 1978). Piaget(1956, 1966) took a psychological universalistposition, with qualifications, and Kohlberg(1969,1971) argued strongly for the invarianceof what he regarded as a cognitive-develop-

    This review was supported by a New York StateHuman Ecology research grant to the first author andby National Institute of Child Health and HumanDevelopment Grant 5 POI HD 03008 to the secondauthor. The authors are grateful to Thomas Achenbachand Sally Styfco for their thoughtful comments on adraft of the article.

    Requests for reprints should be sent to John R. Weisz,Department of Psychology, University of NorthCarolina, Chapel Hill, North Carolina 27514.

    mental sequence rooted in an inherent logicand in universal characteristics of both thenervous system and the environment.

    Of course it is impossible to know that anygiven developmental phenomenon occurseverywhere without exception, since one cannever test all possible exceptions (Popper,1959). However, if one is not to let the claimsof cognitive-developmental theorists go un-challenged, it is important to evaluate theextent to which transcontextual validity (seeWeisz, 1978) has been demonstrated for thePiagetian account of development.

    One approach to assessing such validityacross experimental contexts is to examinedevelopmental sequences across various cul-tures (Buck-Morss, 1975; Simpson, 1974).Another approach, of particular interestbecause of the cognitive emphasis of Piagetiantheory, is to compare groups of children whodiffer markedly in measured intelligence, thatis, groups of mentally retarded and non-retarded children. If children at very differentIQ levels were to show identical Piagetiandevelopmental sequences, then the transcon-

    Copyright 1979 by the American Psychological Association, Inc. 0033-2909/79/8604-0831$00.75



    textual validity of the Piagetian account ofdevelopment would be substantially supported.If retarded and nonretarded children were todiffer in their sequence of development, thenuniversality could hardly be claimed for thePiagetian account.

    A second theoretical issue that has sparkedrecent interest in comparative cognitive re-search is reflected in the ongoing debatebetween proponents of developmental anddifference theories of mental retardation. Thedevelopmental position, set forth by Zigler(1969), is intended to apply to retardedindividuals not suffering from organic impair-ment. Zigler has maintained that the retardedchild passes through cognitive-developmentalstages in the same order as the nonretardedchild, with only two differences: The retardedchild passes through the stages more slowlyand attains a lower upper limit relative to thenonretarded child.1

    A number of theorists hold what Ziglerlabeled the general difference position. Oneaspect of this position is the view that thecognitive development of retarded personsdiffers from that of nonretarded persons inways that go beyond mere differences in rateand ceiling of development. Milgram (1973),for example, maintained that the cognitivelevels, or stages, of retarded children differfrom those of the nonretarded in that theformer are more likely to contain traces ofdevelopmentally earlier levels and are morelikely to show regression to those earlierlevels. (For further discussion of the develop-mental position, the specific difference posi-tions, and the rationale underlying them, seeWeisz & Zigler, in press.)

    This theoretical conflict has generated anew emphasis on comparative research intothe processes (rather than the products) oflearning and reasoning (Weisz, 1977; Weisz &Achenbach, 1975) and into processes of reason-ing, as described in Piagetian theory (seeWilton & Boersma, 1974).

    The growing interest in the pursuit ofdevelopmental universals, and the growingintensity of the developmental versus differencedebate, have thus combined to lend theoreticalforce to research comparing the cognitivedevelopment of retarded and nonretardedpersons along Piagetian lines. This body of

    research has grown rapidly within the pastdecade; it now appears to be substantialenough to serve as a resource in our efforts toanswer the principal question raised by theuniversality issue and the developmental versusdifference debate. This question can be statedin the form of a testable hypothesis.

    Similar Sequence Hypothesis

    An appropriate label seems to be thesimilar sequence hypothesis. The hypothesisholds that during development retarded andnonretarded persons traverse the same stagesin precisely the same order and differ only inrate of development and in the ultimateceiling they attain. To be precise, the develop-mental position (Zigler, 1969, 1971) generatesthis hypothesis only with respect to nonre-tarded and cultural-familial retarded persons(thus excluding, for example, brain-damagedand genetically impaired individuals).2 In

    1 An additional postulate of the developmentalposition is that familial retarded and nonretardedpersons who are equivalent in developmental level(often operationally denned as mental age) do not differin the formed cognitive processes they employ inlearning and reasoning. This particular proposition isnot germane to the present review and consequently isnot discussed here. However, Piagetian evidencebearing on this proposition is being reviewed (Weisz& Zigler, in press).

    2 The reasoning underlying this qualification bearsbrief explanation. The developmental position holdsthat mental retardation can be viewed as a develop-mental phenomenon most appropriately among personswhose retardation does not result from specific physio-logical defects. Such investigators as Benton (e.g.,1962), Cruikshank (e.g., 1967), and Reitan (e.g., 1973)have devoted many years to demonstrating idiosyn-cratic performance characteristics that distinguishbrain-injured individuals from those with intactnervous systems. Furthermore, a number of studiesemploying the specific kinds of problem-solving tasksmost often used in research on the developmental-difference controversy have revealed effects of orga-nicity on retarded children's performance (Balla,Styfco, & Zigler, 1971; Balla & Zigler, 1964; Elkind,Koegler, Go, & Van Doorninck, 1965; Harter, Brown,& Zigler, 1971). In harmony with such findings,proponents of the developmental position have adheredto the two-group approach (see Zigler, 1969), wherebyfamilial retarded individuals are distinguished fromthose suffering from organic impairment (includinggenetic anomalies such as Down's syndrome). There issome disagreement among investigators over the needfor the two-group approach (Ellis, 1969; Milgram,


    cognitive-developmental theory, however, theclaims for the universality of the developmentalsequence appear to be broader. Piaget (1956)held that" the minimum program for establish-ment of stages is the recognition of a distinctchronology, in the sense of a constant order ofsecession" (p. 13). According to Kohlberg(1969), the claim that there is an invariantorder of cognitive stages rests upon anassumed invariance in certain features of theenvironment and of the nervous system andupon "a logical analysis of orderings inherentin given concepts" (p. 355). These inherentorderings are seen as logically essential andas independent of individual differences amongpeople. Kohlberg continued, "The invarianceof sequence in the development of a concept orcategory is not dependent upon a prepatternedunfolding of neural patterns; it must dependupon a logical analysis of the concept itself"(p. 355). Thus, the similar sequence hypothesisas advanced by cognitive-developmental the-orists seems to predict a truly universalordering of stagesan ordering that is thesame for retarded children of all etiologies(including genetic impairment, brain injury,and other neurological anomaly) as it is forall nonretarded children. There is a conserva-tive version of the similar sequence hypothesisthat applies only to familial retarded andnonretarded persons and a liberal versionthat applies to all persons. In the presentarticle we present evidence bearing on bothversions.

    In contrast with both these versions,Milgram (1973) has argued that the retardedchild's cognitive stages differ from those ofthe nonretarded child. In contrast with theliberal version of the hypothesis, Rogers(1977) has described a rationale for (thoughshe has not necessarily endorsed) the hypoth-esis that profoundly (and thus nonfamilial)

    1973). Moreover, there is a strong Piagetian rationalefor applying the similar sequence hypothesis to allpersons, retarded or nonretarded, organically impairedor intact (see the remainder of the paragraph fordetails). In what follows we describe this rationale, andwe go on to review evidence in a manner that bearsdirectly on the conservative, two-group-orientedversion of the hypothesis and on the more liberal versionin which the similar sequence hypothesis is applied toall retarded groups regardless of etiology.

    retarded children have abnormal develop-mental patterns.

    Material Excluded From the Present Review

    The present article is an attempt to synthe-size studies relating to the similar sequencehypothesis. In selecting studies to be reviewed,we excluded studies of reading per se and oflanguage development per se. Although bothareas can be viewed from the perspective ofPiagetian theory, neither is central to thetheory; furthermore, the research in both areasis now so voluminous as to warrant separatereview. We also excluded studies designed toaccelerate cognitive development, since it isnot the purpose of this article to determinewhether retarded or nonretarded children canbe trained more readily.

    The studies we do include in this reviewvary widely in their sampling procedures,their experimental methodology, and theirapproaches to data analysis and reporting.Consequently, the studies differ in their levelof importance vis-a-vis the hypothesis ofparticular interest here. For this reason, wereserve the right to vary the level of detailin which we describe the studies and giverelatively greater space to those that seem toafford the clearest tests of the hypothesis.

    Tests of the Similar Sequence Hypothesis

    Cross-Sectional and Order-of-Difflculty Evidence

    One approach to testing the similar sequencehypothesis is to assess groups of mentallyretarded children at more than one develop-mental level with respect to their performanceon various Piagetian tasks. If the direction ofthe difference in performance from onedevelopmental level to another is the samefor the retarded as for the nonretarded, or ifthe direction is consistent with the develop-mental sequence posited by cognitive-develop-mental theory, then the similar sequencehypothesis is supported. A second generalapproach to testing the similar sequencehypothesis is to rely less upon the develop-mental levels of the groups sampled than uponthe relative-difficulty levels of the varioustasks or behavioral items being employed.Perhaps the simplest, but also the leastinformative, of the variants of this approach


    is to rank order the items with respect to thenumber of subjects who pass each one; ifthis rank ordering of a retarded sample matcheseither the rank ordering of a nonretardedsample or the developmental sequence positedby cognitive-developmental theory, then thesimilar sequence hypothesis is supported,albeit modestly. A more informative type oforder-of-difficulty evidence is the type thatemploys scaling procedures, allowing one todetermine, for example, how many of thechildren who grasp Concept A also graspConcept B and vice versa. Such evidence,when combined with Guttman-type (e.g.,Guttman, 1950) scalogram analyses, canprovide a relatively strong test of the similarsequence hypothesis. The studies reviewed inthis section all employed some type of cross-sectional evidence, order-of-difficulty evidence,or a combination of the two.

    Development in the sensorimotor period.Early evidence bearing on the similar sequencehypothesis was provided by the research ofWoodward (1959, 1961, 1962, 1963). The firstof her studies (1959) focused primarily ona group of 65 institutionalized children andadolescents with a chronological age (CA)range of 7-16 years who were so profoundlyretarded that they failed to attain a basal ageof 2 years on the Terman-Merrill scale.Although the author maintained that thissample excluded cases involving motor orsensory disability, the cases involved a divers-ity of medical problems (e.g., 19 subjectswere epileptic), and 38 of the children were"emotionally unstable." Woodward used threemeans of assessing the sensorimotor stages ofthis heterogeneous group. First, she observedtheir spontaneous mannerisms and theirmanipulation of toys presented individuallyand in a standardized order. Second, eachsubject was presented with three pairs oftasks, each pair tapping one of Piaget's(1953, 1955) last three sensorimotor stages(there are six stages in all). Third, Woodwardpresented each child with a series of objectconcept tasks in which a piece of candy or atoy was first used to attract the subject'sattention and was then withdrawn and con-cealed to varying degrees.

    All but the object concept tasks wereanalyzed in a way that sheds light on the

    similar sequence hypothesis. Each task wasclassified with respect to the Piagetian sensori-motor stage it was designed to represent;then the tasks were ranked ordered with respectto the percentage of subjects passing each.The difficulty level rankings of these 11 itemswere identical to the Piagetian stage levelorder, with one exception: A task involvingcoordination of vision and hearing (Sensori-motor Stage 2) proved to be slightly moredifficult than a task involving manipulation ofobjects (Stage 3); 53 subjects passed themanipulation task, and only 49 passed thecoordination task. Furthermore, when thepossibly insensitive coordination task wasremoved from the analyses, 59 of the 65children passed all of the items at stagesbelow their highest stage level response. Giventhe extreme diversity of this sample, the highincidence of emotional instability, and theapparent tendency of many not to showresponses of which they were actually capable(e.g., some delayed for a half hour beforegrasping an object placed before them), thesedata lend surprisingly strong support to thesimilar sequence hypothesis.

    Recently, Rogers (1977) undertook aninvestigation similar to that of Woodward inseveral respects. The subjects, 40 profoundlyretarded children ranging in age from 8-14years, with IQs below 20, were given a seriesof Piagetian tasks. By means of these tasks,each child's performance was classified intoSensorimotor Stage 3, 4, 5, or 6 in each of fourconceptual domains: object permanence (tasksinvolving searches for a hidden object),spatiality (tasks involving visual anticipationand rotation of objects), causality (tasksinvolving the use of physical prompts andtools, the removal of obstacles, and inferenceas to the cause of a jingling sound inside a box),and imitation (tasks involving the reproductionof both self-initiated and experimenter-initiatedmovements and sounds). Performance withineach of the four domains was analyzed usingscaling techniques, and Guttman's (1950)coefficient of reproducibility and Green's(1956) index of scalability were calculatedfor each scale. The object permanence andimitation tasks formed highly reproduciblescales in the orders hypothesized by Piaget(1955, 1962, 1972). Causality tasks also


    formed a highly reproducible scale, althoughthe item order differed from the predictedsequence in one respect: One Stage 6 itempreceded one Stage 5 item. The authorattributed this irregularity to a poor choice ofStage 6 task (i.e., opening a box to obtain abell when box opening has just been demon-strated to the subject), "since the task usedmight have been accomplished using imitationrather than problem-solving skills" (Rogers,1977; pp. 841-842). Finally, the individualspatiality tasks did not all form highly repro-ducible scales, but when the tasks within eachstage were combined (and subjects werecredited with a stage level for passing one ormore of the tasks from that level), the stagesdid form a highly reproducible scale. Rogersconcluded convincingly that her findingssupport "the invariant sequentiality of sensori-motor stages" (p. 841).

    The preoperational-concrete operational tran-sitionthe Inhelder study of conservation. Oneof the earliest studies bearing on the similarsequence hypothesis was carried out byPiaget's associate, Barbel Inhelder, in theearly 1940s. This study, now published inEnglish (Inhelder, 1943/1968), involved theassessment of conservation of substance,weight, and volume in 159 persons who hadbeen labeled mentally retarded by Swisseducation officials. The sample was extremelyheterogeneous (see Jordan, 1976). Ages rangedfrom 7J-52 years, IQs ranged from 35-104,institutionalized and noninstitutionalized per-sons were included, and the range of etiologiesand physical maladies included such diversestates as "defective environment," rickets,hearing defect, "abandoned," and schizo-phrenia. The procedure involved semistruc-tured clinical interviews with each subject.Since the procedure was not perfectly standard-ized and little in the way of formal dataanalysis was presented, it is difficult toevaluate Inhelder's conclusions, including herreferences to "oscillations" in the reasoningof retarded subjects, discussed later in thisarticle. However, in Piaget's (1968) descriptionof the Inhelder study, he explained that inthe entire sample,

    not one ^individual] understood the conservation ofweight without having the conservation of substance,nor the conservation of volume without both weight

    and substance, while the conservation of substancewas found without the other two, and the conservationof weight was found without the conservation ofvolume, (p. 11)

    Given the marked heterogeneity of the sample,such uniform support for the similar sequencehypothesis is noteworthy.

    Other studies of conservation and relatedconcepts using retarded samples only. Studiesof conservation and related concepts donesince the Inhelder investigation have a bearingon the similar sequence hypothesis, despitethe fact that they only sampled retardedsubjects. Klauss and Green (1972) assessedconservation of number and volume in 27trainable mentally retarded subjects rangingin age from 13-19 years and in IQ from 29-57.These investigators found that volume con-servation presented greater difficulty than didnumber conservation, a finding consistentwith the pattern apparent in the nonretarded.Marchi (1971) tested conservation of mass,weight, and volume in 106 educable mentallyretarded children. Difficulty level evidencesuggested that contrary to Marchi's prediction,the retarded "follow a similar sequence in theacquisition of mass, weight, and volume aspostulated for normals" (p. 6442).

    Roodin, Sullivan, and Rybash (1976) asses-sed qualitative identity, quantitative identity,and equivalence conservation (see Elkind,1967) in 60 institutionalized retarded childrenaveraging 13 years of age and about 47 in IQ.Dyed water was poured from a standard 100 mlbeaker; to test qualitative identity, subjectswere asked, "Is the water in this glass (com-parison) the same water that was in that glass(empty standard)?" To assess quantitativeidentity, subjects were asked, "Is there asmuch water in this glass (comparison) as therewas in that glass (empty standard)?" Toassess equivalence conservation, two standardbeakers were filled with equal levels of water,and the contents of one were then poured intoa comparison beaker; the experimenter thenasked, "Is there as much water in this glass(standard) as there is in this glass (compar-ison)?" Previous reasearch (e.g., Papalia &Hooper, 1971) with nonretarded children hadsuggested that the developmental order forthe attainment of these three concepts wouldbe qualitative identity, quantitative identity,


    and equivalence conservation. In the Roodinet al. study, analyses of the number of con-servers on each task indicated a parallel orderof difficulty.

    In a similar study also employing 60 institu-tionalized retarded children (age range ofapproximately 10-16 years and average IQ ofapproximately 57), McManis (1969c) in-vestigated identity and equivalence conserva-tion with three types of material (Styrofoamballs, clay, and water). Like Roodin et al.(1976), McManis found evidence that thedevelopmental sequence of his retarded sub-jects replicated that of nonretarded children.The notion that identity conservation mustprecede equivalence conservation was sup-ported by the finding that no subject whofailed to achieve identity conservation showedequivalence conservation, whereas 13%-18%of the subjects (precise percentage dependingon the particular task used) displayed identityconservation without equivalence conservation.

    Three studies that examined conservationof number, and number concepts generally, inmentally retarded groups yielded similarfindings, despite some differences in methodol-ogy. Woodward (1961) investigated numericalconcepts of 94 institutionalized individuals(50 adults with average CA of 19 years and44 children and adolescents with averageCA of 12 years) ranging in IQ from 25-73.Tests given to the subjects included assess-ments of their understanding of (a) one-to-onecorrespondence and equivalence of correspond-ing sets, (b) ways of equalizing unequalgroups, (c) seriation, and (d) conservation ofcontinuous quantity (water and sand). Per-formance was scored as indicative of one oftwo preoperational stages or of concreteoperational thinking. When the stage levelassignments were plotted as a function of theIQs (and thus roughly of the mental ages orMAs) of the adult subjects sampled, thetable reflected precisely what would beexpected from the application of Piaget'sstage scheme to nonretarded individuals.

    In the second of the three studies,Mannix (1960) administered eight of Piaget's(1952) number concept tasks to 48 "educa-tionally subnormal" individuals ranging inMA from 5-9 years. The tasks included twotests of additive composition, one test of

    coordination of equivalence relations, twotests of judgment of correspondence betweensets of items, and two conservation tasks(continuous and discontinuous quantities).Responses to these tests were classified intoPiagetian stage levels, a scalogram was con-structed, and the coefficient of reproducibilitywas .94. Mannix's brief report gave littleinformation as to the precise nature of thescale types; but apparently the scalogram wasconsistent with Piaget's stage theory, becausethe author concluded that educationallysubnormal children "pass through the threestages of development described by Piaget"(Mannix, 1960, p. 181).

    The third of these studies on numberconcepts was conducted with 20 institutional-ized mentally retarded persons in New Zealand(CA range of 8-17 years; IQs of 29-65).Singh and Stott (1975) presented thesesubjects with a series of number conservationtasks designed to classify them with respectto three Piagetian number stages: Stage 1child fails to attend to relevant cues and failsto conserve; Stage 2child selectively attendsto only certain relevant cues, can matchperceptually, but cannot conserve; Stage 3child conserves, showing understanding ofinvariance of properties despite transformationin appearance. Data bearing on the similarsequence hypothesis are not reported indetail, but the authors' conclusion is quiteclear: "Retarded children apparently developsequentially in the same order as normals butat a slower rate and at a later CA" (Singh &Stott, 1975, p. 220).

    One other study that used only a retardedsample deserves mention both because of itsscope and because of an important issue itraises. In this study, Lister (1972) assessedsix types of conservation among 115 educa-tionally subnormal pupils in Great Britain.The subjects were aged 8-16 years, and theirIQs ranged from 47-81. Both difficulty levelrankings and a scaling procedure stronglysuggested the following developmental se-quence in the emergence of these types ofconservation: number, substance, length,weight, volume, and area. Although no scalo-gram statistics were calculated, only 6 of the115 subjects showed a scalogram responsepattern inconsistent with the preceding order.


    Lister noted that the order with respect tosubstance, weight, and volume was consistentwith previous Piagetian research, whereas thesuggested order of the remaining attributesdiffered from at least some previous findingswith nonretarded subjects. Her own interpre-tation of the discrepancies was that theyresulted from experiment-to-experiment varia-tions in the specifics of the problems used toassess the various types of conservation. Thisis a very real possibility, and it is one reasonwhy tests of the similar sequence hypothesisthat expose retarded and nonretarded subjectsto the same experimental procedures must beregarded as stronger evidence than experimentsthat test only retarded children and comparethe findings with those of different experi-ments. We now turn to six studies of theformer type.

    Studies of conservation and related conceptsemploying both retarded and nonretarded subjects.Four of the experiments in which the per-formance of retarded and nonretarded subjectswas directly compared were conducted byMcManis (1969b, 1969d, 1969e, 1970). Inone of these, 90 institutionalized retardedsubjects (IQs of 47-73) and 90 nonretardedelementary school children (IQs of 85-115)were tested for conservation of mass, weight,and volume of clay, using Piaget and Inhelder's(1941) "sausage" technique. About half theretarded subjects were organically impaired.Analyses of the mean scores for the conserva-tion tasks indicated that conservation of masswas easiest and conservation of volume mostdifficult for both the retarded and the non-retarded group, providing some support forthe notion that the order of emergence of thesetypes of conservation in groups of bothaverage and below-average IQ is as follows:mass, weight, then volume.

    In another article McManis (1969e) re-ported his assessment of conservation andtransitivity of weight (clay) and length(sticks) in what appears to be the samesample used in his 1969d experiment. Thestudy was designed to test the hypothesis,derived from Kooistra (1964), that for anygiven property (e.g., weight) conservationwill appear developmentally earlier thantransitivity. The results supported this hypoth-esis for both weight and length in retarded

    and nonretarded subjects. McManis (1970)then explored the relations among conserva-tion, seriation, and transitivity (of length)within groups of 80 institutionalized mentallyretarded persons (IQs of 46-72) and 80nonretarded elementary school children (IQsof 85-116). Among both retarded and non-retarded children who showed discrepantperformance on the conservation and seriationtasks, nearly all showed conservation withoutseriation. Among both retarded and non-retarded children who showed discrepantperformance on the seriation and transitivitytasks, nearly all showed seriation withouttransitivity. These findings indicate thatseriation falls developmentally between con-servation and transitivity (at least with respectto the property of length, as measured in thisexperiment) for both retarded and nonretardedpersons.

    In a fourth article based on the same sampleused in two of the preceding studies (McManis,1969d, 1969e), McManis (1969b) testedPiaget's (1952) view that there are threehierarchically ordered stages in the develop-ment of quantitative comparison processes.In the first stage, children are said to consideronly uncoordinated perceptual relations ofgross qualitative equality or difference; inthe second stage, intensive quantity, childrenare said to compare quantities by seriatingthem along more than one dimension (e.g.,width and height) simultaneously; in the thirdstage, extensive quantities, children are saidto be capable of overruling apparent differencesbetween two equal quantities by imposingequal units of measurement upon them.McManis tested his young subjects' perform-ance of these three types of comparison, usingsticks, colored water, and beads. The analysisof scores on these tasks indicated that forboth the retarded and the nonretarded group,gross comparisons were the simplest (theywere passed by nearly all subjects in bothgroups) and extensive comparisons were themost difficult. These findings are consistentwith the view that for children at both IQlevels the developmental order is as follows:gross, intensive, and extensive quantity (theorder posited by Piaget). One other study(McManis, 1969a) should be mentioned in thisconnection. McManis's tests of quantity com-


    parison were given to 140 institutionalizedmentally retarded persons, who were dividedinto equal groups representing different IQlevels (IQs of 30-49 and of 50-69). Theprocedure and analyses were similar to thoseemployed in the preceding study (McManis,1969b). The results indicated, as in thepreceding experiment, that comparison ofgross quantities was easiest and comparisonof extensive quantities was most difficult,regardless of the IQ level of the subjects.

    Three other comparative studies weredesigned to address the problem of order ofevents across the types of conservation. Gruenand Vore (1972) assessed conservation ofnumber (poker ships), continuous quantity(water), and weight (clay rolled into variousshapes) in familial retarded (IQs of 55-80)and nonretarded (IQs of 90-120) public schoolpupils. Both retarded and nonretarded groupswere divided into three subgroups of MA: 5,7, and 9 years. Evidence on developmentalordering was in the form of mean scores forthe three types of conservation, analyzedwithin each MA level. In one set of analyses,conservation judgments alone (i.e., disregard-ing the subjects' verbal explanations) con-stituted the dependent variable. With thiscriterion, performance of nearly all subjects atMA 9 was correct; for the other two MAlevels, both retarded and nonretarded subjectstended to score significantly better on thenumber task than on the quantity or weighttasks. For nonretarded children at MA 7,however, the differences were not significant.The quantity and weight tasks did not differsignificantly in difficulty for retarded andnonretarded children.

    In a second set of analyses by Gruen andVore, the dependent measure was conservationjudgment in combination with the subject'sexplanation for that judgment. Using thiscriterion, there was no significant task effectat the MA 5 level. At the MA 7 level bothretarded and nonretarded subjects did some-what better at the number task than at thequantity and weight tasks, but the differenceswere only significant for the retarded subjects.At the MA 9 level both retarded and non-retarded subjects scored significantly higheron the number than on the weight task andsignificantly higher on the quantity than on

    the weight task. Thus, although order-of-difficulty patterns were similar for retardedand nonretarded subjects, with use of judg-ments alone and judgments plus explanations,task differences tended to be statisticallysignificant among the retarded more oftenthan among the nonretarded. In attemptingto account for this trend, Gruen and Vore(1972) made an interesting point:

    McManis (1969[e]) has suggested that there is atransitional period (MAs of 7-10) in which the variousconcrete operations are obtained and that retardedchildren progress through this period more slowlythan do normal children. If this is true, it would beexpected that the performance of normal children onvarious conservation tasks would vary less from taskto task than that of retarded children. This alsosuggests that retarded children may be ideal subjectsfor investigating the transition process from preopera-tional to concrete-operational thinking, (p. 156)

    We conclude this section of the review bydiscussing two conservation studies by Achen-bach. Building on the work of Charlesworth(1969) and Mermelstein and Shulman (1967),Achenbach (1973) inferred children's identityconcepts with respect to color, number,length, and continuous quantity from theirsurprise reactions to contrived changes inthose properties. For example, to test numberidentity concepts two toy Indians were placedin a box, and when the bottom was openedthree Indians dropped out. Among nonretardedsubjects (M IQ = 116), there were signif-icantly more frequent surprise reactions to achange in color than to changes in the threequantitative properties, a finding consistentwith Piaget's view (see Piaget & Voyat, 1968)that children develop identity concepts forqualitative properties such as color prior tothe emergence of identity concepts for quanti-tative properties such as number, length, andcontinuous quantity. Surprise reactions tochange in color and number were virtuallyidentical in 45 familial and 16 Down syndromeretarded subjects (M IQ = 47). The fre-quencies of surprise reactions to changes ofthe three different quantitative propertiesfor both retarded and nonretarded subjectsare consistent with the findings of others(including Gruen & Vore, 1972) that successfulperformance on conventional length andnumber conservation tasks is simpler than, andthus presumably developmentally prior to,


    success on conventional continuous quantitytasks. Thus, once again we see fairly strongsupport for the view that the sequence ofdevelopmental events for the retarded childis similar to that for the nonretarded child.

    A different type of similarity is illustratedby Achenbach's (1969) study of nonretardedpublic school children (IQs of 94-168) andnonorganically impaired retarded children(IQs of 31-78) from public schools andinstitutions. He assessed children's conserva-tion concepts with respect to length, area, andvolume3 (employing 4 tasks for each of thethree properties) by using optical illusions tocreate discrepancies between the actual andapparent sizes of various stimuli. To testconservation of length, for example, theexperimenter presented each child with abarbell illusion in which a small metal rodthat fit into a groove that just touched theinner edges of two circles was placed intoanother groove that passed through twocircles and touched their outer edges. Theeffect was to make the rod appear longer inthe second position than in the first. Subjectswere then asked whether the rod would fitinto the original groove. An important featureof the study, for our purposes, is that the 12tasks were designed to be free of intellectualdemands in the areas of additivity, numeration,conservation of equivalence, or complex verbalexpressiondimensions along which moretraditional conservation tasks often vary. Thismade it possible to test the contention ofBraine and Shanks (1965a, 1965b) that theattainment of conservation with respect tothe various properties would be parallel if theperformance criteria used were standard acrossthe types of conservation. Consistent with theBraine-Shanks view, Achenbach (1969) founda "total absence of evidence for a horizontaldecalage" (p. 677) in the three types of con-servation, for both the retarded group andthe nonretarded group. In both groups, therewere neither consistent nor significant differ-ences in success rates for length, area, andvolume tasks. This finding, together with thereasoning of Braine and Shanks, suggests thatsome of the order-of-difficulty and scalogram-type evidence reviewed in the precedingparagraphs may be more indicative of differ-ences in the specific requirements of the

    contrived tasks employed than of actualdifferences in the order of emergence of thevarious types of conservation.

    Concepts of time. A number of studies haveaddressed the similar sequence hypothesis incontent areas other than conservation. Inone study on the concept of time, Lovell andSlater (1960) interviewed 50 educationallysubnormal children (IQs not reported) aged8, 9, 10, 11, and 15 years and 50 "average toabove average" children aged 5-9 years. Theinterview included tasks (some were Piaget's)designed to measure concepts of simultaneityand equality of synchronous intervals (e.g.,asking the child to judge whether two dollstraveling at different rates but starting andstopping at the same time actually traveledfor the same amount of time). There were alsotasks involving chronological ordering ofevents and children's concepts of age andinterior time. Little in the way of statisticalanalysis was reported, but nonetheless, Lovelland Slater concluded that the understandingof these five concepts of time follows roughlythe same sequence in retarded as in normalchildren, although the stages in understandingare reached some years later by retardedchildren.

    Concepts of space. Two studies that in-cluded a retarded sample were specificallyconcerned with spatial concepts. In one ofthese, Houssiadas and Brown (1967) sampled40 institutionalized, mentally retarded Aus-tralians (M IQ = 55) who showed no evidenceof mongolism or other specific defects andwho ranged in age from 8-15 years. Thesesubjects were presented with two perspective-taking tasks (one pictorial and one usingmanipulation of actual objects) in which theywere asked to identify their own perspectiveon a perceptual array as well as the perspectiveof another person seated at a different position.Although no statistical analyses were reported,the pattern of passes and failures on thedifferent items was consistent with the viewthat retarded individuals pass through thestages identified by Piaget; that is, first thereis difficulty in identifying one's own perspective

    3 The volume conservation task used by Achenbach(1969) actually tapped what Piaget (has) called con-servation of continuous quantity.


    and that of another, second there is onlydifficulty in identifying how a perceptualarray might look from another position, andthird the individual is able to "coordinateperspectives," identifying not only his ownperspective but that of another person as well.Summing up, Houssiadas and Brown (1967)concluded, "It is clear that the pattern ofpredominant responses follows the samesequence suggested by Piaget, whose data werederived from normal children" (p. 213).

    In a more fine-grained analysis of spatialconcepts, Woodward (1962) tested the sameinstitutionalized retarded group of 50 adultsand 44 children used in the study of numberconcepts described earlier (Woodward, 1961).In this sample, 50% of the adults and 61%of the children showed some type of organicimpairment. The spatial tasks included mea-sures of the ability to reproduce a spatialorder under varying degrees of transformation(e.g., reproducing a circular array of beads ona horizontal rod). Using a similar procedurewith nonretarded children, Piaget and Inhelder(1956) identified seven stages through whichtheir children passed as they improved on thetasks. Woodward (1962) constructed a tableof scale types to assess the comparability ofher results with those of Piaget and Inhelder.Although no scalogram statistics were cal-culated, the great majority of Woodward'ssubjects fit scale types consistent with thedevelopmental sequence posited by theGenevans.

    A second task employed by Woodwardinvolved drawing copies of 21 geometricfigures used by Piaget and Inhelder. Thecompatibility of subjects' scores on thesetasks with a four-stage sequence advancedby Piaget and Inhelder (1956) was demon-strated by the fact that "subjects classifiedby the features of a given stage showed thefeatures of the lower stages in most cases"(Woodward, 1962, p. 31). However, onceagain no scalogram statistics were reported,and 5 of the 14 performance criteria bywhich stage assignments were made wereoutside the appropriate difficulty level forat least some of the subjects. The third taskemployed by Woodward was a "referencepoints" problem in which adult subjects onlywere presented with drawings of a bottle

    tilted at various angles and said to be aboutone-fourth full of water. The subjects' taskwas to pencil in the portion of the bottleoccupied by the water. The performance datapresented for this task were extremely sketchy,but Woodward indicated that the order ofdifficulty of the tasks was the same as thatfound by Piaget and Inhelder. In her overviewof her findings bearing on what we have calledthe similar sequence hypothesis, Woodward(1962) concluded that for her retarded subjects,"The sequence suggested by Piaget andInhelder [for nonretarded children J wasconfirmed for all three spatial concepts thatwere investigated" (p. 35).

    Relative thinking. In investigating the " logicof relations" in children, Piaget (1928) used a"brothers and sisters" problem and a "rightand left" problem. In the former problem,children's understanding of the relation be-tween being and having a sibling was exploredby asking such questions as "George hasthree brothers, Paul, Henry, and Charles.How many brothers has Paul? How manybrothers are there in this family?" In theright and left problem, children were in-structed, "Show me your right hand, your left.Show me my right hand, my left" Lane andKinder (1939) used these two Piagetianproblems with 50 institutionalized retardedindividuals of unspecified etiology who weregrouped, for purposes of data analysis, intofour different IQ levels: 38, 51, 64, and 77.Instead of scalogram statistics, relative levelsof the questions were reported for each IQgroup. These data indicated that the rankordering of difficulty for the 11 questions wassimilar across the different IQ levelsaparallelism consistent with the similar sequencehypothesis.

    Moral judgment. Abel (1941) investigatedmoral judgment in 74 institutionalized "sub-normal adolescent white girls" (aged 15-21years; IQs unspecified). Subjects were ques-tioned about seven brief stories concerningimmanent punishment (the inevitability ofpunishment following a misdeed), retributivejustice (punishment orientation, particularlyof the" eye for an eye" variety), and judgmentsof the gravity of a misdeed (using informationon consequences of the deed and intent of thetransgressor). Mirroring previous findings with


    nonretarded individuals (Lerner, 1937, 1938),Abel's findings were that with increasingmaturity (defined in terms of MA) subjectsgave nonsignificantly greater weight to intentand less weight to consequences in judgingthe gravity of a misdeed and were significantlyless likely to consistently advocate retributivepunishment. Unlike the nonretarded personsin at least some research, Abel's more maturesubjects (MAs of 9-11 years) did not showany less pronounced a belief in immanentpunishment than did her less mature subjects(MAs of 6-8 years). In fact, about 82% ofboth groups showed such a belief, which Abel(1941) attributed to the "constraining" in-stitutional environment "that controls thegirls with threats of immanent punishment"(p. 386). Except for this one anomaly, theAbel data are consistent with the similarsequence hypothesis.

    Stitdies of multiple concepts. We concludethis section on cross-sectional research with adiscussion of studies that have assessedconcepts in more than one conceptual domain.DeVries (1970, 1973a, 1973b, 1974) assesseda variety of Piagetian concepts in bright(M IQ 130), average (M IQ 105), andretarded (M IQ 72); etiologies not reported)children, all enrolled in public schools. Thetasks included the brothers and sisters andright and left problems described earlier;tests of generic and sex identity and of con-servation of mass, number, length, and liquid;interviews on magic and dream concepts;object sorting and class inclusion problems;and a guessing game ("Which hand has thepenny?") designed to reveal the level ofchildren's role-taking skills. Of all the tasksused, data from the guessing game task werepresented in the most complete manner (seeDeVries, 1970). Using an independent sampleof 64 high-IQ children, DeVries (1970)classified behavior on the guessing game withrespect to 10 characteristics (e.g., does notalways hide penny in the same hand). Thesecharacteristics formed a highly reproducibleGuttman-type scale with a reproducibility of.95 and an index of consistency of .66. Thescale was then used with the bright, average,and retarded samples and checked againstKohlberg's (1969) criteria for developmentalsequentiality, namely, (a) mean scale scores

    should increase with age, (b) success on eachindividual scale item should increase with age,and (c) the sequence of items should bejustifiable with a logical rationale based onPiagetian theory. DeVries (1973b) maintainedthat her scale met the third criterion, and herdata (DeVries, 1970) indicate that the firsttwo criteria were met within the bright,average, and retarded groups separately.Similar analyses were carried out with respectto the other 14 Piagetian tasks, with a Guttmanscale constructed for each. Within the averageand retarded groups each scale met Green's(1956) criterion of an index of consistencygreater than .50, and the lowest coefficient ofreproducibility was .94 (DeVries, Note 1).DeVries (1973b) indicated that all of theKohlberg criteria for sequentiality "wereapplied to each ability group (i.e., bright,average, and retarded subjects) separately,and the order of scale items was the same foreach ability group on all tasks" (p. 3).

    Stearns and Borkowski (1969) investigatedconservation of continuous quantity (water)and discontinuous quantity (blocks andmarbles) as well as horizontal-vertical spaceperception in institutionalized retarded in-dividuals (IQs unspecified) ranging in agefrom 7|-27 years. Consistent with Piaget's(e.g., 1964) view (and supporting findings;see Elkind, 1961) that conservation of con-tinuous quantity is more difficult and emergesdevelopmentally later than conservation ofdiscontinuous quantity, Stearns and Borkow-ski found performance on their test of theformer concept to be significantly poorerthan performance on their two tests of thelatter concept. Scores were also highly similarfor the tests of horizontal and vertical frames ofreference; this finding is consistent withPiaget's (see Piaget & Inhelder, 1956) viewthat concepts of the vertical and of thehorizontal are acquired at the same time.

    Finally, we turn to two studies by Lovell andhis colleagues (Lovell, Healey, & Rowland,1962; Lovell, Mitchell, & Everett, 1962).The studies reported few relevant statisticalanalyses, but the diversity of concepts exam-ined makes them worthy of brief attention.In the Lovell, Mitchell, and Everett study,groups of nonretarded and educationallysubnormal individuals (no IQs reported)


    were divided into separate age groups. Theskills investigated included additive andmultiplicative classification (of objects andpictures differing in multiple dimensions),seriation, multiplication of asymmetrical tran-sitive relations, hierarchical classification, classinclusion, and visual and tactile classification.For all tasks the tabled data indicated ageneral improvement in performance withincreasing age level for both normal and sub-normal groups (no significance tests reported).

    In the study by Lovell, Healey, and Rowland,the subjects were again groups of nonretardedand educationally subnormal persons fromspecial schools (IQs unreported) who weredivided into separate age groups. The groupswere presented with 12 of the tasks used byPiaget, Inhelder, and Szeminska (1960) tostudy the child's geometric concepts. Withinnormal and subnormal groups separately,correlation coefficients were calculated thatrelated Piagetian stage levels on the 12 tasksto subjects' age levels. Of the 24 coefficients,23 were significant at the .01 level. In bothLovell et al. studies, the details of subjectselection, experimental procedure, and statist-ical analyses are so skimpy that the findingsmust be regarded as only suggestive. Nonethe-less, although they are not by any meansdefinitive, the data are in harmony with thesimilar sequence hypothesis.

    Summary of the Cross-Sectional and Order-qf-Difficulty Evidence on the Similar SequenceHypothesis

    Thus far we have reviewed 28 studies inwhich cross-sectional and order-of-difficultyevidence is reported in ways that have somebearing on the similar sequence hypothesis.The degree of retardation involved in thesamples ranged from profound to mild; theretarded persons sampled ranged in age fromchildhood to adulthood, were both institu-tionalized and noninstitutionalized, and in-cluded both cultural-familial cases and individ-uals with diverse organic and emotionaldisorders. The nonretarded contrast groups,when employed, ranged from slightly belowaverage to extremely high in IQ. The studiesreported also varied widely in their experi-mental methodology and in their methodsof data analysis. Despite this great diversity

    in methodology and in sample characteristics,the data reviewed show rather consistentsupport for the similar sequence hypothesis,both in its conservative form, which appliesonly to nonretarded and familial retardedpersons (Zigler, 1969), and in its broaderform, in which universality of developmentalsequence is held to be independent of individualsubject characteristics such as organic impair-ment (see Kohlberg, 1969).

    There were very few exceptions to thisgeneralization: (a) Woodward (1959) foundthat among her profoundly retarded subjects1 sensorimotor task out of 11 proved to beslightly more difficult than another task thatPiaget designated as being one sensorimotorstage higher, (b) Among Rogers's (1977)profoundly retarded subjects, a causality taskdesigned to be at Stage 6 proved to be easierthan one of the tasks at Stage 5, whereasindividual spatiality tasks within each stagehad to be combined to yield a highly reproduc-ible scale, (c) In Achenbach's (1973) familialand Down's syndrome retarded sample, sur-prise reactions indicative of color and numberidentity occurred with equal frequency, where-as in his nonretarded sample, color surprisewas significantly more frequent than numbersurprise, (d) Abel (1941) found no significantdecline with increasing MA in belief in im-manent punishment in her institutionalizedfemale retarded sample, a finding that differsfrom some earlier research with nonretardedchildren.

    These four instances of no support for thesimilar sequence hypothesis are exceedinglyminor. They may have resulted, as the authorsof these studies generally suggested fromidiosyncratic (or misinterpreted) properties ofthe tasks selected or from other measurementerrors, and in some cases (e.g., Abel, 1941)they may reflect the suppressive influence ofan idiosyncratic environment that merelydelays the shift from one level of reasoning toanother. Furthermore, in each of the fourstudies, the findings of no support wereoutnumbered by findings supporting the similarsequence hypothesis.

    While noting the strong level of support thatthe reviewed evidence has yielded for thesimilar sequence hypothesis, we must alsonote that such cross-sectional and level-of-


    difficulty evidence, even at its best, can supportonly indirect inference regarding the actualprocess of development. Considerably moredirect and potent inference is possible when aninvestigator observes the same individuals atmore than one point during the course ofdevelopment, that is, in longitudinal fashion.Such research is often expensive and complex,and consequently it is relatively rare, partic-ularly with mentally retarded persons. How-ever, three longitudinal studies have somebearing on the similar sequence hypothesis.We now turn to these.

    Longitiidinal Evidence

    Development in the sensorimotor period. Onelongitudinal investigation was designed byWohlhueter and Sindberg (1975) as an exten-sion of Woodward's (1959) cross-sectionalstudy of sensorimotor development in pro-foundly retarded persons (described earlier inthis article). These investigators conductedmonthly assessments of institutionalized 1-6-year-old profoundly, severely, and moderatelyretarded children (no IQs reported). ThePiagetian object concept tasks used byDecarie (1965) were employed for 1 to 1? yearsor until a child performed at the highest of the10 substage levels for 2 consecutive monthlysessions. Of the principal sample of 49 children,20 had progressed to the highest substage levelby the end of the study; of the remaining 29,10 showed a generally monotonic increase, and9 seemed to be at a plateau, with objectconcept levels the same for most of the 12 ormore sessions. Thus, 39 of the 49 subjectsshowed patterns harmonious with the patternof object concept stages posited by Piaget(1955) and found by subsequent investigatorsusing nonretarded samples. For the remaining10 subjects, however, there was a variabledevelopmental pattern in which substagelevels appeared to rise and fall from sessionto session, ranging over as many as 3 or 4substages during the 12 or more sessions.

    In an effort to determine what character-istics might distinguish this group of atypicalsubjects, Wohlhueter and Sindberg, (1975)examined medical histories and clinical findingsfor their sample; the distinguishing feature ofthe variable group was that the majority ofsubjects "were found to have EEG abnormal-

    ities, especially dysrhythmias or a history ofseizures" (p. 516). This finding raises atleast two possible interpretations with respectto the variable developmental pattern: (a)that individuals with brain anomalies asso-ciated with electroencephalogram (EEG) ab-normalities may show atypical sequences ofdevelopment with respect to the object conceptand (b) that behavioral and attentionalabnormalities in individuals with anomalousEEG patterns make accurate assessment ofobject concept substages difficult.

    One other unusual pattern was noted by theinvestigators, namely, some children seemedto bypass or skip over some of the substages.This apparent skipping phenomenon has beennoted in research with nonretarded childrenas well (see Uzgiris, Note 2). Although thisis an interesting phenomenon, it is difficultto know how often skipped substages mayactually have been traversed by the childrenin the intervals between experimental sessions.Furthermore, both the skipping phenomenonand the variability phenomenon might bebetter understood if we were able to rule outspecific method effects, as could have beendone if a nonretarded sample had beenincluded in this study for comparative pur-poses. Nonetheless, the Wohlhueter-Sindberginvestigation at least raises significant ques-tions about the validity of the similar sequencehypothesis with respect to certain substagesin the development of the object concept.

    A longitudinal study reported by Cicchettiand Sroufe (1976), however, yields strongsupport for the similar sequence hypothesiswithin the sensorimotor period. The studyfocused on the relation between cognitive andaffective development in home-reared Down'ssyndrome infants during the period from4-18 months of age. Sroufe and his colleagues(e.g., Sroufe & Wunsch, 1972) earlier demon-strated that among normal infants there is adevelopmental progression from mirth responseto auditory and tactile stimulation that isphysically intense or vigorous (e.g., ticklingthe baby's chin or saying "Boom!") to mirthresponses to social and visual stimulationthat is progressively more subtle and complex(e.g., the sight of mother sucking on baby'sbottle). At monthly intervals the infantssampled by Cicchetti and Sroufe were pre-


    sented with 15 auditory and tactile items ofthe intense or vigorous type and 15 socialand visual items of the more subtle andcomplex type. As in the earlier research withnormal infants, the Down's syndrome infantslaughed earliest in response to the auditoryand tactile items and latest in response to"the more cognitively complicated social andvisual items" (Cicchetti & Sroufe, 1976,p. 923). The responses of the Down's infants,of course, came months later (in CA) than didthe corresponding responses of normal infants.Smile responses, a more sensitive index ofpositive affect in the Down's syndrome sample,showed the same pattern and revealed evenmore clearly than laughter responses thedevelopmental decline in positive affect arousedby simpler auditory and tactile items as theinfants matured beyond 13 months. Thisinverted-U-shaped developmental pattern alsoresembles earlier findings with normal infants.In stressing the similarity of their findings withDown's infants to those with normal infants,Cicchetti and Sroufe (1976) pointed out thatthe laughter items were similarly ordered forboth groups, "category by category and, inthe main, item by item" (p. 923). Finally, toassess the merits of their claim that theaffective responses they measured were closelyrelated to cognitive development, Cicchettiand Sroufe calculated correlations of indicesof affective expression (e.g., earliest laugh,total amount of smiling to all items, etc.)with the Bayley mental and motors scales andthe Uzgiris-Hunt object permanence andoperational causality scales. All 44 correlationswere statistically significant.

    Therefore, the Cicchetti-Sroufe investiga-tion, unlike the Wohlhueter-Sindberg study,provides uniform support for the liberalversion of the similar sequence hypothesisin which the hypothesis is applied to allretarded children regardless of etiology. TheCicchetti-Sroufe research deserves specialattention because of its unusually carefulmethology and its emphasis on the integrityof the developing infant. The research demon-strates a thoughtful means of assessing thePiagetian hypothesis that affective develop-ment and cognitive development are inter-dependent. In addition, it may help to pointthe way to tests of the similar sequence

    hypothesis in behavioral domains other thancognitive development as it has traditionallybeen construed.

    Research on the preoperational-concrete opera-tional transitionthe Temple longitudinal study.The third longitudinal study reviewed hereis by far the broadest in scope. In this ongoinginvestigation, Stephens and her colleagues(Stephens, 1974; Stephens, Mahaney, &Mclaughlin, 1972; Stephens et al., 1974)have conducted biennial assessments of theperformance of retarded and nonretardedpersons on a variety of Piagetian tasks. Thesample included 75 retarded subjects (IQs of50-75) from special education classes and75 nonretarded subjects (IQs of 90-110) fromthe same Philadelphia schools. In the firstwave of testing, the age range in both subjectgroups was 6-18 years. Results from the firsttwo waves of testing have now been reportedand are discussed here within two contentcategories:

    Moral judgment. The Temple battery in-cluded 11 measures designed to assess threeaspects of moral judgment: (a) the relativeweight assigned to intent versus consequencesin judging the seriousness of a misdeed, (b)awareness of the injustice of punishing anentire group for the acts of only one or a fewmembers, and (c) the ability to judge therelative fairness of various types of punishmentincluding retributive and reciprocal justice.To determine whether judgment along thesethree dimensions follows the same develop-mental course in the retarded as in the non-retarded, Mahaney and Stephens (1974)examined changes in scores on the 11 compo-nent measures over the 2-year period fromthe first to the second wave of testing. Theyfound that on 1 of the intent-versus-conse-quences measures the retarded group showeda nonsignificant decline in score (i.e., theymade slightly less mature moral judgmentsaccording to the scoring criteria adopted bythe authors) and that on 1 of the grouppunishment items nonretarded subjects showeda nonsignificant decline. On the 9 remainingitems the direction of change was the same forboth the retarded subjects and the nonretardedsubjects; this similarity extended to 2 itemson which both groups showed significant de-


    clines in score, raising questions about thescoring of these particular items.

    Inhelder (1943/1968), in a study describedearlier, referred to certain "oscillations" in thereasoning of the retarded. It is of some interestto note that Mahaney and Stephens (who wasthe translator of the Inhelder book) reportedoscillations, that is, instances when "theimprovement which occurred in one area ofmoral judgment was not maintained whenopinions were solicited on another, but similar,situation" (Mahaney & Stephens, p. 137),in both retarded and nonretarded subjects.There is some indication that such oscillationsmay have been somewhat more frequent inthe retarded group. One line of evidence thatsuggests this possibility comes from Mahaneyand Stephens's analysis of change scores forthree separate age levels within both theretarded and the nonretarded samples. Of the29 change scores reported for the nonretardedgroups, 20 were increases (11 significant), 7were decreases (3 significant), and 2 showedno change. Among the retarded subjects, 17change scores were increases (7 significant),11 were decreases (4 significant), and 1involved no change. However, when retardedcomparison groups were staggered in order tobroaden the age difference involved in agegroup comparisons, for example, by comparingPhase 1 6-10-year-olds with Phase 2 12-16-year-olds, the only items showing a decreasewith age were the two that showed a decreasein nonretarded subjects as well. Overall, thereport by Mahaney and Stephens (1974)suggests that although growth in moraljudgment concepts among the retarded maybe "torporific and sporadic" (p. 141), thedirection of development is the same for bothretarded and nonretarded persons.4

    Conservation, classification, symbolic imagery,and formal operations in the Temple study.The Temple investigation (Stephens, 1974;Stephens et al., 1972; Stephens et al., 1974)also included 29 measures of cognitive develop-ment across four broad conceptual domains:(a) conservation (of substance, length, weight,continuous quantity, and volume, as well asterm-to-term correspondence), (b) logic classi-fication (class inclusion and class intersection,and relative thinking measured by the broth-ers-sisters and right-left tests), (c) operativity

    and symbolic imagery (tests involving imag-ined rotations of objects through space,transferring from two to three dimensions, andchanging one's perspective on a stimulus),and (d) combinatory logic (Piaget & Inhelder's1956, combination of liquids task). Explana-tions given by subjects on each of the 29items were scored on a 9-point scale that tookinto account, among other things, the degreeto which the subject wavered between a correctand an incorrect answer and the degree towhich reversibility was shown. Stephens andMcLaughlin (1974) reported on changes inscores on the 29 measures over the 2-yearperiod separating the two waves of testing.They found that the nonretarded group showedimprovement on all 29 measures, with 25statistically significant; the retarded groupalso improved on all 29 measures, with 26statistically significant. This finding indicatesthat the direction of development on thesePiagetian reasoning tasks was similar in theretarded and nonretarded groups. In anotherreport following the second wave of theTemple study (Stephens et al., 1972), thePiagetian reasoning tasks were rank orderedwith respect to the MAs at which 50% of thesubjects (in the retarded and nonretardedgroups separately) made correct responses.As Stephens et al. indicated, the order of diffi-culty for both subject groups was generally con-sistent with previous findings that conservationof substance precedes conservation of weight

    4 Whether this apparent similarity in the direction ofdevelopment is actually a function of an invariant,stagelike progression is thus far an open questionbecause of the questionable nature of the measuresthemselves. In a critique of this portion of the Templelongitudinal study, Kohlberg (1974) maintained thatPiagetian moral judgment measures used in theTemple study do not even warrant detailed longitudinalanalysis, because

    Piaget himself does not consider that his moraljudgment measures yield genuine stages, nor do theypair up with his logical stages in ways compatiblewith his current thinking about cognitive stages . . . .Empirical research confirms the fact that Piaget'smoral stage measures do not meet the criteria ofstructural stages which his logical stages do meet,(p. 142)

    This being the case, it is appropriate to be cautiousabout what one concludes with respect to the moraljudgment portion of the Temple investigation.


    and that conservation of weight precedesconservation of volume. In addition, althoughthe ranks were rather crude because MA levelswere listed in whole years and because not eventhe most ardent Piagetian would expect all29 items to form an orderly developmentalscale, it is interesting to note that our owncalculations yielded a Spearman rho of .634between the rank order given for the retardedgroup and that given for the nonretardedgroup.

    The preceding data are consistent with thesimilar sequence hypothesis, as far as theyhave been taken, but they could be takenconsiderably further. With the one exceptionmentioned in the preceding paragraph, therehas been no apparent effort thus far by theTemple investigators to check their findingsagainst specific developmental stage sequencessuch as the horizontal decalages reported fornonretarded subjects in previous Piagetianresearch (Kohlberg & DeVries, 1971; Nassafat,1963; Siegelman & Block, 1969; Smedslund,1964; Uzgiris, 1968). Moreover, there has beena persistent inclination to report data in termsof group means, rather than in terms of thenumber of individuals (retarded and non-retarded) who show specific developmentalpatterns. This latter type of analysis is theunique province of longitudinal research andcan only be approximated indirectly by scalingprocedures in research of the nonlongitudinalvariety. There is some indication (see Stephens,1974) that efforts to profile individualperformance changes over time and to cross-validate specific vertical and horizontal de-calages will be forthcoming from the Templeinvestigators. Such efforts are needed if theinvestigators are to fully capitalize on thepower of their longitudinal design.

    Status of Evidence on the Similar SequenceHypothesis

    Only 1 of the 3 longitudinal studies reviewedthe Wohlhueter and Sindberg (1975) investi-gation of object concept substagesproducedfindings inconsistent with the similar sequencehypothesis. In that investigation a distinctsubgroup of 10 (out of 49) children showedapparent atypical developmental sequences,and most of these children showed anomalous

    EEC patterns. This finding may indicate thatbrain wave anomalies can be associated withatypical developmental patterns. Alterna-tively, the EEG abnormalities may simply havebeen associated with attentional and otherdeficits that interfered with accurate assess-ment of substage levels in children whose actualdevelopment was consistent with the similarsequence hypothesis. The latter interpretationhas special credence in the area of objectconcept assessment, in which proceduresdemand that the subject sustain attention toan object long enough to seek after it onceit has been removed from the preceptual field.Of the 28 nonlongitudinal studies reviewed,only 4 contained a finding inconsistent withthe similar sequence hypothesis, and in eachof these studies the inconsistent finding wasrelatively minor and was alone among anumber of findings supporting the hypothesis.Furthermore, the questions raised generallyconcerned rather fine-grained steps or sub-stages within horizontal decalages on whichstudies of nonretarded subjects alone havenot always agreed.

    These facts, plus the measurement problemsinherent in these experimental procedures,make the degree of consistency in the findingsof these 31 studies rather striking. Positivefindings have now been reported in conceptualareas that include sensorimotor spatial con-cepts, object permanence, causality, imitation,affective responding, identity and equivalenceconservation (of many properties), seriation,transitivity, moral reasoning, comparison proc-esses (or gross, intensive, and extensivequantities), time, space, relative thinking,role taking, mental imagery, geometric con-cepts, and classification and class inclusion.For the 31 studies spanning this list of concep-tual areas, the great preponderance of theevidence is consistent with the hypothesisthat retarded and nonretarded persons traversethe same stages of development in the sameorder, differing only in the rate at which theyprogress and in the ultimate ceiling theyattain. The hypothesis seems to be generallysupported in studies of retarded individuals,regardless of etiology, with the possibleexception of individuals suffering from pro-nounced EEG abnormalities.


    Quality of the Evidence

    Having said this, we believe it is importantto comment on the quality of the availableevidence and to offer suggestions for improvingit. Cross-sectional data relevant to the similarsequence hypothesis have most often beenpresented in ways that provide only theweakest inferential power. A table displayingthe percentage of subjects at each age levelwho pass each Piagetian item can yield onlya rather faint glimmer of developmentalsequence compared with the informationgenerated when each child is classified withrespect to specific pass-fail patterns, that is,with respect to response scale types. When sucha scaling analysis is combined with calculationof scalogram summary statistics (e.g., Green,1956; Guttman, 1950), the potential powerof the nonlongitudinal design is more fullyutilized.

    Similarly, the bulk of the longitudinal datawe found was presented only in terms of meandifference between experimental groups orchanges in group means or percentages overtime. As Hunt (1974) has noted, reportingonly group summary statistics at Time 1 andTime 2 can mask the fact that some individualsprogressed while others regressed over time.Thus, although it is useful to know that theTime 1 and Time 2 means differed in the samedirection for retarded and nonretarded groups,such information is no substitute for ananalysis of the number of individuals in eachgroup showing specific developmental patternsover time. In both longitudinal and non-longitudinal research aimed at testing thesimilar sequence hypothesis, it makes littlesense to invest the time and energy necessaryto gather potentially relevant data and thenanalyze the data in ways that fail to capitalizeon their full potential.

    Suggestions Toward Improved Research

    These problems, and others to which wereferred earlier in the text, suggest threeprinciples that if widely adopted would sub-stantially improve the quality of evidence onthe similar sequence hypothesis.

    Structuring Direct Comparisons

    A problem with many of the studies reviewedin earlier sections is that their samples included

    only mentally retarded subjects. In those fewinstances in which findings of these studiesdisagree with findings of other studies samplingonly nonretarded subjects, the discrepanciesare difficult to interpret. This is because ofuncertainty over whether the discrepanciesreflect actual process differences between theretarded and the nonretarded or whether thedifferences in experimental methodology acrossstudies are responsible. An obvious way toprevent such difficulties is to expose retardedand nonretarded children within a similarcognitive developmental range to preciselythe same procedure by including both groups inthe same study. To fail to do so is to riskuninterpretable findings.

    Attending to Etiology

    The marked heterogeneity of many of thementally retarded samples described earliersuggests a somewhat opportunistic approachto subject selection or perhaps an approach inwhich the etiology of retardation is simply notregarded as an important factor. Yet, theoret-ical considerations discussed early in thisarticle (see also, Weisz, 1976; Zigler, 1969,1971) point to the need to give special attentionto familial retarded children as opposed tothose suffering from organic impairment orgenetic disorder. Furthermore, Wohlhueterand Sindberg's (1975) report of atypicaldevelopment in a group of children with ahigh incidence of EEG anomalies suggeststhe potential importance of efforts to identifydevelopmentally distinct subgroups within thenonfamilial population. Their analysis illus-trates that subgroup analyses can be usefuleven when they are post hoc.

    Promoting Uniformity

    Finally, there is a clear need for increaseduniformity across studies in the kinds ofstatistical analyses carried out and in theway statistics are reported. Toward this end,we suggest that every cross-sectional studyaddressing the similar sequence hypothesisshould yield data bearing on the followingthreefold question:

    1. Within the retarded and nonretardedgroups do the task items form the same scale,and does this scale show high reproducibility


    (a la Guttman, 1950) and a high index ofconsistency (a la Green, 1956) ?

    2. Do mean scale scores increase with levelof cognitive maturity in each separate subjectgroup (see Kohlberg, 1969) ?

    3. Do mean levels of success on eachindividual item increase with levels of cognitivematurity in each separate subject group (seeKohlberg, 1969)?In longitudinal research, Questions 2 and 3should also be asked in a way that onlylongitudinal investigation permits: Over theperiod spanned by the longitudinal study, inwhat percentage of individual subjects (fromretarded and nonretarded groups) do scalescores and individual item scores (a) increaseeither smoothly or monotonically, (b) remainstable throughout, and (c) show at leastsome declines.

    These recommended questions are designedto promote greater uniformity, and thusgreater comparability, among studies address-ing the similar sequence hypothesis. Inopposition to such uniformity, one might arguethat the degree of consistency in the findingsof the numerous studies reviewed here is allthe more impressive precisely because of themethodological diversity of the studies. Thereis some truth to this argument, but only inthose cases in which findings support thesimilar sequence hypothesis. However, wehave argued that even in those cases apparentgroup similarities in developmental sequencemay result from a failure to ask the mostprobing questions of one's data. It seems clearfrom our review that evidence from the 31studies currently available offers rather con-sistent support for the similar sequencehypothesis; it also seems likely that the bestevidence has yet to be gathered.

    Reference Notes

    1. DeVries, R. Personal communication, February 1,1978.

    2. Uzgiris, I. C. Some antecedents of the object concept.Paper presented at the meeting of the EasternPsychological Association, Philadelphia, April 1969.


    Abel, T. M. Moral judgments among subnormals.Journal of Abnormal and Social Psychology, 1941,36,378-392.

    Achenbach, T. M. Conservation of illusion-distortedidentity: Its relation to MA and CA in normals andretardates. Child Development, 1969, 40, 663-679.

    Achenbach, T. M. Surprise and memory as indices ofconcrete operational development. PsychologicalReports, 1973, 33, 47-57.

    Balla, D., Styfco, S. J., & Zigler, E. Use of the opposi-tion concept and outerdirectedness in intellectuallyaverage, familial retarded, and organically retardedchildren. American Journal of Mental Deficiency,1971, 75, 663-680.

    Balla, D., & Zigler, E. Discrimination and switchinglearning in normal, familial retarded, and organicretarded children. Journal of Abnormal and SocialPsychology, 1964, 61, 664-669.

    Benton, A. L. Behavioral indices of brain injury inschool children. Child Development, 1962, 33, 199-208.

    Braine, M. D. S., & Shanks, B. The conservation of ashape property and a proposal about the origins ofthe conservations. Canadian Journal of Psychology,1965, 19, 197-207. (a)

    Braine, M. D. S., & Shanks, B. The development ofconservation of size. Journal of Verbal Learning andVerbal Behavior, 1965, 4, 227-242. (b)

    Buck-Morss, S. Socio-economic bias in Piaget's theoryand its implications for cross-culture studies. HumanDevelopment, 1975, 18, 35-49.

    Charlesworth, W. R. The role of surprise in cognitivedevelopment. In D. Elkind & J. H. Flavell (Eds.),Studies in cognitive development: Essays in honor ofJean Piaget. New York: Oxford University Press,1969.

    Cicchetti, D., & Sroufe, L. A. The relationship betweenaffective and cognitive development in Down'ssyndrome infants. Child Development, 1976, 47,920-929.

    Cruikshank, W. M. The brain-injured child in home,school, and community. Syracuse, N.Y.: SyracuseUniversity Press, 1967.

    Decarie, T. G. Intelligence and affectivity in earlychildhood. New York: International UniversitiesPress, 1965.

    DeVries, R. The development of role-taking as reflectedby behavior of bright, average, and retarded childrenin a social guessing game. Child Development, 1970,41, 759-770.

    DeVries, R. Performance on Piaget-type tasks of high-IQ,ai>erage-IQ, and low-IQ children. Paper presented atthe meeting of the Society for Research in ChildDevelopment, Philadelphia, Pa., 1973. (ERICDocument Reproduction Service No. ED 086 374/PS007 129) (a)

    DeVries, R. The two intelligences of bright, average, andretarded children. Paper presented at the meetingof the Society for Research in Child Development,Philadelphia, Pa., 1973. (ERIC Document Reproduc-tion Service No. ED 079/102/SE 016 419) (b)

    DeVries, R. Relationships among Piagetian, IQ, andachievement assessments. Child Development, 1974,45, 746-756.

    Elkind, D. The development of quantitative thinking.Journal of Genetic Psychology, 1961, 98, 36-46.

    Elkind, D. Piaget's conservation problems. ChildDevelopment, 1967, 38, 15-27.


    Elkind, D., Koegler, R. R., Go, E., & Van Doorninck,W. Effects of perceptual training on unmatchedsamples of brain-injured and familial retardedchildren. Journal of Abnormal Pscychology, 1965,70, 107-110.

    Ellis, N. R. A behavioral research strategy in mentalretardation: Defense and critique. American Journalof Mental Deficiency, 1969, 73, 557-566.

    Green, B. F. A method for scalogram analysis usingsummary statistics. Psychometrika, 1956, 21, 79-88.

    Gruen, G. E., & Vore, D. A. Development of conserva-tion in normal and retarded children. DevelopmentalPsychology, 1972, 6, 146-157.

    Guttman, L. The basis of scalogram analysis. In S. A.Stouffer et al. (Eds.), Measurement and prediction(Vol. 4). Princeton, N.J.: Princeton UniversityPress, 1950.

    Harter, S., Brown, L., & Zigler, E. The discriminationlearning of normal and retarded children as a functionof penalty conditions and etiology of the retarded.Child Development, 1971, 42, 517-536.

    Houssiadas, L., & Brown, L. B. The coordination ofperspectives by mentally retarded children. Journalof Genetic Psychology, 1967, 110, 211-215.

    Hunt, J. McV. Discussion: Developmental gains inreasoning. American Journal of Mental Deficiency,1974, 79, 127-133.

    Inhelder, B. The diagnosis of reasoning in the mentallyretarded. New York: Day, 1968. (Originally pub-lished, 1943)

    Jordan, V. B. Cognitive development among retardates:Reanalysis of Inhelder's data. Paper presented atthe meeting of the Society for Research in ChildDevelopment, Denver, Colo., 1976. (ERIC Docu-ment Reproduction Service No. EB 121 035/EC082 713)

    Klauss, S. D., & Green, M. B. Conservation in trainablementally retarded children. Training School Bulletin,1972, 69, 108-114.

    Kohlberg, L. Stage and sequence: The cognitive-developmental approach to socialization. In D.Goslin (Ed.), Handbook of socialization theory andresearch. Chicago: Rand McNally, 1969.

    Kohlberg, L. From is to ought: How to commit thenaturalistic fallacy and get away with it in the studyof moral development. In T. Mischel (Ed.), Cognitivedevelopment and epistemology. New York: AcademicPress, 1971.

    Kohlberg, L. Discussion: Developmental gains in moraljudgment. American Journal of Mental Deficiency,1974, 79, 142-146.

    Kohlberg, L., & DeVries, R. Relations betweenPiagetian and psychometric assessments of intel-ligence. In C. Lavatelli (Ed.), The natural curriculum.Urbana: University of Illinois Press, 1971.

    Kooistra, W. H. Developmental trends in the attain-ment of conservation, transitivity, and relativism inthe thinking of children: A replication and extensionof Piaget's ontogenetic formulations (Doctoraldissertation, Wayne State University, 1963). Dis-sertation Abstracts, 1964, 25, 2032. (UniversityMicrofilms No. 64-9538)

    Lane, E. B., & Kinder, E. F. Relativism in the thinkingof subnormal subjects as measured by certain of

    Piaget's tests. Journal of Genetic Psychology, 1939,54, 107-118.

    Lerner, E. Constraint areas and the moral judgment ofchildren. Memasha, Wis.: Banta, 1937.

    Lerner, E. Observations sur le raisonnement moralde 1'enfant. Cahiers pedagogiques experimentaux etpsychologiques de 1'enfant (Vol. No. 11). Geneva:Palais Wilson, 1938.

    Lister, C. The development of ESN children's under-standing of conservation in a range of attributesituations. British Journal of Educational Psychology,1972, 42, 14-22.

    Lovell, K., Healey, D., & Rowland, A. D. Growth ofsome geometric concepts. Child Development, 1962,33, 751-767.

    Lovell, K., Mitchell, B., & Everett, I. R. An experi-mental study of the growth of some logical structures.British Journal of Psychology, 1962, 53, 175-188.

    Lovell, K., & Slater, A. The growth of the concept oftime: A comparative study. Child Psychology andPsychiatry, 1960, /, 179-190.

    Mahaney, E. J., & Stephens, B. Two-year gains inmoral judgment by retarded and nonretarded persons.American Journal of Mental Deficiency, 1974, 79,134-141.

    Mannix, J. B. The number concepts of a group ofE. S. N. children. British Journal of EducationalPsychology, 1960, 30, 180-181.

    Marchi, J. U. Comparison of selected Piagetian taskswith the Wechsler Intelligence Scale for Children asmeasures of mental retardation (Doctoral disserta-tion, University of California, Berkeley, 1970).Dissertation Abstracts International, 1971, 31, 6442A.(University Microfilms No. 71-51, 833).

    Mermelstein, E., & Shulman, L. S. Lack of formalschooling and the acquisition of conservation.Child Development, 1967, 38, 39-52.

    McManis, D. Comparison of gross, intensive, andextensive quantities by retardates. Journal ofGenetic Psychology, 1969, 115, 229-236. (a)

    McManis, D. Comparisons of gross, intensive, andextensive quantities by normals and retardates.Child Development, 1969, 30, 237-244. (b)

    McManis, D. Conservation of identity and equivalenceof quantity by retardates. Journal of Genetic Psy-chology, 1969, 115, 63-69. (c)

    McManis, D. Conservation of mass, weight, and volumeby normal and retarded children. American Journal ofMental Deficiency, 1969, 73, 762-767. (d)

    McManis, D. Conservation and transitivity of weightand length by normals and retardates. DevelopmentalPsychology, 1969, /, 373-382. (e)

    McManis, D. Conservation, seriation, and transitivityperformance by retarded and average individuals.American Journal of Mental Deficiency, 1970, 74,784-791.

    Milgram, N. A. Cognition and language in mentalretardation: Distinctions and implications. In D. K.Routh (Ed.), The experimental psychology of mentalretardation. Chicago: Aldine, 1973.

    Nassafat, M. Etude quantitative sur devolution des opera-tions intellectuelles: Le passage des operations concretesaux operations formelles. Neuchatel, Switzerland:Delachaux et Niestl6, 1963.


    Papalia, D., & Hooper, F. A developmental comparisonof identity and equivalence. Journal of ExperimentalChild Psychology, 1971, 12, 347-361.

    Piaget, J. Judgment and reasoning in the child. NewYork: Harcourt, Brace, 1928.

    Piaget, J. The child's conception of number. New York:Humanities Press, 1952.

    Piaget, J. The origins of intelligence in the child. London:Routledge & Kegan Paul, 1953.

    Piaget, J. The construction of reality in the child. London:Routledge & Kegan Paul, 1955.

    Piaget, J. The general problem of the psychobiologicaldevelopment of the child. Discussions on ChildDevelopment, 1956, 4, 3-27.

    Piaget, J. Play, dreams, and imitation in childhood.New York: Norton, 1962.

    Piaget, J. Cognitive development in children: ThePiaget papers. In R. E. Ripple & V. N. Rockcastle(Eds.), Piaget rediscovered: A report of the Conferenceon Cognitive Studies and Curriculum Development.Ithaca, N.Y.: Cornell University, School of Educa-tion, 1964.

    Piaget, J. Necessity et signification des recherchescomparatives en psychologic g6n6tique. InternationalJournal of Psychology, 1966, I , 3-13.

    Piaget, J. Preface. In B. Inhelder, The diagnosis ofreasoning in the mentally retarded. New York:Chandler, 1968.

    Piaget, J. Piaget's theory. In P. H. Mussen (Ed.),Carmichael's manual of child psychology (3rd ed.).New York: Wiley, 1970.

    Piaget, J. The psychology of intelligence. Totowa, N.J.:Littlefield, Adams, 1972.

    Piaget, J., & Inhelder, B. Le developpement des quantiteschez I'enfant. Neuchatel, Switzerland: Delachauxet Niestl6, 1941.

    Piaget, J., & Inhelder, B. The child's conception of space.New York: Humanities Press, 1956.

    Piaget, J., Inhelder, B., & Szeminska, A. The child'sconception of geometry. New York: Routledge &Kegan Paul, 1960.

    Piaget, J., & Voyat, G. Recherche sur 1'identitfi d'uncorps en d6veloppement et sur celle du mouvementtransitif. In J. Piaget, H. Sinclair, & Vinh Bang(Eds.), Epistemologie et psychologie de I'identite:Etude d'epistemologie genetique (Vol. 24). Paris:Presses Universitaires de France, 1968.

    Popper, K. R. The logic of scientific discovery. New York:Basic Books, 1959.

    Reitan, R. M. Psychological assessment of deficitsassociated with brain lesions in subjects. In J. L.Khanna (Ed.), Brain damage and mental retardation:A psychological evaluation. Springfield, 111.: CharlesC Thomas, 1973.

    Rogers, S. J. Characteristics of the cognitive develop-ment of profoundly retarded children. Child Develop-ment, 1977, 48, 837-843.

    Roodin, P. A., Sullivan, L., & Rybash, J. M. Effectsof a mem