developmental models of supervision: is it development?

8
Professional Psychology: Research and Practice 1987, Vol. 18. No. 3, 209-216 Copyright 1987 hy the Am n Psychological Association. Inc. 0735-7028/87/J00.75 Developmental Models of Supervision: Is It Development? Elizabeth L. Holloway Division of Counseling and Educational Psychology University of Oregon Developmental models of supervision have become the Zeitgeist of supervision thinking and research. Eighteen different models of supervision that refer to developmental principles have been cited in the psychiatric, psychological, and social work disciplines (Worthington, 1984). Numerous empirical studies have been designed to investigate developmental paradigms (Roehlke, 1984). If these models are to continue to influence research and practice, it seems warranted that the underlying assumptions and principles of the proposed models be critically examined. This article contains such a critical examination, and the conceptual and methodo- logical approaches currently applied in the research of this topic are discussed from a develop- mental perspective. The design of theoretical models for the clinical supervision process has been of interest to counseling psychologists and other clinical disciplines for the last 30 years. For 25 of those years the predominant approach has been the extrapolation of counseling theory to the supervisory experience. Supervi- sion models have been identified traditionally by counseling nomenclature—for example, psychodynamic supervision, ra- tional-emotive theory supervision, and behavioral supervi- sion (Goodyear, Bradley, & Bartlett, 1983). The call for a new approach has come from researchers' and educators' recog- nition that a purely clinical model of training is inadequate in explaining the supervisory phenomenon (Hess, 1980; Hol- loway, 1984; Holloway & Hosford, 1983; Holloway & Wam- pold, 1983; Lambert, 1974, 1980; Littrell, Lee-Borden, & Lorenz, 1979; Loganbill, Hardy, & Delworth, 1982; Stolten- berg, 1981). In the last 5 years there has been a significant change in the conceptualization of supervision as theorists have begun to apply descriptions of psychosocial development to counselor trainees' clinical learning in graduate training programs. Eighteen different models of supervision that refer to devel- opmental principles have been cited in the psychiatric, psy- chological, and social work disciplines (Worthington, 1984). Developmental models of individual supervision that have appeared in the counseling psychology literature include those by Littrell et al. (1979), Stoltenberg (1981), Loganbill et al. (1982), and Blocher (1983). Hogan's developmental concep- ELEABETH L. HOLI.OWAY received her PhD from the University of Wisconsin-Madison in 1979. She is currently on the faculty of the Division of Counseling and Educational Psychology at the University of Oregon. She is a counseling psychologist, and her research areas include counselor training, clinical supervision, and counseling proc- ess. She is President-Elect of the Oregon Counselor Education and Supervision Association and is on the Editorial Boards of The Coun- seling Psychologist, The Journal of Counseling Psychology, and Coun- seling and Values. THE AUTHOR THANKS Lloyd Lovell for his comments on an earlier version of this article. CORRESPONDtNCE CONCERNING THIS ARTICLE should be addressed to Elizabeth L. Holloway. Division of Counseling and Educational Psychology, University of Oregon. Eugene, Oregon 97403. tion of supervision appeared in the clinical literature in 1964 and has been referred to extensively in the more recent counseling models. Developmental models for the group su- pervision of counselors have also appeared recently (Sansbury, 1982; Yogev, 1982). Numerous empirical studies have been designed to investigate developmental paradigms (Roehlke, 1984). Developmental models of supervision have become the Zeitgeist of supervision thinking and research. If these models are to continue to influence research and practice, it seems warranted that the underlying assumptions and prin- ciples of the proposed models be critically examined at this stage of their development. In this article, I engage in such a critical examination and discuss from a developmental per- spective the conceptual and methodological approaches cur- rently applied in the research of this topic. Characteristics of the Developmental Models Essentially, five developmental models of individual super- vision have been the focus of thinking and research in the counseling literature; three of these models (Blocher, 1983; Loganbill et al., 1982; Stoltenberg, 1981) have explicitly linked their origins to psychosocial developmental theory. These three models serve as the basis of the conceptual discussion. In the remaining two models, Hogan (1964) and Littrell et al. (1979) characterized the supervisory process, using the terminology and descriptors of development. How- ever, these models do not refer to a particular theory as the origin of their work or to an underlying developmental process that directs the hierarchical progression of clinical learning. Because no explicit reference to traditional developmental theory is claimed, these models are examined only as they have been used as a referent and/or basis for the Stoltenberg, Loganbill et al., and Blocher models. Little attention has been directed at the unique underlying assumptions of the Stoltenberg (1981), Loganbill et al. (1982), and Blocher (1983) paradigms. Because each appears distinct enough to warrant individual scrutiny, the fundamental de- velopmental characteristics of each of the three models are described in the next sections. 209

Upload: others

Post on 01-Jan-2022

1 views

Category:

Documents


0 download

TRANSCRIPT

Professional Psychology: Research and Practice1987, Vol. 18. No. 3, 209-216

Copyright 1987 hy the Am n Psychological Association. Inc.0735-7028/87/J00.75

Developmental Models of Supervision: Is It Development?

Elizabeth L. HollowayDivision of Counseling and Educational Psychology

University of Oregon

Developmental models of supervision have become the Zeitgeist of supervision thinking andresearch. Eighteen different models of supervision that refer to developmental principles have

been cited in the psychiatric, psychological, and social work disciplines (Worthington, 1984).

Numerous empirical studies have been designed to investigate developmental paradigms(Roehlke, 1984). If these models are to continue to influence research and practice, it seems

warranted that the underlying assumptions and principles of the proposed models be critically

examined. This article contains such a critical examination, and the conceptual and methodo-

logical approaches currently applied in the research of this topic are discussed from a develop-

mental perspective.

The design of theoretical models for the clinical supervisionprocess has been of interest to counseling psychologists andother clinical disciplines for the last 30 years. For 25 of thoseyears the predominant approach has been the extrapolationof counseling theory to the supervisory experience. Supervi-sion models have been identified traditionally by counselingnomenclature—for example, psychodynamic supervision, ra-tional-emotive theory supervision, and behavioral supervi-sion (Goodyear, Bradley, & Bartlett, 1983). The call for a newapproach has come from researchers' and educators' recog-nition that a purely clinical model of training is inadequatein explaining the supervisory phenomenon (Hess, 1980; Hol-loway, 1984; Holloway & Hosford, 1983; Holloway & Wam-pold, 1983; Lambert, 1974, 1980; Littrell, Lee-Borden, &Lorenz, 1979; Loganbill, Hardy, & Delworth, 1982; Stolten-berg, 1981).

In the last 5 years there has been a significant change in theconceptualization of supervision as theorists have begun toapply descriptions of psychosocial development to counselortrainees' clinical learning in graduate training programs.Eighteen different models of supervision that refer to devel-opmental principles have been cited in the psychiatric, psy-chological, and social work disciplines (Worthington, 1984).Developmental models of individual supervision that haveappeared in the counseling psychology literature include thoseby Littrell et al. (1979), Stoltenberg (1981), Loganbill et al.(1982), and Blocher (1983). Hogan's developmental concep-

ELEABETH L. HOLI.OWAY received her PhD from the University of

Wisconsin-Madison in 1979. She is currently on the faculty of the

Division of Counseling and Educational Psychology at the University

of Oregon. She is a counseling psychologist, and her research areas

include counselor training, clinical supervision, and counseling proc-

ess. She is President-Elect of the Oregon Counselor Education and

Supervision Association and is on the Editorial Boards of The Coun-seling Psychologist, The Journal of Counseling Psychology, and Coun-

seling and Values.

THE AUTHOR THANKS Lloyd Lovell for his comments on an earlierversion of this article.

CORRESPONDtNCE CONCERNING THIS ARTICLE should be addressed to

Elizabeth L. Holloway. Division of Counseling and Educational

Psychology, University of Oregon. Eugene, Oregon 97403.

tion of supervision appeared in the clinical literature in 1964and has been referred to extensively in the more recentcounseling models. Developmental models for the group su-pervision of counselors have also appeared recently (Sansbury,1982; Yogev, 1982). Numerous empirical studies have beendesigned to investigate developmental paradigms (Roehlke,1984). Developmental models of supervision have becomethe Zeitgeist of supervision thinking and research. If thesemodels are to continue to influence research and practice, itseems warranted that the underlying assumptions and prin-ciples of the proposed models be critically examined at thisstage of their development. In this article, I engage in such acritical examination and discuss from a developmental per-spective the conceptual and methodological approaches cur-rently applied in the research of this topic.

Characteristics of the Developmental Models

Essentially, five developmental models of individual super-vision have been the focus of thinking and research in thecounseling literature; three of these models (Blocher, 1983;Loganbill et al., 1982; Stoltenberg, 1981) have explicitlylinked their origins to psychosocial developmental theory.These three models serve as the basis of the conceptualdiscussion. In the remaining two models, Hogan (1964) andLittrell et al. (1979) characterized the supervisory process,using the terminology and descriptors of development. How-ever, these models do not refer to a particular theory as theorigin of their work or to an underlying developmental processthat directs the hierarchical progression of clinical learning.Because no explicit reference to traditional developmentaltheory is claimed, these models are examined only as theyhave been used as a referent and/or basis for the Stoltenberg,Loganbill et al., and Blocher models.

Little attention has been directed at the unique underlyingassumptions of the Stoltenberg (1981), Loganbill et al. (1982),and Blocher (1983) paradigms. Because each appears distinctenough to warrant individual scrutiny, the fundamental de-velopmental characteristics of each of the three models aredescribed in the next sections.

209

210 ELIZABETH L. HOLLOWAY

Stollenberg's Model

Stoltenberg's (1981) model of counselor development isbased on Hogan's (1964) descriptions of trainee levels ofdevelopment and Hunt's (1975) application of ConceptualSystems Theory (CST; Harvey, Hunt, & Schroder, 1961) tothe teaching environment. Hunt and his associates extensivelyinvestigated the matching of certain learning characteristicsof adolescents with optimal learning environments in class-rooms. Stoltenberg extrapolated Hunt's person-environmentmatching scheme to the supervisory situation. He describedthe counselor's progress across four developmental levels fromneophyte to master counselor. He used dimensions of inter-personal perception, identity, motivational orientation, emo-tionality, and cognitive structural attributes to characterizeeach of these levels of development. Stoltenberg suggestedthat the supervision environment should be designed to op-timize the trainee's learning at each of the four levels oftrainee development. Characteristics of the supervision envi-ronment have been adapted from Hogan's work and indicatethe type of supervisor role and the instructional content to beemphasized at each level.

The emphasis of Stoltenberg's (1981) model on the system-atic matching of instructional environments in supervisionwith learner characteristics is a critical step in advancing theutility of supervision as an instructional strategy; however,the manner in which CST has been used to justify the coun-selor's learning as a developmental process is not consistentwith the theory's underlying constructs. A careful reading ofCST (Harvey et al., 1961) and Hunt's (1975) "matchingmodel" reveals significant discrepancies between this super-vision model and the developmental theory on which it isbased.

The primary construct in CST is the person organismicvariable named conceptual level (CL). CL is a cognitive vari-able that refers to the structural aspects of cognition and inparticular social cognition. CL is described by Hunt (1978) as"a personality characteristic that describes persons on a de-velopmental hierarchy of increasing conceptual complexity,self-responsibility and independence" (p. 78). Of these threecharacteristics, it is conceptual complexity that describes fun-damental perceptual processes of differentiation and integra-tion and determines acts of perceiving, thinking, and judging.The remaining two descriptors reflect a person's interpersonalbelief system. The developmental hierarchy of CST is char-acterized by discrete stages representing qualitative shifts inbehavior. Theoretically, CL is professed to be a global, orga-nismic variable of cognition, although the global nature ofthe construct has been questioned and is not yet empiricallyvalidated (Miller & Wilson, 1979). CL is the construct centralto the theory's structure; it is the governing principle ininformation processing.

The fundamental characteristics of CL have been somewhatdistorted in the Stoltenberg (1981) model. Stoltenberg's modelimplies that counselors in training recapitulate their devel-opment on CL; that is, they begin at low levels of CL uponentry to training and progress through the stages as they gainexperience. Although Stoltenberg stated that "No specific time

table of progress through the four levels has been presented,as this varies significantly from trainee to trainee" (p. 60), themodel describes a sequential and hierarchical developmentalframework that is driven by changes on the dimensions ofsocial cognition and information processing—in essence, CL.In this model there is not sufficient attention given to theinfluence of previously established cognitive structures on thetrainee's progression through the learning experience. Themodel suggests that most trainees will use elementary cogni-tive structures to cope with their new environment, whichthus implies that entry level into a program may result in adevelopmental regression on an organismic construct. Fur-thermore, it suggests that the experience of learning the skillsof counseling may affect developmental change in establishedcognitive structures. If CL is a central construct in personalitydevelopment, what renders the experience of being a coun-selor so powerful that a graduate student abandons previousacquired cognitive structures and resorts to elementary levelsof information processing on entry into counselor training?According to Hunt's (Hunt & Sullivan, 1974) theory, cogni-tively complex students would adapt to the new environmentreadily; different learning environments would be needed bystudents with different levels of CL. Thus the design oflearning environments would be driven by level of CL andnot by level of experience.

Regardless of the inconsistencies between the theory andmodel, CL and Hunt's (1975) matching model of learningmay still have merit in the design of counselor instructionand supervision programs. For instance, entry-level coun-selors could be assessed on CL, and instructional environ-ments could be matched to their cognitive structure in orderto maximize their acquisition of counselor skills. In numerousstudies researchers have investigated the relationship betweenCL and counseling-related tasks. In a recent meta-analysis ofstudies in which the matching model was investigated incounseling, Holloway and Wampold (1986) validated theoverall potency of the matching model in counselor educa-tion. If directors of training programs believe that a higherlevel of CL is either a necessary or a preferred style for thecounselor, then learning environments could be created tofacilitate a trainee's movement from lower to higher levels ofcognitive functioning. In these strategies, in contrast to theStoltenberg approach, CL, rather than the trainee's experiencelevel in the training program, would be targeted as the factorin determining instructional environments, or CL would betargeted for change in training counselors. Of course, theburden then would be to show how CL and trainee proficiencyor client outcome relate.

Loganbill, Hardy, and Delworth 's Model

The Loganbill et al. (1982) model is perhaps the mostcomprehensive developmental model of supervision. It isbased primarily on the theories and assumptions of develop-mental psychology, particularly those of Margaret Mahler,Erik Erikson, and Arthur Chickering. Four assumptions formthe basis for the model: (a) core concepts in developmentaltheory apply to learning to be a professional counselor insofar

IS IT DEVELOPMENT? 211

as the training experience is more than the incremental build-up of skills and is rather "the integrated formulation of atherapist with an identity" (Loganbill et al., 1982, p. 15); (b)there are distinct, sequential, hierarchical, and necessarystages in the development of the counselor Just as "infantslearn to crawl before they learn to walk, though the ultimategoal is not to crawl well. . . . So it is with the supervisee[S]ome of the stages and processes may be very painful, butit is developmentally important for the supervisee to experi-ence them fully" (p. 4); (c) although the stages are sequential,different learning tasks may be at different developmentallevels; and (d) to progress within and between stages, oneassumes "a careful sequence of experience and reflection" (p.15). The model does not just reflect development during thelife of a formal training program but is continuous throughoutone's professional life. To account for this ongoing process,Loganbill et al. suggested that the stage model is actually onein which a counselor "may cycle and recycle through thesevarious stages at increasingly deeper levels" (p. 17).

The model as described has characteristics similar to devel-opmental models of maturation (distinct, sequential, andhierarchical stages) and similar to pluralistic life-span devel-opment theories (continuous recycling of stages on varioustasks, not all of which may be executed at the same develop-mental level; Baltes, 1983). The reliance on both a tnatura-tional and pluralistic view of development to explain traineechange during graduate training has resulted in a ratherunparsimonious and contradictory explanation, creating asituation that forces the reader to question the integrity of thedevelopmental nature of the model.

Rest (1979), a leading theorist in moral development, de-scribed two conditions that refute the basic structure of adevelopmental construct: Individuals do not change, or indi-viduals change in every possible way, disregarding sequential-ly (or upward directionality). Rest, however, described threepermissible types of developmental paradigms: (a) a simplestage model, (b) a modified stage model, and (c) an atomisticmodel. Although there is some similarity between Loganbillet al.'s (1982) model and the first two of Rest's paradigms, itis the atomistic model that seems to be most applicable.

A simple stage model is one in which a particular type ofthinking dominates during a particular period of developmentand is then replaced in a series of stages by qualitativelydifferent systems. The modified stage system is one in whicha stage still represents a qualitatively different system, but anindividual is not assigned one-for-one to a stage. In this systemthe type of organizational system that a person might bringto a problem can be influenced by the content and particularqualities of the problem. However, there remains in this stagesystem an upward trend of increasingly more complexschemes being used more frequently. Change is representedby a shift in the distribution of the use of different types ofresponses. Loganbill et al.'s (1982) supervision model seemsto have characteristics of both of these paradigms. For ex-ample, they likened the counselor's developmental process toa maturational description of learning to walk and at the sametime stated that a person could be at different stages ondifferent tasks. This inherent contradiction makes the alter-

native atomistic model more plausible than the typologicalapproach of the simple or modified stage models.

The atomistic approach is a content-specific model in whichevidence of psychological unity between isolated, discon-nected concepts is sought. Cohesiveness among the conceptsmust be substantiated in order to justify the usefulness of adevelopmental framework over a descriptive approach. Lo-ganbill et al. (1982) proposed that there exists a logical inter-connectedness among content, task, and performance factorsof the trainee's learning. The cohesive force among the eventsand encounters of training is the struggle to form a profes-sional identity. The acceptance of the Loganbill et al. (1982)model as an atomistic developmental structure hinges on theacceptance that the professional identity process is sufficientlypowerful to trigger a developmental event. Although the taskof becoming a counselor is complex indeed, it is difficult tobe convinced that it is any different from learning the skillsof any other professional role.

In conclusion, the Loganbill et al.'s (1982) model is exceed-ingly complex and lacks the elegance and simplicity usuallyexpected of "truth." They adopted such a broad definition toaccount for the trainee's cognitive, affective, and social learn-ing behaviors that it endangers the usefulness of the develop-mental approach. For example, the model refers to threesequential stages and two transitional periods for which "eightkey issues in supervision have been identified.... [T]he su-pervisee is hypothesized to be in any one stage of functioningfor each of the eight issues at a given point in time" (p. 17).This means that a supervisee could be operating at 40 differentpoints of development at any given time. This is compoundedfurther by the model's claim that an individual may cycle andrecycle through this system on each of the eight key issues atdifferent times. The model's expansion of the developmentalparadigm has resulted in such specificity that any behavioralchange is equated with a developmental change. Thus thevalue of the developmental paradigm has been lost. Thisapproach to supervision might be described more accuratelyas an interesting and complex descriptive mechanism forordering some of the events of professional growth and is ananalogy of development rather than developmental per se.

Blocher's Cognitive Developmental Approach

Blocher's (1983) understanding of the supervision processis based on Dewey's (1916) approach to learning and cognitivedevelopmental theory and research related to "developmentalperson perception" (Blocher, 1983, p. 27), which focuses onthe influence of cognitive structures on social perception andjudgment. This approach represents a "constructivist" viewof human cognitive functioning as postulated by Kelly (1963),

Lewin (1935), and Werner (1978). Like Stoltenberg (1981),Blocher was interested in the creation of a supervisory learningenvironment that would optimize a trainee's learning; how-ever, Blocher applied the principles of cognitive developmentto supervision quite differently. Blocher used the knowledgein human cognitive development (including authors such asHarvey etal., 1961; Loevinger, 1976; Perry, 1968; and Piaget

212 ELIZABETH L. HOLLOWAY

& Inhelder, 1969) to postulate the growth and developmentin cognition available to a student of counseling. The devel-opment of the trainee is not arranged in a series of hierarchicalstages; rather, the individual's growth is seen as an idiosyn-cratic process determined by his or her own unique learningstyle and developmental history. In the model there is anassumption that there is a demand for highly complex func-tioning in the counseling situation. Thus the supervisor, whendesigning the learning environment, must focus on the ulti-mate goal of the trainee's acquisition of new, more complex,and more comprehensive schemata for understanding humaninteraction. The creation of optimal person-environment fitsbetween trainee needs and supervisory strategies results insystematic instructional learning. Blocher saw supervision as"psychological education in the fullest and most completesense of the term. It uses psychological content in a systematicway to change ihe psychological functioning of the learner"

(p. 28).At this point one may remark that some of these assump-

tions regarding the need for highly developed cognitive proc-esses and environmental match are very reminiscent of theStoltenberg (1981) model; however, a critical differenceshould also be evident. Blocher (1983) did not assume thatcounselor preparation would create a distinct set of develop-mental stages motivated by the type of task demanded of thelearner. In contrast, he stated:

The ultimate focus of the developmental supervisory model is,

however, on the acquisition of new more complex and more

comprehensive schemas for understanding human interaction.

These schemas in their most general sense are analogous to the

cognitive stage concepts in the Loevinger (1976) or Perry (1970)

developmental models. It should be noted that the goals involv-

ing cognitive change are not viewed narrowly. This kind ofgrowth in schemas for information-processing is viewed as a

centra! ingredient in all interpersonal functioning, (p. 30: em-

phasis added)

Thus Blocher differed from Stoltenberg in his use of cognitivedevelopmental principles as a basis for understanding learnerneeds and the cognitive demands of the counseling situationto devise supervisory instructional strategies that would en-hance learner development toward the application of process-ing schemata by the student. He suggested, not that suchschemata are content specific to counseling or that a uniqueset of developmental stages unfold in the counselor trainingprocess, but rather that counselor preparation and particularlysupervision can provide a context in which cognitive structurein a more global sense can be affected.

Research on the Developmental Models

Since the appearance of the developmental models of su-pervision, there have been numerous research projects inwhich researchers investigated the learning experience ofcounselor trainees. Because an extensive review of this empir-ical literature is presented elsewhere in this issue, I focus onlyon those studies directly relevant to the models reviewed inthe previous section and on the characteristics and results of

these scientific inquiries vis-a-vis a developmental perspective.Seven studies (Heppner & Roehlke, 1984; Hill, Charles, &Reed, 1981; Miars et al., 1983; Reising & Daniels, 1983;Wiley & Ray, 1986; Worthington, 1984; Worthington &Stern, 1985) appeared in the counseling psychology literature,and the researchers explicitly tested one or more of thedevelopmental models that I discuss. These studies are dis-cussed on three dimensions: (a) the experimental questions,(b) the methodological approaches, and (c) the conclusionsthat might be drawn from the results.

The Experimental Questions

Experimental hypotheses in general were focused onchanges that might occur in trainees' or supervisors' percep-tions of the supervisory experience during the trainees' pro-gression from a neophyte to master counselor. Two issues arenoteworthy in the characterization of the hypotheses and arediscussed in this section. First, experimental questions wereposed in order to contrast supervisory experiences amongtrainees at various levels of training. Second, confirmation ofthe developmental paradigm was typically an inferential, posthoc enterprise.

From a developmental perspective, the use of experimentalquestions that focus on differences at various experience levelsis problematic at this point in the scientific inquiry of coun-selor development. Although there are advantages in compar-ing clusters of naturally occurring and rationally chosengroups (usually second-year or predoctoral interns) to deter-mine whether such groups also represent qualitative differ-ences in behavior, there are also recognized hazards. Devel-opmentalists point out that the chosen groups do not neces-sarily represent the actual periods in which change occurs. Inother words, by structuring the hypotheses around nonran-domly chosen times at which one expects to find differences,one risks the possibility of ignoring points of change betweendata collection times. In finding differences the researcherperhaps erroneously concludes that behavior changes occurin a lock-step fashion that is concurrent with existing timestructures. This could be a problem in this body of researchfor two reasons.

First, research regarding characteristics of trainee develop-ment is in its infancy. There is no established literature tojustify the choice of particular points of data collection. Inthe majority of studies' hypotheses, trainees in beginningpractica, advanced practica, and internships are compared.These data collection times could be anywhere from 4 monthsto 5 years apart. The probability that significant trainingevents would occur between these times seems high. It wouldbe more revealing to randomly choose points in time to accessdata, and thus it would be more likely to describe the naturallyoccurring developmental progression. Researchers cannot as-sume that developmental change is necessarily linked to aca-demic programming. Reising and Daniels (1983) identifiedthis problem and designed a study in which they first lookedfor factors that might distinguish among trainees and thendetermined these factors" relationship to different experiencelevels. More studies in which this approach is used are needed

IS IT DEVELOPMENT? 213

to test the underlying developmental structure of thesemodels.

The second problem with investigations of developmentalprogression is that only two studies have been devised to testthe characteristics of a specific developmental model: Reisingand Daniels (1983) tested Hogan's (1964) paradigm, andWiley and Ray (1986) tested Stoltenberg's (1981) model. Inmost cases, experimental questions have tested the influenceof experience on trainees' perceptions of supervision. Anyevidence of a developmental structure has been inferred byresearchers looking for similarities between trainees' charac-teristics and those qualities described in the various models.Consequently, a post hoc inferential process has been usedprimarily in judging the validity of the developmental process.In future research, experimental questions must relate moreexplicitly to the models presented in the literature if investi-gators are interested in remarking on their validity.

Methodology

The methodology in the studies under discussion heretended not to be developmentaUy sensitive. Three factorsaccount for this situation: (a) traditional cross-sectional de-signs, (b) types of measurement devices, and (c) sources ofdata collection.

Cross-sectional designs. At present, the most obviousproblem in supervision research is the absence of longitudinaldata to investigate developmental change. Hill et al.'s (1981)study is the only investigation in which a longitudinal ap-proach was used, and they asked only as an exploratory partof the project that students report what changes in counselingskills and what personal changes they experienced duringgraduate training. The necessity of longitudinal data in study-ing an adult behavioral change was explained aptly by Baltes(1983):

Individuals live in a changing biocuitural context,... the expla-

nation of long-term processes is apt to involve complex historical

paradigms, and . . . long-term processes are likely candidates for

explanatory discontinuity rather than simple cumulative causal

explanations.... (p. 101}

The current reliance on cross-sectional data is not an ade-quate approach to the developmental issue for the followingreasons. First, researchers have not gained access to the his-torical and cultural development of the trainee; therefore, theeffect of existing personality structures such as ego, cognitive,and moral levels cannot be used to understand contemporaryobservations. Reising and Daniels (1983) concluded in theirstudy that assessment of basic personality structures are anecessary step in furthering the understanding of the trainee'sexperience. Second, the lack of information on intraindividualchanges across the course of a training program seriouslyweakens a developmental explanation of behavior changebecause any observed group change does not necessarily rep-resent a behavior change for a particular individual nor doesit describe the particular pattern of change experienced by anindividual. Third, long-term processes are involved in adultdevelopment, and empirical investigations have typically been

limited to approximately 4-year time spans or that time duringwhich the training program starts and finishes. In using cross-sectional data, one assumes that there are no selection biasesin sampling, no differences between cohort groups enteringgraduate programs at different times, and no historical effectsthat may affect groups differentially. These assumptions can-not be made within the context of graduate training programs.Clearly there are changes in curriculum, instructional person-nel and perhaps theoretical orientation within the same pro-gram across several years.

Longitudinal studies would mitigate some of these prob-lems, but unfortunately they include their own deficits suchas testing effects, sampling bias, and generation effects. Moreinnovative methodological strategies that combine cross-sec-tional and longitudinal techniques have been proposed in thedevelopmental psychology literature (Achenbach, 1978).These techniques include longitudinal sequential, cross-sec-tional sequential, and time-lag sequential designs. The detailsof these designs are available in other sources (e.g., Achen-bach, 1978). and it is enough to say here that each strategyhelps to control for effects that are attributable to cohort,point of measurement, and date of measurement in differentmanners; each strategy has its own methodological advantagesdepending on the experimental question being posed.

Methods of measurement. Methods of measurement mayalso be jeopardizing a developmental understanding of thesupervision phenomenon. In these studies, with the exceptionof Hill et al. (1981), measurement involves the use of struc-tured self-report questionnaires to assess trainee and supervi-sor perceptions. The use of self-report techniques is a standardpractice in developmental research; however, in initial inves-tigations, preferred methods include (in increasing order ofpreference) open-ended questionnaires, structured and semi-structured clinical interviews, and direct observation (Rest,1979). The use of structuredquestionnaires before more open-ended techniques unnecessarily biases the results in favor ofthe experimenters' understanding of what they should findand emphasizes the comparison of subjects on quantitativechanges.

Sources of data. Sources of data collection are a criticalfactor in the design of these supervision studies.' Investiga-tions have been restricted to the context of the supervisoryrelationship. Specifically, researchers have examined trainees*perception of their needs in supervision (Heppner & Roehlke.1984; Hill etal., 1981: Reising & Daniels, 1983: Worthington.1984; Worthington & Stern. 1985), supervisors' perceptionsof their supervisory practice with different levels of trainees(Miarset al,, 1983: Wiley & Ray, 1986; Worthington & Stern,!985), trainees'judgments of supervisors" behaviors in super-vision (Heppner & Roehlke. 1984; Worthington, 1984; Wor-thington & Stern, 1985), and supervisors'judgments of train-ees' behaviors (Wiley & Ray, 1986; Worthington & Stern,1985). From these vantage points they attempt to understandboth issues of the developmental model: the trainee's changingand the supervisor's changing supervision strategies to accom-

1 Worthington (1987) presents a complete review of study charac-teristics in Table 2.

214 ELIZABETH L. HOLLOWAY

modate the trainee. However, an examination of the sourceof these perceptions reveals that they are relevant primarilyto the supervision relationship. Trainees and supervisors areasked to remark on events in supervision that reflect emo-tional, relationship, and professional needs. An assumptionundergirding this approach to data collection is that thesupervisory event captures the essence of the trainee's devel-opment and is the best source of information on this phenom-enon. Although the supervisor's observation of the traineemay be valuable as a reflection of his or her own behaviorand judgment of the trainee in supervision, it is confoundedby the supervisor's own experience of the relationship. Simi-larly the trainee's judgments of his or her own training needsare influenced by the relationship. Investigators need to tapthe trainee's changes as a counselor more directly, but not by

assuming that changes as a supervisee necessarily representthose changes as a counselor. If the course of the supervisoryrelationship is the object of study, then current sources ofinformation may be relevant and informative. However, be-cause it is the trainee's development as a counselor that is theprofessed target of interest, contexts of counselor behaviormust also be assessed.

Trainees' perceptions of supervisors' behaviors and char-acteristics have also been studied. These can be seen as anindirect reflection of the trainee's developmental stage as wellas more obviously a comment on the supervisor's actualchange. In the former case it is inferred that trainees willchange their judgments of different supervisory behaviors inaccordance with their developmental level. Clearly, one'sperceptions do change through growth; however, with currentresearch designs, trainees' perceptions are confounded bysupervisors' characteristics and actual supervisors' behaviors.As a consequence of confounding trainee characteristics andqualities of the relationship, results cannot be used to confirmdevelopmental change of the trainee. In addition, predomi-nant use of self-report strategies that rely on perceptions ofself must be augmented with actual observations of supervi-sory and counseling processes.

Results

The overall results of the studies examined are relativelyconsistent and give marginal support for a developmentalmodel. However, because of the previously discussed meth-odological characteristics of these studies, one must exercisecaution in drawing conclusions regarding the developmentalimplications of the results.

The predominant finding regarding trainee characteristicsacross various levels of trainee experience indicate that differ-ences exist only between very beginning-level and intern-leveltrainees and that these differences center on relationshipcharacteristics (Heppner & Roehlke, 1984; Miarset al., 1983;Reising & Daniels, 1983; Wiley & Ray, 1986; Worthington,1984; Worthington & Stern, 1985). For example, initial-leveltrainees appear to require more support and encouragementin supervision, whereas interns demonstrate increasing inde-pendence from the supervisor (Hill et al., 1981; Reising &Daniels, 1983: Wiley & Ray, 1986; Worthington, 1984; Wor-

thington & Stem, 1985) and more pronounced interest inhigher level skills and personal issues affecting counseling(Heppner & Roehlke, 1984; Hill et al., 1981; Worthington &Stern, 1985). Certainly these results do not provide confir-mation of a sequential change process distinguished by aseries of hierarchical stages.

Evidence of supervisory change across experience levelsparallels findings regarding trainee change. In the three studiesin which researchers examined supervisors' perceptions oftheir behaviors with trainees of different experience levels,only distinctions between entry-level trainees and interns werefound (Miars et al., 1983; Wiley & Ray, 1986; Worthington& Stern, 1985). Supervisors reported that they recognizedmore variable need during the relationship with beginning-level students, whereas interns' supervisory relationships weremore stable (Worthington & Stern, 1985). Supervisors sawthemselves providing different learning environments for en-try-level and intern-level trainees (Miars et al., 1983; Wiley &Ray, 1986).

In summary, the results regarding trainee and supervisorychange across levels of experience were disappointing from adevelopmental perspective. The empirical discovery that moresupport may be needed early on in training and that moresophisticated skills are the focus in later learning is easilyunderstood from a task mastery perspective. Goulet's (1973)comment regarding learning progression is an example of thisperspective:

[V]ery little new information is actually acquired in a learning

experiment since subjects usually have the basic requisites of task

mastery already available in their repertoire of habits and

skills.... [L]earning not to respond in a particular fashion is

likely just as responsible for performance changes with practice

as the two mechanisms which are most commonly identified

with the phenomenon of learning, i.e., the acquisition of a

specific habit or skill, and the selection or use of an appropriate

skill or strategy available in the subject's behavioral repertoire,(pp. 281-282)"

The question at issue is whether elaborate developmentalmodels are needed in order to explain the research findings.

Noteworthy in these studies was the discovery of moderat-ing variables that have exhibited a relationship with supervi-sory experience. Wiley and Ray (1986) and Reising andDaniels (1983) found that only supervised practica experienceand not nonsupervised counseling experience correlated withdifferences in supervisory needs of trainees. Hill et al. (1981),in their examination of students' acquisition of counselingskills, concluded that previous counseling experience did notaccelerate students' clinical progression in the training pro-gram. These findings imply that changes in the trainee dependmore on the particular context of the supervisory relationshipthan on the actual process of establishing a professionalidentity. Additional variables of interest are Worthington's(1984) discovery that supervisory differences across training

levels varied with geographical location, and Worthington's(1984) and Reising and Daniel's (1983) findings that theoret-ical orientation correlated with the supervisor's report ofbehaviors. These results suggest that a general model of trainee

IS IT DEVELOPMENT? 215

development may not be governing the supervisory experi-ence but rather that other aspects of training such as philos-ophy of training programs and characteristics of supervisorsmay be the most influential factors on trainees' learningexperiences.

Although these researchers are interpreting their results astentatively supporting a developmental model, lack of devel-opmental-specific methodology, confinement to the supervi-sory experience as a source of information, predominant useof structured self-report questionnaires, and lack of evidenceof distinct, sequential stages in trainees' growth reflect theprematurity of such claims.

Conclusions and Alternative Explanations

Any developmental theory must not only describe but alsoexplain. Although the developmental models of supervisionprovide a framework with which to describe some of thebehaviors typically observed by clinical supervisors, they donot provide an adequate developmental explanation from thestandpoint of developmental psychology. Critical to each ofthese models is the context of the supervisory relationship asthe primary environmental arena in which the developmentof the trainee is evidenced. Although the power and signifi-cance of supervision should not be underemphasized in thetraining program, supervision is only one component of acomprehensive training approach, and indeed the trainingprogram is only one aspect of the student's life. How are thedevelopmental shifts that are postulated in these models evi-denced (a) outside of the supervisee role, (b) in the counselorrole, (c) in the student role, and (d) in other roles of theperson's life? If indeed these changes are seen as qualitativeshifts central to a counselor's growing professional identity,then they must also be influencing other contexts of life. Infact, it has been frequently argued in counseling approachesthat the professional identity is a part of and integrated intothe personal identity, insofar as without such congruencybetween professional behavior and self, the counselor lacksauthenticity and consequently potency in the counseling re-lationship. Thus professional identity cannot be held separatefrom other areas of one's self, and changes must be consideredoutside of the supervisory situation in order to validate thedevelopmental nature of the trainee's learning process. Nei-ther the models nor the research sufficiently addresses thisconsideration and thus has not yet dealt with the underlyingconstructs of a developmental model.

Although it is not surprising that researchers in the fieldhave chosen to conceptualize counselor training from a de-velopmental perspective (after all, most psychologists are ed-ucated to think in terms of personality structure and change),there are some equally appealing and more heuristic ways inwhich to approach the understanding of the trainee's learningexperience other than developmental paradigms. I entertaina few of these ideas, not because I am convinced that devel-opmental modeling in supervision is misguided, but ratherbecause it is so intuitively attractive that I fear it will not besufficiently challenged before psychologists are convinced ofits validity. As Hess (in press) commented in a recent paperon supervisee growth, "these (developmental) schemas are

heuristic devices to help understand the student but may bea function of shared fictions by the authors."

Alternative explanations for trainee change may be that thesupervisory relationship itself creates a trainee's initial vulner-ability and final independence. In other words, the trainee'sfeelings are not intrinsic to becoming a counselor or establish-ing a professional identity but are a result of being in anintensive, evaluative, ongoing, and demanding relationship.In most formal relationships, particularly of inequitablepower, there are feelings of vulnerability and requests forspecificity of role expectations expressed by the subordinatepartner. As trainees become more accustomed to the role ofsupervisee and as they progress toward learning a profession,they naturally become more confident and focus on the tasksof counseling rather than on the degree of support in thesupervisory relationship. Worthington and Stern's (1985) re-search on the supervisory relationship at different levels oftraining could be very adequately interpreted from this frame-work.

Another approach to understanding current observationsof a trainee's learning experience is an instructional one. Thepsychiatric literature on supervision has referred to the resi-dent's clinical growth in terms of a learning experience (Bar-nat, 1974; Cohen & DeBetz, 1977; Fleming, 1953; Fleming& Benedek, 1964; Gaoni & Neumann, 1974; Schlessinger,1966). In addition, the literature in counseling psychology hasrecently promoted the instructional mission and characteris-tics of the supervisory relationship (Abbey, Hunt, & Weiser,1985; Friedlander& Ward, 1984;Goodyearetal., 1983; Hess,1980; Holloway, 1982: Holloway & Wampold, 1983): thusthere is conceptual and empirical support for a pedagogicalunderstanding of the student's clinical training. Knowledgein educational psychology regarding instructional strategiesbased on characteristics of the learning task, student's level ofskill acquisition, and student's learning style all seem to havedirect relevance to issues dealt with in the tutorial of super-vision. Abbey et al.'s (1985) use of Kolb's (1984) experientiallearning cycle is an excellent example of the application ofinstructional models to counseling and supervision.

It is not clear that what occurs in clinical training is betterdescribed by various developmental models than by suchalternative conceptions as those used in learning theory andother models to explain the acquisition of other cognitive andaffective skills. It remains incumbent upon developmentaliststo substantiate that a structural, qualitative, and predictablechange occurs as a result of training to be a counselor.

References

Abbey, D. S., Hunt, D. E., & Weiser, J. C. (1985). Variations on a

theme by Koto: A new perspective for understanding counseling

and supervision. The Counseling Psychologist, 13, 477-501.

Achenbach, T. M. (1978). Research in developmental psychology:

Concepts, strategies, methods. New York: Free Press.

Bakes, P. B. (1983). Life-span developmental psychology: Observa-

tions on history and theory revisited. In R. M. Lerner (Ed.),

Developmental psychology: Historical and philosophical perspec-tives (pp. 79-110)! Hillsdale, NJ: Erlbaum.

216 ELIZABETH L. HOLLOWAY

Barnat, M. R. (1974) Some characteristics of supervisory identifica-tion in psychotherapy. Psychotherapy: Theory, Research and Prac-

tice, 11, 189-192.Blocher. D. H. (1983). Toward a cognitive developmental approach

to counseling supervision. The Counseling Psychologist, 11(1), 27-34.

Cohen, R. J.. & DeBetz. B. (1977). Responsive supervision of the

psychiatric resident and clinical psychology intern. American Jour-

nal of Psychoanalysis, 37, 51-64.

Dewey, J. (1916). Democracy and education. New York: Macmillan.

Fleming, J. (1953). The role of supervision in psychiatric training.

Bulletin of the Menninger Clinic. 17, 157-169.Fleming, J., & Benedek, T. (1964). Supervision: A method of teaching

psychoanalysis. Psychoanalytic Quarterly, 33, 71-96.

Friedlander, M. L., & Ward, G. W. (1984). Development and vali-

dation of the supervisory styles inventory. Journal of Counseling

Psychology, 31, 541-557.

Gaoni, B., & Neumann, M. (1974). Supervision from the point of

view of the supervisee. American Journal of Psychotherapy, 23,

108-114.

Goodyear, R. K., Bradley. F. O., & Bartlett, W. E. (1983). An

introduction to theories of counselor supervision. The Counseling

Psychologist. 11(1), 19-20.

Goulet. L. R. (1973). The interfaces of acquisition: Models and

methods for studying the active, developing organism. In J. R.

Nesselroadc & H. W. Reese (Eds.), Life-span developmental psy-

chology: Methodological issues (pp. 281-298). New York: Aca-

demic Press.

Harvey, O. J., Hunt, D. E., & Schroder, H. M. (1961). Conceptual

systems and personality organization. New York: Wiley.

Heppncr, P. P., & Roehlke, H. J. (1984). Differences among super-

visees at different levels of training: Implications for a develop-

mental model of supervision. Journal of Counseling Psychology,30. 252-262.

Hess, A. K. (Ed.) (1980). Psychotherapy supervision: Theorv. research,

and practice. New York: Wiley.

Hess, A. K. (in press). Growth in supervision. Stages of supervisee

and supervisor development. I'he Clinical Supervisor.

Hill. C. E.. Charles. D., & Reed, K. G. (1981). A longitudinal analysisof changes in counseling skills during doctoral training in counsel-

ing psychology. Journal of Counseling Psychology, 28, 428-436.

Hogan, R. A. (1964). Issues and approaches in supervision. Psycho-

therapy: Theory, Research and Practice, 1, 139-141.

Holloway, E. L. (1982). Interactional structure of the supervision

interview. Journal of Counseling Psychology, 29, 309-317.

Holloway. E. L. (1984). Outcome evaluation in supervision research.

The Counseling Psychologist, 12(\), 167-174.

Holloway, E. L., & Hosford, R. E. (1983). Towards developing a

prescriptive technology of counselor supervision. The Counseling

Psychologist. 11(1), 73-77.

Holloway, E. I,., & Wampold. B. E. (1983). Patterns of verbal

behavior and judgments of satisfaction in the supervision interview.

Journal of Counseling Psychology, 30. 227-234.

Holloway, E. L., & Wampold, B. E. (1986). The relation between

conceptual level and counseling-related tasks: A meta-analysis.

Journal of Counseling Psychology, 33, 310-319.Hunt. D. E. (1975). The B-P-E paradigm for theory, research and

practice. Canadian Psychological Review. 16, 185-197.

Hunt, D. E. (1978). Theorists are persons, too: On preaching what

you practice. In C. Parker (Ed.), Encouraging student development

in college. Minneapolis: University of Minnesota Press.

Hunt, D. E., & Sullivan. E. V. (1974). Between psychology and

education. Hinsdale, IL: Dryden.

Kelly, G. (1963). A theory of personality. New York: Norton.

Kolb, D. A. (1984). Experiential learning. Englewood Cliffs, NJ:

Prentice-Hall.

Lambert, M. J. (1974). Supervisory and counseling process: A com-parative study. Counselor Education and Supervision, 14, 54-60.

Lambert, M. J. (1980). Research and the supervisory process. In

A. K. Hess (Ed.), Psychotherapy supervision: Theory, research, and

practice (pp. 423-450). New York: Wiley.Lewin, K. (1935). A dynamic theory of personality. New York:

McGraw-Hill.

Littrell, J. M., Lee-Borden. N., & Lorenz, J. (1979). Developmentalframework or counseling supervision. Counselor Education and

Supervision, 19, 129-136.

Loevinger, J. (1976). Ego development: Conceptions and theories. San

Francisco: Jossey-Bass.

Loganbill, C., Hardy, E., & Delworth, U. (1982). Supervision: A

conceptual model. The Counseling Psychologist, 10(1), 3-42.

Miars. R. D., Tracey, T. J., Ray, P. B., Cornfeld, J. L., OTarrell, M.,

& Gelso. C. J. (1983). Variation in supervision process across

trainee experience levels. Journal of Counseling Psychology, 30,

403-412.

Miller, A., & Wilson, P. (1979). Cognitive differentiation and integra-

tion. A conceptual analysis. Genetic Psychology Monographs, 99,

3-40.

Perry, W. G. (1968). forms of intellectual and ethical development inthe college years: A scheme. New York: Holt, Rinehart &Winston.

Piaget, J., & Inhelder, B. (1969). The psychology of the child. New

York: Basic Books.

Rest. J. R. (1979). Development in judging moral issues Minneapolis:

University of Minnesota Press.

Reising, G. N., & Daniels, M. H. (1983). A study of Hogan's model

of counselor development and supervision. Journal of Counseling

Psychology, 30, 235-244.

Roehlke, H. J. (1984). Research on levels of trainee experience. Paperpresented at the annual meeting of the American Psychological

Association, Toronto, Canada.

Sansbury, D. L. (1982). Developmental supervision from a skills

perspective. The Counseling Psychologist, 10(\), 53-58.

Schlessinger, N. (1966). Supervision in psychotherapy. Archives of

General Psychiatry, 15, 129-134.

Stoltenberg, C. (1981). Approaching supervision from a develop-

mental perspective: The counselor complexity model. Journal of

Counseling Psychology, 28, 59-65.

Werner, H. (1978). Developmental processes: Heinz Werner's selected

writings (S. S. Barlen & M. B. Franklin. Eds.). New York: Inter-

national Universities Press.

Wiley. M. O'L.. & Ray, P. B. (1986). Counseling supervision by

developmental level. Journal of Counseling Psychology, 33, 439-

445.Worthington, E. L. (1984). An empirical investigation of supervision

of counselors as they gain experience. Journal of Counseling Psy-chology, 31, 63-75.

Worthington, E. L. (1987). Changes in supervision as counselors and

supervisors gain experience: A review. Professional Psychology:

Research and Praclice, IS, 189-208.

Worthington, E. L., & Stern, A. (1985). The effects of supervisor and

supervisee degree level and gender on the supervisory relationship.

Journal of Counseling Psychology, 32, 252-262.

Yogev. S. (1.982). An eclectic model of supervision: A developmental

sequence for beginning psychotherapy students. Professional Psy-

chology. 13. 236-243.

Received March 24, 1986Accepted July 2, 1986