the two disciplines of scientific...

THE TWO DISCIPLINES OF SCIENTIFICPSYCHOLOGY1

LEE J. CRONBACH

University of Illinois

NO man can be acquainted with all of psy-chology today, as our convention programproves. The scene resembles that of a cir-

cus, but a circus grander and more bustling thanany Barnum ever envisioned—a veritable week-longdiet of excitement and pink lemonade. Three daysof smartly paced performance are required just todisplay the new tricks the animal trainers havetaught their charges. We admire the agile paper-readers swinging high above us in the theoreticalblue, saved from disaster by only a few gossamerthreads of fact, and we gasp as one symposiastthrusts his head bravely between another's sharptoothed jaws. This 18-ring display of energies andtalents gives plentiful evidence that psychology isgoing places. But whither?

In the simpler days of psychology, the presiden-tial address provided a summing-up and a state-ment of destination. The President called the rollof the branches of psychology—praising the growthof some youngsters, tut-tutting patriarchally overthe delinquent tendencies of others—and showedeach to his proper place at the family table. Myown title is reminiscent of those grand surveys, butthe last speaker who could securely bring the wholeof psychology within one perspective was Dashiell,with his 1938 address on "Rapprochements in Con-temporary Psychology" (15). My scope must befar more restricted.

I shall discuss the past and future place withinpsychology of two historic streams of method,thought, and affiliation which run through the lastcentury of our science. One stream is experimen-tal psychology; the other, correlational psychology.Dashiell optimistically forecast a confluence of thesetwo streams, but that confluence is still in the mak-ing. Psychology continues to this day to be lim-ited by the dedication of its investigators to one orthe other method of inquiry rather than to scien-tific psychology as a whole.

1 Address of the President at the Sixty-Fifth Annual Con-vention of the American Psychological Association, NewYork, New York, September 2, 1957.

A stream of thought is identified by many fea-tures: philosophical underpinnings, methods of in-quiry, topical interests, and loci of application. Theexperimental and correlational streams have allthese aspects, but I am concerned with them asdisciplines within scientific psychology. The job ofscience is to ask questions of Nature. A disciplineis a method of asking questions and of testing an-swers to determine whether they are sound. Sci-entific psychology is still young, and there is rapidturnover in our interests, our experimental appa-ratus and our tests, and our theoretical concepts.But our methods of inquiry have become increas-ingly stable, and it is these methods which qualifyus as scientists rather than philosophers or artists.

THE SEPARATION OF THE DISCIPLINES

The experimental method—where the scientistchanges conditions in order to observe their conse-quences—is much the more coherent of our twodisciplines. Everyone knows what experimentalpsychology is and who the experimental psycholo-gists are. Correlational psychology, though fullyas old as experimentation, was slower to mature.It qualifies equally as a discipline, however, be-cause it asks a distinctive type of question and hastechnical methods of examining whether the ques-tion has been properly put and the data properlyinterpreted.

In contrast to the Tight Little Island of the ex-perimental discipline, correlational psychology is asort of Holy Roman Empire whose citizens identifymainly with their own principalities. The disci-pline, the common service in which the principali-ties are united, is the study of correlations pre-sented by Nature. While the experimenter is in-terested only in the variation he himself creates,the correlator finds his interest in the already exist-ing variation between individuals, social groups,and species. By "correlational psychology" I donot refer to studies which rely on one statisticalprocedure. Factor analysis is correlational, to besure, but so is the study of Ford and Beach (23)

671

672 THE AMERICAN PSYCHOLOGIST

relating sexual behavior to differences along thephylogenetic scale and across the cultural spectrum.

The well-known virtue of the experimental methodis that it brings situational variables under tightcontrol. It thus permits rigorous tests of hypothe-ses and confident statements about causation. Thecorrelational method, for its part, can study whatman has not learned to control or can never hope tocontrol. Nature has been experimenting since thebeginning of time, with a boldness and complexityfar beyond the resources of science. The correla-tor's mission is to observe and organize the datafrom Nature's experiments. As a minimum out-come, such correlations improve immediate decisionsand guide experimentation. At the best, a Newton,a Lyell, or a Darwin can align the correlations intoa substantial theory.

During our century of scientific psychology, thecorrelators have marched under many flags. In per-haps the first modern discussion of scientific methodin psychology (1874), Wundt (54) showed how"experimental psychology" and "ethnic psychology"(i.e., cross-cultural correlations) supplement eachother. In one of the most recent (19S3), Bindraand Scheier (4) speak of the interplay of "experi-mental" and "psychometric" method. At the turnof the century, the brand names were "experi-mental" and "genetic" psychology, although ex-perimenters were also beginning to contrast their"general psychology" with the "individual psychol-ogy" of Stern and Binet.

In 1913, Yerkes made the fundamental point thatall the correlational psychologies are one. His namefor this branch was "comparative psychology."

Although comparative psychology in its completeness nec-essarily deals with the materials of the psychology of infant,child, adult, whether the being be human or infra-human;of animal or plant [!]—of normal and abnormal individ-uals; of social groups and of civilizations, there is no reasonwhy specialists in the use of the comparative method shouldnot be so distinguished, and, if it seems necessary, labelled(55).

Even in advocating research on animals (56),Yerkes is emphatic in defining the goal as cor-relation across species. In France, la psychologiccompare continues to include all of differentialpsychology; but in America, as Beach (2) haslamented, comparative psychology degenerated intothe experimental psychology of the white rat andthereby lost the power of the correlational disci-pline.

Except for the defection of animal psycholo-gists, the correlational psychologists have remainedloosely federated. Developmental psychologists,personality psychologists, and differential psycholo-gists have been well acquainted both personally andintellectually. They study the same courses, theydraw on the same literature, they join the samedivisions of APA.

Experimental and correlational psychologists,however, grew far apart in their training and inter-ests. It is now commonplace for a student to gethis PhD in experimental psychology without gradu-ate training in test theory or developmental psy-chology, and the student of correlational branchescan avoid experimental psychology only a little lesscompletely. The journals of one discipline havesmall influence on the journals of the other (14).Boring even dares to say (5, p. 578) that there isa personality difference between the fields: the dis-tinction being that correlational psychologists likepeople!

Certainly the scientific values of psychologistsare sharply divided. Thorndike (9, 44) recentlyasked American psychologists to rate various his-toric personages by indicating, on a forced-choicequestionnaire, which have made the greatest con-tributions to psychology. A factor analysis of theratings shows two distinct factors (Figure 1). Onebipolar factor (irrelevant to our present discussion)

LABORATORYDATA

/tab. EnpeA

QUANTI-TATIVE

/Differ- \( enllol +}

>VERBAL

Global. Theory ,

FIELD DATA

FIG. 1. Factors accounting for esteem of leaders in psy-chology by American psychologists (based on correlationspresented by Thorndike, 44, corrected for attenuation andrefactored).

THE Two DISCIPLINES OF SCIENTIFIC PSYCHOLOGY 673

ranges from verbal to quantitative psychologists.The other factor has at one pole the laboratory ex-perimenters like Stevens, Dodge, and Ebbinghaus,and at the opposite pole those like Binet, May, andGoodenough who collect and correlate field data. Apsychologist's esteem for the experimenters is cor-related — .80 ( — 1.00, corrected for attenuation)with his esteem for scientists who use correlationalmethods.

There was no such schism in 1913 when Yerkesstated the program of correlational psychology.Genetic psychology and experimental psychologywere hard at work on the same problems. Termandemonstrated in his 1923 presidential address (43)that the mental test was within the tradition ofexperimental, fundamental research in psychology,and had quotations to show that the contemporaryexperimentalists agreed with him. Wells and God-dard, in 1913, had been asked to lecture on mentaltests within the Holy Temple itself, the Society ofExperimental Psychologists. And, in 1910, theHigh Priest Titchener had said:

Individual psychology is one of the chief witnesses to thevalue of experiment. It furnishes the key to many, other-wise inexplicable differences of result, and it promises toallay many of the outstanding controversies. . . . There canbe no doubt that it will play a part of steadily increasingimportance (46).

But when Terman spoke in 1923, the commonfront had already been fatally breached. Watsonhad announced that experimental treatment couldmake and unmake individual differences at will,thus stripping them of scientific importance. Thur-stone had taken the first firm stride in the oppositedirection:

I suggest that we dethrone the stimulus. He is only nomi-nally the ruler of psychology. The real ruler of the domainwhich psychology studies is the individual and his motives,desires, wants, ambitions, cravings, aspirations. The stimu-lus is merely the more or less accidental fact . . . (45,p. 364).

The personality, social, and child psychologists wentone way; the perception and learning psychologistswent the other; and the country between turnedinto desert.

During the estrangement of correlational and ex-perimental psychology, antagonism has been no-tably absent. Disparagement has been pretty wellconfined to playful remarks like Cattell's accusa-tion that the experimental psychologist's "regardfor the body of nature becomes that of the anato-

mist rather than that of the lover" (7, p. 1S2), orthe experimentalist Bartlett's (1, p. 210) satire onthe testers emerging from World War I, "chantingin unaccustomed harmony the words of the oldjingle

'God has a plan for every manAnd He has one for you.' "

Most correlationists have done a little experi-menting in the narrow sense, and experimentershave contributed proudly to testing work under war-time necessity. But these are temporary sojournsin a foreign land. (For clear expressions of thisattitude, see 5, pp. S70-S78 and 52, p. 24.)

A true federation of the disciplines is required.Kept independent, they can give only wrong an-swers or no answers at all regarding certain im-portant problems. It is shortsighted to argue forone science to discover the general laws of mind orbehavior and for a separate enterprisewith individual minds, or for a one-wayence of personality theory upon learnin

oncerneddepend-theory.

Consider the physical sciences as a parallel. Phys-ics for centuries was the study of general laws ap-plying to all solids or all gases, whereas alchemyand chemistry studied the properties and reactionsof individual substances. Chemistry was once onlya descriptive catalogue of substances andtechniques. It became a systematic science whenorganized quantitative studies yielded principles toexplain differences between substances ani to pre-dict the outcomes of reactions. In consequence,Mendeleev the chemist paved the way forphysicist, and Fermi's physics contributes

analytic

Bohr theto Law-

hemistryrence's chemistry; the boundary betweenand physics has become almost invisible.

The tide of separation in psychology ha|s alreadyturned. The perceiver has reappeared in perceptualpsychology. Tested intelligence and anxiey appearas independent variables in many of tb: currentlearning experiments. Factor analytic studies havegained a fresh vitality from crossbreeding with clas-sical learning experiments (e.g., 18, 22). [ Harlow,Hebb, Hess, and others are creating a truly ex-perimental psychology of development. And stu-dents of personality have been designing subtle com-binations of experimental and correlations! method(see, for example, 29) which may ultimately proveto be our parallel to the emergence of physicalchemistry.


CHARACTERIZATION OF THE DISCIPLINES

In the beginning, experimental psychology was asubstitute for purely naturalistic observation ofman-in-habitat. The experimenter placed man inan artificial, simplified environment and made quan-titative observations of his performance. The ini-tial problem was one of describing accurately whatman felt, thought, or did in a defined situation.Standardization of tasks and conditions was re-quired to get reproducible descriptions. All ex-perimental procedures were tests, all tests were ex-periments. Kraepelin's continuous-work procedureserved equally the general study of fatigue and thediagnosis of individuals. Reaction time was impor-tant equally to Wundt and to Cattell.

The distinctive characteristic of modern experi-mentation, the statistical comparison of treatments,appeared only around 1900 in such studies as thatof Thorndike and Woodworth on transfer. Theexperimenter, following the path of Ebbinghaus,shifted from measurement of the average mind tomeasuring the effect of environmental change uponsuccess in a task (51). Inference replaced estima-tion: the mean and its probable error gave way tothe critical ratio. The standardized conditions andthe standardized instruments remained, but thefocus shifted to the single manipulated variable,and later, following Fisher, to multivariate manipu-lation. The experiment thus came to be concernedwith between-treatments variance. I use the word"treatment" in a general sense; educational andtherapeutic treatments are but one type. Treat-ment differences are equally involved in comparingrats given different schedules of reinforcement,chicks who have worn different distorting lenses,or social groups arranged with different communi-cation networks.

The second great development in American ex-perimental psychology has been its concern withformal theory. At the turn of the century, theoryranged far ahead of experiment and made no de-mand that propositions be testable. Experiment,for its part, was willing to observe any phenomenon,whether or not the data bore on theoretical issues.Today, the majority of experimenters derive theirhypotheses explicitly from theoretical premises andtry to nail their results into a theoretical structure.This deductive style has its undeniable defects, butone can not question the net gains from the accom-panying theoretical sophistication. Discussions of

the logic of operationism, intervening variables, andmathematical models have sharpened both the for-mulation of hypotheses and the interpretation ofresults.

Individual differences have been an annoyancerather than a challenge to the experimenter. Hisgoal is to control behavior, and variation withintreatments is proof that he has not succeeded. In-dividual variation is cast into that outer darknessknown as "error variance." For reasons both sta-tistical and philosophical, error variance is to bereduced by any possible device. You turn to ani-mals of a cheap and short-lived species, so that youcan use subjects with controlled heredity and con-trolled experience. You select human subjects froma narrow subculture. You decorticate your subjectby cutting neurons or by giving him an environ-ment so meaningless that his unique responses dis-appear (cf. 25). You increase the number of casesto obtain stable averages, or you reduce N to 1, asSkinner does. But whatever your device, your goalin the experimental tradition is to get those embar-rassing differential variables out of sight.

The correlational psychologist is in love with justthose variables the experimenter left home to for-get. He regards individual and group variationsas important effects of biological and social causes.All organisms adapt to their environments, but notequally well. His question is: what present charac-teristics of the organism determine its mode and de-gree of adaptation?

Just as individual variation is a source of embar-rassment to the experimenter, so treatment variationattenuates the results of the correlator. His goal isto predict variation within a treatment. His ex-perimental designs demand uniform treatment forevery case contributing to a correlation, and treat-ment variance means only error variance to him.

Differential psychology, like experimental, beganwith a purely descriptive phase. Cattell at Hop-kins, Gallon at South Kensington, were simply ask-ing how much people varied. They were, we mightsay, estimating the standard deviation while thegeneral psychologists were estimating the centraltendency.

The correlation coefficient, invented for the studyof hereditary resemblance, transformed descriptivedifferential research into the study of mental or-ganization. What began as a mere summary sta-tistic quickly became the center of a whole theoryof data analysis. Murphy's words, written in 1928,


recall the 'excitement that attended this develop-ment:

The relation between two variables has actually beenfound to be statable in other terms than those of experiment. . . [Moreover,] Yule's method of "partial correlation" hasmade possible the mathematical "isolation" of variableswhich cannot be isolated experimentally. . . . [Despite thelimitations of correlational methods,] what they have al-ready yielded to psychology . . . is nevertheless of suchmajor importance as to lead the writer to the opinion thatthe only twentieth-century discovery comparable in impor-tance to the conditioned-response method is the method ofpartial correlations (35, p. 410).

Today's students who meet partial correlation onlyas a momentary digression from their main work instatistics may find this excitement hard to compre-hend. But partial correlation is the starting placefor all of factor analysis.

Factor analysis is rapidly being perfected into arigorous method of clarifying multivariate relation-ships. Fisher made the experimentalist an expertpuppeteer, able to keep untangled the strands tohalf-a-dozen independent variables. The correla-tional psychologist is a mere observer of a playwhere Nature pulls a thousand strings; but hismultivariate methods make him equally an expert,an expert in figuring out where to look for thehidden strings.

His sophistication in data analysis has not beenmatched by sophistication in theory. The correla-tional psychologist was led into temptation by hisown success, losing himself first in practical predic-tion, then in a narcissistic program of studying histests as an end in themselves. A naive operation-ism enthroned theory of test performance in theplace of theory of mental processes. And prema-ture enthusiasm2 exalted a few measurementschosen almost by accident from the tester's stockas the ruling forces of the mental universe.

In former days, it was the experimentalist whowrote essay after anxious essay defining his disci-pline and differentiating it from competing ways ofstudying mind. No doubts plagued correlationistslike Hall, Gallon, and Cattell. They came in onthe wave of evolutionary thought and were buoyedup by every successive crest of social progress orcrisis. The demand for universal education, the de-velopment of a technical society, the appeals fromthe distraught twentieth-century parent, and finallythe clinical movement assured the correlational psy-

2 This judgment is not mine alone; it is the clear consen-sus of the factor analysts themselves (see 28, pp. 321-32S).

chologist of his great destiny. Contemporary ex-perimentalists, however, voice with ever-increasingassurance their program and social function; andthe fact that tonight you have a correlational psy-chologist discussing disciplinary identities impliesthat anxiety is now perched on his windowledge.

Indeed, I do speak out of concern for correla-tional psychology. Aptitude tests deserve their finereputation; but, if practical, validated proceduresare to be our point of pride, we must be dissatisfiedwith our progress since 1920. As the ExecutiveCommittee of Division 5 itself declared this year,none of our latter-day refinements or innovationshas improved practical predictions by a noticeableamount. Correlational psychologists who foundtheir self-esteem upon contributions to theory canpoint to monumental investigations such as theStudies of Character and The Authoritarian Per-sonality. Such work does throw strong light uponthe human scene and brings important facts clearlyinto view. But theories to organize these facts arerarely offered and even more rarely solidified (30;31, p. 55).

POTENTIAL CONTRIBUTIONS OF THE DISCIPLINES

TO ONE ANOTHER

Perhaps it is inevitable that a powerful newmethod will become totally absorbing and crowdother thoughts from the minds of its followers. Ittook a generation of concentrated effort to movefrom Spearman's tetrad equation and Army Alphato our present view of the ability domain. It tookthe full energies of other psychologists to move fromS-R bonds to modern behavior theory. No doubtthe tendency of correlationists to ignore experimen-tal developments is explained by their absorptionin the wonders and complexities of the phenomenatheir own work was revealing. And if experimen-talists were to be accused of narrow-minded concen-tration on one particular style and topic of research,the same comment would apply.

The spell these particular theories and methodscast upon us appears to have passed. We are freeat last to look up from our own bedazzling treasure,to cast properly covetous glances upon the scien-tific wealth of our neighbor discipline. Trading hasalready been resumed, with benefit to both parties.

The introduction of construct validation into testtheory (12) is a prime example. The history ofthis development, you may recall, was that the


APA's Committee on Psychological Tests discoveredthat available test theory recognized no way of de-termining whether a proposed psychological inter-pretation of a test was sound. The only existingtheory dealt with criterion validation and could notevaluate claims that a test measured certain psy-chological traits or states. Meehl, capitalizing onthe methodological and philosophical progress ofthe experimenters, met the testers' need by suggest-ing the idea of construct validity. A proposed testinterpretation, he showed, is a claim that a testmeasures a construct, i.e., a claim that the testscore can be linked to a theoretical network. Thisnetwork, together with the claim, generates predic-tions about observations. The test interpretation isjustified only if the observations come out as pre-dicted. To decide how well a purported test ofanxiety measures anxiety, construct validation isnecessary; i.e., we must find out whether scores onthe test behave in accordance with the theory thatdefines anxiety. This theory predicts differences inanxiety between certain groups, and traditional cor-relational methods can test those predictions. Butthe theory also predicts variation in anxiety, hencein the test score, as a function of experience orsituations, and only an experimental approach cantest those predictions.

This new theory of validity has several verybroad consequences. It gives the tester a start to-ward the philosophical sophistication the experi-menter has found so illuminating. It establishesthe experimental method as a proper and necessarymeans of validating tests. And it re-establishes re-search on tests as a valuable and even indispensableway of extending psychological theory.

We may expect the test literature of the futureto be far less saturated with correlations of testswith psychologically enigmatic criteria, and farricher in studies which define test variables by theirresponsiveness to practice at different ages, to drugs,to altered instructions, and to other experimentallymanipulated variables. A pioneering venture in thisdirection is Fleishman's revealing work (21, 22) onchanges in the factorial content of motor skills as afunction of practice. These studies go far beyonda mere exploration of certain tests; as Ferguson hasshown (19, 20), they force upon us a theory whichtreats abilities as a product of learning, and atheory of learning in which previously acquiredabilities play a major role.

Perhaps the most valuable trading goods the cor-

relator can offer in return is his multivariate con-ception of the world.

No experimenter would deny that situations andresponses are multifaceted, but rarely are his pro-cedures designed for a systematic multivariateanalysis. The typical experimental design and thetypical experimental law employ a single dependentvariable. Even when more than one outcome ismeasured, the outcomes are analyzed and inter-preted separately. No response measure, however,is an adequate measure of a psychological construct.Every score mixes general construct-relevant vari-ance with variance specific to the particular meas-uring operation. It is all right for the agriculturistto consider size of crop as the fundamental variablebeing observed: that is the payoff for him. Ourtask, however, is to study changes in fundamentalaspects of behavior, and these are evidenced onlyindirectly in any one measure of outcome.

The correlational psychologist discovered longago that no observed criterion is truly valid andthat simultaneous consideration of many criteria isneeded for a satisfactory evaluation of performance.This same principle applies in experimentation. AsNeal Miller says in a recent paper on experimentswith drugs:

Where there are relatively few facts it seems easy to ac-count for them by a few simple generalizations. . . . As webegin to study the effects of a variety of drugs on a numberof different behavioral measures, exceptions and complexitiesemerge. We are forced to reexamine and perhaps abandoncommon-sense categories of generalization according to con-venient words existing in the English language. As new andmore comprehensive patterns of results become available,however, new and more precise generalizations may emerge.We may be able to "carve nature better to the joint" andachieve the simplicity of a much more exact and powerfulscience (32, pp. 326-327).

Theoretical progress is obstructed when one re-stricts himself to a single measure of response (34).Where there is only one dependent variable, it ispointless to introduce intervening variables or con-structs. When there are many response variables,however, it is mandatory to subsume them underconstructs, since otherwise we must have a separateset of laws for every measure of outcome. Dealingwith multiple response variables is, as Miller says(33), precisely the problem with which the factoranalysts have been concerned. Factor analysis, bysubstituting formal for intuitive methods, has beenof great help in locating constructs with which tosummarize observations about abilities. It is rea-


sonable to expect that multivariate treatment of re-sponse measures would have comparable value inexperimental psychology.

Experimenters very probably have even more togain from treating independent variables as a con-tinuous multivariate system. The manifold treat-ment categories in a Fisherian design are establisheda priori. In agriculture, the treatment dimensionsthe farmer can manipulate are obvious: fertilizer,water, species of seed, and so on. In a more basicscience, we require genotypic constructs to describesituations, constructs like the physical scientist'stemperature and pressure. The conditions the psy-chologist most easily manipulates—stimulus form,injunction to the subject, strength of electric shock—are not chosen because we intend to apply thesespecific conditions when we get around to "con-trolling behavior." They are used because theseconditions, we hope, embody scientifically usefulconstructs.

The experimenter has no systematic way to clas-sify and integrate results from different tasks ordifferent reinforcers. As Ferguson remarks (20, p.130; see also 19, p. 100): "No satisfactory meth-odology has emerged for describing particular learn-ing tasks, or indicating how one task differs fromanother, other than by a process of simple inspec-tion." We depend wholly on the creative flair ofthe theorist to collate the experiments and to inventconstructs which might describe particular situa-tions, reinforcements, or injunctions in terms ofmore fundamental variables. The multivariatetechniques of psychometrics are suited for preciselythis task of grouping complex events into homo-geneous classes or organizing them along major di-mensions. These methods are frankly heuristic, butthey are systematically heuristic. They select vari-ables with minimal redundancy, and they permit usto obtain maximum information from a minimum ofexperimental investment.

In suggesting that examining treatment condi-tions as a statistical universe is a possible way toadvance experimental thinking, I am of course echo-ing the recommendations of Egon Brunswik (6,esp. pp. 39-58). Brunswik criticized the Fisherianexperimenter for his ad hoc selection of treatmentsand recommended that he apply the sampling prin-ciples of differential psychology in choosing stimuliand conditions. A sampling procedure such asBrunswik suggests will often be a forward step, butthe important matter is not to establish laws which

Heart Dtatolie Systolic ilConduof-Rate B.R RR 11 anoe

2 -

FIG. 2. Mean response to four stressors expressed in termsof resting standard scores (data from 50).

apply loosely to a random, unorganized collectionof situations. The important matter is to discoverthe organization among the situations, so that wecan describe situational differences as systematicallyas we do individual differences.

Research on stress presents a typical problem oforganization. Multivariate psychophysiological dataindicate that different taxing situations have differ-ent effects. At present, stressors can be describedand classified only superficially, by inspection. Acorrelational or distance analysis of the data groupstreatments which have similar effects and ultimatelypermits us to locate each treatment within a con-tinuous multidimensional structure having con-structs as reference axes. Data from a recentstudy by Wenger, Clemens, and Engel (50) maybe used as an illustration. Figure 2 shows themeans of standardized physiological scores underfour different stress conditions: mental arithmetic,

FIG. 3. Multivariate diagram showing similarity betweenfour stressors.


a letter association test, hyperventilation, and a coldpressor. The "profiles" for the four conditions arevery significantly different. I have made a distanceanalysis to examine the similarity between condi-tions, with the results diagrammed in Figure 3.There is a general factor among all the treatments,which distinguishes them from the resting state,and a notable group factor among three of them.According to these data, a mental test seems to in-duce the same physiological state as plunging one'sfoot into ice water!

Much larger bodies of data are of course neededto map the treatment space properly. But the apt-ness of an attempt in this direction will be apparentto all who heard Selye's address to the APA lastyear. His argument (40) that all stressful situa-tions lead to a similar syndrome of physiologicalchanges is strongly reminiscent of Spearman's argu-ment regarding a general factor linking intellectualresponses. The disagreement between Selye andother students of stress clearly reduces to a quanti-tative question of the relative size of specific andnonspecific or general factors in the effects of typicalstressors.

APPLIED PSYCHOLOGY DIVIDED AGAINST ITSELF

Let us leave for the moment questions of aca-demic psychology and consider the schism as itappears in applied psychology. In applied psychol-ogy, the two disciplines are in active conflict; andunless they bring their efforts into harmony, theycan hold each other to a standstill. The conflict isespecially obvious at this moment in the challengethe young engineering psychology offers to tradi-tional personnel psychology.

The program of applied experimental psychologyis to modify treatments so as to obtain the highestaverage performance when all persons are treatedalike—a search, that is, for "the one best way."The program of applied correlational psychology isto raise average performance by treating personsdifferently—different job assignments, different ther-apies, different disciplinary methods. The correla-tionist is utterly antagonistic to a doctrine of "theone best way," whether it be the heartless robot-making of Frederick Taylor or a doctrinaire permis-siveness which tries to give identical encouragementto every individual. The ideal of the engineeringpsychologist, I am told, is to simplify jobs so thatevery individual in the working population will be

able to perform them satisfactorily, i.e., so that dif-ferentiation of treatment will be unnecessary. Thisgoal guides activities ranging from the sober to thebizarre: from E. L. Thorndike and Skinner, huntingthe one best sequence of problems for teaching arith-metic, to Rudolf Flesch and his admirers, reducingParadise Lost to a comic book. If the engineeringpsychologist succeeds: information rates will be soreduced that the most laggard of us can keep up,visual displays will be so enlarged that the mostmyopic can see them, automatic feedback will pre-vent the most accident-prone from spoiling the workor his fingers.

Obviously, with every inch of success the engineerhas, the tester must retreat a mile. A slight reduc-tion in information rate, accomplished once, reducesforever the validity and utility of a test of abilityto process data. If, once the job is modified, themyopic worker can perform as well as the man with20/20 vision, Snellen charts and orthoraters are outof business. Nor is the threat confined to the in-dustrial scene. If tranquilizers make everybodyhappy, why bother to diagnose patients to determinewhich treatments they should have? And if tele-vised lessons can simplify things so that every fresh-man will enjoy and understand quantum mechanics,we will need neither college aptitude tests nor finalexaminations.

It is not my intention to warn testers about loom-ing unemployment. If test technology is not greatlyimproved, long before the applied experimentalistsnear their goals, testing deserves to disappear. Mymessage is my belief that the conflicting principlesof the tester and the experimenter can be fused intoa new and integrated applied psychology.

To understand the present conflict in purposes,we must look again at historical antecedents. Pas-tore (36) argues with much justice that the testersand classifiers have been political conservatives,while those who try to find the best common treat-ment for all—particularly in education—have beenthe liberals. This essential conservatism of person-nel psychology traces back to the days of Darwinand Spencer.

The theory of evolution inspired two antagonisticmovements in social thought (10, 42). Darwin andHerbert Spencer were real determinists. The survi-val of the fittest, as a law of Nature, guaranteedman's superiority and the ultimate triumph of thenatural aristocrats among men. As Dewey put it,Spencer saw "a rapid transit system of evolution


. . . carrying us automatically to the goal of per-fect man in perfect society" (17, p. 66). Men varyin their power of adaptation, and institutions, bydemanding adaptation, serve as instruments of natu-ral selection among men. The essence of freedom isseen as the freedom to compete for survival. ToSpencer, to Galton, and to their successors down tothe present day, the successful are those who havethe greatest adjustive capacity. The psychologist'sjob, in this tradition, is to facilitate or anticipatenatural selection. He seeks only to reduce its cru-elty and wastage by predicting who will survive inschools and other institutions as they are. He takesthe system for granted and tries to identify who willfit into it. His devices have a conservative influencebecause they identify persons who will succeed inthe existing institution. By reducing failures, theyremove a challenge which might otherwise force theinstitution to change (49).

The experimental scientist inherits an interpreta-tion of evolution associated with the names of Ward,James, and Dewey. For them, man's progress restson his intelligence; the great struggle for survival isa struggle against environment, not against competi-tors. Intelligent man must reshape his environ-ment, not merely conform to it. This spirit, thevery antithesis of Spencerian laissez-faire, bred to-day's experimental social science which accepts noinstitution and no tradition as sacred. The individ-ual is seen as inherently self-directing and creative.One can not hope to predict how he will meet hisproblems, and applied differential psychology istherefore pointless (39, p. 37).

Thus we come to have one psychology which ac-cepts the institution, its treatment, and its criterionand finds men to fit the institution's needs. Theother psychology takes man—generalized man—asgiven and challenges any institution which does notconform to the measure of this standard man.

A clearer view of evolution removes the paradox:

The entire significance of the evolutionary method in biol-ogy and social history is that every distinct organ, structure,or formation, every grouping of cells or elements, has to betreated as an instrument of adjustment or adaptation to aparticular environing situation. Its meaning, its character,its value, is known when, and only when, it is considered asan arrangement for meeting the conditions involved in somespecific situation (16, p. 15).

We are not on the right track when we conceive ofadjustment or adjustive capacity in the abstract. Itis always a capacity to respond to a particular treat-

ment. The organism which adapts well under onecondition would not survive under another. If foreach environment there is a best organism, for everyorganism there is a best environment. The job ofapplied psychology is to improve decisions aboutpeople. The greatest social benefit will come fromapplied psychology if we can find for each individ-ual the treatment to which he can most easily adapt.This calls for the joint application of experimentaland correlational methods.

INTERACTION OF TREATMENT AND INDIVIDUAL

IN PRACTICAL DECISIONS

Goldine Gleser and the writer have recently pub-lished a theoretical analysis (11) which shows thatneither the traditional predictive model of the cor-relator nor the traditional experimental comparisonof mean differences is an adequate formulation ofthe decisions confronting the applied psychologist.Let me attempt to give a telescoped version of thecentral argument.

The decision maker has to determine what treat-ment shall be used for each individual or each groupof individuals. Psychological data help a college,for example, select students to be trained as scien-tists. The aim of any decision maker is to maxi-mize expected payoff. There is a payoff functionrelating outcome (e.g., achievement in science) toaptitude dimensions for any particular treatment.Figure 4 shows such a function for a single aptitude.Average payoff—if everyone receives the treatment—is indicated by the arrow. The experimentalistassumes a fixed population and hunts for the treat-ment with the highest average and the least varia-bility. The correlationist assumes a fixed treatment

PAYOFF

Mean

APTITUDE

FIG. 4. Scatter diagram and payoff function showing out-come as a function of individual differences.


PAYOFF

<-. Give tB- x GivetA—>

APTITUDEFIG. 5. Increase in payoff as a result of selection.

and hunts for aptitudes which maximize trie slopeof the payoff function. In academic selection, headvises admission of students with high scores on arelevant aptitude and thus raises payoff for the in-stitution (Figure 5).

Pure selection, however, almost never occurs.The college aptitude test may seem to be intendedfor a selection decision; and, insofar as the individ-ual college is concerned only with those it accepts,the conventional validity coefficient does indicatethe best test. But from a societal point of view, therejects will also go on into other social institutions,and their profit from this treatment must be weighedin the balance along with the profit or social con-tribution from the ones who enter college. Everydecision is really a choice between treatments. Pre-dicting outcome has no social value unless the psy-chologist or the subject himself can use the infor-mation to make better choices of treatment. The

PAYOFF

Give Treatment A -••>

PAYOFF

APTITUDEFIG. 6. Payoff functions for two treatments.

APTITUDEFIG. 7. Payoff functions for two treatments.

prediction must help to determine a treatment forevery individual.

Even when there are just two treatments, the pay-off functions have many possible relationships. InFigure 6 we have a mean difference between treat-ments, and a valid predictor. The predictor—though valid—is useless. We should give everyoneTreatment A. In Figure 7, on the other hand, weshould divide the group and give different treat-ments. This gives greater payoff than either treat-ment used uniformly will give.

Assigning everyone to the treatment with thehighest average, as the experimentalist tends to rec-ommend, is rarely the best decision. In Figure 8,Treatment C has the best average, and we mightassign everyone to it. The outcome is greater, how-ever, if we assign some persons to each treatment.The psychologist making an experimental compari-son arrives at the wrong conclusion if he ignores theaptitude variable and recommends C as a standardtreatment.

Applied psychologists should deal with treatmentsand persons simultaneously. Treatments are char-acterized by many dimensions; so are persons. Thetwo sets of dimensions together determine a payoffsurface. For any practical problem, there is somebest group of treatments to use and some best allo-cation of persons to treatments. We can expectsome attributes of persons to have strong interac-tions with treatment variables. These attributeshave far greater practical importance than the at-tributes which have little or no interaction. In di-viding pupils between college preparatory and non-college studies, for example, a general intelligencetest is probably the wrong thing to use. This test,

THE Two DISCIPLINES or SCIENTIFIC PSYCHOLOGY 681

PAYOFF

Give Treatment C >

tc PAYOFF

Give te*Givetc4 Give tA

APTITUDE APTITUDE

FIG. 8. Payoff functions for three treatments.

being general, predicts success in all subjects, there-fore tends to have little interaction with treatment,and if so is not the best guide to differential treat-ment. We require a measure of aptitude which pre-dicts who will learn better from one curriculum thanfrom the other; but this aptitude remains to be dis-covered. Ultimately we should design treatments,not to fit the average person, but to fit groups ofstudents with particular aptitude patterns. Con-versely, we should seek out the aptitudes which cor-respond to (interact with) modifiable aspects of thetreatment.

My argument rests on the assumption that suchaptitude-treatment interactions exist. There is,scattered in the literature, a remarkable amount ofevidence of significant, predictable differences in theway people learn. We have only limited success inpredicting which of two tasks a person can performbetter, when we allow enough training to compen-sate for differences in past attainment. But we dofind that a person learns more easily from onemethod than another, that this best method differsfrom person to person, and that such between-treatments differences are correlated with tests ofability and personality. The studies showing inter-action between personality and conditions of learn-ing have burgeoned in the past few years, and theliterature is much too voluminous to review in pass-ing. Just one recent finding will serve in the way ofspecific illustration, a study done by WolfgangBohm at Vienna (38, pp. 58-59). He showed hisexperimental groups a sound film about the adven-tures of a small boy and his toy elephant at the zoo.At each age level, a matched control group read averbatim text of the sound track. The differencesin average comprehension between the audiovisual

and the text presentations were trivial. There was,however, a marked interaction. For some reasonyet unexplained, a general mental test correlatedonly .30 with text learning, but it predicted filmlearning with an average correlation of .77.3 Thedifference was consistent at all ages.

Such findings as this, when replicated and ex-plained, will carry us into an educational psychologywhich measures readiness for different types ofteaching and which invents teaching methods to fitdifferent types of readiness. In general, unless onetreatment is clearly best for everyone, treatmentsshould be differentiated in such a way as to maxi-mize their interaction with aptitude variables. Con-versely, persons should be allocated on the basis ofthose aptitudes which have the greatest interactionwith treatment variables. I believe we will findthese aptitudes to be quite unlike our present apti-tude measures chosen to predict differences withinhighly correlated treatments.

THE SHAPE OF A UNITED DISCIPLINE

It is not enough for each discipline to borrowfrom the other. Correlational psychology studiesonly variance among organisms; experimental psy-chology studies only variance among treatments. Aunited discipline will study both of these, but it willalso be concerned with the otherwise neglected inter-actions between organismic and treatment variables(41). Our job is to invent constructs and to forma network of laws which permits prediction. Fromobservations we must infer a psychological descrip-tion of the situation and of the present state of theorganism. Our laws should permit us to predict,

3 Personal communication.


from this description, the behavior of organism-in-situation.

There was a time when experimental psycholo-gists concerned themselves wholly with general,nonindividual constructs, and correlational psychol-ogists sought laws wholly within developmental vari-ables. More and more, nowadays, their investiga-tions are coming to bear on the.same targets. Onepsychologist measures ego involvement by a person-ality test and compares the behavior of high- andlow-scoring subjects. Another psychologist height-ens ego involvement experimentally in one of twoequated groups and studies the consequent differ-ences in behavior. Both investigators can test thesame theoretical propositions, and to the extent thattheir results agree they may regard both proceduresas embodiments of the same construct.

Constructs originating in differential psychologyare now being tied to experimental variables. As aresult, the whole theoretical picture in such an areaas human abilities is changing. Piaget (37) corre-lates reasoning processes with age and discovers adevelopmental sequence of schemata whose emer-gence permits operational thought; Harlow (24)begins actually to create similar schemata in mon-keys by means of suitable training. It now becomespossible to pursue in the controllable monkey en-vironment the questions raised by Piaget's uniquecombination of behavioral testing and interviewing,and ultimately to unite the psychology of intelli-gence with the psychology of learning.

Methodologies for a joint discipline have alreadybeen proposed. R. B. Cattell (8) has offered themost thorough discussion of how a correlationistmight organize data about treatment and organismsimultaneously. His factor analytic procedures areonly one of many choices, however, which modernstatistics offers. The experimenters, some of them,have likewise seen the necessity for a united disci-pline. In the very issue of Psychological Reviewwhere the much-too-famous distinction between 5-1?and R-R laws was introduced, Bergmann andSpence (3) declared that (at the present stage ofpsychological knowledge) the equation R = f (S)must be expanded into

R = f (S, T, D, I)

The added variables are innate differences, motiva-tion, and past experience—differential variables all.Hull (26, 27) sought general laws just as did

/' Organism ^\ at present / '

,'' Predicted "\Response /

****. *•**

FIG. 9. Theoretical model for prediction from historic data.

Wundt, but he added that organismic factors canand must be accounted for. He proposed to do thisby changing the constants of his equations witheach individual. This is a bold plan, but one whichhas not yet been implemented in even a limited way.It is of interest that both Hull (27, p. 116) andTolman (47, p. 26) have stated specifically that fortheir purposes factor analytic methods seem to havelittle promise. Tucker, though, has at least drawnblueprints of a method for deriving Hull's own indi-vidual parameters by factor analysis (48). Clearly,we have much to learn about the most suitable wayto develop a united theory, but we have no lack ofexciting possibilities.

The experimenter tends to keep his eye on ulti-mate theory. Woodworth once described psycho-logical laws in terms of the S-O-R formula whichspecifically recognizes the individual. The revisedversion of his Experimental Psychology (53, p. 3),however, advocates an S-A-R formula, where Astands for "antecedent conditions." This formula-tion, which is generally congenial to experimenters,reduces the present state of the organism to an in-tervening variable (Figure 9). A theory of thistype is in principle entirely adequate to explain,

. \( Predicted \~\ Response /

**-^_ ^*

FIG. 10. Theoretical model for prediction fromahistoric data.

THE Two DISCIPLINES or SCIENTIFIC PSYCHOLOGY 683

Organism Nat present / \x..^ f*' \ ^_

\ v/'' Predicle, v f Predicted ^' 'V Response /

FIG. 11. Theoretical network to be developed .by a uniteddiscipline.

predict, and control the behavior of organisms; but,oddly enough, it is a theory which can account onlyfor the behavior of organisms of the next generation,who have not yet been conceived. The psychologistturns to a different type of law (Figure 10) when-ever he deals with a subject whose life history hehas not controlled or observed in every detail. Atheory which involves only laws of this type, whilesuitable for prediction, has very limited explanatoryvalue. The theory psychology really requires is aredundant network like Figure 11. This networkpermits us to predict from the past experience orpresent characteristics of the organism, or a combi-nation of the two, depending on what is known.Filling in such a network is clearly a task for thejoint efforts of experimental and correlational psy-chology.

In both applied work and general scientific work,psychology requires combined, not parallel, laborsfrom our two historic disciplines. In this commonlabor, they will almost certainly become one, witha common theory, a common method, and commonrecommendations for social betterment. In thesearch for interactions we will invent new treatmentdimensions and discover new dimensions of the or-ganism. We will come to realize that organismand treatment are an inseparable pair and that nopsychologist can dismiss one or the other as errorvariance.

Despite our specializations, every scientific psy-chologist must take the same scene into his field ofvision. Clark Hull, three sentences before the endof his Essentials oj Behavior (27, p. 116), voicedjust this need. Because of delay in developingmethodology, he said, individual differences haveplayed little part in behavior theory, and "a sizeable

segment of behavioral science remains practicallyuntouched." This untouched segment contains thequestion we really want to put to Nature, and shewill never answer until our two disciplines ask it ina single voice.

REFERENCES

1. BARTLETT, F. C. Fifty years of psychology. Occup.Psychol, 1955, 29, 203-216.

2. BEACH, F. A. The snark was a boojum. Amer. Psy-chologist, 19SO, 5, 115-124.

3. BERGMANN, G., & SPENCE, K. W. The logic of psycho-physical measurement. Psychol. Rev., 1944, SI, 1-24.

4. BINDRA, D., & SCHEIER, I. H. The relation betweenpsychometric and experimental research in psychol-ogy. Amer. Psychologist, 1954, 9, 69-71.

J5V BORING, E. G. History of experimental psychology.(2nd ed.) New York: Appleton-Century-Crofts, 1950.

6. BRUNSWIK, E. Perception and the representative designof psychological experiments. Berkeley: Univer.California Press, 1956.

7. CATTELL, J. McK. The biological problems of today:Psychology. Science, 1898, 7, 152-154.

8. CATTEIX, R. B. Factor analysis. New York: Harper,1952.

9. CLARK, K. E. America's psychologist?. Washington,D. C.: APA, 1957.

10. CORWIN, E. S. The impact of the idea of evolution onthe American political and constitutional tradition.In S. PERSONS (Ed.), Evolutionary thought in Amer-ica. New Haven: Yale Univer. Press, 1950. Pp.182-201.

11. CRONBACH, L. J., & GLESER, GOLDINE C. Psychologicaltests and personnel decisions. Urbana: Univer, Illi-nois Press, 1957.

12. CRONBACH, L. J., & MEEHL, P. E. Construct validityin psychological tests. Psychol. Bull, 1955, 52, 281-302.

13. CRONBACH, L. J., & NEFF, W. D. Selection and training.In Com. on Undersea Warfare Panel on Psychologyand Physiology, Human Factors in Undersea Warfare.Washington, D. C.: Nat. Res. Coun., 1949. Pp.491-516.

14. DANIEL, R. S., & LOUTTIT, C. M. Professional problemsin psychology. New York: Prentice-Hall, 1953.

15. DASHIELL, J. F. Some rapprochements in contemporarypsychology. Psychol. Bull., 1939, 36, 1-24.

16. DEWEY, J. Studies in logical theory. Chicago: Uni-ver. Chicago Press, 1903.

17. DEWEY, J. The influence oj Darwin on philosophy andother essays. New York: Holt, 1910.

18. EYSENCK, H. J. Reminiscence, drive, and personalitytheory. J. abnorm. soc. Psychol., 1956, 53, 328-333.

19. FERGUSON, G. A. On learning and human ability.Canad. J. Psychol, 1954, 8, 95-112.

20. FERGUSON, G. A. On transfer and human ability.Canad. J. Psychol, 1956, 10, 121-131.

21. FLEISHMAN, E. A. Predicting advanced levels of pro-ficiency in psychomotor skills. In Proc. Sympos. on


Human Engng. Washington, D. C.: Nat. Acad. Sci.,1956. Pp. 142-151.

22. FLEISHMAN, E. A., & HEMPEL, W. E., JK. Changes infactor structure of a complex psychomotor test asa function of practice. Psychometrika, 1954, 19,239-252.

23. FORD, C. S., & BEACH, F. A. Patterns of sexual behav-ior. New York: Harper, 1952.

24. HARLOW, H. F. The formation of learning sets. Psy-chol Rev., 1949, 56, 51-65.

25. HARLOW, H. F. Mice, men, monkeys, and motives.Psychol. Rev., 1953, 60, 23-32.

26. HULL, C. L. The place of innate individual and speciesdifferences in a natural-science theory of behavior.Psychol. Rev., 1945, 52, 55-60.

27. HULL, C. L. Essentials of behavior. New Haven: YaleUniver. Press, 1951.

28. LAUGIER, H. (Ed.) L'analyse factorielle et ses applica-tions. Paris: Centre National de la Recherche Sci-entifique, 1955.

29. LAZARUS, R. S., & BAKER, R. W. Personality and psy-chological stress—a theoretical and methodologicalframework. Psychol. Newsletter, 1956, 8, 21-32.

30. MCCANDLESS, B. R., & SPIKER, C. C. Experimental re-search in child psychology. Child Develpm., 1956,27, 75-80.

31. MCCLELLAND, D. C. Personality. In P. R. FARNS-WORTH (Ed.) Annu. Rev. Psychol., 1956. Stanford:Annual Reviews, 1956. Pp. 39-62.

32. MILLER, N. E. Effects of drugs on motivation: Thevalue of using a variety of measures. Ann. N.Y.Acad. Sci., 1956, 65, 318-333.

33. MILLER, N. E. Liberalization of basic S-R concepts:Extensions to conflict behavior and social learning.In S. Koch (Ed.), Psychology: A study of a science.Vol. II. General systematic formulations, learning,and special processes. New York: McGraw-Hill,in press.

34. MILLER, N. E. Objective techniques for studying mo-tivational effects of drugs on animals. In E. TRA-BUCCHI (Ed.), Proc. Int. Sympos. on PsychotropicDrugs. Amsterdam, Netherlands: Elsevier PublishingCo., in press.

35. MURPHY, G. An historical introduction to modern psy-chology. (3rd ed.) New York: Harcourt, Brace, 1932.

36. PASTORE, N. The nature-nurture controversy. NewYork: Kings Crown Press, 1949.

37. PIAGET, J. Psychology of intelligence. M. PIERCY andD. E. BERLYNE (Trans.). London: Routledge andKegan Paul, 1950.

38. ROHRACHER, H. Aus der wissenschaftlichen Arbeit desPsychologischen Institutes der Universitat Wien.Wiener Z. Phil., Psychol, Pddag., 1956, 6, 1-66.

39. SCOON, R. The rise and impact of evolutionary ideas.In S. PERSONS (Ed.), Evolutionary thought in Amer-ica. New Haven, Yale Univer. Press, 1950. Pp. 4-43.

40. SELYE, H. Stress and disease. Science, 1955, 122, 625-631.

41. SHEN, E. The place of individual differences in experi-mentation. In Q. McNEMAR and M. A. MERRILL(Eds.), Studies in personality. New York: McGraw-Hill, 1942. Pp. 259-283.

42. SPENGLER, J. J. Evolutionism in American economics.In S. PERSONS (Ed.), Evolutionary thought in Amer-ica. New Haven: Yale Univer. Press, 1950. Pp. 202-266.

43. TERMAN, L. M. The mental test as a psychologicalmethod. Psychol. Rev., 1924, 31, 93-117.

44. THORNDIKE, R. L. The psychological value systems ofpsychologists. Amer. Psychologist, 1954, 9, 787-790.

45. THURSTONE, L. L. The stimulus-response fallacy inpsychology. Psychol. Rev., 1923, 30, 354-369.

46. TITCHENER, E. B. The past decade in experimental psy-chology. Amer. J. Psychol., 1910, 21, 404-421.

47. TOLMAN, E. C. The determinants of behavior at achoice point. Psychol. Rev., 1938, 45, 1-41.

48. TUCKER, L. R. Determination of parameters of a func-tional relation by factor analysis. ETS Res. Bull.,1955, 55, No. 10.

49. TYLER, R. W. Can intelligence tests be used to predicteducability? In K. EELLS et al., Intelligence and cul-tural differences. Chicago: Univer. Chicago Press,1951. Pp. 39-47.

50. WENGER, M. A., CLEMENS, T. L., & ENGEL, B. T. Auto-nomic response patterns to four stimuli. Unpublishedmanuscript, 1957.

51. WOODWORTH, R. S. Dynamic psychology. New York:Holt, 1918.

52. WOODWORTH, R. S. Experimental psychology. NewYork: Holt, 1938.

53. WOODWORTH, R. S., & SCHLOSBERG, H. Experimentalpsychology. (2nd ed.) New York: Holt, 1954.

54. WUNDT, W. Principles of physiological psychology.Vol. 1. (5th ed.) E. B. Titchener (Trans.) NewYork: Macmillan, 1904.

55. YERKES, R. M. Comparative psychology: A question ofdefinitions. /. Phil. Psychol., and sci. Methods, 1913,10, 580-582.

56. YERKES, R. M. The study of human behavior. Science,1914, 29, 625-633.

the two disciplines of scientific...

Documents