The Methodology of Behavior Study
THE ULTIMATE object of scientific study is prediction, for with prediction we can have control. This is best accomplished by the experimental scientist. The chemist, for example, can predict and control, within limits, because he has learned by experience that certain materials in certain situations always behave in the same way. He can prepare his situations, introduce his materials and get uniform reactions. He is able to measure influence because his materials are stable and he can control all the influences reaching them, or if they change and enter combinations he is able to measure the changes and record the combinations and again predict. The scientist is able to determine a limited number of laws — that under given conditions given results will invariably follow. He is not, however, able to give a complete causal explanation of any phenomenon. He cannot, for example, explain completely why a certain wild rose bloomed under a certain hedge at a certain moment. In order to do this it would be necessary to begin with the formation of the material world, determine every force and measure every influence in the universe in the order of their reciprocal action down to the present moment.
The complete determination of the causation of any act of human behavior would be a task not less impossible than this. The chemist deals with elements which are relatively simple, while the behaviorist deals with actions which are in turn based on incommensurable physiological conditions — an incredibly complicated integration of endocrines, enzymes, blood chemicals, chromosomes, various nervous systems, behaving as a whole. Moreover, the material of the chemist is static, does not change from time to time, while the material of the behaviorist (the
(554) human organism) is itself evolving. The individual is changing, under influences which cannot be measured. His response in situations changes with periods of physical, mental and emotional maturation and as result of experiences in an endless variety of preceding situations. The student of behavior can therefore not hope to establish even the limited number of laws possible in the case of the exact scientist. He may hope to be able to determine that in certain situations certain reactions will usually follow. He will be able to make inferences but probably unable to establish laws. This would imply, then, not a complete but an adequate causal explanation of behavior.
It is desirable, therefore, to set up, if possible, a methodological procedure in behavior studies which will fix some limits to the behavior universe, as the scientist fixes some limits to the material universe, and at the same time give data for an adequate prediction and control.
At the close of chapter XI we presented a picture of the "inner [physiological] environment " as drawn by Claude Bernard — the environment in which the organism lives — and compared it with the outer environment —the world of outer space in which the organism behaves. The physiological system may be regarded as representing what the organism wants. If we could read the organism completely in terms of its morphology and chemistry we could predict what it wants. In his epoch—making study of the integrative action of the nervous system Sherrington points out that we cannot understand the reflexes unless we appreciate what they mean for activity. " The reflex action cannot be really intelligible to the physiologist until he knows its aim."  The structure represents, then, implicit behavior; the overt behavior is the process of satisfying the wants.
Among all the intricacies of the physiological system there are two major features of far-reaching consequences for behavior. One of them relates to the basic appetites and contains the so-called hunger and sex drives, representing the conditions of organic continuity — nutrition and reproduction. The other relates to the presence in the organism of certain preformed tendencies to behave in specific action patterns, whereby the organism is more or less predestined by its internal structure to behave in given ways. These unlearned action tendencies are the so-called instincts. While we do not wish to emphasize the importance of the " internal environment " for behavior studies — we wish, in fact, to
(555) minimize it —it is methodologically important to appreciate. the role of these drives and instincts in their relation to the socially more important field of learned behavior.
In the following description of the gastric contractions of the newborn we see the mechanism of the hunger drive:
We have now made observations on a number of new-born infants, and on two pups, born 8-10 days before term, with results showing that the empty stomach at birth and in the prematurely born exhibits the typical periods of tonus and hunger contractions of the adult, the only difference between infant and adult being the greater frequency and relatively greater vigor of these periods in the young. In the case of the two pups, and in some of the infants, the observations were made before their first nursing. It is thus clear that in the normal mammal the gastric hunger mechanism is completed, physiologically, and is probably active some time before birth .
Psychological interest has recently centered on the relation of the unlearned to the learned factors in behavior. In the scale of evolution below man unlearned behavior is the prevalent form of activity. Insects may be hatched, traverse their life cycle of predestined activities, feed, reproduce and expire within the duration of an hour. In his brilliant work Wheeler has exemplified the complicated unlearned activities of ant societies, and Herrick  has shown how several great action systems (insects, birds, mammals, man) approach different termini of structure and activity. Each reaches the limit of efficiency and perfection set by its morphological type and becomes static.
In contrast, man is distinguished in the organic scale by the feebleness of his unlearned action tendencies. He is obliged to learn almost everything and consequently is characterized by a learned habit system developed through his experiences.
Watson was the first to experiment with new-born children, taking them when they were an hour or more old, with a view to determining the presence of unlearned behavior tendencies. He was able to find only three instinctive reactions:
Fear. What stimulus apart from all training will call out fear responses; what are these responses, and how early may they be called out? The principal
(556) situations which call out fear responses seem to be as follows: (1) To suddenly remove from the infant all means of support, as when one drops it from the hands to be caught by an assistant . . . (2) by loud sounds. . . . The responses are a sudden catching of the breath, clutching randomly with the hands (the grasping reflex invariably appearing when the child is dropped), sudden closing of the eye-lids, puckering of the lips, then crying; in older children possibly flight and hiding (not yet observed by us as " original reactions"). In regard to the age at which fear responses first appear, we can state with some sureness that the above mentioned group of reactions appears at birth. .
Rage. In a similar way the question arises as to what is the original situation which brings out the activities seen in rage. Observation seems to show that the hampering of the infant's movements is the factor which apart from all training brings out the movements characterized as rage. If the face or head is held, crying results, quickly followed by screaming. The body stiffens and fairly well-co÷rdinated slashing or striking movements of the hands and arms result; the feet and legs are drawn up and down; the breath is held until the child's face is flushed. In older children the slashing movements of the arms and legs are better co÷rdinated, and appear as kicking, slapping, pushing, etc. These reactions continue until the irritating situation is relieved, and sometimes do not cease then. Almost any child from birth can be thrown into a rage if its arms are held tightly to its sides; sometimes even if the elbow joint is clasped tightly between the fingers the response appears. .. .
Love. The original situation which calls out the observable love responses seems to be the stroking or manipulation of some erogenous zone, tickling, shaking, gentle rocking, patting and turning upon the stomach across the attendant's knee. The response varies. If the infant is crying, crying ceases, a smile may appear, attempts at gurgling, cooing, and finally, in slightly older children, the extension of the arms, which we should class as the forerunner of the embrace of adults.
If we now give our attention to two of these behavior factors, the physiologically based hunger and the psychologically based love, we see that each has eventually far-reaching behavior implications. We may assume that hunger will eventually lead to pursuit of game, manual skill, mechanical inventions, foresight, economy, property, conflicts, war, the state, etc. Similarly the instinct of love, expressing itself first in response between mother and child, will later function as the sex drive and have
( 557) a long train of social consequences—courtship, marriage, family—with profound influence on behavior.
We have from Watson's list two other unlearned behavior reactions, anger and fear. He does not name curiosity, but it will appear later with further maturation of the organism, and lead to adventure, exploration, scientific interest, etc. There will also appear a train of other behavior traits which have been called instincts: Gregariousness (the herd instinct), acquisitiveness, instinct for self-preservation, instinct for workmanship, the artistic instinct, instinct for dominance (" will to power "), instinct of rivalry, etc. It would be possible to multiply this list almost indefinitely with the aid of a dictionary (pity, compunction, gratitude, religion, etc.) — a galaxy of psychological entities interacting by some sort of magic. These are probably all learned " attitudes," not instincts, but they are based on organic conditions, body chemistry and maturation of structure. These or the unlearned reactions may not be present today and present tomorrow, when the organism is riper, and they may appear only as a reaction to specific stimuli. Morgan, for example, describes the behavior of a moor-hen chick which had been swimming and practicing for some weeks but had never dived. On the appearance of a rough-haired pup it dived for the first time in its life. Complicating the situation we have the endocrines, the enzymes, blood chemistry, mentioned above, providing a class of endogenous chemical-electrical stimuli. Thurstone has even elaborated the point that, properly speaking, all stimuli are located within the organism itself, that the hunger contraction, for example, not the food, is the stimulus. Finally, we have the question of consciousness, whether it exists, when it appears, whether it is a special endowment or a linking-up of various reflexes, etc.
The traditional interpretations of behavior have worked from this approach and with these data. Focusing on " instincts," " consciousness," " original nature," they attempted to explain why the organism behaves in given ways in view of its internal nature and structure, and the attempt has led to a great deal of controversy and much confusion. On the contrary, we find that all the programs which we outlined in the preceding chapter are behavioristic. They ignore largely questions of
(558) the organic causation of behavior, the " why " of behavior reactions, and limit themselves to the observation, measurement and comparison of behavior manifestations —how the individual behaves in specific situations. This is precisely what the scientist does. He has learned to limit his problems to conditions which he can measure. He does not inquire why his materials behave in given ways but how they behave in given situations.
We are not anxious to discourage behavior studies from the standpoint of the mechanisms of the organism. On the contrary, it is very useful to have the data provided by Carlson on hunger and Watson on the " instincts," etc. In interpretation it is necessary to work with hypotheses, which are heuristic devices employed in the search for meaning — to be abandoned if the data do not provide a sufficient number of corroborations. The hypotheses should be as many as possible and for this purpose the data of the " inner environment," the unlearned and learned " attitudes," " norms," " values," " goals," etc., are useful. Sherrington's work on the nervous system, for example, will hardly be without value to the behaviorist, since the total situation contains the physiological and neurological. (In social studies, as in disputations, the inclination is great toward the " all or none " principle of interpretation.) But if we take a social situation involving the physiologically based hunger, and the psychologically based love of Watson's statement, a situation where a child is placed at birth in his mother's arms, and trace the reactions of both for a period of time, we shall find that measurable interpretations can be made in terms of the behavior expressions but not in terms of the behavior mechanisms.
To exemplify this we take some data from a study of new-born children made in Vienna. Working in the hospitals Hetzer and Tudor-Hart  divided 126 children into 9 groups of 14 each, the first group containing children 3 days old and under, and the last group containing those 4 to 5 months old, and experimenting with sound-stimuli, they observed the rate at which the child learns to separate out and give attention to the human voice among other sounds. All the children noticed all the sounds (striking a porcelain plate with a spoon, rattling a piece of paper, and the human voice) sometimes, but the reaction of
(559) the new-born to noise; in the first weeks was far more positive than the reaction to the voice, even to loud conversation: 92% of frequency to "ear-splitting" noises and 25% to the excited voice. But in the the week the proportion was about the same, and in the fourth week the reaction was more frequent to the voice. Here we begin to see the behavior of the child humanized by the prominence and function of mother or nurse in the situation. A process of conditioning has been going on — the human voice and feeding simultaneously. The voice has gained a significance over other sounds in the feeding complex, but at first the person speaking or the loudness or softness of the tone makes no difference. The voice has been associated with feeding, and angry tones have not yet been associated with punishment. The first specific reaction to the voice is a puckering of the lips, which appears in the third week. This is a pre-social reaction because it is not associated with any definite person, merely with a voice, any voice — a voice among other noises. The speaking person does not exist for the child. The voice stimulates the saliva reflex and if feeding does not follow the child will cry.
In the meantime the mother or nurse has held the child in her arms, stroked it, changed it, warmed it and made it comfortable, in addition to feeding it, but it is not until the second month that the child recognizes her as a person. At that age he may interrupt his nursing to look at his mother and smile. He has begun to identify the mother and associate her with a feeling of well-being. The authors call the smile the first social reaction of the child.
At this point the mother and child are involved in an intimacy. It is the first social relationship in the developmental history of the child, and it grows out of the hunger contractions and the mother's response. As this intimacy continues it has been observed that some bad habits may have been established —that the child may have become "spoiled." The surprising number of problem children at early age levels and their prevalence in the whole population was indicated in document 42 (p. 88).
Taking up these problems, it will be possible to interpret the behavior of the child at this early age in terms of his " original nature " — the physiologically based hunger and the instinctive love response, leading to a fixation on the mother up to a certain point, but no further. As a result of the intimacy we may have a habit system in which response is
(560) overemphasized. It is expressed eventually in clinging to the mother crying when separated, jealousy, emotional outbursts, etc. The child is then able to use these reactions as power devices to control the mother Through the tantrum he can secure petting, candy, or anything lit wants. We have in mind the case of a child unable to sleep except on hip mother's body in a certain position, of another who deliberately crawled on the bed in order to wet it and spite the mother. The birth of another child introduces a crisis and is frequently the occasion of a train of mi:; behavior to secure attention.
Psychiatrists have located a large class of maladjustments in this field — disappointment, inferiority-feelings, frustrated expectations and ambitions. " Between the bed-wetting and tantrums of childhood at the one extreme and the dilapidated conduct of the senile dement, come a fascinating series of clinical pictures representing the failures to meet the tests of life at each of the seven ages of man." These cases may be interpreted as meaning that the behavior patterns formed on the response level, beginning in the arms of the mother, have persisted beyond infancy and that they have proven unsuccessful means of adaptation to life.
In the meantime the hunger and its satisfaction has resulted in growth and the organism is integrating (musculature, nervous system, glands, etc.) for performances — for the pursuits, explorations, conflicts, skills, goals, careers of adult life in a society containing more enmity reactions than love reactions. As growth and integration progress the motor activities become more diversified, and we have play and curiosity and exploration. At this point there begins to be a hampering of the child's movements by the mother which, as Watson's description has shown, provokes resistance. Consequently some confusion arises in the attempt to trace causation. If " obstinacy," " negativism," " destructiveness," " tantrums " (document 42) are forms of naughtiness designed to hold the attention and provoke the response of the mother, they appear to be also "performance" expressions, a fight with the mother when she attempts to hamper the child's movements in her effort to conform him to a code. We have at this point mixed motives, as when the small boy gave his mother a good-night embrace and at the same time wished he could strangle her. When the attitudes of response and
( 561) resistance coexist and function intermittently we have all " ambivalence " of the emotions. Boner, for example, records a hospital case where a woman wept on hearing of the death of her child, but in the intervals of her sobbing it was noticed that she was smiling slyly.
It will be seen that it has been possible up to this point to interpret the child's behavior somewhat successfully from either of the two standpoints, that of " original nature," or that of " situation," but if we should continue the attempt to interpret his behavior in terms of " original nature " we should have to fall back on the pseudo-instincts enumerated above and to assume differences in constitution, in blood chemistry, in the operation of endocrine organs, in the preponderance of this or that " instinct " and we should run into endless speculations and have after all no program of treatment. These speculations formed the content of the older psychological, sociological and educational literature.
The behavioristic or situational approach, on the other hand, ignores or minimizes instincts and original nature and studies behavior reactions and habit formation in a great variety of situations comparatively. It assumes that whatever can be learned about original nature will be revealed in its reactions to these various situations. We regard this approach as the only one capable of giving a rational basis for the control of behavior which may be a substitute for the common sense, preceptual, ordering-and-forbidding type of control which has been traditional and which, to the degree that it had efficiency in the past, has now broken down.
If we take, for example, the " spoiling " in document 42 we see that this may originate in a variety of situations. Frank has enumerated some of the situations relating to the regulation of physiological tensions which may be the occasion of good or bad habit formation:
The infant is confronted with his first tensional problem shortly after birth, when physiological hunger appears, in the rhythmic contractions of the stomach accompanying the fall of blood sugar. These contraction tensions are usually relieved by the maternal ministrations at recurrent intervals, but the child must learn to sustain these tensions or to diffuse them (by crying, fretting, or other overt activity) until the feeding period arrives. This problem demands a physiological adjustment, with a regularization of nutritional process, uniform utilization of blood sugar over the period between feedings, and the concomitant functional adjustments, as well as the learning of the
(562) overt activity described. To a considerable extent the child must. also learn the unvarying sequence of processes involved in digestion, at least to the ex tent of developing a straightforward sequence of metabolic processes, a task not so automatic as many assume, as any mother will testify, since only I he sucking and swallowing reflexes are prepared. .. .
Indeed, the regularization of the nutritional processes and the management of hunger tensions is a major problem for the infant and young child. The number and variety of " feeding problems," both physiological and psycho logical, exhibited by young children is ample evidence for this statement mid for the importance assigned it in the development of personality.
The hunger tensions and their management are usually made more acute for the child at the time of weaning and of the introduction to solid food. For then the infant is faced with the necessity of learning to use novel stimuli to relieve his hunger tensions. If he is wisely handled at these times, he may achieve a wholesome solution, but the clinical records reveal an astonishing number of children who have made the transition with difficulty. This problem is more or less a prototype of subsequent tensional problems, in that the child, who has learned to use a certain stimulus to relieve his tensions, is now required to relinquish that stimulus and to learn to use other stimuli. This substitution of a novel stimulus for the customary (or biological) stimulus is, of course, one of the essential features of learning, and it describes the physiological learning as well as the psychological... .
This process of substituting a new for an accustomed stimulus is difficult largely because the period of learning involves a prolongation of the tensions, the release of which is the focus of the child's activity. In starting a child on solid food, for example, he is offered substances for which he has no learned method of handling; indeed, he scarcely has learned to swallow non-liquid food, and so is inclined to reject it. While he is tentatively tasting and trying to swallow the solid food, the unrelieved hunger tensions increase his irritability toward the strange food and also his efforts to obtain the customary liquids. He must sustain these tensions until he has learned to swallow the solid food and discovered its use as a stimulus to relieve his hunger tensions. This learning can be greatly facilitated by wise handling which assists the child in meeting this trying situation, by giving him such soothing and reassurance as will enable him to endure these prolonged tensions. .
The elimination of waste through the urine and faeces occurs as a reflex in the young infant, when the accumulating pressure tensions release the bladder sphincter or the anal sphincter. Sooner or later the parents attempt to teach the child continence and so present him with a new set of tensional problems. He must learn to sustain these accumulating tensions until the appropriate time and place for their release is presented. This learning involves a gradual
(563) raising of the threshold of the sphincters so that they will hold against the increased pressures and, more difficult, also learning to use these intravesicular pressures as a stimulus to the overt activity of seeking the appointed place for their release. The child must also learn to regularize his eliminations to a large extent... .
As we shall see later, the child is increasingly faced with this problem of learning to respond to present situations in terms of their consequences, which means that his behavior must become increasingly instrumental, or, in other words, he must learn to sustain tensions while he is achieving the duly sanctioned opportunities and means for their release.
In early infancy . . . the infant will [normally], when slightly fatigued, go to sleep, thereby achieving the release of accumulated muscular tension and the restoration of the depleted physiological processes. As the child grows older, however, this almost automatic slumber may, and usually does, disappear, largely because the child cannot readily get rid of the tensions accumulated during waking hours. The child then may have to learn how to release his muscular tensions for sleeping by developing a method of relaxing when put to bed. Not all children do learn this release, especially those who are allowed to become too fatigued. Even when they do sleep, they fail to achieve a wholesome relaxation. We must, therefore, include sleeping in our inventory of tensional problems facing the child.. . .
One process in the child which is functionally complete and efficient at birth is the sympathetic reaction or so-called " emotional response." Under stimulation, such as shock and the blocking of activity, the sympathetic division of the vegetative nervous system is stimulated into action, thus initiating a series of physiological changes: a quick visceral spasm, followed by a progressive relaxation of visceral tone, accelerated pulse and respiration, alternation in circulation from visceral to peripheral, and release of glycogen from the liver into the circulation. All of these changes are in the nature of preparations for the exertion of flight or fight, and they operate to raise the tonicity of striped muscle and to make available the energy resources of the organism. This condition of " panic," however, is inhibitive of any discriminative or adjustive reaction to a situation, and so is rarely useful in social behavior where learned patterns of response must be employed.
The liability to this " panic " and to the release of the suddenly available energy in activities such as retreat from contacts with other persons, violent attacks upon persons or things, and so on, gives rise to tensional problems which we may group under the term " emotional reactions." The essential features of this problem, considered from the viewpoint of the child learning to live in group life, are, first, to inhibit the progressive development of the sympathetic reaction beyond the initial visceral spasm, so that he may escape
(564) that cumulative " panic " and its expression in non-sanctioned behavior; second, to learn some form of motor activity more or less adequate to such emotional-producing situations, so that he can deal with those situations when-ever they recur. Whatever increased tonicity may be evoked by such situations will then be channeled into some form of overt activity specifically addressed to the requirements of that situation. . . . Moreover, the child does not. ordinarily learn either to inhibit his sympathetic reaction (or panic) or to develop an adequate motor pattern, such as a technique addressed to the situation, except with the aid and collaboration of others. For if, concurrently with the emotional stimuli, some soothing or reassuring stimuli, either auditory or tactual or both, are received, the progressive development of the sympathetic reaction may be checked and, with further practice, inhibited or at least restricted to the initial visceral spasm or " start." . . .
The importance of this problem of emotional management arises from the fact that failure to inhibit the " panic " reaction may operate to compromise the child's future learning in situations of the same or similar character as the panic-producing situation. For after a child has once been thrown into panic, a recurrence of the original situation (or portion thereof) will revive the panic and to that extent render learning of motor techniques impossible. So again the intervention of other persons appears necessary to enable a child to escape from his emotional reactions. The threshold to emotional stimuli can be raised by practice, as we see in the various occupations and professions where individuals are taught to handle emotion-producing situations with a specific technique, such as soldiers, firemen, butchers, undertakers, surgeons, nurses, and so on. What we call the secularization of life (or progress) is just this development of techniques for meeting situations which previously evoked reactions of an emotional type. .. .
[Problems arise also when upon weaning the] child will have to learn to use auditory and visual substitutes for the tactual intimacies of infancy, and thereby will be directed to the world of things and people around him, wherein his increasing strength and mobility favor an ever widening exploration. Thus he gradually learns to use these new opportunities for achieving adjustment and so begins the process of his socialization.
We do not enumerate the further steps in " tensional management " which, according to Frank's description, are involved in " the approach to the person and property of others," etc., but in this statement we have a research program in terms of behavior reactions in specific situations, where the effect on behavior formation is measurable when taken com-
( 565) -paratively. Experimientation and observation (with the use of control groups) would determine the situations favorable to the learning of tensional management.
We have seen that this approach has as yet had a very limited application to human society. The child welfare institutes are working most systematically, the sociologists have made a beginning and we have noted certain important specific projects (Freeman, Richards, Healy and Bronner, etc.). Taken in connection with the whole body of problems and programs which we have been considering this approach suggests certain applications, limitations and extensions, which we shall now consider.
The aim of scientific research is to determine that under certain conditions certain results will follow in certain proportions. We have pointed out above that the student of human behavior is not able to set up a situation in which there is a sufficient degree of control to produce true experimental methods of the type of those in the chemical laboratory. He has not been able to hold other factors constant, while he measures the influence of the variation of some particular factor, and everywhere the complications of the data have led to difficulties in the way of objective analysis. The approaches that have been made from the morphological, physiological and psychometric points of view have attempted to isolate some specific part of the human being from the behavior complex and to relate this specific part to the total remaining part. The isolated factor may be quite simple, as, for example, some product of body metabolism such as the hydrogen ion concentration in the saliva, or quite complicated, as abstract intelligence or mechanical abilities. The measurement of a specific factor, if it be simple enough, can often be done with great accuracy. The investigation of its relationship to other behavior variables, however, becomes a very complicated matter.
Where the total situation is so complicated, the interrelations so numerous and measurement so necessary, the method will evidently be very intimately related to statistical procedure. Although it is impossible to set up real experimental control for the solution of a problem, if groups of individuals roughly similar in a large number of attributes can be studied in varying situations the specific type of behavior resulting may be compared, statistically, for the different situations and inferences drawn as to the relative effects of the situations on the behavior.
( 566) A study of this sort may often give results that arc very good approximations to the experimental type of situation. This is well illustrated in Freeman's study of foster children (p. 342). A group of children whose IQ's were known at the time of placement had been taken from an environment lacking in certain respects and placed in an environment enriched in these respects. A number of their siblings had been left in the old environment. A number of unrelated foster siblings were brought up with them in the new environment. Inferences were drawn on the basis of the statistical relationship found between the IQ's of this group of children and their own siblings as compared with the relationship found between their IQ's and those of their foster siblings. These data were then referred to the relationship that had been found to exist in other groups of siblings brought up in the same environment. The difficulty of interpreting a study of this kind arises, of course, from the fact that the situation that was allowed to vary was of so complicated a nature. The factors that make for a " good " or a " bad " environment may be largely a matter of judgment and it is impossible to measure an " environment " completely. A more narrowly defined study of the same kind is that of the effect of nursery school training on IQ variability. Here children are selected on the basis of a large number of likenesses in environment and the major difference then becomes nursery school training. The changes in IQ over a period of time for the two groups are compared.
A converse application of statistics in lieu of experimentation is seen in studies of the criminal. Here the problem is to find by how many measurable qualities the criminal is differentiated from those who do not commit crimes. Groups of criminals are matched with groups of non-criminals in certain respects and the significance of any differences found in other measurable respects is determined statistically (Goring). Of course, in all such studies clear-cut, definite results are seldom obtained because of the complexity of what is being measured, and because of the large part that unmeasurable factors play in proportion to those that can be measured. For instance, in the study of the criminal the most adequate approach that has been made, from the point of view of the use of statistics as leading to a situation that approximates experimentation, is a study that compares a group of young delinquents in institutions with boys of the same age, social class, etc., who have not become delinquent. Both groups were given intelligence tests, tests for me-
(567) -chanical aptitudes, tests to determine psychoneurotic responses, and data were collected as to nationality, occupation of parents, size of family, room space per person in the home (Slawson). But it must be obvious that very important aspects of the environment are probably not touched by these measurements, and likewise important aspects of personality make-up are not included. A study of this sort may be quite objective, give verifiable results and lead to guarded and careful inferences as to factors important in the etiology of crime, but it will very probably give a quite inadequate basis for the understanding of crime. In any interpretative study, by selecting out only those factors which are at the moment capable of quantitative expression, there is a necessary overweighting of those factors as against factors not readily expressed quantitatively. With regard to the factors measured, provided they are in a comparable form, an estimate may be made of their relative importance as compared with each other. That is, relationships within the group of measured factors will be accurately defined, but factors which cannot be measured readily (or at all) will receive no attention.
In some of the approaches which we have studied this premature quantification of the data is quite obvious, as in the studies that depend on ratings of traits. A large number of persons will be judged on a trait, say, aggressiveness. This may be done by ranking them from the most aggressive to the least aggressive, assigning numerical values for degrees of aggressiveness, or describing, in behavior terms, " degrees " of aggression thought to be equal distances apart, having the judgments made in behavior terms and then assigning values. These methods are all full of pitfalls. The instrument of measurement, i.e., a human " judge," is erroneous and inconsistent. It is never certain how much of the judge and how much of the subject appears in the actual judgment or " measurement." Tabulations resulting from these judgments are difficult to interpret. Statistical manipulations of the data, the application of complicated methods which have definite meaning only when applied to data of a strictly defined and limited character, are absurd. (A good deal of the data in mental measurements, a large part of that in personality and temperament measurement and most of that in psychiatric measurement has this fault in greater or less degree).
Another difficulty often found in these investigations is that one part of the problem can be measured directly, in genuinely quantitative
( 568) terms, but in the comparisons and correlations that must be made other parts of the problem will be of this pseudo-quantitative sort. For example, a perfectly objective study of individual differences in speed of handwriting may be made. The investigator, however, wishes to study not speed of handwriting, which seems to have little general importance, but general " speed of movement " and " speed of decision " which seem to be weighted with great social significance. Instead, however, of making studies which would show the various relationships of various sorts of speed in the same individuals (which would probably take years of work) this handwriting test will be called a test of " speed of movement," and an attempt will be made to validate it by correlating it with the judgments various people may make of the speed of movement and decision of the individuals taking the tests. This is a short-cut method, based on the assumption that judgments will be made on the basis of recalling observations of " speed of movement " in a large number of natural situations. Assuming that the problem was a good one, that there is a relation between speed of handwriting and speed of decision (which may be doubted) the investigator wants to prove too much in too little time and the results will have little scientific value.
Another variety of this procedure is brought about by an oversimplification of the problems of human behavior. This is frequently seen in persons who have approached the field from another field where high scientific standards prevail, say, from the field of biochemistry. The investigator may have worked out and applied an accurate method of determining certain biochemical states. He assumes (rightly) that personality, temperament, the emotions, etc., have a biochemical base and wishes to work out the relationship of his very accurately determined biochemical index to personality or temperament or " behavior." And in study after study we find him accepting subjective, grossly inconsistent " ratings " of the personality and behavior factors, correlating them with his very accurately defined index and giving interpretations that seem to assume an equally scientific basis for both of the correlatives. This procedure is probably partly due to the na´vetÚ often found in research workers in their approach to the problems of fields other than their own, particularly the more intangible social and psychic fields, but it is partly due also to the lateness of development of objective studies in these fields so that the choice for the investigator from another field is this imperfect sort of correlation or none at all.
Among the approaches which we have reviewed, the psychometric has had the advantage over the other~ in that it has had two full decades of development during which interest and stimulation in the field have been intense. Not only has the very wide application of intelligence tests resulted in a standardisation and development of norms of performance, but it has led to the accumulation of many valuable data on the concomitants of performance on these tests. These concomitants have been found to be occupational, geographical, " cultural " (in the sense of superior material environment), racial and educational (in the sense of the amount of schooling received). This wide application of the tests and the definite knowledge of so many of the concomitants of the results have pointed the way for the development of controls and for the study of probable causation. This has led to the possibility of determining the effect of certain of these concomitant factors on variations in others — notably studies such as Freeman's, where the effect of changed environment on the IQ variability was tested. It has also led to important studies of IQ in relation to delinquency, whereby control has been exercised by equalizing delinquent and non-delinquent groups for certain of these concomitant factors. It will be seen also that so long as the psychometrists clung to the idea that they were dealing with " original nature " in their test results, and that original nature was unmodifiable, that the responses were concomitants of differing original natures, and that being unmodifiable (or so only within narrow limits) original nature was producing the differing responses, very little good behavior material was evolved. When, however, they utilized this knowledge of concomitants to set up controlled experiments, put aside for a moment the question of the urmodifiability of original nature and changed individuals about from one situation to another and recorded the actual changes in their intellectual behavior, interesting light was thrown on " original nature " by the situational study.
This methodological scheme, comprising the development of norms to which deviating groups may be referred, and the study of concomitant factors, has been best worked out and applied in the psychometric field. This means that, for the behaviorists, the materials of the psychometrist are in a much more usable shape than are those of the other fields, although studies of cases, life-histories, etc., indicate that these other fields probably play a more important part in the determination of behavior than does the psychometric.
It is desirable, therefore, that other behavior fields should analyze their materials in terms of the situation. In the personality and psychiatric fields, for example, the difficulty has been that most of the studies have been made from the point of view of the inner life outward, i.e., rather than studying behavior in a variety of situations as a means of inferring drives, instincts, emotions, etc., the instincts, emotions, etc., have been assumed to have a reality of their own and behavior has been studied in terms of them. There has been a tendency to pre-determine what " types " of reactions a set-up would bring out, and obscurity has been the general result. The really fruitful studies have been those that have been based on widespread observation and objective recording of behavior in varying situations, and it is this type of study that leads to the possibility of the development of controls. That is, reactions are first studied in the more " natural " situations and the factors involved in, and concomitant with, these situations are brought out in the behavior study. Then more controlled situations can be evolved which will allow for and rule out as many of the concomitant interfering factors as possible. Through studies of this sort we learn how people behave and from them we can then infer why people behave as they do.
The studies that are going on in the child welfare research institutes are particularly valuable in this respect because the nursery school represents a variety of situations in which social interaction may be studied, and because the early age level makes possible the study of the beginnings of certain behavior patterns as they are related to the situations conditioning them. It is obvious that there are great gaps in all these studies: children at the earliest age levels are not even yet being studied, from this point of view, or not being studied widely enough to give valuable results. They are not being followed through over long enough periods of time. The " situational " studies are incomplete and uneven — it has been possible to make too few studies of the home situations, for instance.
We are of the opinion that verification, through statistics, is an important process in most of the fields of the study of human behavior. Relationships can be indicated, various processes can be evaluated, if the data are in a form where statistical methods may legitimately be applied, and if the interpretations keep within the limitation of the assumptions on which the methods were based. Probably the greatest distrust of statistics has come through the unwise manipulations of data
( 571) that are often made, through the expression in terms; of great precision of results obtained when complicated formulae are applied to very inexact data, and through the totally erroneous assumption on the part of many statisticians that the statistical results tell all that. Call he told about the subject.
What is needed is continual and detailed study of case-histories and life-histories of young delinquents along with the available statistical studies, to be used as a basis for the inferences drawn. And these inferences in turn must be continually subjected to further statistical analysis as it becomes possible to transmute more factors into quantitative form. Statistics becomes, then, the continuous process of verification. As it becomes possible to transmute more and more data to a quantitative form and apply statistical methods, our inferences will become more probable and have a sounder basis. But the statistical results must always be interpreted in the configuration of the as-yet unmeasured factors and the hypotheses emerging from the study of cases must, whenever possible, be verified statistically.
The behavior document (case study, life-record, psychoanalytic confession) represents a continuity of experience in life situations. In a good record of this kind we are able to view the behavior reactions in the various situations, the emergence of personality traits, the determination of concrete acts and the formation of life policies, in their evolution. Perhaps the greatest importance of the behavior document is the opportunity it affords to observe the attitudes of other persons as behavior-forming influences, since the most important situations in the development of personality are the attitudes and values of other persons. In the document prepared by Shaw (p. 97) we have seen the determination of behavior partly by institutions, taken as situation, and partly by behavior of others, taken as situation.
It has been strongly objected, especially by the adherents of the school of " behaviorism," that this introspective method has no objectivity or validity. What they mean is that these records will not reveal the mechanisms of behavior, the process of consciousness, what is going on inside of us when we think and act, and with this we are in agreement. But the unique value of the document is its revelation of the situations which have conditioned the behavior, and concerning this there can be no doubt.
There may be, and is, doubt as to the objectivity and veracity of the
( 572) record, but even the highly subjective record has a value for behavior study. A document prepared by one compensating for a feeling of inferiority or elaborating a delusion of persecution is as far as possible from objective reality, but the subject's view of the situation, how he regards it, may be the most important element for interpretation. For his immediate behavior is closely related to his definition of the situation, which may be in terms of objective reality or in terms of a subjective appreciation —"as if " it were so. Very often it is the wide discrepancy between the situation as it seems to others and the situation as it seems to the individual that brings about the overt behavior difficulty. To take an extreme example, the warden of Dannemora prison recently refused to honor the order of the court to send an inmate outside the prison walls for some specific purpose. He excused himself on the ground that the man was too dangerous. He had killed several persons who had the unfortunate habit of talking to themselves on the street. From the movement of their lips he imagined that they were calling him vile names, and he behaved as if this were true. If men define situations as real, they are real in their consequences.
The total situation will always contain more and less subjective factors, and the behavior reaction can be studied only in connection with the whole context, i.e., the situation as it exists in verifiable, objective terms, and as it has seemed to exist in terms of the interested persons. Thus, the behavior records of the child clinics are contributing important data by including the child's account of the difficult situation, the often conflicting definitions of this situation given by parents, teachers, etc., and the recording of such facts as can be verified about the situation by disinterested investigators.
In the field of psychiatry the context becomes particularly significant, and it is desirable to have here a multiplication of records showing how situations are appreciated and motivate behavior, but the records should be made not without regard to the factual elements in the situation. To the degree that the psychiatric cases are approached from the standpoint of the total situation it will appear that the problems of behavior taken all together assume an aspect of totality. The unfortunate separation of the " abnormal " from the " normal " in behavior studies will disappear, and the abnormal, pathological and criminal behavior reactions will appear not as " disease " but as socially (and individually) undesirable behavior reactions in given situations, and from this standpoint
( 573) they will lend themselves more readily to study from the behavioristic standpoint.
The situational approach, utilizing statistical methods and the life-record, is capable of throwing light on many problems whose etiology remains obscure. For example, in the literature of delinquency we find under the heading "causative factors" such items as the following: Early sex experience, 18% for boys and 25% for girls; bad companionship, 62% for both sexes; school dissatisfaction, 9% for boys and 2 % for girls; mental defect, 14%; premature puberty, 3%; psychopathic personality, 14%; mental conflict, 6.5%; motion pictures, 2 %, etc. Now it is evident that many young persons have had some of these experiences without becoming delinquent, and that many mentally defective persons and psychopathic personalities are living at large somewhat successfully without any record of delinquency; some of them are keeping small shops; others are producing literature and art. How can we call certain experiences " causative factors " in a delinquent group when we do not know the frequency of the same factors in a nondelinquent group? In order to determine the relation of a given experience to delinquency it would be necessary to compare the frequency of the same experience in the delinquent group and in a group representing the general non-delinquent population. It is obviously absurd to claim that feeblemindedness or psychopathic disposition is the cause of crime so long as we have no idea of the prevalence of these traits in the general population. Similarly, the Oedipus complex (mother fixation) and Electra complex (father fixation) are weighted by the Freudians and made prominent sources of the psychoneuroses and of delinquency, whereas the clinical records show a multitude of cases where children with behavior disturbances are either indifferent to the parents or directly hate them. Again, with regard to economic factors as cause of crime we find, for example, in the records of the White-Williams Foundation of Philadelphia (an organization dealing primarily with nondelinquent children) the same unfavorable economic conditions, broken homes, etc., which are usually assigned as " causative factors " in the studies of delinquency, but in this case without delinquency. The simple expedient of using a control group would aid in clarifying this question of causative factors.
With the same utilization from the situational standpoint of statistical method, life-record and control groups, we can study with advantage the
( 574) various projects and programs which we have reviewed in these pages, measure their influence and test their claims — juvenile courts, boys' clubs, boy and girl scouts, parent education associations, social agencies, recreations; also the family, the gang, the daily press, commercialized pleasure, etc.
As a more specific example, we have seen that the school is tending to assume responsibility for the "whole child," to be converted, at least for the lower age levels, into a behavior-forming situation. It is taking over the activities formerly carried on by other organizations and performing them better. We have described the varieties of these school programs —those originating externally, as in the child guidance clinics, the projects of some of the community organizations, etc., those originating within the school, as directed towards curriculum and activity changes, the introduction of behavior specialists, and the participation of the teachers in behavior-modification schemes — and it would be possible by a series of surveys to study their efficiency comparatively, let us say, with reference to maladjustment and delinquency.
In the preceding chapter we have seen that the sociologists have made territorial surveys and have determined localities in cities which evidently contain characteristic behavior-forming influences. Their work up to the present has amounted to a preliminary definition of the situation. They know, for example, that in one local area 37% of all boys of juvenile court age have been in the juvenile court (or before the police) and in another area not 1% of the boys. Their further task is to measure, with the aid of personality documents and control groups, the specific behavior-forming situations within these areas.
We have seen also that cities and larger localities differ among themselves as behavior-forming situations. Healy and Bronner estimated, for example, that their failures in Chicago were 50% and in Boston only 21 %. The difference is certainly not due in the main to differences in juvenile court procedure, but to differences in the attitudes of the population, and this in turn to differences in the configurations of social influence. We examined in some detail the city of Cincinnati from this standpoint, and have been impressed with Rochester as a characteristic behavior-forming situation, particularly with the prominence of the school in the total configuration of influences. It is desirable that these cities, or others, should be studied comparatively with the aid of the same technique.
With the progress of our studies of (he various behavior-forming situations we may hope to approach the still more obscure problem of mass behavior — the participation of whole populations in common sentiments and actions. This is represented by fashions of dress, mob action, war hysteria, the gang spirit, mafia, omertÓ, fascism, popularity of this or that cigarette or tooth paste, the quick fame and quick infamy of political personalities, etc. We are unable to define this total situation satisfactorily, but it involves the interaction of language and gesture and gossip and print and symbols and slogans and propaganda and imitation, and seems, more than anything else, the process eventuating in the formation of the distinctive character of communities, nationalities and races. The process itself may be described as a series of definitions of situations whereby behavior norms are established.
In the same connection (while we do not advocate anthropological and historical studies as remunerative behavioristic studies in themselves and are of the opinion that the past contains no models on which we may build in the present) it would be useful to extend our studies of this situational character to the large cultural areas, to the contemporaneous races and nationalities, in order to understand the formation of behavior patterns comparatively, in their most general and particular expressions, and appreciate the capacity of human nature to work under various and widely contrasted habit systems. Furthermore, behavior studies within these wide limits may be expected to reveal comparatively and in the most general way the situations within which particular maladjustments (delinquency, crime, the psychoneuroses) tend to appear, and the situations and habit systems unfavorable to their appearance, or, more positively, the situations within which the activities are integrated about particular interests, leading to pursuits, r˘les and careers.