WW ~ #— ___—— — —_’——— —_’— —__.—— _—_—— — __._——— #4— ___—— I — ~—_— __’__’—— ______.—. __.__’ ’_— ‘_——— ”—4 __—— —_————— _—_'— __'_.—— _—_.__4- _——_ ._____—— ______—— F———‘ ___’_— MC‘MS Ii)? {F‘Q'E’RAJEROUP é‘NTERACTIO-N UNKSER CGN’L‘RTEQNSa 0F SURVEILLANCE. ANS? ANQNYMITY Thesfis fa? Hm bagraa c? M. 6.. MtCfiK‘éAN STM‘E U‘N§‘s’ERSiTY :‘V‘iawin Di. Laavy Wfifii SSSSSS WWW ~, E j " * 5:51:33“ 31 try ROOM USE ONLY ‘ ‘ V '2 A ‘ ‘ ' ‘94. - in glut; ‘4 '~ .3 arr-1" ABSTRACT MODES OF INTRA-GROUP INTERACTION UNDER CON- DITIONS OF SURVEILLANCE AND ANONYMITY BY MARV IN D. LEAVY Submitted to the College of Social Science of Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Sociology and Anthropology 1964 Marvin D. Leavy ABSTRACT The main purpose of this thesis is to present and to explore some research implications of symbolic interaction theory for the study of collective (crowd) behavior. This study attempts to see Whether this standpoint regarding human interaction is heuristically viable and promising as a con- ceptual approach to the analysis of intra-group and intra- crowd relationships. If the research carried out here pro- duced findings in expected directions given our control over small groups in an experimental setting, then the potential of this approach would be enhanced. Being in the “context of discovery," the study's general focus is upon the effects of an experimentally con- trolled condition of surveillance upon aspects of interaction observed in a small-group laboratory setting. Face-to-face interaction as characterized by George Herbert Mead stresses certain qualities, prime among them being (a) reciprocity in taking the roles of others into account, (b) sharedness of norms structuring the situation, (c) perceived.mutual sur- veillance, (d) interpersonal evaluation, and (e) a concern for particular others‘ remarks based on proficiency in recall- ing and identifying them. In ideal-typical contradictinction to this mode of small group interaction is a mode of interpersonal relation- ship drawn from a modified natural history approach to crowd l Marvin D. Leavy 2 behavior. The characteristics listed above as profiling small group interaction would fall at low levels within crowds. It is posited that perceived feelings of surveillance or of anonymity will play critically differentiating roles in whether small five-person laboratory groups (in whiCh condi- tions of normative restraint reduction are equivalently in- duced) are activated into collectivities marked by small group or by crowd characteristics. Three general hypotheses were formulated and tested to determine if the vantage point explicated would appear to be of promise: 1. There is a positive intercorrelation between reciprocity, norm-sharedness, perceived surveillance, per- ceived evaluation, and concern with particular others in a small gathering of persons. Small gatherings of persons will tend to cluster at high levels or at low levels of these in- dependently conceived variables in situations of ex peri- mentally induced normative restraint reduction. 2. A higher level of engrossment (i.e., an inability to recollect and identify particular remarks theoretically correspondin to a heightening of attention in a focal object of attention will obtain in gatherings under the condition of anonymity than in gatherings under the experimental condi- tion of surveillance. This finding will obtain only if there is a higher level of perceived anonymity among "groups“ in the anonymity condition. 3. The engrossment level of groups under study will associate most closely with crowds i.e., with groups evidenc- ing low levels of reciprocity, norm-sharedness, perceived surveillance, perceived mutual evaluation, and concern with particular others. These hypotheses were tested in a small group laboratory setting, research design and methodological procedures derived but revised somewhat from a study conducted by Festinger, Pepitone, and.Newcomb. The variables and their relationships Marvin D. Leavy 3 were conceptualized and Operationalized. Pearson Product- moment correlations between dimensions of intra-group relationships did not indicate a clustering of modes of relationship; the t-score difference between the Engrossment score means for the anonymity and surveillance groups reSpect- ively was tabulated, compared, and found non-significant (perceived anonymity was not significantly higher in the anonymity condition than under surveillance): and the rank— order correlations between engrossment and levels of crowd- like relationships between “group members“ were computed and found to be either non-significant or contrary to expectations. Possible interpretations of these findings were con- sidered and the implications of the research endeavor for future investigations suggested. MODES OF INTRA-GROUP INTERACTION UNDER CON- DITIONS OF SURVEILLANCE AND ANONYMITY BY MARVIN D. LEAVY A THESIS Submitted to the College of Social Science of Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Sociology and Anthropology 1964 To Bob Potter - who first encouraged me to think along sociological lines. TABLE OF CONTENTS Chapter I. THEORETICfAL POSITIOI-q O O O O O O O O O O O O A. The Symbolic Framework of Human Interaction. B. The Natural History Approach to Collective Behavior. C. Two patterns of Intra-group Relationship A Reformulation. D. Normative Control. E. Surveillance. II. REVIEW OF RELEVANT RESEARCH . . . . . . . . A. The small Group Literature. B. The Study of Festinger, Pepitone, and Newcomb. III. THE DESIGN OF RESEARCH . . . . . . . . . . . A. Conceptual EXpectations. B. The Logic of Proof C. Methodology. D. Operational Hypotheses. IV. PROCEDURES OF THE RESEARCH . . . . . . . . . A. The Anonymity Condition. B. The Surveillance Condition. V. BEHAVIORAL DATA AND ITS ANALYSIS . . . . . . A. The Findings. B. Statistical Interpretations. VI. INTERPRETATION AND IMPLICATIONS . . . . . . A. Substantive Interpretation. B. Consequences for Future Research. APPENDIX A: Post-Session Questionnaire . . . . . . . BIBLIOGRAPHY iii Page 17 25 36 44 51 Table I. II. III. LIST OF TABLES Mean score levels of Dimensions of Intra- Group Interaction by Experimental and Control Groups . . . . . . . . . . . . . . . Correlation Matrix of Dimensions of Intra- Group Interaction . . . . . . . . . . . . . . Mean Number of Identification Errors per Group 0 O O O O O O O O O O O O O O O O O 0 Mean Engrossment Scores . . . . . . . . . . Rank-Order Correlations of Engrossment Level and Dimensions of Intra-Group Interaction. . iv Page 44 48 48 50 I. THEORETICAL POSITION George Herbert Mead's delineation of the twin-born processes of mind and self within social milieux has been a valuable and influential framework with which to account for and to systematize the social develOpment of humans.1 Mead stressed Speech and significant symbols in this develOpment. His pragmatic, process-stressed conceptual scheme has served as the groundwork for much, if not the plurality, of present- day social psychological thinking regarding the nature of the ties of interdependence uniting persons in a society.2 As the focus of this thesis is upon the dimensions of relationships among humans and their attendant behavior, it is not neces- sary to trace the emerging mind and self processes so notably discussed elsewhere.3 lMead, G. H. Mind, Selngand SocietyL Charles Morris (ed.), (ChicagozUniversity of Chicago Press, 1934). 2Mead's work is the fountainhead of the symbolic inter- action position. His continued and widespread impact may be found in the works of Blumer (see particularly his ”Society as Symbolic Interactionism“ in Arnold Rose (ed.), Human Nature and Social Processesy (Boston: HoughtoneMifflin Co., 1962)—and TrSocial Psychology” in E.P. Schmidt (ed.), Man and Society (N.Y.: Prentice-Hall, 1937). The textbooks of Lindesmith and Strauss, Social;gsycholoqy (N.Y.: Holt, Rinehart and Winston, 1956): Tamotsu Shibutoni, Society and Personality (Englewood Cliffs, N.J.: Prentice-Hall, 1962) reflect this impact as the works of such social psychologists as Kuhn, Kinch, Goffman, Bolton, and Rose. See footnote 42, 43, and 44. OMead, 9p. cit. Parts II and III: also Ansebn Strauss (ed.),zhg Social Psychology of George Berbert Mead (Chicago, Ill.:University of Chicago Press, 1956) and particularly C. PSfeutze, The Social Self (N.Y.:Bookman Associates, 1954) for cogent tracings of this develOpment. 2 The processes of interaction among socialized humans are integrated into a mode or sequence of activity Mead called the social act.4 The association of men as Characterized in the social act involves men reciprocally “reSponding” to the actively interpreted meanings of others; not as reSponding to naturally-eliciting stimuli or signs triggering the re- lease of reaponses.5 Imaginative activity, then, is of paramount importance in human interaction. Moreover, the ability develOped by humans to reSpond to their own gestures (through role-taking) enables men to share one another's experience. Therefore, behavior is not simply social when it is a response to others but rather when it has incorporated within it the behavior of others.6 The human being comes to respond to himself as he imagines others reSponding to him, and in so doing he shares the conduct of others. These last few ideas may be summarized by asserting that an individual may act socially toward himself, dust as he imagines others acting toward him. From this symbolic interaction position, this is tantamount to saying that men develOp and possess selves. The emergence of self-awareness, self-indication, IMead,‘gp.lg;§. See also Bernard.Meltzer, The Social Psychology of Geggge Herbert Mead (Kalamazoo, Mich., Division of Field Services of Western Michigan University, 1959) for a succinct presentation of these stages. 5The best elaboration of this point is found in Herbert Blumer, “Attitudes and the Social Act,“ Social Problems, III (1955): pp. 59-65. Also see his "Society as Symbolic Interaction," in Arnold Rose (ed.), Human Nature and Social Prgcess (Boston: HoughtoneMifflin & Co., 1962). 6Meltzer, Op, cit., p. 14. 3 and consequently self-control are critical phases in man's constructions of his behavior. This position of Mead and his elaborators (most notably Blumer) as to the characterization of face-to-face encounters stresses certain qualities of that interaction. These dis- tinctive features constitute a model type of relationship be- tween reciprocal “selves,” eadh viewing an “other,“ each possessing an image of that “other“ as well as of one's"self,“ and each communicating in terms of those images and the mean- ings imputed to the “other's” actions. Humans acting in this manner are behaving as social beings. The main stream of symbolic interaction theory has maintained a bi-polar ideal typology of intra-group relation- ship of individuals.7 The first mode of interaction is typified by the social act as outlined above. One prototype of this mode of interaction would be the customary small group setting. The second mode of relationship between individuals de—emphasizes the processes of the reciprocal incorporation of the “other” into one's acts and its prototype is the “acting crowd.“ This characterization of crowds compared with 7Actually, a third type of relationship between per- sons that distinguishes a collectivity has been given atten- tion to - the mass. However, this mode is hardly a web of relationships at all: persons perform separately: the privacy of one's actions is reinforced by a perceived leak of inepec- tion of one's behavior. The characteristics of mass relation- ships are not at issue here. 4 small groups in terms of the relationships‘between persons grew out of a natural history approach to collective behavior.8 The natural history approach attempts to identify the forms of interaction in unstructured large groups, to establish a generic classification (taxonomy) of collectivities based on the mode of interpersonal relation acted out therein and to isolate analytically the natural line of formation of these collectivities.9 All three of these purposes are touched upon to some extent in this thesis. Concern with the cri- teria distinguishing and circumscribing collective behavior from other forms of interaction is evident throughout the literature, in fact never seems far from the forefront of 10 attention. Not all adherents to the natural history tradition 8The Natural History Approach'has been so labeled.by Kurt and Gladys Lang in their discussion of viewpoints regard- ing collective behavior Collective Dynamics (New‘York: Thomas Crowell Co., 1961). This approach is classically represented in the thought of Herbert Blumer in ”Collective Behavior“ in A.McC. Lee (ed.), New Outline of the Princi les of Sociolo (N.Y.: Barnes and Noble, 1961), pp. 167-224. Another modern representative of this approach is found in Turner and.Killion, Collective Behavior (Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1957). The chief rubric uniting this approach is a non- evaluative (neither progressively nor pathological) value orientation toward collective phenomena. Also see this author's The Curgent Scientific Status of Thggry_in Collective Behavior, mimeograph, 196 3. 9H. Blumer. “Collective Behavior,“ in E. Gittler (ed), Review of Sociolggy: Analysis of a Decade (N. Y. iJohn Wiley, 1957). pp. 130-131. loBrown reviews this boundary-making proclivity in ”Mass Phenomena“ Handbook of Social Psychology, Gardner Lendzey (ed. ), Vol. ,(Reading, Mass. : Addison-Wesley, 1954),pp.833-877. Smelser hypercritically talks on these concerns in the opening chapter of his Theory of Collective Behavior (N. Y.: The Free Press of Glencoe, 1963). We shall not treat the various writers' criteria for differentiating types of crowds. 5 share the same criteria for differentiating the two broad 11 classes of human interplay. Nevertheless, a general picture of crowd behavior can be presented that does justice to all adherents of the approach. Crowds are gatherings of indivi- duals which are not institutionally sanctioned, focus upon a common object, and their behavior has been variously termed volatile, transitory, and spontaneous.12 "Social unrest" is the precipitant factor in the emergence of crowd behavior. Social unrest is described phenomenologically as urges to act in the absence of specific goals, leading to aimless, random behavioral states of frustrated anxiety: excited feelings of vague apprehension, alarm, or irritability compounded by increasing suggestibility. The upset of normative social restraints among persons leads to “circular reaction“ which becomes the new’mode of relation between them. The first indicator of circular reaction is “milling“ (a pure circular reaction): a more intense form is ”collective excitement”: and finally, “social contagion” which Blumer depicts as a rapid, unwitting, and non-rational 11Blumer advocates distinguishing collective behavior from small group behavior, on the one hand, and from cultural- ly prescribed behavior, on the other. His distinction between small group interaction and crowd relationships is eSpecially pertinent to this thesis. The two are distinguished, for Blumer, by psychological characteristics, communicative pattern, and patterns of mobilization for action. See Collective Behavior”in Gittler (ed.), pp. 127-135. Turner and Killian 22; gig. is similar but Lang and Lang, 92; gi£., stress the transformations of crowd processes which differ from other groupings. 12Lang and Lang. "Processes of Collective Dynamics,‘ in Arnold Rose (ed.), Op. cit., pp. 343-345. 6 dissemination of impulses stimulated beyond the prior two indicators of "circular reaction.“ The new'mode of relation between persons is one in which individuals are less self- aware and less self-controlled. The acting crowd (one of four types of elementary collective groupings for Blumer) is described as Spontaneous, impulsive, non-traditional, and therefore not bound to cultural expectations. Individuals in such crowds, for Blumer, are "emotional,“ less personally sensitive, more uninhibited as to claims of privacy, and the aggregates they form are relatively unstructured and amorphous. From this interpretation, it is clear that the acting crowd is marked by a mode of relationship between persons Operating at a less symbolic level than the mode of relationships in small groups.1 3 This bleak schism is inveighed against heartily. Crowd-like behavior is not completely alien from the prin- ciples underlying “normal“ group behavior. The grounds for the distinctions made do not do justice to the continuous .nature of the variables which might more realistically depict and separate the one class of behavior from the other. It is argued that the same dimensions may be employed to conceive each mode of intra-group relationship. It is suggested that a false dichotomy has been built up of stereotypical concepts depicting, on the one hand, 13This paraphrase is adapted from Blumer's "Collective thavior” in Lee, Basic Principles. . . 7 organized and institutionalized behavior and, on the other hand, non-institutionalized collective behavior. For example, crowds are often generally labeled as suggestible. If this is true, why is it so difficult to break up a crowd intent on some aim? Revolutionary mobs do not immediately follow orders to take out their pent-up tension in a cross-country bicycle race. Crowds can not be considered as generically destructive either. In terms of the maintenance of cultural values which solidify communities of peOple, vigilante mobs and lynching parties are not destructive but are manifestly integrative in their evasion of institutionalized justice. Bizarre, irrational, and ultra-emotional behavior of persons also can not be defining characteristics of crowds. These psychological symptoms are not peculiar to persons in crowd gatherings as any witness to a family argument will attest. Crowds are thought by many to be the upshot of no rational calculation, i.e., form Spontaneously with no leaders pre— cipitating or guiding the course of action. This is blatant- ly at odds with reality. If crowd behavior is so unstructured institutionally then crowd participants all the more seek cues as to how to organize their behavior. Leaders serve this function. Perhaps suggestibility (see above) has been attri- buted to crowds for the very reason that persons in ambiguous, situations so desperately seek some suggestion as to a definition of the situation and a plan of action. It can not be held that only “lower-class” or "trashy" elements of the pepulation are susceptible to crowd activation. This 8 imputation reflects the low regard that in general has been attached to crowd phenomena usually to the extent that crowd behavior is considered as deviant and not as integral to societal maintenance. The natural history tradition can not be charged with this emotive bias. However, suggestibility, destructiveness, wild and hallucinatory behavior, Spontaneity, and recruitment of less intelligent peOple have all on occasion been ascribed to crowd activity as incidental characteristics by prOponents of this approach. Underlying these imputed attri- butes is the presumption that, within crowds, self-control is diminished and that calculated, sypholicly formulated action tails off considerably. In this thesis, these pseudo- criteria of crowds are abandoned. Rather, new discriminatory indices are needed which measure differences in types of joint human endeavor. Con- ceptual and quantitative dimensions are needed to understand the mechanisms which underly all human group behavior. It will be far more productive to focus upon the dimensions characterizing relationships between persons in gatherings to see which clusters of scaled qualities mark off interpersonal relations in small-group settings which contrast with those. in crowd settings. The dimensions range on continua Which at their extremes should be the defining characteristics of each mode of relationship. In the next section, such a theoretical reconception is presented. The dimensions of (a) reciprocity (in taking the role of the other), (b) sharedness of norms, (c) perceived surveillance, (d) interpersonal evaluation and 9 (e) a concern with particular other's statements are related and their covarying presence is schematized as the key differ- ence between small group and crowd relationships. What are some characteristics of interaction which are critical to its continuation or stability? In the small group, persons communicate responsively and re3ponsibly with other persons, each person presuming the other to possess a self formed much as his own. Behavior is reciprocal i.e., ideally one's Speech calls out in the "other“ the impulses to act (the meaning of one's speech) that it has for oneself. Their behavior is role—oriented: each is taking the role of the other into account in enacting one's “own“ role. In be- ing aware of the status or rank of the “other“ as well as other facets known of the ”other's“ personality, a person sensitively monitors and interprets the Speech of that parti- cular person in terms of the actions and intentions (past, present, or future) attributed to him. Humans in interaction are evaluating and assessing each “other" as well as the contents (meanings) of “other's“ statements. Explicit and implicit norms are reciprocally held which order and constrain their behavior. Knowledge of the norms is shared even if not Jointly adhered to. Because of continuous symbolic imagery and self-indication of events, humans can modify or adjust their behavior as interaction proceeds in alignment with others. Humans then take into account the differentially important behavior of others in a small group that permits face-to-face interaction and construct their comments in light 10 of the way in which they expect or wish particular others to act or react. Moreover, they can adjust their conceptions and their behavior in accord with redefined situations and redefined self-images. In sum: in the prototypical small group, behavior is reciprocal, role-oriented, evaluated and appraised as are the behavers: norms are shared: and attention and concern is given to “other" in order to continually con- struct one's actions. In contradistinction to this mode of symbolic inter- action in the group is a mode of interpersonal behavior dis- cussion of which can be based on the less pronounced presence of the same dimensions. This mode is a type of relationship manifested by persons in “acting“ crowds. The crowd mode is one characterized by a kind of intra-group relationship in which the persons are engpossed, not so much in one another nor in the formulation of their own remarks, as in a focal 14 The persons are very object of attention or an objective. much in symbolic contact but the nature of their interrelation is such that the individuals are not reciprocally taking any particular other into account in order to coordinate action. No other person in the collectivity is being singled out for special attention. The focus of attention demands the concern 14Turner and Killion, op. cit., eSpecially stress the “magnetizing" object of attention. pp. 83-87. ll of each in a parallel co-acting fashion rather than it being a “social object”15 exacting a joint interpretation and a coordinated line of behavior. Also, there is less assessment of each other's distinctive actions. Since there is less self-indication, less concern and involvement with "others," and less assessment of others, there is correSpondingly less deliberate adjustment in the gestures or actions made. Behavior,once started, tends to be consummated unless impeded, not upon reflection, but by external social control. Ideally, this mode is the antithesis of the first mode of interaction. This mode of relationship often is termed ”mass-like“ but it would seem more valuable to confine the term “mass" to the private actions of individuals which are anonymous and which are carried out in isolation rather than in the presence of others. This absence of even shoulder- to-shoulder contact should set apart mass behavior from crowd behavior.16 To repeat then the distinguishing features of the relationships amongst persons in crowds: reciprocity (in taking the roles of others into account), mutual evaluation, and concern with particular ”others“ are present in significant- ly lesser degree than in small groups. Two other dimensions 15For the symbolic interaction position, no environ- mental “stimulus” effectively encroaches upon human activity until it is defined by persons in terms of possible plans of action toward it. Usually these definitions emerge consensually, i.e. the meanings of objects are socially derived. See Blumer, “Society as Symbolic Interaction“. 16See above footnote 7. 12 offered here as crucial to distinguishing small group from crowd relationships must be discussed in greater detail. These two are nature of normative control and the degree of perceived surveillance. What prompts a newly formed collectivity to take on small group characteristics or crowd characteristics? Two context variables having bearing - normative control and per- ceived surveillance - will be reviewed. One vantage point in social psychology views sets of norms as internalized by per- sons during socialization: thereby social control of behavior is incorporated into self-control.l7 From this perSpective, certain general norms are construed by persons to be in play and binding as part of the situations which they encounter. These regulative guides to behavior are shared and held as long as there is no obvious evidence that deviance from the 18 Consensus is held by participants norms can be satisfying. in face-to-face interaction and this consensus structures the manner in which individuals gpa persons serve obligations to others and make legitimate claims upon others to do the same. 17Lendesmith and Strauss. Social Psychology (N.Y.: Holt-Dryden, 1956), ch. 14. 18A recent assertion of this is found in John Finley Scott, The Internalization of Norms, Paper submitted for read- ing at The American Sociological Association (Los Angeles, 1963). This proyiso is consistent with the continual reorienting personsffh while interacting. 13 The norms constrain and yet provide a framework permitting leeway in the formation of personal actions. What happens, however, to intra-group relationships when norms defining a situation become ambiguous or are reversed? When no institutionalized norms exist or are shared, persons struggle to structure and "define" a situation (as long as no threat is perceived in doing so by the persons) in accord with an accepted definition with its corresponding norms. However, when the prevailing norms are conceived as actively dissatisfying (perhaps because of the inequity of reward perceived to accrue to persons) or as downright dis- functional to the attainment of goals, peOple will disaffili- ate from them. It is commonly accepted that, in the crowd relationship, a sense of personal obligation to "other” is diminished.19 To outsiders (gig g gig a crowd), normative control does not appear as extensive nor as constraining. Behavior appears to be more uninhibited. This apparent lack of constraint (when traditional norms are perceived by parti- cipants as dissatisfying or as dysfunctional) is not because norms are absent but is due to the emergence of new norms which temporarily (7) replace traditional ones. "Old" norms are suspended in the crowd in favor of new (or alternative) patterns of normative control. This new set of norms may appear as binding - even legitimate in the 19Turner and Killion, op. cit. 14 newly defined situation permitting them - yet not as con- straining to the participants because they coincide with what the actors ”desire“ to do.20 The key difference between the nature of normative control in small groups and in crowds is that, in the latter, the norms are not reciprocally shared. The emergent norms develOp in the process of crowd activation. They are held and acted in accord with separately by persons pip parallel to the unstably defined situation. The parti- cipants are too engrossed in the situation to be aware of either normative sharing or of normative constraint. It will be noted that nothing about the content of norms distinguishes crowds and small groups. It is not the type of belief held which is a sure-fire predictor of inter- personal action. The concentration here is upon the nature of normative control tying persons together in a gathering. A national emergency may provoke a situation which is fairly rationally discussed by the Joint Chiefs of Staff in closed session or which may trigger widespread collective panic. The kind of relationships between persons in a gathering is influenced by the manner in which the norm is held by inter- actors. 20Newcomb, T."Social Psychological Theory: Integrat- ing Individual and Social Approaches,” in Hallander and.Hunt (eds.).” W Sociatl Psxcholosy (New York: Oxford University Press, 1963) “What happens...is that what the individual wants to do and what.he perceives as demanded by his role come to be identical," p. 17. 15 The second major condition distinguishing the mech- anisms that differentiate crowd behavior from small group behavior is that of perceived anonymity of action as opposed to the perceived surveillance of one's action. Crowd'be- havior tends to be anonymous. Crowd participants (a) do not feel that their personal actions are under scrutiny or even under the watchful evaluation or concern of "others." Con- comittantly, the participants are (b) not surveying the Specific actions of particular others in the Situation. This reciprocal anonymity of individual action in crowd engross- ment is linked to (c) the lack of reciprocal evaluation of performance by participants. Assessment of others need not be reckoned with since it is presumed by each that all are too engrossed with an objective or an object of attention to be singling out the behavior of other persons. This recog- nition, of course, is rarely an articulated one. Even more important to this argument, persons in crowds feel anonymous as individuals to external view or appraisal. External anonymity refers to the sensing of a situational condition in which the crowd itself is unassail- able i.e., participants are not under critical in3pection out- side of the crowd boundaries, or, if they are, one is not personally subject to judgment. In this collectivity, one's personal reSponsibility is not on trial: one's act will be lost in the crowd. Anonymity as conceived here refers pri- marily to this perceived absence of either (a) external l6 surveillance of one's “own" behavior or of (b) concern over personal reSponsibility for one's behavior in the crowd. Behavior is literally not owned by the self.21 Anonymity hfi doesArefer then to an exclusively intra-group aspect. This sense of external anonymity of one's behavior is critical to, and correlated with, the other dimensions of crowd behavior. Likewise, perceived surveillance of one's actions is associ- ated with the dimensions of the small group mode of inter- action. A number of dimensions have been underscored which, theoretically, are anticipated to exist at different levels in small groups and in crowds. Reciprocal role-taking is viewed as a key activity distinguishing small group inter- action and crowd inter-relations of persons. A symbolic interactionist framework has been adOpted which argues that low role-taking in gatherings is related to low concern with particular others. These two characteristics plus low norm- ative consensus and low perceptions of being evaluated by others are marked in gatherings drawn together in parallel fashion to a provoking object of attention, i.e., in crowds. Persons will not feel under scrutiny within and particularly from outside such groups. 21The quickest way (police chiefs report) to break up a panty raid of students is for an official looking plains- clothesman to walk about taking down names of students who are participating. This serves as a jostling reminder that one will be held accountable at a later date for one's behavior. II . REV IEW OF REL ENAN T RESEARCH The research undertaken here intends to explore the feasibility of tapping the dimensions of intra-group inter- action stressed in this thesis. More Specifically, the in- fluence of surveillance as opposed to anonymity upon the mode of interaction within small gatherings is being investigated. What better place than the small group literature to dis- cover whether or not our characterization of “small group" and “crowd" relationships has received any empirical study and support? There is almost no directive aid to be found in the literature of small group research that has c0ped with the problems and distinctions discussed here. Small groups have been a focal research site for the study of group dynamics, morale, interpersonal influence, task solution, sociometric (friendship) choice, communication networks, and leadership.22 22Hare, A.P. Handbook of Small Group Research (N.Y.: The Free Press of Glencoe, 1962) and Hare, Borgotta, and Bales, Small Gropps: Studies in Socipl Interaction (N.Y.: KnOpf & Son, 1955) are the two most comprehensive sources reviewing the range of research employing the small group as a research Site and as a unit of study pp; ceneris . l7 18 Many of the studies in these areas have treated aspects of intra-group relationships but none seem to have focused attention upon the nature of interaction under different normative settings or under experimental conditions of anony- mity or surveillance. Nearly all social-psychological studies undertaken in the small group setting have basically presumed a paradigm of social interaction akin to the depiction of "small group“ relationships eXpressed in this work. They have taken such a model of interaction as given. Crowd behavior has not been the subject of extensive empirical study, particularly in laboratory settings under experimentally imposed conditions. PrOportionally little attention has been devoted to collective processes and the small group setting, by and large, has been by-passed a locus 23 for their analysis. Exceptions are studies by Polansky, Lippet, and Redl: Swanson: and Festinger, Pepitone, and Newcomb.24 23There are many reasons for this general neglect. The unexpectedness of collective manifestations is probably over- rated but is nonetheless true enough. Few funds are allocated for the theoretical study of collective behavior processes in laboratory Situations. Finally, current methodology is not geared to behavior which is considered ephemeral, unpredictable, and even dangerous. 24Polansky, Lippet, and Redl. "An Investigation of Behavioral Contagion in Groups," Human Rglptions, (3) 1950, pp. 319-348: G.E. Swanson, “A Preliminary Laboratory Study of the Acting Crowd,” Amepican Sociological Review, (5) 1953, pp. 522-533: Festinger, Pepitone, and Newcomb, "Some Conse- quences of De-Individuation in a Group,“ Journal of Abnormgl and Social PsyChology, (47) 1952, pp. 382-389. It might be worthwhile to review Swanson's study of "acting" crowds. Swanson tested a number of hypotheses drawn from many theoretical sources. Swanson conceived of a crowd as a particular form of social organization for the collective solution of problems. This form of organization might appear in any size of 19 Since this research will adapt certain of the methodological tools and procedures from a small group laboratory study con- ducted by Festinger, Pepitone, and Newcomb that study must be presented in some detail. The general problem that Festinger pp pl faced was one of determining the personality experiences attendant to the submergence of individuals into group situations leading to "individuated" and ”de-individuated” types of groups.25 Festinger pp pl's theoretical base appears extremely similar collectivity two persons or more. The behavior of an "acting crowd“ (a depiction drawn from Blumer) aims at manipulating the environment external to the selves of the members. Swanson's experimental design varied the extent to which Subjects were acquainted with a task and the extent to which they were ac- quainted with each other. Thus, four SXperimental conditions were created, non-acquaintance both with task and associates predicted as the most likely setting for crowd behavior to emerge. Small experimentally composed groups were set to solv- ing a complicated game requiring a high coordination of effort. Swanson drew from LeBon's contention that in an “acting crowd” the following characteristics prevail: suggestibility, high volume of communication, communication directed toward the group rather than toward Specific members, low volume of self- oriented needs expressed, fixity of role behaviors, less clarity of differences in role behavior, less efficient task performance, less satisfaction with other group members, a lower recall of task related events and a lower recall of non-task events. Groups composed of those who were unacquainted with the task and each other were predicted to score lower on the average than other three types of groupings on these variables and, except for the latter two, they did so. All of the pre- dictions were supported at 95% level of confidence except for the twohypotheses predicting lower recall of non-task events and recall of task related events. 25Festinger pp $1.1.” op. cit., pp. 382-383. This divi- sion is based upon the types of satisfactions obtainable by group members under these two modes of relationship. 20 to and demonstrates the continued influence of LeBon's notions of the rationality and normative constraint of persons when "in isolation“ but whom, when congregated, lose their criti- cal faculties and normative restraints.26 Central to this decline is Festinger pp 31's assumption that the lowering of normative (internalized) constraints is a satisfying state of affairs. Although their Operationalizations of types of social relationship for "individuation" and “de—individuation“ in a group setting will be maintained, it is necessary to re- formulate their entire thesis so that the relationship of variables is in line with the symbolic interactionist schema used here. After presenting a brief resume of the Festinger study, new questions will be posed consistent with this schema. "Anyone who observes persons in groups and the same persons individually is forced to conclude that they behave differently in these two general kinds of situations.“27 According to Festinger, individuals, when alone, are much more restrained as inferred from their behavior than in cer- tain types of group situations in which conditions are salutary for their obtaining culturally sanctioned release 26G. LeBon. The Crowd (1895). (New York: The Viking Press, 1960). 27Festinger p3; p_l_., op. cit., p. 383. 21 from the stringencies of prescribed and proscribed behavioral dictates. Under such conditions of a general reduction in normative restraint, Festinger hypothesizes that group members will be marked by less singling out of the other in that group setting for particular attention. This inability to identify the behaviors of others Festinger labels as a group condition of De-individuation.28 In opposition, Individuation is a type of group setting in which one's self and others in a gathering are more highly normatively controlled. Persons are involved with others in the situation. Self-esteem is the satisfaction endemic to such a setting. Evaluation, of other's performances, a high ability to single out other's behavior all under socialized restraint profile this mode of relationship.29 Festinger investigated twenty-three groups (N 4-7) under identical laboratory conditions in order to test this general hypothesis. De-individuation was Operationalized, for a group, as the number of identification errors made by the Subjects after a forty minute discussion period as to who said what during that forty minute discussion period. These scores were corrected by a memory factor. Fifteen 28Ibid.. pp. 382-384. 29Ibid. 22 statements were read back to the subjects (10 that actually were made during that period and five that were not made): each person was asked to recall if the statement were made and, if so, who made it (each person had a name card plainly visible to the others). A memory error was made when a sub- ject reported, on his protocol, that a statement was made when actually it was not or if he indicated that a statement was not made when indeed it was. An identificationjwas one in which the statement was attributed to the wrong subject. Procedurally, reduction of normative restraints was induced in each of the groups by reading the results of a ficticious study describing how the overwhelming majority of a large representative sample of undergraduates registered, under psychiatric analysis, hatred or hostility toward either or both parents. Furthermore, the statement pointed out that those most unwilling to discuss the matter or most vehemently denied having such impulses possessed the most deep-seated hatred. The statement generally reported that those not diagnosed as hostile felt that they were a bit hostile and were quite willing to discuss their feelings cpenly. Such conditions Festinger believed highly conducive to the reduc- tion of normative restraints. Festinger theorized that groups high in negative statements (adjusting for positive commentary) would make high numbers of identification errors as a group - i.e., a correlation Should exist between reduc- tion of normative restraints and "de-individuation" in small 23 gatherings. Their analysis of results showed (in corrected form) a Significant correlation (r = .57: p >.01) between de— individuation and reduction of normative restraints among the groups in the study.30 Many serious criticisms can be aimed at Festinger pp pi's conception of variables and processes. There seems to be confusion in Festinger's thinking as to whether "individua- tion" and "de—individuation" refer to group or to individual processes. The two terms are used in both senses.31 Secondly, the Operationalization of normative restraint reduc- tion is open to serious question. Finally, throughout the logic of proof, Festinger temporaly confuses the sequence of 32 processes he is analyzing. The group relationship of 30The preceding three paragraphs are a summary of proceedings and findings of Festinger pp gl's study. The correlation obtained was a corrected one in that it excluded from analysis one group‘s results which diverged Sharply from the others. 31Persons are “individuated in groups“ and individua- tion "is a group phenomenon.“ See pp. 383, 384, 389. 32De‘individuation as pre-condition: p. 383. "...under conditions where the member is not individuated in the group, there is likely to occur for the member a reduction of inner restraints against doing certain things. In other words, ... many behaviors... impossible because of the existence within himself of restraints become possible under conditions of de- individuation within a group.“ See also, p. 389 (summary). Normative restraint reduction as pre—condition: p. 384. "...create conditions in which there was a strong force acting on the members to engage in some behavior against Which there were strong inner restraints. Under such conditions some groups would befable to create de—individuation Situations than others.""(ItaIics mineT 24 de-individuation purportedly leads to the reduction of norma- tive restraints. Operationally, however, normative restraint reduction is first induced. Only after the level of this reduction is measured (i.e., counting negative statements made) can “de-individuation“ be measured. III. THE DESIGN OF RESEARCH The precise logical implications of symbolic inter— actionism have rarely been drawn out for small group Situa- tions and thus have rarely been applied to laboratory study. Even less so have they been applied to the analysis of crowd behavior at this level. The research carried out here was not intended to replicate and confirm the theoretical bases of Festinger pp pl‘s study. A different orientation, entail- ing a revision in framework, methodology, and terminology, has been adopted. At the same time, a definite corrSSpondence is detectable between the poles of individuation-de—individuation and small group-crowd interaction. Nevertheless, a design was constructed which would probe the expectations that follow from selected principles and premises of symbolic interaction- ism. These expectations must first be articulated. First, do we empirically observe a promising correla- tion of variables we have designated as, on the one hand, characterizing crowd relationships and, on the other hand, characterizing small group relationships? An inquiry into the characteristics of the relationships between persons in groups would hypothesize an actual conjoining between predicted 25 26 sets of variables that had been theoretically related. Only if this were so, could the two modes of relationships be regarded as fruitful distinctions. The matrix of pairs of relationships between reciprocality, norm-sharedness, per- ceived surveillance, mutual evaluation, and concern with particular others Should indicate ten inter-correlations. Secondly, if a situation was defined by the partici- pants in a setting as one in which they do not feel personally accountable for their acts beyond the confines of the gather- ing (perceived external anonymity), then it is tentatively offered that such groups of persons would rank high on a valid measure of de-individuation (or engrossment as it has been called here)33 in the situation. Anonymity, as induced and as perceived by participants, is an important context variable promoting engrossment when normative restraint is reduced. Anonymity (particularly from outside surveillance) is a facet of crowd—like relationships between persons. Thirdly, it is submitted that engrossment in the situation is to some extent associated with low operating levels of reciprocity (of role-taking), perceived evaluation by others, and concern with others. Moreover, norms will be held in parallel fashion rather than shared and anonymity rather than surveillance will be perceived by persons by "groups“ engrossed in a situation. Low engrossment levels 33"Engrossment" seems to escape many of the misleading connotations of the term “de-individuation.“ Engrossment also connotes an object or activity attracting the attention of some person or group. 27 would be associated with higher levels of reciprocity (of role-taking): norm-sharedness: perceived evaluation by others: perceived surveillance of behavior: and concern with particu- lar others. Therefore, the same dimensions denote both small group and crowd interaction: the two modes of relationship differ in the degree or level of presence of the dimensions. A research design resembling Festinger pp pl's design was employed to gauge the worthwhileness of these three con- ceptual expectations. Before formally stating the hypotheses and Operationalizing the variables used in this research, a formal logic of proof tying the three expectations together must first be elaborated and clarified. First, if each per- son in a small group is taking the role of particular others into account, is singling out others, then it can be said that there is reciprocity (of role-taking) in that inter- action. In such interaction, concern with particular others maintains self-awareness and other-awareness. Such awareness of contrast between self and other diminishes when individual attention is focused a tOpic or object of attention. If persons are not reciprocally engaged in accountable concern with each other, it would follow that evaluation of one's or of “other's“ behavior is not in the forefront of attention: one is freer from differentiating and comparing (and therefore ranking) one's behavior with others. One is less under norms of conventional obligation in regard to others. Norms zip p vis an object are held in a mass-like individualistic manner. 28 These characteristics then Should, along with perceived anonymity, cluster consistently at approximately similar levels at one extreme (high) characterizing small groups and at the other (low) extreme characterizing crowds. Secondly, anonymity, is a condition which can be manipulated so that perceived freedom from outside inSpection (surveillance) will be more likely sensed by a collectivity of persons. Among "groups“ to which identical inducements are given to engage in behavior in which a norm (posited as traditional) may permissively be suSpended for the purposes of their interaction it would be presumed that perception of outside surveillance would curtail felt freedom from account- ability. This would be a factor (i.e., perceived surveillance) which would maintain reciprocal interaction along structured lines and curb emergence of a new pattern of inter-relation- ship y$§., an ”acting" crowd. Third: Finally, regardless of the condition of im- posed anonymity or surveillance, the actual Engrossment level of a group should roughly corrSSpond to the level of crowd- like relationships manifested within them. A construct conceived as the mean number of identifi- cation errors per person made by each group (corrected by a memory-error factor) was the index for Engrossment. In- ability to identify the source of remarks Should relate to groupings at the Crowd-relationship extreme. Stemming from the above, the following hypotheses were formed: 29 1. There is a positive correlation between recipro- city, norm—sharedness, perceived surveillance, perceived evaluation, and concern with particular others in a small gathering of persons. Small gatherings will tend to be characterized by consistent levels of these variable inter- actional tendencies in Situations of experimentally induced normative restraint reduction. 2. A higher level of engrossment (i.e., an inability to identify particular persons' remarks, theoretically corresponding to a heightening of attention in a focal object of interest) will obtain in groups under the condition of anonymity than will groups under surveillance. This only obtains if there is a higher level of perceived anonymity among groups in the anonymity condition. 3. The engrossment level of a group associates most closely with crowd-like gatherings i.e., with groups evidenc- ing low levels of reciprocality, norm-sharedness, perceived surveillance, perceived evaluation, and concern with parti- cular others. One major recognition is incorporated into these hypotheses. The subjective definition of the situation by a collectivity's participants is more crucial in accounting for their actions than is the experimental manipulation of condi- tions as “structured" by one outside of the collectivity. Unless experimentally manipulated variables are perceived or acknowledged in the experimentor's intended ways by the per- sons under study then the determinative influence of a social structural condition can not be traced or realized. 30 Personal awareness is more than the channel whereby social structural conditions impinge upon one, but since awareness itself is selectively perceptive, ”awareness“ of a condition often pp the construction of conditions. It was necessary to Operationalize each of the concepts employed in this research. Although statistical levels of probability of outcome were pre-set, any rigid logic of proof must be abandoned to a later stage of inquiry. This study is an eXploratory reconnaissance to draw out the value of the hypotheses representative of a symbolic interactionist approach, not to critically test them. I. Constructs of Relationship to Focal Object (and to Others) High engross- ment level Conceptual Definition A heightened inability to identify the re- marks of particular persons correSponding to heightened interest in a focal object Operationalization The high average number per person of identifi- cation errors in re- calling which persons uttered particular statements from a sample of fifteen statements, 10 of which were actually made during a 40 minute discussion (corrected for memory errors)34 34 This Operationalization for engrossment is identical with Festinger pp 31's operationalizationcf de-individuation. The mean number of memory errors is subtracted for each person from his mean number of identification errors in order to remove memory as a factor confounding the measure of inability to identify particular Speakers. Low engross- ment level II. Experimental Conditions Anonymity Surveillance III. 31 Maintenance of ability to Single out and correctly attribute Specific remarks to particular persons Conceptual Definition Freedom from account- ability of one's acts to outside assessment Sensed inSpection of one' s behavior by others outside of the group A low average number of the above Operationalization The experimental pro- cedures comprising the research situations (See Procedures) Presence of and playing during research Situation of a tape recorder - micrOphone in center of group semi-circle~ Dimensions of Interaction - Operationalized as the mean score on a post-session questionnaire item tapping the Specific dimension of: A. Reciprocality-mutual taking into account the roles and per- sonalities of particular others in face-to-face relationship. OPERATIONALIZATION: Number 3. How'much did you find yourself taking what the others said here into consideration in making your comments? Very much so Quite a bit Somewhat Only slightly Not at all B. Sharedness of norms - consensus as to the explicit and implicit norms guiding the interaction. OPERATIONALIZATION: Number 4. How'much would you say that the expressed feel- ings of others here ”influenced“ your statements in the direction of agreement? C. 32 A good deal A little bit Hardly at all My views REjected were in- otherviews depend- SXpressed ent of reacted others against them. Supyeillance - a recognition that one's actions under in- Spection and that one's self is being appraised. OPERATIONALIZATION: Number 5. To what degree (if any) did you believe that what you might say could be held against you? A strong Some enduring Thought it Only vaguely Did not concern concern conceivably occur to throughout might be me Evaluativeness (Evaluation) - A perception of mutual assess- ing of behavior among persons in order to compare and to rank them on some personal performance criteria or to rate personal adherence to values. OPERATIONALI ZATION : Number 8. To what degree did you feel that others might disapprove of you if you did not contribute to the discussion? Strongly A definite Somewhat Only a little No such possibil- feeling ity existed Concgpp withyparticular others - Perceived awareness of and retention of the peculiar and distinguishing actions of particular other persons (as opposed to OPERATIONALIZATION: Number 9. To what extent do you think the others here were keeping in mind the stated feelings of each particular person in the group? Most were to More so than No more 55' A little Not very a high degree usual with than usual less so much at new persons than in all. most small groups 33 The above five dimensions are continuous, rangdng conceptually from each's full fledged existence (5) to dia- metric absence (1). Methodologically, they are scalar, rang- ing from high (small group mode) to low (crowd mode) levels on a Likert-like scale - an interval scale. The five intra- group characteristics are tapped by items each constituting its own scale. The reSponseS of persons are translated into scores, averages for the group calculated and compared. The experimental condition (surveillance) and control condition (anonymity) plus the High engrossment-Low engrossment relationships are discrete in conception. The statistical tests for the three hypotheses are (1) Pearson Product- Moment correlations (2) t-tests of significance of difference between means (couched as null hypotheses of no difference for very small samples) and (3) Spearman-Brown Rank-order Co- relations (rho). Because it is difficult to say on a ratio scale where 123 engrossment ends and where high engrossment begins, the rank order correlation was considered a more honest approach to determining associations between engrossment and the dimensions of intra-group interaction. The variables to be correlated did not range along the same interval scales. This reason, too, prompted use of the Rank-order correlation. The study was conducted as a small group laboratory research. The method employed was controlled, non-participant observation. Two paper and pencil "test forms" were admin- istered: a recall-check form and a post-session questionnaire. 34 The observation system was a near complete transcript of a discussion group's behavior recorded in a forty minute dis- cussion period. These two indices (See Appendix A) were administered to each person in each discussion group. The recall form was the means for tabulating the person's identification and memory errors for the fifteen statements read back to him. The post-session questionnaire was constructed to tap the variable dimensions of intra-group relatedness. Its reli- ability was not pre—tested or tested. However, the validity of these indices of intra-group relationships would have to be cautiously considered. The face validity was checked successfully by one other graduate student in social psycho- logy. However, empirical validity posed problems that this piece of research does not purport to surmount. Sampling procedures for selecting groups were as follows: nine groups of five persons each (undergraduate males under age 28) were recruited from an Introductory Social Psychology class at Michigan State University. Assign- ment of these men was largely on the basis of free-time availability of a period of one and one-half hours that the potential "subjects" indicated that they could be present to participate in a small group study. As well as being practical, this method would seem as random as any other means of assignment. If after ten minutes after the appointed time only four of the five had appeared, the experiment began. 35 If the fifth party then arrived he was not allowed to partici- pate although he might be assigned to a later group. IV. PROCEDURES 0? THE STUDY ANONYMITY CONDITION35 The procedures in the Anonymity Condition (control) were arranged in the following manner: A sample of nine groups of undergraduate males volun- teered for the small group study. Groups were composed of five persons and each group “run“ separately. AS persons arrived at the discussion room, they were seated around a conference table and were engaged by the observor in mildly cheerful small talk. This procedure was adopted to prevent excessive prediscussion among participants which might have introduced undesirable factors. The setting was plainly free from "outside“ observation. When all persons had arrived, the eXperimentor (who sat off to the side), directed them to print their first names on cards so that each could be identified by the others in the group, and then proceeded to read aloud the following statement. The alleged survey read and its findings are, needless to say, entirely ficticious. "The following statement represents a summary of an important research project that has recently come to the attention of psychiatrists and social scientists concerned 35The procedures of The Anonymity Condition were adapted, with minor revision, from Festinger p§_g;}s procedures. 0pI cit., pp. 384-387. 36 37 with problems of personal adjustment among students. Al- though the results are demonstrably reliable, it is believed that additional implications can be brought to light by hav- ing small groups of students discuss their personal views relating to these results. ”A highly representative sample of 2365 students (1133 female and 1232 male) on 14 campuses, from all social- economic classes and several nationality backgrounds, was subjected to an intensive three-week psychiatric analysis consisting of repeated depth interviews and a battery of sensitive diagnostic tests. The results Show without question that 87 per cent of the sample possessed a strong, deep- Seated hatred of one or both parents, ranging from general— ized feelings of hostility to consistent fantasies of violence and murder. A finding of further Significance was that those individuals who at first vehemently denied having such hostile impulses or who were unwilling to discuss their per- sonal feelings in the matter were subsequently diagnosed as possessing the most violent forms of hostility. In other words, conscious denial, silence, or embarrassment were found to be almost sure signs of the strongest kind of hatred. Of the 13 per cent in whom no trace of hostility was found, the great majority thought they probably hated their parents and were willing to discuss every aSpect of their feelings with the investigator. “In summary, 87 per cent were found by modern psychia- tric techniques to possess deep-seated resentments and 38 hostilities toward one or both parents. Individuals in this category who most vigorously denied that they had such feel- ings revealed, at the conclusion of analysis, the strongest degree of hatred. Thirteen percent were found to be free of such aggressive impulses. Most of these individuals at first thought they were basically hostile and were interested in discussing their feelings toward their parents freely. ”Would you discuss in detail amongst yourselves your own personal feelings toward your parents in the light of these results. Try to analyze yourself in such a way as to get at the basic factors involved.” The participants were each given a copy of the above statement and were asked to start discussing the matter. The discussion lasted 40 minutes. The discussion material was designed to create condi- tions in which the phenomenon of engrossment might occur. The particular topic was chosen because it was felt that most pe0ple would have restraints against eXpressing hatred of their parents. In other words, to the degree that the parti- cipants accepted the statement, they would experience a more or less strong pressure to reveal negative feelings toward their parents. Observation Methods During the 40 minute discussion an observer cate- gorized statements in terms of whether they reflected positive or negative attitudes toward parents in the present 39 or the past: positive or negative attitudes of others toward their parents: theories about parent-child relationships: and whether they expressed concern with the interaction of group members and the discussion procedure. Each contribution to the discussion was categorized and recorded next to the name of the person who made it together with the length of the con- tribution in seconds. Pauses which lasted for 20 seconds or longer were also recorded. (After ”running“ the first group, these codings were made from the full transcript after each session.) The Recording of Statements (L To obtain a measure oprerson's ability to identify who had said what in the discussion, the Observor recorded, as nearly verbatim as possible, 10 statements made by the group members during the discussion. The following criteria were strived for in selecting these statements from the dis- cussion: l. The content of the statement was distinct enough to permit identification of the person who made it, i.e. the statement should be as dissimilar as possible from those made by other group members. 2. The statement was about a sentence in length. 3. The statement was grammatically coherent. 4. The 10 statements came from as many group members as possible. 40 5. The 10 statements were distributed over the en- tire 40 minute discussion period. At the end of the discussion and While the observor was selecting the 10 statements, the group was given a ten minute break in Which to fill in an adjective Check list and to relax silently. After making certain that the name cards were visible and separating participants to prevent c0pying a form was distributed and the following instructions were read: “I am going to read off some statements that were made in the discussion and some that were not made. If you do not remember the statement hav- ing been made, place a check in the first column next to the appropriate number. If you remember the statement, but off hand you do not recall who made it, place a check in the second column next to the appropriate number. Finally, if you recall who made the statement write the first name of that person in the third column. This is not a memory test of any kind, and there is no need to guess.“ The statements were then read off in the same temporal order in which they were made in the discussion. InterSpersed among the 10 statements, in constant order for all groups, were 5 statements that were not made in the discussion. These were included so as to provide a basis for comparing "Identification Errors,” i.e., errors in recalling who said what, with "Memory Errors,“ i.e., errors in remembering the content of the discussion. The participants' reSponses, to the statements were scored in the following way. If, on any of the 10 statements actually made in the discussion, the person failed to recall 41 who had made it or if he attributed the statement to the wrong person, he was given an error. As with “identification errors” these “memory errors“ were averaged for the group. In recording the statements frequently it was diffi- cult to meet the criteria for selection of statements. The statements recorded varied greatly in their "identifiability." Sometimes a statement would be recorded and, later on, others would make very similar statements, thus making the identifica- tion ambiguous and difficult. Sometimes, when the discussion was proceeding rapidly, it was not possible to record the statement accurately and consequently the recorded statement would be quite different from what was actually said. To c0pe with these difficulties some of the statements were eliminated from the analysis when there were good grounds for believing that they were poor statements. The specific criteria used to eliminate a statement were as follows: In groups of five persons a statement was eliminated if: (a) all or all but one person made errors on it, or (b) all but two made errors and the person who made the statement erred himself. In groups of four persons a statement was eliminated if all or all but one, including the person who made the statement made errors on it. The number of identification and memory errors, if changed by this elimination, was adjusted for each group. 1,“, 42 The Measurement of Intra-Group Relationships A post-session questionnaire (See Appendix A) in- cluding items to measure the levels of reciprocity, norm- sharedness, mutual evaluation, perceived surveillance, and concern with particular others was administered immediately after the recall form. Numerical weights of from 1 to 5 were assigned each alternative. Average scores on these dimen- sions were calculated for each group. Explanation to the Subjects After the questionnaire had been administered, the general purposes of the study were explained in detail to the group. The confidential nature of personal remarks was assured. Groups were told that the data presented for the discussion tepic were entirely ficticious and the reasons for using it, together with the rest of the procedures, were dis- cussed with them. Sufficient time was apent in this manner with each group for them to leave with a good understanding of the eXperiment. They were also asked not to tell others about the experiment since the researcher did not want future participants to know what was going to happen. As far as is known, the participants kept faithfully silent about it. SURVEILEANCE CONDITION The procedures under surveillance were the same as under the anonymity condition with the following exceptions: 43 A tape recorder was present (off to the side yet not hidden) and a micrOphone placed in the middle of the semi- circle of persons. Nothing was said about this equipment. The tape recorder was turned on after the initial statement was read. The researcher asked the subjects to introduce themselves immediately after “flipping“ the switdh to record. The reel ran for nearly the entire forty minute discussion period. This procedural change constituted the induction of feeling that what might be said in the gathering could be re- corded, be preserved and held up to later hearing. This pro- cedure comprised the condition which could encourage the perception that one's remarks could be subject to “external" inSpection beyond the confines of the gathering. V. BEHAVIORAL DATA AND ITS ANALYSIS Tables I through V summarize the research findings pertinent to the hypotheses under investigation. In Table I are the mean item scores on the dimensions of intra-group interaction for groups under the Surveillance and under the Anonymity conditions. (The Questionnaire is reproduced in Appendix A). Table I. Mean Score Levels of Dimensions of Intra-Group Interaction by EXperimental Groups. M w m Surveillance Anonymity Dimension (N = 4) (N = 5) Reciprocity 3.20 3.43 Norm-sharedness 3.26 3.12 Surveillance 1.58 1.52 Perceived Evaluation 2.29 2.56 Perceived Concern with 3.41 3.90 Particular Others The differences between all scores under the two conditions are non-significant at pl>.05. Thus, “members" of groups in both conditions indicated that they considered others "Some- 1 what“ in making their comments. It is interesting that the eXpressed feelings of others influenced participants comments 44 45 "in the direction of agreement hardly at all" (Norm-shared— ness). It must be remembered here that these are checked responses to questions and not necessarily valid reports of the tone of interaction. More telling was the common comment for both types of group that the degree of feeling that "what was said might be held against you“ veered between being "vaguely" held to "not occurring" to persons at all. CorreSpondingly, perceived evduation as measured by "degree to which you thought others might disapprove of you” averaged between “only a little“ to “somewhat.“ However, perceived concern with particular others as gauged by the degree to which the "stated feelings of others were kept in mind“ tended to be "more so than usual with 'new' persons" although this tendency was not clear-cut. Hypothesis I. There is a positive correlation between reciprocity, norm-sharedness, perceived surveillance, per- ceived evaluation, and perceived concern with others in a small gathering of persons. Small gatherings will tend to be characterized at consistent levels, on these dimensions. Moreover, they will tend to be characterized at consistently high levels (as in the small-group mode) or at consistently low levels (as in the crowd mode) of these variables in situ- ations of eXperimentally induced normative restraint reduc- tion. The results of Table I invalidate the contention of significant differences between levels of interaction pro- duced by induced conditions of normative restraint reduction. 46 As to the inter-correlations between dimensions, Table II shows the Pearson Product moment correlations between the ten possible pairs of dimensions: Table II. Correlation Matrix of Dimensions of Intra—Group Interaction. Dimensions Recip- Norm- Perceived Perceived Concern with rocity shared- Evaluation Surveil- Particular ness lance Others Reciprocity .50 .57 .76* .66 Norm- sharedness -.23 .58 .ll Perceived Evaluation .69* .18 Perceived Surveillance .77* Concern with Particular Others *Significant at p )~.05. Calculating the significance of these intercorrela— tions at p .05 and with n-2 (i.e., 7) degrees of freedom, any correlation larger than .67 would be statistically signi- ficant. Table II indicates that perceived Surveillance is significantly correlated with Reciprocity, Perceived Evalua- tion, and with Concern with Particular Others. Surveillance's correlation with norm-sharedness falls at .58. Seven of the ten inter-correlations are associated in the predicted direc- tions at levels exceeding .50. It would appear that the more 47 feeling manifested within the groups that what was said "might be held against" persons (beyond the confines of the group), the more persons in such groups felt that "others were keeping in mind" what was said, were feeling that dis- approval could result from non-participation in discussion, and "were taking into consideration what others were saying in making their comments." At first glance, it might appear anomalous for feelings that one's behavior is under scrutiny (perceived surveillance) to be associated with anticipated disapproval upon non-participation in discussion (perceived evaluation) but there is no incongruity here. An "expert“ with a wide reputation who appears in a symposium with lay peOple may feel that he is called upon to contribute dis— prOportionally - that he is being watched egg eXpected to participate. The two feelings are not at cross-purposes. The sharedness of norms is meagerly associated with Perceived Evaluation (-.23) indicates this tendency: the more permissable one feels it is to keep quiet in the gathering (low evaluativeness) the more it is felt that the eXpressed feelings of others “influenced" one's statements toward the group norm. However, this correlation was not significant. The extent to which persons experience that “others are keep- ing in mind" every one's remarks does not appear to be associated with either the "sharedness of norms“ or with anti- cipated experienced disapproval if one were not to contribute to the discussion. 48 HYPOTHESIS II. A higher level of engrossment (i.e., an inability to identify particular persons' remarks theoreti- cally corresponding to a heightening of attention in a focal object of interest) will obtain in groups under the condition of anonymity than groups under the Surveillance condition. This will only obtain if, indeed, there is a higher level of perceived anonymity among groups in the experimental condition of anonymity. Table III. Mean Number of Identification Errors per Group. Anonymity Surveillance (N=5) (N=4) 7.6 8.0 Table IV. Mean Engrossment Scores. Anonymity Surveillance (N=5) (N=4) 1.04 .56 t-score difference==.767 (non-significant at p:>.10). Table IV indicates the difference in engrossment be- tween groups in the surveillance condition and in the anonymity condition.' The t-score difference indicates that, although the difference in engrossment level between the two conditions is in the predicted direction, this difference for 49 the size of samples compared is non-significant. Comparing the gross number of identification errors averaged under the anonymity condition and under the surveillance even more indicates no difference in powers of singling out the Speaker of particular remarks as influenced by the researcher's arrangement manipulation of conditions. Since Table I has shown that there is no difference in perceived surveillance between the two types of groups, and therefore the non- significant difference in engrossment levels is not surprising. Until anonymity is perceived by groups there can be no ex- pectation that engrossment (the inability to single out the source of a statement made during interaction) should in- crease. The hypothesis is not invalidated but awaits more precise Specification of intervening variables before it can be critically tested. HYPOTHESIS III. The engrossment level of a group will associate most closely with crowd-like gatherings, i.e., with groups evidencing low levels of reciprocity, norm- sharedness, perceived (external) surveillance, and perceived concern with particular others. Table V shows, by means of rank-order correlations (Spearman-Brown) the measure of association between engrossment and the reSpective variables of intra-group interaction: The hypothesis would predict negative inter- correlations between engrossment and the variables of 50 interaction.36 None of the correlations are significant except for the correlation of engrossment and Perceived Con- cern with others (which is directly counter to expectations.) It must be remembered, however, that when the number of groups is less than ten, tests of significance are of doubt- ful value since normality for such small samples can not be presumed. Table V. Inter-Correlations (rho) of Engrossment Level and Intra-Group Dimensions of Interaction. W Recip- Norm- Perceived Perceived Concern with rocity Shared— Evaluation Surveil- particular ness lance others Engross— ment .26 -.15 .09 .23 .63* *Significant at p .10 (non-significant at p .05). Further analysis of the findings relative to this hypothesis (and to the others) will be presented in the clos- ing chapter. 3bEngrossment was also correlated with measure of Attractiveness. Festinger ep al predicted that "de-individuated" groups will find their groups “more attractive” than will indivi- duals in "individuated“ groups. Thisrypothesis was taken over in this study. A positive association was predicted between Engrossment and the reSponse on the same questionnaire item Festinger used. A correlation of only .05 was obtained. The "attractiveness“ item: "Frankly, how much would you like to return for further discussions on similar tOpics with the same group (assuming your schedule to be free?)" Eefinitely Fairly Strong Don't Care Fairly Strong Definitely would like desire to re- one way or desire not to not want to turn the other return to return VI. INTERPRETATION “SD IMPLICATIONS The main purpose of this thesis was to explore some research implications for the study of crowd behavior. A rationale was presented for believing that symbolic inter- action theory might be a heuristic predictive scheme for interpreting and clarifying the relationships existing be- tween persons in small groups and in crowds. A research de- sign was devised and undertaken to probe selected hypotheses. The more Specific purpose of this research was to come to grips with how personally defined conditions of anonymity and surveillance within a small gathering differentially affect intra-group relationships between persons. Following a recap- itulation<1f'the study's hypotheses a summary of findings will be made. Substantive interpretation of the statistical findings must be tenuous yet it can be argued that the symbol- ic interactionist framework has been "put to a test.“ Hypothesis I. There is a positive correlation be- tween reciprocity, norm-sharedness, perceived surveillance, perceived evaluation, and perceived concern with others in a small gathering of persons. Small gatherings will tend to be characterized at consistent levels, on these dimensions. 51 52 Moreover, they will tend to be characterized at consistently high levels (as in the small-group mode) or at consistently low levels (as in the crowd mode) of these variables in situations of experimentally induced normative restraint reduction. Hypothesis II. A higher level of engrossment (i.e., an inability to identify particular persons' remarks theoretic- ally correSponding to a heightening of attention in a focal object of interest) will obtain in groups under the condition of anonymity than groups under the Surveillance condition. This will only obtain if, indeed, there is a higher level of perceived anonymity among groups in the experimental condition of anonymity. Hypothesis III. The engrossment level of a group will associate most closely with crowd-like gatherings, i.e., with groups evidencing low levels of reciprocity, norm- sharedness, perceived (external) surveillance, and perceived concern with particular others. Table V shows, by means of rank-order correlations (Spearman-Brown) the measure of association between engrossment and the reSpective variables of intra-group interaction. Hypothesis I: among the dimensions pinpointed as characterizing Small group interaction, surveillance and reciprocity each correlated most consistently with the other four dimensions of interaction tapped, All three of the significant product-moment correlations involved surveillance. 53 The wide range of variability found among the ten inter- correlations (from r = .23 to .77) is attributable to two or three "deviant cases" as Table II indicates. The existence of a general factor binding the dimensions "together" can not yet be denied. More precise Specification of these dimensions coupled with a larger sampling representation are necessary before more valid and reliable measurements of them, and the association between them, can be attained. There must be more conclusive evidence before this “system“ of variables (or single variables within the “system") can be accepted or discarded. Hypothesis II: Both in the measurement of identifi- cation errors per group and in the construct of engrossment (the number of identification errors per group's average member minus his [Ehe group's average member47 number of memory errors) no significant difference was found between groups under anonymity and under surveillance. The t-score difference between engrossment under those two types of con- dition was .77 (non-significant at.p>>.10). This leads to the belief that a group member's ability to Single out the Spokesman of particular views made during a discussion is un- related to structurally conducive conditions of surveillance 37 the expectation and anonymity. As indicated earlier, (higher engrossment under Anonymity condition) has not been disconfirmed . The hypothesis will not have been put to crucial test until there is a significantly higher level of perceived anonymity among members of those groups under the 37.5.3232: Po 4*9. 54 anonymity condition. Hypothesis III: The statistical findings relative to this hypothesis are the most discouraging to the researcher's expectations. Negative correlations would be predicted be- tween a group's level of engrossment and the reSpective dimen- sions of intra-group interaction gauged. Except for the correlation between engrossment and perceived concern with particular others (at variance with expectation), the data indicate only a relative absence of association between en- grossment and reciprocity (r = .26), norm-sharedness (r = -.15), Surveillance (r = .23) and perceived Evaluation (r = .09). However, the dataeppear to show that the feeling within and relating persons in the groups that "others were keeping in mind the stated feelings of each particular persons in the group" during a discussion is related (contrary to eXpectation) to high inability to recollect who made Specific remarks during that discussion. This may only demonstrate the weakness of the instruments used here to produce the desired relation- ships. It is obvious that some other behavioral consequence of heightened inability to single out the Speaker of Specific remarks must be measured in order that empirical validity for the engrossment index may be gauged. As for this hypothesis, as the first, the dynamic inter-relations between these dimensions of intra-group interaction is the goal: the face validity and independent isolation of conception for each relevant dimension must be improved. One suggestion: The Operations involved in constructing the engrossment score for a group used in this study Should not be confused with the 55 sequence of processes actually marking the emergence of crowd relationships. It would be desirable to construct the Operations so that they approximate or trace the theoretical sequence in some way. Another source of evocative ideas worthy of consider- ation were the subjective feelings acquired by the observor during the course of the nine group sessions. Confrontation with “real-life” small groups (even in laboratory situations) raised unforeseen problems egg insights which required both onethe-Spot adjustment and later reflection . Hunches develOped during the running of groups were probed after administration of the post-session questionnaire. Researchers can profit from such observation. These hunches were crude and amorphous and usually were not confirmed during these post-session conversations. Two gambits on the part of per- sons in groups seemed to take a relatively uniform pattern regardless of eXperimental Condition. First, a common tack for a group was to debunk the study reported to them. This renunciation of the methods or results of the psychiatric study read to them permitted a less glaringly incongruous (in light of the ficticious study's results) upholding of their love of their parents. Secondly, it was interesting to note that far more common than an outright positive assertion of the "I lpye my parents" variety were the less enthusiastic and ambivalent variations on a theme that; “But I can't believe that I hate my parents..., I'd know it if I did... I don't hate them." 56 Beside eXplaining some of the purposes of this research, post— session probing most often revealed that the subjects were unclear as to the purposes of the research (desirable) and that often, indeed, the presence of the observor‘was for- gotten. Although such a comment can not be taken at face value such a perception by participants would be ideal for the research. ESpecially under the surveillance condition (Presence of tape recorder), groups reported that they did not think that the researcher was recording their remarks yerbatim. One key relationship in this study possibly over- shadowed the meticulous arrangement of conditions and con- founded the researcher-"subject" role-relationship in this setting. Each group's participants were members (during the academic term the study was conducted) of an Introductory Social Psychology class taught by the researcher. This role- relationship of teadher-student may well have carried over and been the central role-definition of the situation made by the observor: non-involvement in the discussion, use of non- evaluative remarks in the instructions, and the assurance of confidentiality to be given their actions. Therefore, even though most participants displayed an ostensibly blase atti- tude toward being "tested," their remarks could have been in- fluenced by awareness of the teacher-student relationship. This awareness could have, in some way, impeded the partici- pant's recall of the particular participants who had earlier voiced the statements read to them after the discussion. 57 Likewise, the post-session questionnaire, although represented as having no right or wrong answers, could still have been construed as a test situation by the participants,38 sub- sequently affecting the results. Throughout this study careful consideration has been given to distinguishing, on the one hand, the calculated con- ditions imposed - Surveillance and Anonymity - and, on the other, the perceived conditions of relationship between the participants - e.g., perceived anonymity and perceived sur- veillance. This conceptual distinction has been matched by their independent and distinctive measurement as well. This research recognizes that the perceived and defined nature of a Situation more directly guides behavior than do structural- ly imposed conditions which can only be attempts or frame- works inducing such perception.39 This point is a rebuttal to those critics who object to small group studies carried out under “laboratory“ condi- tions. Many have questioned the relevance and similarity to “reallife” situations that contrived small group studies in 40 laboratory conditions can approximate. Ordinarily, such 38Item number 6 on the post-session questionnaire asked whether persons felt if they were being tested on something or other. Well over half indicated “possibly" or “yes.” 39Blumer, "Attitudes and the Social Act,” loc. cit. 40Ironically, symbolic interactionists themselves have been key critics. 58 studies do not attempt to approximate holistically dynamic small groups in the first place. Usually the objective is to eXplore or determine the effects of one variable upon another as isolated and controlled for extraneous (for the purposes of the study) variables. Control over extraneous conditions can generally be accomplished only by equating them as well as possible. Critics voice the objection that the underlying principles of “everyday“ interaction somehow are suspended when humans are placed in "artificial" (gig) and unaccustomed- to settings. This assumption is patently false. The small group researcher must accept the ”verisimili- tude" of the laboratory Situation for the rigorous examination of behavioral processes under various controlled conditions. Often in doing so, the researcher may fail to heed his own admonitions that persons are still acting as persons who are not failing to work out definitions of the research setting. The researcher may so constrict his view upon selected aSpects of small group process that his vieWpoint at to what is going on is "out of focus.“ But the participants are not viewing the setting in the same way as is the researcher.41 The researcher would not want them to. The persons are not becom- ing "subjects" or "humans in general" but remain Operating as persons: relating their personal identities to the Situation, 4lFor a stimulating discussion of this from a psycho- IOgist's standpoint, see Joan Crisswell, "The Psychologist as Perceiver," in Tagiuri and Petrull% Person Peggeption and Interpersonal Behavigg (Stanford, Cal.: Stanford University Press, 1958). 59 checking their definitions with others of what is required of them, etc. In short, they may not be gearing their behavior along the lines structured by the researcher. For example, the rules or manners g 13 the social act do not become sus- pended unless they, the participants (and not the researcher), have judged or concluded that new ground rules need to be enacted. The laboratory setting, then, is very real to the participants and their behavior in such settings will be chosen for its deemed apprOpriateness. In this study, the role definition of student and counter-role of teacher may have overshadowed both the researcher-subject role-relationship (which in itself would have been desirable) but also the participant-participant relationships sought in which the reduction of normative restraint would have been more likely. Crowd behavior would be more likely to ensue from normative restraint reduction if the surveillance of “teacher“ were not an image inhibiting the participants. Encouraging a particular definition of a situation to be held on a group wide-basis is a delicate process at best and even if definitions are eagerly sought in ambiguous situations (such as small group researches) they are skeptical- ly and tenuously held. Every small group researcher must face this problem anew in his work. The gamut of possible interpretations that can be formed by groups should not be ignored by the researcher: blinders only help one when one is certain where he is going. In the study of collective behavior processes at the small group level this can not be 60 said to be the case. In light of the foregoing warnings and perSpective, it is possible to reinterpret the findings obtained in the study. The firstand third hypotheses were generally uncon- firmed. How is this to be accounted for? Some reasons have already been suggested. The tape recorder's presence and Operation obviously was not a strong enough intervening variable i3 itself to pro- voke a feeling of being surveyed. The participants' inter- pretation of the situation did not seize upon its presence as necessitating a crucial change in relationships between them. It must be remembered that most gatherings will not fit (and do not fit) the extreme ideal type of either the "small group" or the "crowd." To expect that the groups studied in this research should fall into one or the other category merely upon a possible reduction of perceived norm- ative restraint is to ignore the reinforcing influences which bolster "small group" Operation. What is more, crowd relationships and small group relationships do not cover every conceivable type of relation- ship that can characterize a physical gathering of persons. How would the interaction of persons engaged in a “mysterious" ritual in a lodge meeting be characterized when the partici— pants have performed their reSpective roles tens of times? In any case, this is a defense of the small group as a method- ologically legitimate locale for the illumination of small crowd processes, small group processes, and other theorized modes 61 of relationship between personsthat occur in small gatherings. The feasibility of the symbolic interaction vantage point in generating distinctions between types of inter- personal relationships in crowds and in small groups was not at critical stake here. Nevertheless, some consequences for future research in this area can be offered. There are real differences between what has been typed as small group inter- action and as crowd relationships. Whether or not this form- ulation will correSpond most predictively and most heuristical- ly among competing explanatory schemes to actual crowds and actual small group can only be decided by further research. One possible line of research to crowd behavior could be to View the relationships between persons as they partici- pate in a "naturally occurring" collective event. Often the source of subjective Opinion, events such as fiestas (Mardi Gras) or crowds viewing a fire or joining in an impromptu (7) snake dance at a football rally could help substantiate, as well as provoke, hypotheses. The difficulties underlying such research hardly needsto be mentioned. However, such obser- vation should not be ignored. A more promising approach, derived from a symbolic interaction vantage point as was this study, to the behavior 42 of persons in gatherings has been offered by Goffman. Al- though he does not concentrate upon crowd processes or build 42Goffman, E. Behavior in Public Places (New‘York: McMillan, 1963). 62 a framework for conceptualizing them, his work offers leads a? to the understandingArelationships of persons in unstructured Situations as well as in focused gatherings. Perhaps the most promising (if difficult) attack upon the dynamic symbolic inter-personal transformations crowd participants undergo and undertake (if persons g9 differ as small group members and as crowd participants) would be to 43 obtain valid measures of self-attitudes or of "significant ll others44 during activation of crowd relationships. Nearly impossible during an actual holocaust, panic, or hostile mob action, at least the small group again would be an experimental setting in which such measures could be taken in order to detect differences and linkages between (a) mode of relation- Ship between persons and (b) differences (or changes) in the salience or priority of self-attitudes or "significant others.“ ‘ 43Kinch, J. ”A Formalized Theory of the Self-Concept", Amegican Jogrnal of Sociology, 68 (1963), pp. 481-486. Winifred Kuhn and Thomas McPortland, “An Empirical Investiga- tion of Self-Attitudes,“ American Sociological Review, 19 (1954), pp. 68-78. 44"Significant Other” a term extended from Mead's version of human develOpment, was coined by H. S. Sullivan, The Inter-personal Theogy of PsychiatgyI 1953. To graSp the meaning of the term might best be done by looking at the questions that researchers have used to tOp it: e.g. “Which person is it most important for you that he or she be proud of you?" APPENDIX A: POST-SESSION QUESTIONNAIRE 63 . '9 v . _ .'f I": .1 _, ‘IlV A - I. . a) '5‘ . i k ' ‘J I - ‘ a _ ‘ ' -v ‘ -v P ‘ L ' r . ' 0 V “ a d ) I ~ 1’ ,ery 0,1L 18h nwu n. 253;} rare ewe ‘ . .. _ , .! ~:0 I .. » 1 ,(V .' -- . 5 : - I'. ,1 . .2 .1 . . , A u 1’, . . - \ ‘. Yr u -" ““ .C’v ~- I ~" t ,1 "1 i ’ ‘H {'u" - ‘ ' I ' 'I he . . -. 'h‘r, vr'vx... 1 - ~f~ \._._-_ .-- _ nu" swarm .. It ...».b u... :v -- . 1 an I‘a . a > : n M_ .. . . ‘ . u .- ‘ - 1 ) J - -r f ‘ H "u ’ V ‘ 1':- II‘) 1 - v I Q 0 I ' 1 i A . . 1 a. ‘w‘ ". f 1‘ TA - ‘ I ‘ ~ - _ "‘ "-' - v. ‘ I‘.‘ ' ~ g. . ‘ ... ‘ . a . . . . . . . .. .1 v -., ,- ‘._ 1 . _ _ H .‘.. ‘ f - l ’ ‘ ‘f‘ ’ V A ' 1 »-, . ‘1' A . .1 : - ' ‘ ‘ ‘ l7 ‘ . ' ' ‘ ‘ ' T -» 1, ' . ¢ , 1‘ " ' " ‘ ‘ r in ' c-.- v -- .. r: .. . t . ’ . . » I ‘7‘ t " . _ ' . _ ‘1‘ , _ p g. .‘.. ‘ . , , .A '- _ u ; . ’ _ . ._ r " ',' ‘ k b . a , I" . , ‘ ‘ , ‘ M ' 1 . v : .' - . A , ~ v V 1, _ . a \ I I ’2 '. ~ .' 1 - ' ‘ ' ‘i’. ' . . ‘in I , 7 ‘ " I :- . . 7 p I I _ . . . . .. 1 ' 1‘ , - I , ~ I ‘f - ‘ . - z '1 . , ‘. '. r .< : ‘ ~ t. { ‘ ‘ 4‘ 9". I 4 1 ‘ . . ... ,‘ , . ’ l ‘ I. ' ' ' n v _ v"'\ I. ' g, If . '4 (A. . ' - :1 ' . M In ‘ .3 ~ . Q a n. 4 y. :1 .-I I, ‘ 3‘ ‘ ‘ '; ‘ ‘ ' v ‘ J ‘P ‘ ._ - v v ' ‘ ' I ‘ I . ~ .1 " ... ‘ lLr I '3 Z - "1' . 4 I z . .L . . , ~ v - . H “ Lin. 11.2.- ' ‘- ' ' ' ‘ ‘ x ! I ‘ é . I'. I‘ “v" V: g " I . v '0 " ‘M ' a‘ 0 . ~« . . ...—... . ‘ _. . .-- .. M _ ‘ y , .~. ‘3: "a . ' . - I I ' ' ' . ‘ .-. .I 1'. -.." L} a», '. i '1 .i -' 's- n . _ 2 9 .' ‘ ‘ ' ’ - .“ .- ‘ ‘. ‘ 1' 1. . .l‘ 3' .’ .T . - " y . ‘ ‘ 5 ~ -. ‘ _ u . , . -. . n .' .3-»l) .5»... -A )fi. - -' « Q r ' « r. ‘ . u ‘ ..‘, , 1 .. l A" I.) v. ‘ ‘1 f 1 v: \ I , .4 I. fl, ,-, . t r; .3 w. , 1.. . ' . .' J'. . I , \ \ _~ 1 . .....-r .. 1‘54. -. a! 7'- .5 1'; v . “ - ‘ ' u . I I; . , C .V 1 v .y 5' ' .‘ . , . . ~ ‘ ‘ y , ‘ I ‘ e , ‘ . n V .I I . ‘ .4 x A; ‘ - _ J . A J ‘ ., ' v I . . ' A.I - > ' 1 . V . I . I ‘ l .J B IBL IOG RAPE-TY Blumer, H. ”Social Psychology“ in E. P. Schmidt (ed.), £2? and Society. N.Y.: Boston-Hall, 1937. . “Collective Behavior“ in A. McClung Lee (ed.). New Outline of the Principles of Sociology, N.Y.: Barnes and Noble, 1951. . "Attitudes and the Social Act." Social Problems (III, 1955), 59-65. . "Collective Behavior“ in E. Gittler (ed.), Review of Sociology: Anglysis of a Decadey N.Y.: John Wiley, 1957. . "Society as Symbolic Interaction“ in A. Rose (ed.), Human Nature and Social Process, Boston, Mass: Houghton-Mifflin & Co., 1962. Brown, Roger. ”Mass Phenomena," in G. Lindzey (ed.), Handbook of Social Psychology, Vol. II. Reading, Mass.: Addison-Wesley, 1954. Cresswell, Joan. “The Psychologist As Perceiver" in Taguire and Petrullo (edS.) Person Perspective and Interpersonal Behavior, Stanford, Cal.: Stanford University Press, 1958. LeBon, Gustave. The Crowd (1895). New York: The Viking Press, 1960. Festinger, L., Pepitone, A., and Newcomb, T. "Some Consequences of De-Indivudation in a Group." qurnal of Abnormal and Social Psychology (47, 1952), 382-389. Goffman, E. Behayipr in Public Places, N.Y.: McMillan Co., 1963. Hare, A. P. Handbook of Small Group Research, N.Y.: The Free Press oqulencoeL I962. Hare, A.P., Borgotta, E., and Bales, R. Small Groups: Studies_in Social Interaction. N.Y.: KnOpf and Son, 1955. Kinch, J. "A Formalized Theory of the Self-Concept,” American Journal of Sociology, (68, 1963) 481-486. 64 65 Kuhn, Alfred and McPartland, Thomas. "An Empirical Investigation of Self-Attitudes,“ American Socio- logipal Review, (19,1954), pp. 64-78. Lang, Kurt and Lang, Gladys. ”Society as Symbolic Inter- action," in A. Rose (ed.), Human Nature and Social Process. Lang, Kurt and Lang, Gladys. Collective Dynamics (New York: Thomas Crowell Co., 1961). Lindesmith, Arthur and Strauss, Anselgn. Social Psychology, N.Y.: Hold-Dryden, 1956. Mead, G. H. Mindy Self and Society, Charles Morris (ed.)., Chicago, Ill.: U. of Chicago Press, 1934. Meltzer, Bernard. The Social Psychology of George Herbert Mead (Kalamazoo, Mich.: Division of Field Service of Western Mich. U., 1959). Newcomb, Theodore. "Social Psychological Theory: Integrating Individual and Social Approaches" in Hollander and Hunt (eds.). Current Perspectives in Social Psycho- logy, N.Y.: Oxford U. Press, 1963. Pfeutze, Carl. The Social Self, N.Y.: Bookman's Associates, 1954. Polansky, Lippet and Redl. "An Investigation of Behavioral Contagion in Groups," Humag Relations,(3, 1950.) Scott, John Finley. “The Internalization of Norms," A Paper read to the American Sociological Association, Los Angeles, California, 1963. Shibutani, Tamotsu. Society and Personality, Englewood Cliffs, N.J.: Prentice-Hall, 1962. Smelser, Neil. Theory of Collective Behavior, N.Y.: The Free Press of Glencoe, 1963. Strauss, Anselah (ed.), The Social PsycholOgy of G. H. Mead. Chicago, Ill.: U. of Chicago Press, 1956. Swanson, Guy. "A Preliminary Laboratory Study of the Acting Crowd," Amerigan Sogiological Reviefi (5, 1953), pp.422-433. Turner, Ralph and Killion, Arthur. Collective Behavior, Englewood Cliffs, N.J.: Prentice-Hall, 1959. HICHIGQN STQTE UNIV. LIBRRRIES 3129301 1058207