THE UTKUYY OF REFERENCE GROUP T'HEORY IN STUBYENG LEGISLATIVE BEHAVLOR: THE MKHKMN HOUSE (33‘ REPRESENTATWES Thesis For H13 Degree of M. A. MECHE'GAN SYATE UNJVERSWY kahn W. Scale 39‘65 114E518 LIBRARY y UniVersit " 1* Michigan State 0 6 6 7 \. I.. .II \I \l“ \\ \‘I‘I‘u‘ ‘\ ‘ltllillhl‘l‘ll I‘u“|.| “I‘m 1“.“ \u‘l. “ 0504 \\\\\\\\\ “WW BWLQ W \\ ABSTRACT THE UTILITY OF REFERENCE GROUP THEORY IN STUDYING LEGISLATIVE BEHAVIOR: THE MICHIGAN HOUSE OF REPRESENTATIVES by John W. Soule This study of the 1963 Michigan House of Representatives was undertaken to explore the legislative process from the perspective of contemporary reference group theory. The particular concern here is to discover those reference groups whom the Representatives con- sider most significant - their "significant others". By determining those individuals or groups that act as important points of refer— ence, it is possible to predict how they define various situations they encounter in the legislature. A personal interviewing technique was devised, and over 90 per cent of the representatives were contacted. Two questions were put to each representative: "Whose opinion of you as a legislator is most important to you?” and "What members of the House do you dis- cuss legislation most frequently with?" The first question deter— mined the most significant reference group of each representative, and the second was used to discover the leaders in the House. On the basis of these questions one chapter focuses on the reference group orientations of the legislators, and another on the impact these orientations have on a behavioral phenomenon - legislative leadership. The study showed: 1. Four generic classes of "others" were recognized as dominant reference groups by representatives in the Michigan House of Representatives. The groups perceived in their John W. Soule order of priority were interest groups, constituency, other legislators, and political party. No representative re- garded administrative agencies or the governor as a salient reference point. The particular reference group orientations held by individ- ual representatives were not significantly related to party affiliation, education or occupation. Instead, reference group orientations appeared to stem from the individual's basic perspectives prior to his legislative experience. One political variable did shape representatives' orienta— tions: those from competitive districts perceived their constituency saliently more frequently than representatives from ”safe" districts. In relation to the three other reference groups, political party was not recognized as an important reference point. Minority party members (Democrats) did, however, regard their party as important more often than majority party members (Republicans). Leadership was in the hands of a number of representatives. Majority party members occupied most of the informal leader- ship positions, Probably due to their formal positions of authority accrued by their majority status. Some of the variables related positively to leadership were a relatively high degree of seniority, law and agri- cultural occupations, and formal leadership positions. Still, none of these variables distinguished leaders from John W. Soule non-leaders as well as reference group orientations. Most relevant to the functioning of the House was the finding that all save one of the leaders held interest groups as their most salient reference points. THE UTILITY OF REFERENCE GROUP THEORY IN STUDYING LEGISLATIVE BEHAVIOR: THE MICHIGAN HOUSE OF REPRESENTATIVES By John W. Soule A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Political Science 1965 ACKNOWLEDGMENTS The writer wishes to express his appreciation to many per- sons who have made this thesis possible. Probably the first debt owed by the writer is to Stephen Monsma with whom the interview- ing of Representatives was shared. Professor Carl Couch gave much needed moral support and theoretical direction that is so necessary in writing a thesis at so large a university as Michigan State. In serving as Committee Chairman, Professor Robert Scigliano acted continually through the long drafting stages as an excellent correspondent and critic. Finally, a note of thanks is owed to my wife, Lucinda, who, like all good wives, aided her husband by offering him continual encouragement throughout all stages of the project. J. W. S. ii TABLE OF CONTENTS Page ACMO‘IJLEDGMENTS.0.00.00...IOOOOCOOOOOOOOOOOOO0.00.00.00.00... ii LIST OF TABLESOO oooooooooooooooooooooooooooooooooooooo 0.00000 iv Chapter I. INTRODUCTION. 0 o ooooooooooooooooooooooooooooooooo o o 0' 1 II. REFERENCE GROUP ORIENTATIONS OF MICHIGAN REPRESENTATIVES.. ......................... ........ 14 "Other Legislators" as Significant Others Interest Group Orientations Political Party Orientations The Constituency-Oriented Legislator Summary III. LEGISLATIVE LEADERSHIP...... ..... . .......... ........ 42 Reference Relationships Education Occupation Length of Service Summary IV. CONCLUSION.. ................................ ........ 50 BIBLIOGRAPHY.. .......................... . ............... ..... 66 iii a q .- c '. 1 ‘ v - s g .. . Table II. III. IV. VI. VII. VIII. IX. XI. XII. II. III. IV. V. VI. LIST OF TABLES Chapter II Length of Service and Choice of Reference Groups.... ooooo o.ooooooooooooooooooooooooooooooo Political Party Affiliation and Other LegiSlator orientationSOOOOOOOIOOOOOOOOOOOOOOOOO Formal Education and Other Legislator orientations.0...OOOOOOOOOOOIOOOOOOOOOOO0.00.... Occupations and Other Legislator Orientation...... Political Party Affiliation and Interest Group orientationOOOCOOOOOOOOOOOOIOOOODO0.00.0000...O. Formal Education and Interest Group Orientation... Occupations and Interest Group Orientations....... Length of Service and Interest Group Orientation.. Political Party Membership and Choice of REference GIOUPSQooooooooooooooooooooooooo000.00 The Effect of Safe and Close Seats on Reference Group ChOiceooooooo0000000000000-000.000.000.00. Formal Education and Constituency Orientation..... Occupations and Constituency 0rientations......... Chapter III Leadership of the Michigan House of Representatives The Leaders and Their Formal Positions............ Reference Groups of Representatives and Leaders... Formal Education of Representatives and Leaders... Occupation of Representatives..................... Length of Service in the House.................... iv Page 19 20 22 22 26 27 27 28 30 33 37 37 45 45 48 50 51 54 CHAPTER I INTRODUCTION Students of representative bodies have only recently recog- nized that they, like all social scientists, are engaged in the study of human behavior. Individual cognitive actors compose repre- sentative bodies, and they must be considered as Such before explan— atory research can begin.1 Legislative behavior then is human be- havior and is subject to all the difficulties and complexities that accompany the study of humans. The recognition of this human factor in political affairs has given rise to what is currently known as the behavioral approach. Contemporary social psychologists recently have made progress in the study of human behavior that is pertinent to the study of representative bodies. The uniqueness of humans consists of the kind of symbolic interaction in which they engage.2 This symbolic interaction refers to the peculiar and distinctively human phenome- non, in which humans define each other's actions instead of reacting in a simple stimulus-response fashion. To explain, predict, or 1For a similar argument, see J. Wahlke, "Behavioral Analysis of Representative Bodies," Essays on the Behavioral Study of Politics, ed. by A. Ranney (Urbana: University of Illinois Press, 1962), pp. 173-191. 2See Arnold Rose's excellent chapter "A Systematic Summary of Symbolic Interaction Theory," in A. Rose (ed.), Human Nature and the Social Process (Boston: Houghton Mifflin Co., 1962), pp. 3-20. l 2 even describe human behavior, one must be acutely conscious of this step of interpretation which is present between the stimulus and the response. The resulting reaction is based on the meaning assigned to the stimulus, not to the stimulus act itself. A simple diagram indicates this. stimulus interpretation response Fig. 1 Most previous research on representative bodies ignores this conception of a social act. Such research appears to fall into sev- eral distinct types.3 One variety has ascertained characteristics ot the memberships of representative bodies: age, sex, length of service, occupation, religion, ethnic origin, social class, etc.4 The assumption is made that these antecedent variables somehow deter- mine the structure and performance of legislative bodies. Whether or not these attribute characteristics are related to behavior is rarely tested and the reader is often left with purely descriptive data about which he can only infer the circumstances and conse- quences. The most tenable argument in favor of these statistical 5 . presentations is presented by Aydelotte , who sees "revealing 3See J. Wahlke, op. cit.; see also N. Meller, "Legislative Behavior Research," Western Political Quarterly, Vol. 13 (1960), p. 134. Both provide excellent reviews of the research that has been conducted on representative bodies. Meller's is the more comprehensive. 4N. Meller, op. cit., provides a listing and discussion of the most important of these studies, pp. 141—143. He notes, for example, that the characteristics studies include the cranial capacity of representatives' heads. 5W. O. Aydelotte, "A Statistical Analysis of the Parliament of 1841: Some Problems in Methodology," Bulletin of the Institute of Historical Research, Vol. 27 (1954), p. 141. correlations” in them. A second, more popular, variety of research has been the macrocosmic examination of the actions of the total aggregates. These studies attempt to measure the influence of factors thought to be determinants of collective decisions. For instance, the pres- sures exerted by constituency,6 party,7 pressure groups,8 and state delegations,9 have been examined in hopes of assessing the amount of influence exerted and the channels through which it flows. Such research offers a meritorious picture of the constellation of forces operating at that time, but it is limited to the particular legis- lature being studied at a specific time. Being fluid and unstruc- tured as they are, the legislative forces operating in a particular legislative system are continually changing and vary considerably from one system to another. Both varieties of research use what might be called the "clay man" model. The first type of study described above assumes the social class-interest model, where man acts according to the norms and values of his class or other status. The second type of study 6J. Turner, Party and Constituency: Pressures on Congress (Baltimore: Johns Hopkins Press, 1951). See also D. MacRae, Jr., "The Relation Between Roll Call Votes and Constituencies in the Massachusetts House of Representatives," American Political Science Review, Vol. 46 (1952), pp. 1046-1055. 7A. L. Lowell, "The Influence of Party Upon Legislatures in England and America," Annual Report of the American Historical Association, Vol. 1 (1901), pp. 321-343. 80. Garceau and C. Silverman, "A Pressure Group and the Pressured: A Case Report," American Political Science Review, Vol. 48 (1954), pp. 672-691. 9D. Truman, "State Delegations and Voting Alignments," American Political Science Review, Vol. 50 (1956), pp. 1023-1035. 4 most frequently postulates the pressure—politics model, in which representatives behave in direct proportion to the political forces supposedly being exerted upon them. Both types seem to this writer to exemplify the naive Watsonian behaviorist notion that humans act in a simple stimulus-response fashion; both tend to ignore the crucial definition, or interpretation, that the representative has of the political forces in his world. In other words, political parties, interest groups, constituencies, ideologies, and leadership take on meaning and value for the researcher only when viewed through the eyes of the representatives. Their perceptions then be- come the raw materials for study. Are representatives aware of the particular forces in their political world that cause them to define situations as they do? It is generally recognized by social scientists that stable cue-giving mechanisms are not operating on most issues facing American legisla- tors.10 unlike the British M.P., the American legislator is rela— tively free to make his own decisions. It is not likely that voting behavior in American legislatures is absent minded, or unconscious; on the contrary, it would seem to be one of the most significantly conscious acts possible. The voting situation is often imperfectly defined, full of ambiguity, and requires reflective, deliberative, thought—out responses that are subject to scrutiny of a variety of others. No response is ready-made, because the cue-giving mechanisms are vague and ambivalent. It can therefore be postulated that an American legislator is poignantly aware of the process by which he makes a decision. Whether and in what manner he will choose to 10For example, see E. Shils, "The Legislator and His Environ- ment," University of Chicago Law Review, Vol. 18 (1950—51), pp. 571-584. 5 explain his method is another matter, but the point remains that he should be able to do so. Theoretical Orientation One of the major weaknesses in much of the previous research on representative bodies is the lack of a sound and unified theory that would allow general knowledge to accumulate.11 The lack of a consistent model of the individual representative, or the represen- tative body itself, is a major reason for this weakness. It would be well then to define the terms and make clear the theoretical orientation used in this study at the outset. The concept "reference group" did not gain currency among social scientists until Herbert Hyman used the term in a study published in 1942.12 Since that time, the concept has been popularized and used in a number of different ways. Generally speaking, it refers to the perspective selected by the individual for the guidance of his behavior, especially in a situation where a choice has to be made. It may be a group of which the individual is a member, or it 13 . ~ . may not. For the purposes of this paper, a reference group is an 11 1953). 12H. H. Hyman, "The Psychology of Status," Archives of Psy— chology, No. 269 (1942), pp. 128-174. See D. Easton, The Political System (New York: Alfred Knopf, l3See T. Newcomb and W. Charters, ”Some Attitudinal Effects of Experimentally Increased Salience of a Membership Group," G. Swanson, E. Hartley, T. Newcomb, (eds.) Readings in Social Psychology (New York: Henry Holt Co., 1952), pp. 115-430. A membership group is more frequently referred to by the individual than are groups which 'the individual is not a member. 6 audience whose presumed perspective is used by the actor as his frame of reference. In some fashion each of us performs for some kind of audience; our conduct is oriented toward certain people (real or imagined) whose judgment is deemed important. In this sense a reference group is the source of an individual's values,14 and perspectives.15 Sherif, Hartley, and Newcomb have also employed the concept in this sense.16 The utility of the concept "reference group" rests in part on its being able to explain behavior that is oriented toward audiences that are not obviously represented on the scene, or toward audiences that are prolixly represented. The latter case is apt to be found true in the legislative arena, where numerous individuals or groups are possible audiences for the representative. More Specifically, the utility of reference groups lies in their ability to point out how a representative consistently defines a succession of situations. By identifying the audience to which he plays, it is likely that one might predict his definition of various situations, and thereby, his behavior. There are difficulties present in dealing with the concept "reference group" that might profitably be illuminated. The term 14H. H. Kelley, ”Two Functions of Reference Groups," T. Newcomb, G. Swanson, and E. Hartley, (eds.), op. cit., pp. 410—414. 15T. Shibutani "Reference Groups and Perspectives," American Journal of Sociology, Vol. 60 (1955), pp. 562-569. 16M. Sherif, An Outline in Social Psyghology (New York: Harper Brothers, 1948), pp. 105-106, 123, 150—155. See T. Newcomb, op. cit., pp. 220—232. See also E. Hartley, "Psychological Problems of Mul- tiple Group Membership," in J. Rohrer and M. Sherif (eds.) Social Psychology at the Crossroads (New York: Harper Brothers, 1951), pp. 371-386. 7 implies a number of people (group), which in actuality a reference relationship might well be constituted by one person.17 Another greater difficulty may be the number and complexity of situations the representative faces daily. Depending on the particular bills being considered, the audiences very likely change. Reference groups, furthermore, are relative in that they form a continuum from high to low with some of the low ones having a negative value. Rose has pointed out that negative reference relationships seem to occur when a person is forced into having the relationship against his personal values, and is obliged to act in accord with those others' expectations.18 In the case of a representative holding antipathy toward Negroes, yet having a sizeable block of them in his constitu— ency, a negative reference relationship may well develop in which he consciously acts contrary to the wishes of that group. Various studies have emphasized the fact that how a human be- haves depends on the kind of person he regards himself as being. Eulau, for example, found that the manner in which a man votes de- pends more upon the social class with which he identifies himself rather than on the class a scientific observer might place him.19 The point such research demonstrates is that one's perspectives, or frame of reference, does not conform to ”objective reality”. At this point, "reality" takes on a purely subjective meaning and 17A. Rose, op. cit., p. 11. Rose points out here some of the misconceptions concerned with the concept "reference group" including the fact that often one person constitutes an important reference point. 181bid., pp. 10-14. 19 H. Eulau, "Identification with Class and Political Behavior," Public Opinion Quarterly, Vol. 20 (1956), pp. 515-529. 8 conduct is directed by symbolic representations of the social world. It is crucial then that each individual's frame of reference be investigated and not imputed. Despite the obvious presence of certain individuals and groups in all legislative systems (e.g., political parties, floor leadership), their significance is dependent on the legislators' perceptions of them. (The following diagram illustrates this.) the multiplicity of political the individual representative's groups operating in represen- 1 frame of reference tative bodies 4a- 2 his behavior his definitions of various situations aArrow #4 completes the social act and shows the individual representative's behavior influencing the behavior of others. Fig. 2 Closely related to the concept "reference group" are "role" and "role-taking". For our purposes a role is a collection of pat- terns of behavior which make up a meaningful social unit in society. Role-taking is the imaginative construction of others' roles. When people seek to "make sense" or to understand someone's behavior, they are generally seeking to ascertain the role of which the observed actions are a part. Role-taking allows people to understand each other thus allowing for mutual expectations and then co-ordinated behavior.20 Before reference relationships can occur, role-taking 20See R. Turner, "Role-Taking and Reference Groups," American Journal of Sociology, Vol. 61 (1955-6), pp. 622—636. This excellent article points out the complexities of the concept "role-taking" and its relation to reference groups. must be involved. The particular standpoint the role-taker assumes in relation to a reference group determines the merit he assigns to the values and attitudes of that group. Due to the variety of standpoints a legislator may assume in taking the roles of numerous individuals and groups, it seemed advisable to concentrate on his "significant others" rather than on the entire gamut of his reference relation- ships. The method used to obtain the legislators' significant others, or most salient reference groups, will be discussed shortly. It is helpful at this point to recall Parsons and Shils' defi- nition of a social system, in order to fit this conceptual framework to the study of American legislatures. A social system is a system of actions of individuals, the principal units of which are roles and constellations of roles. It is a system of differentiated action, organized into a system of differentiated roles. Internal differen- tiation, which is a fundamental property of all social systems, requires integration. Different individuals within a social system, like a legisla- ture, have different roles to play, but the assortment of differen- tiated roles are integrated and co-ordinated with each other through individuals' ability to role-take. Co-ordination of indi- vidual roles is dependent on fairly accurate role-taking; this is vital if a social system is to attain its objectives. Individual legislators then perform a variety of roles that are both differentiated from and co-ordinated with those performed by other legislators. The legislature as a social system should, 21T. Parsons and E. Shils, Toward a General Theory of Action (Cambridge: Cambridge Press, 1959), p. 197. 10 furthermore, be conceived as the agency which transforms "inputs" (demands upon the legislature from within and from outside of the immediate system) into "outputs" (legislative actions).22 Inputs enter into the legislature where they are perceived and interpreted with the results becoming outputs. The raw materials for this study are the perceptions individual representatives have of some of the inputs. These raw materials were gathered from the study of the Michigan House of Representatives, Second Session, 1963. Using the interactionist model that legisla- tors are role-taking actors, this study will examine how legislators define and interpret certain stimuli and specifically how this is related to the intra-legislative socialization process and to legis- lative leadership. For these purposes, the following hypotheses were formulated at the outset: 1. Since the importance of political parties, constituencies, other legislators, and interest groups has been ascertained by many students of representative bodies, it was hypo- thesized that these groups would constitute the legislators most salient reference groups or significant others. a. Newer members will have other legislators as reference points more often than will those members with more tenure. In adjusting to the norms of legislative life the new member must pay particularly close attention to his colleagues. Since many norms are not formalized in a rule book, but are unwritten, the only method new 22J. Wahlke, H. Eulau, W. Buchanan, L. Ferguson, The Legislative System: Explorations in Legislative Behavior (New York: Wiley and Company, 1962). 11 members have of ascertaining these norms is through the symbolic reconstruction of other legislators' roles--i.e., through role-taking. Legislators with more seniority will have interest groups as salient points of reference more often than will newer members. After the legislator accumulates seniority and finds himself seated in the more impor- tant committees, a variety of interest groups will likely recognize his newly acquired roles; hence they will attempt to have more contacts for the purpose of influencing him. The legislator with seniority is also in a position where he is more likely to be personally familiar with full-time lobbiests. Since the Republicans occupy the formal positions of power due to their majority status, they receive more attention from special interest groups than Democrats. Thus interest groups should act as reference points more often for Republicans than for Democrats. The minority party Democrats will perceive their party as a salient reference group more often than will Repub- licans. Michigan Democrats have been traditionally conceived of as standing for a relatively cohesive set of principles and representing ecologically homogeneous constituencies. Republicans on the other hand are often portrayed as a conglomoration of liberals and conserva- tives representing heterogeneous communities. Demo- crats, therefore, should find it much easier to identify 12 with their party than Republicans. e. Members from both parties who are from competitive districts will have their constituencies as salient reference groups more often than other members. Since the threat of removal is a potent one, we can assume those representatives from "swing" districts would be much more conscious of their constituents than those representing "safe" districts. 2. Legislative leadership can best be perceived in a functional manner in terms of reference group theory. It was hypoth- esized that those legislators playing informal leadership roles will have a high degree of agreement on what group or groups constitute significant others, due to the similar- ity of their role expectations. a. Variables commonly employed in research on legislative bodies as independent predictors (e.g., age, education, occupation) will not be significantly related to leader- ship. If we assume leadership to be a behaviorial or functional phenomenon, we can postulate the importance of attitudinal variables (i.e., reference groups) in determining behavior patterns over and above that of attribute variables.23 b. Informal leaders will hold formal leadership positions. 23G. Schubert, Judicial Behavior: A Reader in Theory and Research, (Chicago: Rand McNally Company, 1964), pp. 5-6. Research in judicial behavior indicates that prediction of behavior based on attitudinal data is much more successful than that based on attributive data. l3 Informal leadership is operationally defined as being regarded as a frequent source of information about proposed legislation. Since legislators holding formal positions of leadership encourage other members to interact frequently with them about proposed legisla- tion by acting as focal points in the communication networks it is expected that informal leadership will correspond closely with formal leadership. Most important demands, or inputs, on a state legis- lature come from outside the system. Leaders, there- fore, will tend to go outside their immediate social system in their choice of significant others. Since interest groups are frequently depicted as the major sources of legislative inputs, it is expected that they will constitute the leaders' most significant reference points. CHAPTER II REFERENCE GROUP ORIENTATIONS OF MICHIGAN REPRESENTATIVES Stephen Bailey's "vector analysis" is probably the most widely accepted theory which encompasses the entire legislative process.1 This general theory portrays legislative policy making as "the re- sult of a confluence of factors streaming from an almost endless number of tributaries".2 The problem in using the vector model in studying legislative behavior lies in the implicit assumption that these tributaries or forces move along a one-way street. As stated earlier, the researcher cannot assume what he observes as an inves- tigator to be identical with what the actor observes. Let us, then, proceed to let the legislators guide us back down the one—way street from the other direction. Behavior becomes socially meaningful only when it is attached to a social role. As observers, we can define another's behavior only as it conforms to our expectations of his role. It is, of course, possible to infer someone's role from his conduct, but the import of that behavior is that the role is defined prior to it, and that the action was performed in accordance with it.3 Crucially 1Bailey's study is an excellent example of the way a case study may be used as a fertile source for hypotheses concerning the legis- lative process. S. Bailey and H. Samuel, Congress at Work (New York: Henry Holt Company, 1952). 2Ibid., p. 236. 3H. Eulau also uses this approach in dealing with "role theory" in J. Wahlke, et 31., op. cit., pp. 241-242. 14 15 related to role performance is the individual's ability to role- take, discussed earlier. The intent in this chapter is to identify and describe a legis- lator's role-taking in terms of his most salient or significant ref- erence groups in the course of his participation in the legislative process. While it is likely that the legislator has many points of reference, depending on the situations encountered, one is probably more pervasive than others in determining the way he defines various situations. A representative, for example, may assume a party- oriented role if his party is his most significant reference group. He may then introduce legislation consistent with the principles of his party, attempt to obtain party support in committee, mobilize party support on the floor, refuse to compromise with members from the other party, and perhaps go down to defeat on the final roll- call. The first task is to discover those relevant others in the legislative arena who represent significant others for the legis- lators. These significant others, through their expectations of the legislator, and towards whom he has expectations as well, do much to define his role performance in the legislature. On the basis of previous research on legislative bodies, it has been postulated that other legislators will act as significant reference points for many legislators, because it is they who determine the unwritten "rules of the game", the priority of group tasks, and the collective formulation of public policy. In addition to other legislators, certain other individuals and groups make up the diversified clientele, who are also deeply 16 involved in the legislative processes. Party activists, constit- uents, and pressure group members undoubtedly maintain expecta- tions concerning the legislators' behavior that cannot be utterly disregarded. The general role of a representative, nevertheless, would seem to allow him leeway in choosing his own significant others among these clientele. By exploring the relationships between Michigan Representatives and their role-taking standpoints in regard to generic classes of clientele groups, much light might be cast on how "legislative socialization" takes place. Some basic differences in the two par- ties may be illuminated, and several common political variables can be tested in order to explore why legislators regard one group more significantly than another. Does, for example, coming from a com- petitive district influence the legislator's role-taking standpoint toward his constituents? "Other Legislators" as Significant Others Group behavior consists of co-ordinating individual lines of action, which means that the members generally agree on certain patterns of behavior that are acceptable as well as those that are unacceptable. Such mutual expectations are not always communicated 4 in any "rule-book". In other words, the legislative arena is not 4Although not of central concern here, it was found in inter- viewing the Michigan Representatives that virtually everyone was aware of certain unwritten "rules of the game". There was consider- ably less consensus on what these rules were, or how members make it difficult for those who do not follow them, yet there existed e- nough agreement to see that certain common sanctions were present and often were employed against ”mavericks". For a list of 42 un- written rules of the game in four state legislatures, see J. Wahlke, etc., op. cit., pp. 134-140. 17 one in which the actors can do as they freely wish. The process by which a legislator becomes aware of the norms operating in the legislature may be called "legislative socializa- tion".5 In order for the legislative actor to align his actions with those of other legislators, he must ascertain what the others are doing and what they intend to do. By taking the role of indi- vidual legislators, or the role of the entire group (Mead's "generalized other"),6 the member can understand the meaning and direction of others' actions, thereby allowing him to align his own actions with theirs. New members (first and second termers) are likely to use other legislators as salient points of reference whether or not the system is complex, but due to the numerous twists and turns in the legis- lative process, he would appear to be doubly aware of his immediate others (other legislators). In becoming socialized into the legislature, a new member would seem to follow a set of precedents much like those found in the judicial process. It is postulated, therefore, that the new members in Michigan's House of Representatives will hold other leg- islators as significant others, moreso than will members with 5See G. Romans, The Human Group (New York: Harcourt Brace, 1950); in Homans' words a norm is "an idea in the minds of the mem- bers of a group, an idea that can be put into the form of a state- ment specifying what the members should do, ought to do, are ex- pected to do under given circumstances". (p. 341). 6Mead discusses the "generalized other" in C. Morris (ed.) Mind, Self, Society (University of Chicago Press, 1934), pp. 154-156. Although this explanation of group behavior may appear unduly long- winded, it encompasses what Mead thought were the fundamental fea- tures of all group action. By taking the role of the "generalized legislator", the actor becomes related to the legislative whole just as "an atom is related to a molecule". (p. 156). 18 greater seniority. In order to test this supposition, 99 of the 110 Representa- tives were asked the open-ended question in personal interviews: ”Whose opinion of you as a legislator is most important to you?"7 The most salient response was assumed to indicate the representa— tive's most significant audience or reference group. The answers fell quite easily into one of the four following categories: (1) other legislators (consisting of any number), (2) political party (on the local, state or national levels), (3) constituents, or (4) interest groups (again, as a generic class rather than particular groups such as unions, farmers, manufacturers, etc.).8 By using four generic classes of others we gain the advantage of comparability in that these same classes of others are undoubtedly present in other state legislatures. Since the study is exploratory, it was also felt that it was best to use general classes of reference groups and that future research would use these general orientations as a framework from which to pinpoint specific reference groups of legislators. It is true however that all those legislators who 7A11 representatives were interviewed on the House floor for an average period of twenty minutes. This question preceded all others so as to avoid structuring the answer to it in any way. The legis- lators were not asked to funnel their responses into prearranged categories; the question was completely open-ended. A probe was some— times necessary: "What individual's opinion, or what group's opinion, of you as an individual legislator is most important to you?" 8Although 99 members were interviewed, only 94 usable answers were received to this question. A fifth category was formed by the response: "My own opinion is most important to me". Since only a few of these responses were received, and since this study was not concerned with narcissism, which this answer tends to indicate, this category was eliminated and the representatives' second answer used. In all other cases, the legislator's first responses was used. 19 were oriented toward constituency, for example, did not have the same thing in mind. Some referred to the "peOple in my district," others referred to the "groups in my district," and still others to ' This same problem of validity, or what these "friends back home.' orientations meant to the respondents, is present in each of the four orientations. It is important that the reader keep in mind that these orientations refer to Eypgg of reference groups rather than specific groups. It is a fair assumption, however, that dif- ferences within any single reference group orientation are much less than those between orientations. To test the hypotheses concerning legislative socialization, representatives were classified as new (first and second termers), intermediate (third and fourth termers), and senior (those in their fifth term or more). The data in Table I indicate the relation- ships between legislators' length of service and their significant others. The mean number of terms served by other legislator oriented members was 3.6 as compared to 3.9 terms for legislators holding other orientations. TABLE I.--Length of Service and Choice of Reference Groups Reference Groups New Members Intermediate Senior Members Members No. Z No. Z No. Z Other Legislators 8 25 10 40 7 19 Interest Groups 10 31 6 24 16 43 Political Parties 4 13 4 16 2 5 Constituents 10 31 5 20 12 23 Total* 32 100 25 100 37 100 *Reference groups were recorded for 94 of 110 representatives. i 20 While the chart shows that newer members paid closer atten- tion to other legislators than did the older members, it was the representatives with a medium length of service that regarded other legislators most saliently. We might infer that a legislator must serve several terms before he comes to appreciate the importance of his colleagues. More will be said about the representatives' selec- tion of reference groups different from "legislative others" pre- sently. The 25 Representatives who perceived other legislators saliently were separated in order to explore their relationships with several variables commonly used in legislative analyses. Data in Table 11 indicate that Republicans were more likely to perceive other legis- lators significantly than were Democrats. TABLE II.v-Political Party Affiliation and Other-Legislator Orientations Reference Group Democrats Republicans No. Z No. Z Other legislator-oriented 7 16 18 37 Non other legislator—oriented 38 84 31 63 Total 45 100 49 100 Some reason for this might lie in the slim six member majority the Republicans had in the House (58 Republicans, 52 Democrats). If they were to maintain control on roll calls and other matters, the Republicans needed to cooperate closely with one another. The Republicans also controlled the formal positions of leadership, which may have contributed to the increased importance of interaction 21 among Republican Representatives. As will be demonstrated in the next chapter, the Democrats had a more highly centralized informal leadership structure than did the Republicans, which may have caused the latter to be more aware of a number of other legislators. Perhaps the most striking finding is that both new and ex- perienced members had a great deal of freedom in their choice of significant others, providing they stayed within the boundaries of these four groups. The data indicate that the representatives take the roles of a variety of relevant others implicated with them in "others" was legislative activity, but that any of the four groups of sufficiently implicated in the legislative process to constitute. salient reference points for a number of legislators. The data do not show that any drastic changes occur in the representatives' role—taking vieWpoints over the period of time served in the House. From this kind of study it is, indeed, diffi- cult to measure attitudinal change. A longitudinal approach in the form of case studies is perhaps better suited to measure the change in legislators' perspectives toward their relevant others over a period of time. While 27 percent of Michigan's Representatives held other legis- lators as their most salient points of reference, we must neverthe- less challenge Eulau's brash supposition that "the most immediate significant others in a legislator's environment are, of course, 9 other legislators". Other legislators are the most immediate oth- ers of all representatives, but it by no means follows that they are 9J. Wahlke, H. Eulau, L. Ferguson, and W. Buchanan, op. cit., p. 242. 22 significant others. In an effort to explore some other factors which might be re- lated to other legislator orientations, we looked at educational and occupational data. Table III and Table IV include data on the formal education and occupations of those legislators who had other legislator orientations. Legislators with this orientation have TABLE III.--Formal Education and Other Legislator Orientations Education Other-Legislator Non-Other Legislator Orientation Orientation No. Z No. Z College Degree 13 52 25 37 Beyond High School 9 36 30 43 Less than High School 3 l2 14 20 Total 25 100 69 100 TABLE IV.--Occupations and Other-Legislator Orientationa Occupations Other-Legislator Non-Other Legislator Oriented Oriented No. Z No. Z Agriculture 3 12 17 25 Real Estate, Insurance 3 12 14 20 Labor Unions 0 0 4 6 Law b 6 24 3 4 General Business 13 52 29 42 Unknown 0 0 2 3 Total 25 100 69 100 aThe data were derived from personal interviews and Michigan Legislative Handbook, 1963-1964, compiled by B. Kenyon and N. Philleo (Lansing, Michigan: Speaker, Hines and Thomas, Inc., 1963). bIncludes managers, wage earners and public servants. 23 somewhat more formal education than other members, although not enough to postulate a casual relationship between formal education and other legislator orientations. A substantially higher per- centage of those legislators who had other legislators as their significant others were lawyers than were the legislators who held other reference group orientations. The professional training law- yers have had appears to make them more concerned with their peers in the legislature than with individuals and groups outside the system. Ostensibly we have found that about a fourth of Michigan's Representatives are likely to perform roles oriented toward other legislators. Exactly what does this mean in behavioral terms? It probably means these representatives are more easily influenced by other legislators in floor debate; that they are most effectively influenced by their colleagues rather than outside pressures brought to bear by constituents and lobbyists. These are, of course, hypo- thetical statements that need empirical investigation. Further re- search would be interested in learning more about the category "legislative others". Does this reference group, for example, en— compass all the relationships between the legislator and his friend— ship group, his fellow committee members, the House leadership, or his party leadership in the legislature? Here we must be content to describe the four general reference group orientations. In the process, several variables suggest themselves as premises that might well be underlying causes for a particular orientation. These variables will be tested in connection with each particular orientation. 24 Interest Group Orientations Rousseau's view that all conflicting private interest, by their very existence, were inimical to the public interest has for the most part given way to the belief that interest groups collectively constitute a legitimate force in the formulation of public policy. This is not to say, however, that the legitimacy of pressure group activity is universally admitted. In the words of Belle Zeller: Before a legislature this mighty array of pressure groups clashes. . . . The legislature should serve as an impartial referee in seeing that these opposing groups do not take advantage of one another . . . but more chaplthat it should protect the interest of the general public. Not only has interest group activity undergone normative analy- ses, but it has also been subject to empirical theorizing. Arthur Bentley is probably best known for this theoretical treatment of interest group activity,12 and David Truman for a practical demon- stration of it.13 DeSpite the empirical attention political scien- tists have given to relating interest group activity to the legis- lative process, little is known about how or why the legislator is is influenced by pressure groups. Instead, the pressure groups 10Organized interest groups are active in all developed Western political systems. For evidence of this, see J. LaPalombara, "The Utility of Interest Group Theory in Non-American Field Systems," Journal of Politics, Vol. 22 (1960), pp. 29-49. See also H. Ehrmann, Interest Groups on Four Continents (Pittsburgh, Pa.: University of Pittsburgh Press, 1958). 113. Zeller, Pressure Politics in New York (New York: Prentice Hall Co., 1937), p. 264. 12A. Bentley, The Process of Government (Bloomington, Indiana: Principia Press, 1949). 13D. Truman, The Governmental Process (New York: Alfred A. Knopf, 1960). 25 are analyzed in terms of their size, geographic distribution, co- hesiveness, characteristics of leadership, strategic position, wealth, and resources, and so on.14 As can be inferred from the earlier chapter, the notion that these critical characteristics impinge on passive legislators is much over-simplified. A more meaningful approach to the extent which interest groups influence legislative behavior would necessarily include the legis- lator's own perceptions. In this paper we are concerned with legis- lators' orientations toward interest groups as a generic class of significant others, not with their affiliations with particular groups. The basic issues are, of course, to what extent legislators recognize the demands of pressure groups and to what factors a legislator might owe his recognition. In the Michigan House, 32 legislators (34 per cent) of the 94 members interviewed were found to regard interest groups as salient points of reference. While this group was about evenly divided be- tween Republicans and Democrats (Table V), it does not follow that members from both parties were oriented toward identical pressure groups. Although no data was specifically collected concerning positive or negative attitudes of legislators toward pressure groups, it was 14See, Ibid., pp. 158-188. 151m respect to the question on the interview schedule, legis- lators often named specific interest groups. The generally held notion in Michigans politics is that Republicans are closely linked with agricultural groups (e.g., Grange, American Farm Bureau), and the Democrats with labor groups (e.g., A.F.L.-C.I.O., United Mine Workers). This notion was found to be correct: Republicans did in fact tend to name agricultural groups as their significant reference groups while Democrats tended to choose union groups. 26 TABLE V.——Political Party Affiliation and Interest Group Orientation Reference Groups Democrats Republicans No. Z No. Z Interest group—oriented 15 33 17 35 Non-interest group oriented* 3O 67 32 65 Total 45 100 49 100 *This group includes those legislators who regard political parties, other legislators, or constituents more saliently than interest groups. assumed that if a legislator chose an interest group or interest groups, saliently, it indicated a positive relationship. It was found in the Wahlke study of four state legislatures that those legislators who were favorably oriented toward interest group activity had a higher degree of formal education than did leg- islators who were less favorably oriented.16 The data in Table VI do not bear out this conclusion in the Michigan House. Those legis- lators who had interest groups as their most salient points of reference had slightly less formal education than did the membership as a whole. Table VII compares the occupations of the total membership with those of the interest group-oriented legislators. With the exception of the lawyers, who were underrepresented among the 16J. Wahlke, H. Eulau, L. Ferguson and W. Buchanan, op. cit., see Chapter 14. In this study positive, neutral and negative atti- tudes legislators had toward interest groups were ascertained. Those legislators who regarded interest groups favorably (facilita- tors) are very likely similar to those legislators that this study has called "interest group oriented legislators." 27 TABLE VI.--Formal Education and Interest Group Orientation Education Interest Group Non Interest Group Oriented Oriented No. Z No. Z College Degree 13 41 25 40 Beyond High School 11 34 28 45 Less than High School 8 25 9 15 Total 32 100 62 100 TABLE VII.--Occupation and Interest Group Orientations Occupations Interest Group Non Interest Group Oriented Oriented No. Z No. Z Agriculture 8 25 12 19 Real Estate, Insurance 7 22 10 16 Labor Unions 2 6 2 3 Law 0 0 9 15 General Business 14 44 28 45 Unknown 1 3 l 2 Total 32 100 62 100 , l7 . . pressure—group oriented, there were no significant differences between the occupations of the total membership and those legisla- tors having interest groups as salient points of reference. Inasmuch as the data here do not measure the particular interest 17It has been suggested that lawyers, by virtue of their train- ing and the nature of their job, are likely to have close contacts with interest groups. D. Derge, "The Lawyer as a Decision Maker," Journal of Politics, Vol. 21 (1958), pp. 408-433. 28 group affiliations of the legislators, little can be said about the ' or agents, for their extent to which they act as "ball carriers,‘ occupational groups. It is important, however, to remember that actual membership in a formal sense is not identical with psycho- logical membership. Reference groups, as pointed out earlier, are associated wholly with psychological identifications; formal member- ship is by no means necessary for a group to act as a reference point for an individual. TABLE VIII.--Length of Service and Interest Group Orientations Representatives Median Number of Terms Interest Group Oriented 4.1 Total Membership 3.5 Interest groups were held as a significant reference group by members having a longer record of service in the Michigan House (Table VIII). Although attitudinal change pg£_§g_is not measured in Table VIII, the data indicates that interest group activity is more recognizable and, thereby, more appreciated by experienced legislators. It suggests also that interest groups may focus their efforts on more experienced members who have likely attained impor- tant leadership and committee positions. Other than length of service, the political variables commonly used in studying legislative behavior were not associated with 18See D. Truman, op. cit., pp. 338-40. He points out that group pressures need not be overt, but are often exerted by the legisla- tors' own preconceived attitudes. These confusing, but real, circum- stances cause the concept "group pressure" to have a limited value. 29 interest group orientations. A previous study of four state legis- latures found that this role orientation "is associated with none of the demographic characteristics for which data about legislatures are most often collected and tabulated--not with socio-economic status, not with urban-rural composition of districts, not with legislators ages, nor with their religions."19 That over one third of Michigan's Representatives chose inter- est groups as their most salient reference points is of no small significance, although it is a far cry from Bentley's theorizing that the legislature's pply_function and purpose is to accommodate clashing interest groups.20 Much more theorizing and empirical re- search is needed to explore the relationship between legislators and interest groups before the real impact of interest group activity can be assessed. Fruitful research along this line, of course, con- sists of examining interest group activity through the perceptions of the legislators. Political Party Orientations Whether a legislator is a Democrat or a Republican probably constitutes the most widely recognized division of political roles in society. This acknowledged recognition, however, does not in- volve corresponding expectations of behavior appropriate to the labels. The term "party" has probably as many different meanings in American politics as there are politicians. In Congressional 19J. Wahlke, H. Eulau, L. Ferguson, W. Buchanan, op. cit., p. 329. 0 2 A. Bentley, op. cit., pp. 111-120. 30 voting behavior, party label is the single most reliable indicator, but "it is admittedly somewhat less than perfect."21 On the state level the phrase just quoted is even more of an understatement of the relation between voting behavior and party affiliation.22 The simple facts are that on both the state and national levels the party label has severe limitations in enabling one to predict legislative behavior. Individual legislators at both levels dissent frequently from the views of their party col- leagues on matters of major and minor importance. In short, there appears to be little consensus among legislators as to what behavior is called for in performance of a partisan role. Table IX shows the two-party breakdown in terms of the most TABLE IX.--Political Party Membershipand Choice of Reference Groups Reference Group Democrats Republicans No. Z No. Z Political Party 7 l6 3 6 Other Legislators 7 16 18 37 Interest Groups 15 33 17 35 Constituents 16 35 11 22 Total 45 100 49 100 salient reference groups chosen by the individual representatives of each party. Only ten out of the ninety-four representatives inter- 21D. Truman, "The State Delegations and the Structure of Party Voting in the United States House of Representatives," American Political Review, Vol. 50 (1956), pp. 1023-1045. 22W. Keefe, "Comparative Study of the Role of Political Parties in State Legislatures," Western Political Quarterly, Vol. 9 (1956), pp. 726-742. 31 viewed chose their party as a significant reference group. Although the figures are small, over twice as many Democrats as Republicans perceived their party saliently. The clear conclusion is that Michigan Representatives place their political parties in a posi- tion of secondary importance in relation to the other three commonly perceived reference groups. Only eleven per cent give priority to party.23 Michigan is often thought of as having a healthy two-party system. The party balance in the House during the 1963-64 session was very close (58 Republicans, 52 Democrats). Despite the long Republican dominance of both Houses of Michigan's legislature, the statewide elections are becoming increasingly competitive. Each party appeals to distinct geographic reaions and socio-economic groups. The Democrats in Michigan's legislature have benefitted from the national increase in their party strength. The differences between the two parties in Michigan can be summarized as follows: the Republicans are a rural and suburban, high status party, while the Democrats are an urban, lower status party with great appeal to minority groups (eSpecially Negroes). This, of course, is some- what of an oversimplification, but it outlines the major differences. Studies of roll-call voting behavior in state legislatures that are similar to that found in Michigan sometimes conclude: 3This conclusion is subject to challenge in light of the possi- bility that Michigan legislators may perceive their party as an im- portant reference group taken for granted. This possibility is al- ways a real one, especially in the interviewing situation. Legisla- tors may feel that their party is so obviously important that they fail to mention it in answering questions in the interview. 32 In many of the American states with a strong two-party system, during sessions when party balance in the legislature is not one-sided, partisan considerations influence to a significant degree the legislators' voting on issues about which there is some measure of disagreement. Such conclusions are generally tempered by qualifications concerning the degree of "controversy" over the issues, constituency consider- ations, and whether the governor is of the same party as the majority in the legislature. Evidence resulting from this study of legislators' reference groups strongly warns against assigning such a high degree of in- fluence to political parties in state legislative processes. Party influence is usually assigned by describing the correlations between voting behavior and party affiliation; this ignores the legislators' interpretations of the situation. If, as this study suggests, state legislators do not perceive political parties as important factors in defining their various roles, then a considerable amount of explaining is needed to give meaning to the finding that the party label is a reliable indicator of roll-call voting behavior. No significant differences (Table IX) in the two parties were found to exist in Michigan through the use of reference group theory. As generic classes, the four reference groups were commonly per- ceived by members from each party. The majority party tended to see other legislators as important more often than did the Democrats. 24M. Jewell, "Party Voting in American State Legislatures,” American Political Science Review, Vol. 49 (1955), p. 791. 25Due to the exploratory nature of this study, the underlying assumption is left untested here. That assumption has been borne out as correct in other studies noted in the preceding chapter. The assumption is, of course, that individual's significant others effect his definition of situations and objects, and thereby, his behavior. 33 The Democrats, on the other hand, perceived constituency and party as reference points more often than did Republicans. Several variables have seemingly been found to influence the extent a legislator is willing to vote with his party that may also have an effect on his choice of significant others. MacRae found, for example, that the political contest in the constituency, depend- ing on whether the district is "close" or "safe," has a distinct bearing on legislative voting behavior.26 From MacRae's finding it might be hypothesized that a legislator from a "close"district would be more apt to depend on his party for support, hence perceive party as an important reference point, than a legislator from a "sure" district. On the other hand, a legislator from a "close" district might attempt to disassociate himself from his party in the hope of appealing to luke-warm partisans of independents. TABLE X.--The Effect of Safe and Close Seats on Reference Group Choice Reference Groups Legislators from Close Legislators from Safe Seats (number) Seats (number) Political Party 0 10 Interest Group 2 30 Other Legislators 3 22 Constituents 4 23 Total 9 85 26 D. MacRae, Jr., "The Relation Between Roll Call Votes and Constituencies in the Massachusetts House of Representatives," American Political Science Review, Vol. 46 (1952), pp. 1046-1055. MacRae found that the political contest in a constituency has more effect on roll call voting behavior than socio-economic characteris- tics of district. 34 The data in Table X compare reference group choices of legis- lators from competitive districts with reference group choices of legislators from "safe" districts. A "close seat" was defined on the basis of the past four general elections in Michigan. If a seat changed party hands once in the four elections between 1956 and 1962, it was considered competitive or "close."27 It was felt this criterion would best distinguish "close" from "safe" districts in that a seat changing party hands would appear to be highly visi- ble to the legislator. Other means were available, but none would be so distinguishable from the legislator's point of view. DeSpite the fact that on a statewide level Michigan is a two-party state, there were only nine seats in nine districts that were "close," on the basis of the previous definition. There were 86 districts with a total of 110 seats; 17 districts were multi-member, contributing a total of 41 seats.28 The difficulty in quantitatively analyzing these data is ob- vious in the face of so few competitive seats. Percentages have a poor foundation. Table VIII, nevertheless, indicates a negative re- lationship between the legislators who held their party as a sig- nificant reference point and those legislators from "close"districts. None of the representatives from competitive districts mentioned party. More will be said about the effect of political contests on reference group orientations in the next section. No tables are 27Information on electoral contests was taken from Michigan's Secretary of State, Official Canvass of Votes (1956, 1958, 1960, 1962), compiled by J. Hare (Lansing, Mich.: Speaker-Hines and Thomas, Inc.). 28 In the four past elections considered here, none of these multi-member districts was divided. 35 included in this section concerning the length of service, formal education and occupations of those legislators because of the small number holding this orientation. It can be noted that (1) the mean length of service for legislators with this orientation was 2.8 terms compared to 3.9 terms for other members, and (2) no signifi- cant differences emerged in occupational or educational backgrounds between this group and the rest of the members. It was the newer members who were party oriented. This section has only crudely examined the relation between legislators and their political parties. A much more SOphisticated approach is necessary to explore what a party orientation means to the state legislator. Does it mean, for example, that a party- oriented legislator only votes with his party? If so, this ignores other important activities performed by the legislators and defines "party role" very narrowly. Is he active in party caucuses? Does he hold some or all of his party's principles? Is a party-oriented legislator primarily concerned with opposing the programs of the other party? Is the legislator concerned with party on the local, state, or national level? All these questions are pertinent to the usefulness of party label as a variable in the study of legis- lative behavior. At present, it will have to suffice to say that a simplified application of reference group theory suggests that polit- ical parties are of minor consequence in relation to other clientele groups in the eyes of Michigan Representatives. 36 The Constituency-Oriented Legislator The assumption that a legislator behaves in such and such a manner because he represents a unique geographic area composed of citizens with unique interests is often made by students of repre- sentative bodies.29 The empirical question involved here is not whether the district is "typical" or "atypical" by any standard, but whether the legislator perceives it as such. If a legislator's constituency is a salient point of reference for him, then he might well subjectively seek to discover its "interests" in detail. In Michigan, 29 per cent of the representatives regarded their constituents as significant others--l6 were Democrats and 11 were Republicans. Next to interest groups, constituents were the most frequently chosen generic group of ”others" by the representa- tives. While more Democrats (36 per cent) chose constituents as their most significant reference groups than Republicans (22 per cent), the figures cannot support the general notion that Democrats "are interested in the people," while the Republicans "are concerned with principles." Other variables possibly indicating why legislators were con- stituency oriented are presented in Tables XI (education) and XII (occupations). This group of legislators tended to have an average amount of education; over half of them had taken some classes since graduation from high school. Legislators engaged in business also tended to be overrepresented in the constituency oriented group. Neither the data on educational or occupational background, however, 29For example, see D. MacRae, op. cit. 37 TABLE XI.--Formal Education and Constituency Orientation Education Constituency Non'Constituency Oriented Oriented No. Z No. Z College Degree 10 37 28 42 Beyond High School 14 52 25 37 Less than High School 3 ll 14 21 Total 27 100 67 100 TABLE XII.--Occupations and Constituency Orientations Occupation Constituency Non-Constituency Oriented Oriented No. Z No. Z Agriculture 5 19 15 22 Real Estate, Insurance 3 ll 14 21 Labor Unions 2 7 2 3 Law 1 4 8 12 General Business 16 59 26 39 Unknown 0 O 2 3 Total 27 100 67 100 indicated significant differences between the group of constituency oriented legislators and other members. Competitive political contests in the constituencies have been discussed in relation to party orientations, where it was indicated the legislators who won "close" seats tend to ignore their party. Table X shows that four out of the nine representatives from "close" districts chose their constituencies as significant reference points. 38 Again the analysis suffers from few "close" seats, but the data do suggest that there is a positive relationship between electoral com- petition in the constituency and the legislator's perception of it as a salient reference point. This finding is neither new nor surprising as the potent sanction of removal, which is real and present in competitive dis- tricts, can hardly be ignored by any legislator who likes his job. As one Michigan Democrat from a "close" district candidly put it: The opinion of my constituents is most important to me. My district is traditionally Republican and I got in here on a real squeaker, so I try to keep close tabs on my constituency. The length of legislative service is often discussed by politi- cal scientists in relation to district-identification. The belief has arisen that the more tenure a representative has, the more he "grows" away from his constituency. ’Earlier in this chapter (Table I), data were presented which belie this notion. While 31 per cent of the first and second termers perceived their constituency sali- ently, so did 23 per cent of those serving six or more terms. The mean number of terms served by this group was 4.0 compared to 3.7 terms for other members. This, perhaps, explains why the "old- timers" still present were able to become re-elected time after time. In Congress, legislators are often separated by great physi- cal distances from their constituents, thus accounting for the alienation from their "people."30 In state legislatures, geographic distance is greatly reduced and many legislators return home every - weekend. 30D. Matthews, U. S. Senators and Their World (New York: Alfred A. Knopf, Inc., 1960), pp. 76, 219, 221. 39 Similar to the reference group orientations discussed above, the relationship between the state legislator and his district needs further exploration. Most districts, for example, are pluralistic in nature and composed of a multitude of interests. Certain of these interests are bound to be given priority by the representative at the expense of others, due to his frame of reference. It might appear to some observers, this writer for one, that it is doubtful if a legislator is aware of the number of persons per dwelling unit or the percentage of home ownership in his district. As pointed out earlier, though, this study does not attempt to get at the way each legislator "operationalizes" constituency. Whatever variables are used in the future to pinpoint the legislator-district relation- ship, they must be viewed from the legislators' perspectives if they are to be used fruitfully. Summary Through the use of the personal interview it was found that 86 per cent of Michigan representatives regard four generic groups of others as significant points of reference: other legislators, in- terest groups, political parties, and constituencies. Reference group orientations toward these four classes of relevant others is not surprising to anyone familiar with American legislative processes. It has been demonstrated that representatives who perceive oth- er legislators saliently tended to be legislators with a medium amount of experience. Appreciable changes in the priority legisla- tors assigned to reference groups did not seem to occur over the 40 period of time served in the House which suggests that significant others represent a stable base from which to predict future behavior. Interest groups increased slightly in importance for the representa- tives as seniority increased. Republicans tended to be more concerned with other legislators than Democrats, probably because of the benefits accrued by their majority status. Democrats were more concerned with their constit- uencies than Republicans. Interest groups appeared as the single most important reference points for legislators of both parties. Political parties were not important reference groups for either the Republicans or the Democrats, although party assumed more impor- tance for the latter. Districts giving rise to close political contests found their representatives more interested in them as an important reference point. Instead of turning to their party for aid, the legislators from competitive districts turned directly to their constituents. This indicates support for the hypotheses that legislators from competitive districts are more likely to break party ranks in their voting behavior. To a limited extent, length of service and the competitiveness of electoral contest were the most suggestive variables for explain- ing why legislators might hold certain reference group orientations. Several other variables commonly used in the study of representative bodies (education, party affiliation and occupation) proved inconclu- sive in suggesting any relation between the legislator and his general role orientation. It is, perhaps, helpful here to remind the reader that verbali- 41 zation in the interview situation is behavior of a real sort, and that the legislators' verbal interpretations are highly likely to constitute premises underlying what some call more "observable” behavior. Social scientists would be in a sorry plight if they assumed that all verbal behavior prior to action was false; this would doom behavior to irrational outbursts. Although knowing an individual's significant others allows us a great deal of insight into the general and specific roles he performs, it does not mean that his behavior can be predicted in every situation. Reference group orientations do, however, appear to be stable foundations in suggesting how a legislator will conduct himself and give meaning to his behavior. CHAPTERHIII LEGISLATIVE LEADERSHIP On the surface, all representatives are equal in that their votes are equal. No one save their constituents can in a sense order them to do anything--and even then after a considerable time lag. They cooperate with one another only insofar as they find it exped- ient. Opposed to this surface view is the more realistic informal one. A legislature is a group and, like all groups, operates in ac- cordance with a complex set of norms, or expectations of behavior. A representative body, moreover, is designated a task that can be performed only cooperatively, which necessitates internal organiza- tion, divisions of labor, and leadership. This section was undertaken in order to show the superiority of the reference group approach over that of the background attribute approach in distinguishing or separating legislative leadership from the total membership. Our theory is that explanation and predic- tion are most successful between adjacent classes of attribute, at- titudinal, and behaviorial variables. That is, behavior can best be explained by using attitudinal data, and attitudes can be best explained by attribute data. Since leadership is a function of be- havior, it is best analyzed from the attitudinal point of view, 1T. Parsons and E. Shils, op. cit., p. 211, hold this view when they Speak of "integration" and ”coordination." 42 43 i.e., from reference group orientations as opposed attributes such as education, occupation, party affiliation and length of service. In order to test the hypotheses concerning leadership in the Michigan House of Representatives, it was necessary to define and identify the leaders. The term "leadership" has been interpreted in a great variety of ways.2 For the purposes of this thesis, leaders are those individuals to whom other representatives look for guidance and direction. In order to be consistent with the in- teractionist point of view put forth earlier, a leader can never be identified on the basis of "legitimacy" or formal position. He is a leader only to the extent that others perceive him as such. It should be noted, however, that formal positions of authority (e.g., speaker, floor leader), which are depicted as legitimate, may often effect the perceptions of others. This will be clarified shortly. An interviewing technique was employed in which 106 out of the 110 members were contacted. Each member was asked with whom he most frequently discussed proposed legislation in the House.3 It was felt 2M. Fuller, op. cit., Chap. 1, grapples with the problem of re- lating leadership and power and exposes many of the various interpre- tations. 3Specifically, the question was worded: "Who are the members in the House with whom you most often discuss proposed legislation?" The first two, or the two most salient responses, were tabulated to. determine leadership. The question of whether discussion can be equated with leadership or influence is a serious one. Katz and Lazarsfeld have pointed out that discussion may occur between "equals", with neither person influencing the other. E. Katz and P. Lazarsfeld, Personal Influence (Glencoe, Illinois: The Free Press, 1955), pp. 145-158. However, discussions in the legislature about proposed legislation differ considerably from those involving people discussing movies and fashions. Legislators are formally engaged in 44 that this question would identify the functional leaders in the House. Most representatives claimed they talked to all their colleagues about proposed legislation, thus it was often necessary to probe by emphasizing to whom they spoke pp§£_frequently about legislation. Table I shows the selection and breakdown of leaders by name and party label on the basis of the number of times each representative was chosen as a discussant. The twelve members named most fre- quently were selected and considered as the leaders. A number of individuals made up the leadership in the Michigan House of Representatives. That is, no one person stood out as Eh; leader. There did, however, appear to be a definite difference in leadership patterns between the Republicans and Democrats. The highest ranking Democrat was named as a discussant twelve times, while the other two Democrats qualifying as leaders were selected only five times as discussants. The Republican leaders were much more gradually dispersed, indicating less centralized leadership in the hands of one person. The informal leaders tended to occupy formal positions of authority. The data in Table II indicate that formal and informal discussing legislation for a purpose; the primary purpose of the leg- islative leaders is to generate support for proposed legislation. The writer, furthermore, found in the 30 odd hours spent in inter- viewing legislators, that most discussions on proposed legislation were clearly motivated and initiated by an attempt to influence. 4While the size of the leadership group was arbitrarily set, it was likely that only members widely perceived as leaders were included. Almost all the representatives were listed as a discussant at least once; thus the continuum formed was a very gradual one on which the leaders occupied the top twelve places. 45 TABLE I.--Leadership of the Michigan House of Representatives Leader Party No. of Times Selected As A Discussant Kowalski D 12 Folks R 12 Green R 11 Engstrom R 11 Waldron R 8 Whinery R 6 Wurzel R 6 Cobb R 5 Brigham R 5 Montgomery D 5 Gibbs R 5 Ryan D 5 TABLE II.--The Leaders and Their Formal Positions Leader Party Positiona Kowalski D Minority Floor Leader Folks R Chairman--General Taxation Committee Green R Speaker Engstrom R Chairman--Ways & Means Committee Waldron R Majority Floor Leader Whinery R Chairman--Judiciary Committee Wurzel R Chairman--Education Committee Cobb R Chairman--Agriculture Committee Brigham R Chairman--Towns & Counties Committee Gibbs R Chairman--State Affairs Committee Montgomeryb D . . . . . . . . . . . . . . . . . . Ryan D 00.000000900000000 aSee M. Fuller, op. cit., Chap. 111, for a good discussion of the functions of these various positions. bBoth Democrats had important committee assignments (Montgomery and Ryan). 46 positions of leadership correspond closely.5 It is likely that by definition of their formal positions, officers force other repre- sentatives to interact informally with them if they are to perform integrated roles in the House. This would explain the close re- lationship between formal and informal leadership. After identifying the leaders, it was possible to explore the more important questions concerning the relative characteristics of leaders and any commonly held reference groups. Is education an important criterion used by other representatives in selecting leaders? Does being a businessman or lawyer suit a representative for a leadership role? Must he have a long record of service in the House? To what audiences do leaders play? Do leaders' reference group orientations distinguish them more clearly from the total mem- bership than attribute variables? Such questions involve explanatory or causal research. The difficulties involved are great, as is recognized by all scientists who strive toward the controlled experiment.6 There is no pretense made here of controlling all the variables. The previous questions will not be answered in an unequivocal fashion. First, only several of the recognized variables will be tested. Although these variables are important to this researcher, they may not seem to be so to another. Primarily, for reasons of expediency and parsimony, con- 5Fuller also found in her study of legislative leadership that formal and informal leadership corresponded closely. M. Fuller, op. cit., Chap. III. 6See W. Goode and P. Hatt, Methods in Social Research (New York: McGraw Hill Co., 1952), Chap. VII, for an account of the difficulties involved in using John Stuart Mill's classical design of logical proof. 47 stituency make up and voting behavior will be ignored, although they obviously play a vital part in leadership formation and function. Emphasis will be placed upon reference groups; although education, occupation, seniority and age will be explored. Reference Relationships Reference relationships or reference groups which were dis— cussed in the preceding chapter are again useful in discussing leg- islative leadership. The same method was used to determine signifi- cant reference groups as was described previously. It was hypothe— sized at the outset of the study that the leaders would have a high degree of agreement on their choice of significant others, and that interest groups would constitute the basis for this agreement. By virtue of their general roles the leaders are expected to be con- cerned with individuals and groups outside their immediate system. Table III shows that 90 per cent of the leaders who answered the question regarded interest groups as their most significant ref- erence group. Only one leader chose an audience other than interest groups. Although interest groups were the most popular choice as a reference group by the total membership, among the leaders, they assumed an even greater significance. These data support the hypotheses that a leader is expected to know what is going on out- side the House, and that he is supposed to be familiar with the inputs generating from interest groups. Much can be perceived con- cerning the impact interest groups have on the legislative process from these findings. Group theorists have long assigned primary importance to interest groups, but research has not always supported 48 TABLE III.--Reference Groups of Representatives and Leaders Total a Index of Over- Reference Group Membership Leaders representation No. Z No. X Other Legislators 25 27 l 10 .37 Interest Groups 32 34 9 90 2.65 Political Parties 10 10 O O .00 Constituency 27 29 O O .00 Total 94 100 10 100 .... aThe significant others of 16 legislators were not ascertained, two of whom were leaders. their claims. Garceau and Silverman, for example, found Vermont legislators to be surprisingly unable to identify pressure groups and pressure group leaders.7 The data here suggest that the total membership can afford to be ignorant of group pressures because the leadership takes on the primary task of converting group demands into legislation. Due to the negative value many legislators assign in- terest groups, it is no wonder the leaders do not articulate to the membership how various groups stand on particular issues. Hence, much of the membership remains happily ignorant of the role they play in the group struggle. Education There are generally two Opposing views concerning the relation- ship between formal education and political leadership. The kind O. Garceau and C. Silverman, "A Pressure Group and the Pres- sured," American Political Science Review, Vol. 48 (1954), 672-691. 49 of political leadership inferred by these two arguments may not correspond exactly to leadership in the Michigan House of Repre- sentatives, but it is close enough to merit mention. One argument holds that the greater the politician's education, the more likely it is that he will be perceived as an egghead, dreamer, or idealist; hence he will be less apt to be regarded as a leader.8 It is said, on the other hand, that a political leader needs the broader know- ledge provided by a high level of education if he is to grasp the multitude of complex forces converging on him and utilize them for advantage.9 The data in Table IV indicate that neither argument provides the correct rationale, at least in the case of the Michigan House of Representatives.10 Three-quarters of the leaders had taken classes beyond high school; a quarter of the leaders had college degrees. All of the leaders, in other words, had taken some classes beyond their high school degree. Despite the fact that all leaders had a fairly high degree of formal education, they did not 8D. McKean, Party and Pressure Politics (New York: Houghton Mifflin Co., 1949), p. 254. McKean holds this view, although he writes that the prejudice toward higher education decreases as the politician moves from local levels to national and international ones. 9R. Michels held this position in Political Parties, trans. E. Paul and C. Paul (Glencoe, Ill.: Free Press, 1949), p. 72. 10 These findings diverge from M. Fuller, op. cit., p. 18. This and other differences between the conclusions reached here and those reached by Fuller may be explained by the different time periods studied, and differences in the selection of the leaders. The leadership group here is much larger than that in Fuller's study of leadership. 50 TABLE IV.--Formal Education of Representatives and Leaders Level of Education All Members Leaders Index of Over- representation No. Z No. Z High School or less l8 l7 0 0 .00 Beyond High School 44 42 9 75 1.8 College Degree or Beyond 43 41 3 25 .61 Unknown 5 o o o o o o o o 8This index was used by N. Rogoff and D. Matthews. See N. Rogoff, Recent Trends in Occupational Mobility (Glencoe, 111.: Free Press, 1953): and D. Matthews, op. cit., Appendix A. If the representatives were selected on a chance basis, we would expect those social characteristics to be the same as those possessed by the universe from which they were drawn. The ratio: Z of leaders possessing tribute A Z of total membership possessing tribute A represents the index of overrepresentation. An index smaller than 1.0 means the attribute is underrepresented; an index of 2.0 in- dicates twice the eXpected pr0portion based on chance. This index refers then to both the overrepresentation and underrepresentation. Hence the leaders are overrepresentative of the entire membership on the attribute "Beyond High School," but substantially under- representative on the attribute "High School or less." stand out from the total membership on this basis. Leaders cannot be clearly distinguished from other legislatures on the basis of educational background; they appear to have "just enough, but not too much" education. Occupation Although not highly educated, by virtue of their leadership 51 TABLE V.—-Occupations of Representatives Profession Representatives Leaders Index of Over— ' representation No. Z No. Z Real Estate, Insurance 20 18 1 8 .44 Labor Unionsa 5 5 1 8 .44 Agriculture 21 20 5 42 2.2 Law b 9 8 2 17 2.1 General Business 53 49 3 25 .51 Unknown 2 .. . .. .... aOne member was a lawyer who worked for a labor union. He was classified under labor unions. bGeneral Business included management, wage earners, and public servants. abilities the group of leaders might be expected to have held jobs at the top of Michigan's occupational hierarchy. Generally speak- ing, all legislators are stereotyped as successful lawyers and businessmen. The leaders could be expected to conform to this pic— ture even more readily than the average representative if success outside the legislature is equivalent to that inside the legislature. Such was not the case here. Table V shows more than twice as many leaders were engaged in agriculture than should have been expected. Lawyers, too, were overrepresented, while those engaged in business, real estate and insurance were underrepresented in the leadership group. These findings are not entirely surprising when they are put in the context of Michigan's general political scene. The Republi- cans, who occupy the lion's share of the leadership positions, draw 52 heavy support from out-state rural areas. The Democrats receive traditional support from urban areas where organized labor is strongest. It is not surprising to note that the highest ranking Democratic leader was a spokesman for the U.A.W.-C.I.0., while the top two Republican leaders were farmers. It is not particularly startling to find that the leaders reflect the general interests of their constituencies and parties. The second most striking feature in the occupational histories of the leaders was the overrepresentation of lawyers. An attorney is trained in verbal manipulation and performs the everyday role of practicing this art. Lasswell and McDougal entertain the thesis that lawyers are in the strategic position of legal informer to "every responsible policy maker of our society," and by being in this position they "create policy."11 By virtue of their skills, personal contacts, and their profession (a lawyer can afford to take the pig; to politic), they are well suited for political roles in a leadership capacity. The representatives in Michigan tend to support this hypothesis. From this data alone, one might conclude that a Michigan rep- resentative has twice the opportunity of becoming a leader if he is a farmer or lawyer. Caution should be used in accepting such a con- clusion on the basis of so few leaders and without consideration of other factors. 11H. Lasswell and M. McDougal, "Legal Education and Public Policy," in H. Lasswell, Analysis of Political Behavior (London: Routledge and Kegan Paul, 1948), p. 27. 53 Length of Service Leadership roles in legislatures are not easy to perform.12 A reasonable length of time in the House would seem imperative for the leaders to acquire an adequate understanding. Both formal rules and informal folkways must be learned which demand that a new mem— ber behave in certain ways that are difficult to circumvent. The new member who bucks these norms finds himself unable to get his legislation through the House.13 This rationale would lead us to conclude that leaders must serve for relatively long periods in the House before they can occupy leadership status. About half the total membership had served three terms or less, while no one of the twelve leaders had served less than four terms. Although the figures in Table VI show that leaders have served a relatively long period of time in the Michigan House of Representa- tives, this does not mean that a premium is placed on seniority. The seven most senior members were not leaders. Seniority appears to be important in leadership formation; alone, though, it is not sufficient to raise a representative to leadership status. About one half (16) of the legislators who served five terms or more (37) perceived interest groups saliently, and 11 of these 16 were desig- nated as leaders by other members. Thus, the role of a leader in the Michigan House called for a combination of seniority and interest group orientations. 12See Matthews, op. cit., Chap. VII, for an excellent analysis of the complexity of party leadership in the U. S. Senate. 13Many of the representatives interviewed agreed that this was the best way to enforce the unwritten "rules of the game." 54 TABLE VI.—-Length of Service in the House I7 - . Total Membership Leaders Number of Terms Served No. Z No. Z One 23 21 .. .. TWO 14 13 o o o o I: Three 15 14 .. .. I Four 13 12 3 25 Five 17 15 2 17 Six 6 5 .. .. Seven 4 4 2 17 ‘ Eight 9 8 3 25 ' Nine or More 9 8 2 17 Mean number of terms for leaders 6.5 Mean number of terms for total membership 3.4 Summary Although there exists in every legislature a set of informally recognized norms as well as formal ones, there must also be a set of judges or leaders who guide the group, who settle intra-legisla- tive squabbles, maintain order, act as arbiters, expedite business, and promote stability and predictability.14 In performing these functions legislative leaders play a crucial part in transforming inputs into outputs. Through personal interviews with Michigan representatives, Bentley's idea that "leadership is not an affair of the individual leader . . . it is fundamentally an affair of the group . . . pomp and circumstance are but details,"15 was found to be correct. No 14J. Wahlke, H. Eulau, L. Ferguson, and W. Buchanan, op. cit., Chap. VIII. 15A. Bentley, op. cit., p. 223. 55 individual stood out as pp; leader in the Michigan House. The Republicans who outnumbered the Democrats by only 58 to 52, placed four times as many men in informal leadership positions as did the Democrats. This was caused by numerous Democrats crossing party lines to "discuss proposed legislation," presumably because Republi- cans heldthe formal positions of power by definition of their majority status. Leadership for the Democrats appeared to be some— what more centralized than Republican leadership; the minority floor leader ranked much higher as a discussant than any other Democrat. No general picture of a leader emerged in terms of their ages, occupations, or educations. Leaders did have more seniority than the average member. The one variable which stood out, and one which undoubtedly has a more functional impact on the legislative process, was the significant others or reference groups of the leaders. They were unified in their high regard for interest groups. CHAPTER IV CONCLUSION Nine years ago an article which appeared in one of the major journals of social science critically discussed the shortcomings of variable analysis. Variable analysis is that which seeks to re- duce human activity (i.e., legislative activity) to variables and their relations. Contemporary legislative research has for the most part disregarded the warnings sounded in this article. Students of legislative bodies have increasingly come to rely on this scheme of analysis. The fact is that sephisticated forms of variable analysis have become the norm for proper research procedures. The pitfalls present in current variable analysis are perti— nent to this study of the Michigan House of Representatives. In— deed, they are crucial to it. Legislators, and all human beings for that matter, act on the basis of various definitions, or interpre— tations, they have of the world around them. That world consists of objects - job, education, elections, people, institutions, tele- vision, political party, and so on - each of which has a meaning assigned to it by individuals, on the basis of which they act to- ward it. As Blumer accurately puts it: In our activities we went our way by recognizing an object to be such and such, by defining situations with which we are 1H. Blumer, "Social Analysis and the Variable," American Socio- logical Review, Vol. 21 (1956), 683-690. This article was of great value in the preparation and thought that went into this paper. 56 57 presented, by attaching a meaning to this or that event, and when need be, by devising a new meaning to cover something new or different. This is done by a group of individuals acting in concert . . . We can and, I think, we must look upon human group life as chiefly a vast interpretative process in which people, singly and collectively, guide themselves by defining objects, events and situations which they encounter. This crucial process of interpretation and definition which was . . 3 . . . diagrammed earlier cannot be ignored by students of legislative be— havior. The presumed variables (e.g., political party, interest groups, constituency, other legislators) that designate situations, directly or indirectly confronting legislators, operate through this process of interpretation. In other words, Conventional vari- ables in the legislative processes cannot be treated as independent variables upon which some form of legislative behavior (e.g., voting behavior) is a dependent variable.~ Research must concentrate on what lies between the so-called "independent" and "dependent" vari— ables. That process of interpretation lies between the presupposed independent variable and the behavior representing the dependent variable; there is no independent variable Operating in the legisla- tive arena which automatically exercises a set amount of influences on a dependent variable. The real problem in using variable analysis in the study of legislative behavior seems to lie in the tacit assumption that the independent variable predetermines its interpretation. Now it is almost universally acknowledged that no object, situation, or event Zlbid., p. 686. 3The premise here as discussed in Chapter I has been that while studies ignoring the process of interpretation may be able to predict certain aspects of legislative behavior with some degree of certainty, they do little to aid us in understanding how and why this prediction is successful or unsuccessful. 58 carries its own meaning; they have their meanings conferred upon them through the process of definition and interpretation.4 The assumption that independent variables determine their own meaning is legitimate only if a stabilized interpretation occurs in which individuals already have constructed meanings which are immediately applied to the variable. If this stable interpretative process is present, then variable analysis could, practically speaking, disre- gard it. Do legislators, then, have a common interpretation and def- inition of the forces impinging on the legislature, both from within and without? 0n the basis of this study of the Michigan House of Representa- tives, the answer must be an emphatic "no". By rather crudely op- erationalizing the reference group theory, representatives were found to choose their significant others from among four different groups of relevant others. While this finding does not allow us to demonstrate anything conclusive about the interpretative process of legislators, it does indicate that there are several different points of reference being used in defining various situations. Each individual legislator interviewed recognized the impor- tance of one group or a portion of one group. This group was de- fined as his significant others with the logical conclusion being, of course, that he defined various stimuli, or legislative inputs, in terms of the way that group would define them. For example, if 4H. Blumer, op. cit., contends that this interpretative process is a formative and creative one in its own right, and that it cannot be measured in terms of an intervening variable. That this conception is anathema to behavioral social scientists, who base their research on regularities and patterns, is obvious. It is much too early yet to substantiate Blumer's contention. 59 a legislator perceived legislative others most significantly, and he was asked by a certain interest group to sponsor and support a bill that he knew was unpopular among the leaders in the legislature, it is likely that he would refuse. Very likely this same legisla- tor would do nothing in violation of legislative norms. A legislator who regarded his political party as his most salient reference group would probably support it in the committee, in caucuses, on and off the floor, and on roll-call votes. Con- stituency considerations that were inimical to his party's interests would likely receive short shrift. The reader can follow this line of reasoning for each of the four different reference group orien- tations. Each orientation could serve as a fruitful foundation for further research in order to test the hypothesized consequences of a particular perspective. This attack allows the researcher to investigate the entire phenomenon of legislative behavior, rather than one particular aspect such as voting behavior. It allows for the omniscient and omnipo- tent process of interpretation to be taken into account, without which, analyses are doomed to simple descriptions. Most important, the researcher is allowed to insert something measurable between an independent variable such as a close election and a dependent variable, as for example, roll-call voting behavior. No claim is made here that the entire set of interpretation has been accounted for by merely identifying legislators' most salient reference groups. A beginning has been made, but a more subtle and sophisticated approach is needed. One stand-out problem evident here is that this study is merely a "snapshot" which ignores the 60 dynamic nature of the interactionists' theory of reference groups. Only one static aspect of the role—taking process has been examined. A related problem deals with the concept of "significant oth- ers". Each Michigan Representative indicated a predominant refer- ence group at the time of the interview. The concept of significant others wrongly implies an immutable relationship with the legisla- tor. In actuality, an individual's significant others may change as the situations change. If the composition of his significant others changes, his interpretation of stimuli likely changes, which ultimately may lead to what the original investigator might call irrational or inconsistent behavior. The only way out of this dif- ficulty is to postulate that while an individual's reference groups may vary considerably from situation to situation, his significant others remain fairly stable. The legislator, furthermore, may react to a mixture of reference groups rather than one particular group. Another confusion often results concerning role-taking when one assumes that by taking the role of another, the role—taker adopts the standpoint of that other. This difficulty has a direct bearing on reference group behavior. Such an assumption leads the observer to believe that this role—taking relationship will result in behavior that is altruistic or sympathetic to that group.5 5R. Turner exposes this fallacious view, "Role Taking and Ref- erence Group Behavior,” American Journal of Sociology, Vol. 61 (1955-56), pp. 316-328. This is an excellent analysis of the sub- tleties and nuances present in operationalizing reference group theory. In it, Turner also throws out a thought provoking scheme in which relevant others can be classified into different categories depending on the role-taking standpoint. 61 Nothing can be farther from the truth. Successful "confidence men", for example, owe their success to their ability to assess ac- curately the feelings and attitudes of others, while completely avoiding any identification with those attitudes. There appears to be a great deal of similarity between the role-taking of a confi- dence man to that of a practicing politician. Like the "con-man", the legislator must possess an extraordinarily sensitive empathetic ability; he needs to know what his party expects, and what other legislators think. He takes the roles of these various groups by putting himself imaginatively in their places; yet, due to their conflicting expectations, he is forced to avoid personal identifica- tion (empathy versus sympathy). By virtue of the general expecta- tions concerning the legislator's role, he must role-take vocifer- ously with considerably accuracy in many directions, but he is also forced to maintain a separation of personal identities if he is to be a successful compromiser. The standpoint the role-taker assumes in relation to the refer- ence group is apt to vary considerably. A legislator, for example, may take the role of another in order to adopt his values and per- spectives.6 In this case the legislator may be said to have a "psy- chologically functioning membership"7 in that group. In other situations a legislator might use a reference group as a point of 6For discussion of reference groups as a source of perspectives, see T. Shibutani, "Reference Groups and Perspectives," American Journal of Sociology, Vol. 60 (1955), pp. 562-569. 7This phrase is Newcomb's. See "Social Psychological Theory" in Social Psychology at the Crossroads, ed. J. Roher, M. Sherif, op. cit., p. 48. 62 comparison in making self-judgments. Still a third possible stand- point in role-taking may occur when the role-taker sees his role performance being evaluated by relevant others. He attends to the expectations members of this group hold toward him, to the extent of the merit he assigns to their evaluations. No personal identifi— cation is necessarily present. The nuances present in reference group behavior, when coupled with its sister concept of "role-taking", can be seen to be many and complex. In a general way, the findings of this study have been dis-cussed with some of the difficulties that prevent a more con- clusive analysis. What, specifically, have we learned of the Michi- gan House of Representatives? Concisely, the findings of this study are: 1. Four generic classes of "others" were recognized as domi- nant reference points by representatives in the Michigan Legislative process. The groups perceived in their order of priority were interest groups, constituency, other leg- islators, and political parties. No Representative re- garded administrative agencies or the governor as a salient reference point. 2. The particular reference group orientations hald by indi- vidual representatives were not significantly related to party affiliation, education or occupation. Reference group orientations appeared to stem from the individual's basic perspectives prior to his experience in the House. 63 3. One political variable did shape legislators' orientations; representatives from competitive districts perceived their constituency as salient more frequently than did represen— tatives from "safe" districts. 4. In relation to the three other reference groups, political party was not recognized as an important reference point. Minority party members were, however, more apt to regard their party as important than were majority party members. 5. Leadership was in the hands of a number of representatives. Majority party members occupied most of the informal leader— ship positions, probably due to the formal power accrued by their majority status. 6. Some of the variables related positively to leadership were a relatively high degree of seniority, law and agri- cultural occupations, and formal leadership positions. None of these variables were significantly related to leadership in comparison to reference group orientations. 7. Most relevant to the functioning of the House was the find- ing that all save one of the leaders held interest groups as their most salient point of reference. While it would be intellectual suicide to project sweeping generalities from this localized study, several things do emerge that are of general significance. By using generic classes of reference groups, the study obviously gains a wider applicability than if specific groups were used. Each of the four reference group orien- tations are likely to be present in other state legislatures. One factor of general importance brought forth in this study is the 64 seemingly small role political parties play in the decision-making processes of state legislators. Many voting studies have pre- viously depicted party as the single most reliable indicator of legislative voting. Secondly, while administrative agencies may be quite important to members of Congress, their importance is negligible in the eyes of state legislators. Nor was the chief executive chosen as a reference point by any Michigan representatives. Both findings indicate a sharp difference between national and state politics. Interest group theory and research has been primarily concerned with the national government. This study suggests that interest group activity is also of great importance at the state level. In- terest groups were the most frequently chosen reference point by Michigan Representatives and what is more important, interest groups were chosen by 90 per cent of the leaders as their most significant reference group. Further studies such as this, using a more refined reference group theory, could add significantly to our knowledge of legisla- tive behavior. By utilizing the four main reference group orien- tations discussed above as a foundation, further research could develop a technique to examine more closely the kinds of relation— ships legislators have with each group. The task is no easy one, in that arduous hours of interviewing are required. Certainly it is time to stop analyzing aggregate data in hopes of finding variables that "predict" legislative behavior. To understand legislative behavior, we must go back to the individual 65 and learn how he constructs his political world. Those who con— tinue to study the legislative processes with simple theories and complex techniques remind one of the man discovered carefully searching the ground with a magnifying glass under a street light at night. When asked what he was doing, he replied that he had lost something of great value back down the street where there was no street light, and that it was much easier to look for it where there was a light. BIBLIOGRAPHY Books Bailey, S. and Samuel, H., Congress at Work. New York: Henry Holt and Company, 1952. Bentley, A., The Process of Government. Bloomington, Indiana: Principia Press, 1949. Ehrmann, H., Interest Groups on Four Continents. Pittsburgh, Pennsylvania: University of Pittsburgh Press, 1958. Fuller, M., Leadership in the Michigan Legislature. East Lansing, Michigan: Masters Thesis, Department of Political Science, Michigan State University, 1957. Goode, W. and Hatt, P., Methods in Social Research. New York: McGraw Hill Company, 1952. Cross, 8., The Legislative Struggle. New York: McGraw Hill Company, 1953. Homans, G., The Human Grogp. New York: Harcourt Brace, 1950. Matthews, D., U. S. Senators and Their World. New York: Random House, 1960. McKean, D., Party and Pressure Politics. New York: Houghton Mifflin Company, 1949. Mead, G., Mind, Self, and Society. Chicago: University of Chicago Press, 1934. Michels, R., Political Parties. Translated by E. Paul and C. Paul. Glencoe, Illinois: Free Press, 1949. Newcomb, T. and others (eds.), Readings in Social Psychology. New York: Henry Holt and Company, 1947. Parsons, T. and Shils, E., Toward a General Theory of Action. Cambridge: Cambridge Press, 1959. Ranney, A. J. (ed.), Essays on the Behavioral Study of Politics. Urbana: University of Illinois, 1962. 66 67 Rose, A. (ed.), Human Nature and the Social Process. Boston: Houghton Mifflin Company, 1962. Sherif, M., An Outline in Social Psychology. New York: Harper Bros., 1948. Truman, D., The Governmental Process. New York: Alfred A. Knopf, 1960. Turner, J., Party and Constituency: Pressures on Congress. Balti- more: Johns Hopkins Press, 1951. Wahlke, J., Eulau, H., Ferguson, L. and Buchanan, W., The Legislative System. New York: Wiley and Company, 1962. Zeller, B., Pressure Politics in New York. New York: Prentice Hall Company, 1937. Articles and Periodicals Aydelotte, W., "A Statistical Analysis of the Parliament of 1841: Some Problems in Methodology," Bulletin of the Institute of Historical Research, Vol. 27 (1954), 140-149. Blumer, H., "Social Analysis and the Variable," American Sociology Review, Vol. 21 (1956), 683-690. Couch, 0., "Self Attitudes and Degree of Agreement with Immediate Others," American Journal of Sociology, Vol. 63 (1958), 314-323. Derge, D., "The Lawyer as a Decision Maker," Journal of Politics, Vol. 21 (1958), 408-433. Eulau, H., "Identification with Class and Political Role Behavior," Public Opinion Quarterly, Vol. 20 (1956), 515-529. Garceau, O. and Silverman, C., "A Pressure Group and the Pressured: A Case Report," American Political Science Review, Vol. 48 (1954), 672-691. Hyman, H., "The Psychology of Status," Archives of Psychology, Vol. 269 (1942), 121-174. Jewell, M., "Party Voting in American State Legislatures," American Political Science Review, Vol. 49 (1955), 773-791. Keefe, W., "Comparative Study of the Role of Political Parties in State Legislatures," Western Political Quarterly, Vol. 9 (1956), 726-742. 68 Kinch, J., "A Formalized Theory of the Self-Concept," American Journal of Sociology, Vol. 68 (1963), 411-420. LaPalombara, J., "The Utility and Limitations of Interest Group Theory in Non-American Field Situations," Journal of Politics, Vol. 23 (1960), 101-112. Lowell, A., "The Influence of Party upon Legislatures in England and America," Annual Report of the American Historical Association, Vol. 48 (1954), 672-691. MacRae, D., "The Relation Between Roll Call Votes and Constituencies in the Massachusetts House of Representatives," American Political Science Review, Vol. 46 (1952), 1046-1055. Meller, N., "Legislative Behavior Research," Western Political Quarterly, Vol. 13 (1960), 134-146. Shibutani, T., "Reference Groups and Perspectives," American Journal of Sociology, Vol. 60 (1955), 562-569. Shils, E., "The Legislator and His Environment," University of Chicago Law Review, Vol. 18 (1950—51), 571-584. Truman, D., "State Delegations and the Structure of Party Voting in the united States House of Representatives," American Political Science Review, Vol. 50 (1956), 1023-1045. Turner, R., "Role Taking and Reference Groups," American Journal of Sociology, Vol. 61 (1955-56), 622-636. Public Documents Chase, F. and Philleo, N., Michigan Legislative Handbook, 1962-3. East Lansing: Speaker-Hines and Thomas, 1962. Michigan, Secretary of State (J. Hare), Official Canvass of Votes, 1956, 1958, 1960, 1962. HICHIGQN STATE UNIV. LIBRQRIES I! I mm: umnuu ugummu 312 3105047 60