g m Satan %. .a . {“9" ~3:: an (.93 .: “F“ my“. at mp 033910.”: I“ "are“. .~.-. AN ;&fw.§.Lme w: z; ’=- i z“, 51‘";; ,w 2: s: 5. “NM“ EC“? \‘313 R‘AZ. {2 “~ m .3 H.135 }:C‘fi 1‘ .11. 33:12:? W" hi..- ‘ifiifi- CHECA N SCUM"? UNKVCRSE‘FY u .féliq’ti‘lfg J“~f‘:‘. Fif‘ WHEEL!” £1131? 13 “W“ W \\\\\\\\ \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\i 3 1293 10402 4124 LIBRARY ‘ Michigan State University -——— .ANULNALYSIS OF SELECTED ROLE CONFLICT VARIABLES By Jane- lk Rilehert A,THESIS Submitted to Michigan State University in partiel fulfillment of the requirements for the degree of MASTER OF ARTS Department of Sociology end AnthropoIOgy 1961 ACKNOWLEDGMENTS I wish to thank Dr. John C. Howell for general assistance throughout the course of this study. The statistical advice of Mr. LN. Bhargava is sincerely appreciated. The advice, inspiration, comments and criticisms of Dr. Howard Ehrlich of The Ohio State University have been.particularly helpful. Finally, I would like to thank'mv wife, Carol, for her constant patience and understanding. ii I. II. III. V. TABLE OF CONTENTS INTRODUCTION 0 O 0 e e e e e e e e e e e O 0 e e e e Analysis of Diverse Operationalizations of R01. Conflict Veriablcs 0 e e e e e e e e e e TEE-OBI“.OOOOOOOOOOOOOOOOOO. Research Setting and Procedures . . . . . . . . The Selected Role Conflict Variables . . . . . RELATIONSHIPS OF CIHE ROLE CONFLICT VARIABLES AND REPORI'ED PROBABLE BEHAVIOR . O O O O O O I O O O O O Interrelationships Among the Role Conflict Viriables e e e e e e e e e e e e e e e e e e Comparison with Previous Role Conflict Investigations 0 e e e e e e e e e e e e e e INTRODUCTION OF EGO'S PREFERENCE AS A ROIE CONFLICT VAR-”BIB . O O O O O O O O O O O O O O O O O O O O O Simultaneous Combination of Social and PersonalDimensions............. WYCOOOOOOOOOOOOOOOOOOOOO Conclusions and Suggestions for Future Research SEIECTED BIBIIIOGRAPHY O O O O O O O O O O O O O O O O O O O O APENDIXOMOOOOOOOOOOOOOOOOOOOOO... ”PEEK m 0 O O O O O O O O O O O O O O O O O O O O O O 0 iii PAGE 12 13 15 19 38 145 53 55 58 65 INTROWCTION During the past decade there has been a pronounced revival of in- terest in the concept of role. A.result of this heightened interest has been a plethora of conceptual schemata for the more generic concept of role and for concepts derivative therefrom. Unfortunately these necessary attempts at theory construction have been undertaken with little or n9 subsequent empirical verification. the theories have been untestable or’have remained untested. Pert of the explanation.fcr this paucity of empirical work lies in the imprecise manner in ihich the concept of role has been defined; part of it lies in the assuptions which role theorists have made. liemsn and Hughes state that mnny times the concept of role is used bereft of any attempt to define it. Often the concept has been treated as a given. Ihen the concept has been defined, the definition has often remained peculiar to the particular definer. Because the con- cept of role has been knit into the conceptualizations of investigators identified with diverse disciplines, the concept mhen.defined at all, has been definitionally rm- to the particular needs of the investigator.)- Recently, hoeever, students have attempted to lend precise defi- nitions to the concept of role and its derivatives. These definitions appear not to be bound to one particular field of inquiry but are useful to students of various disciplines. Such definitional precision is a first step in the direction of adequate theory construction and its eventual empirical testing.2 2 there is one particular assmsptien which has blocked fruitful inquiry of the concept of role. It has been seemed that consensus exists among embers of society or among groups of role definers as to what is expected of individuals who occupy particular positions in social systems. It was held that both the role definers and the indi- viduals whose positions are being defined are in essential agreement concerning the role behavior apprOpriate to a given situation. l'hat such consensus is not the case has been aptly pointed out by Gross and his associatee.5 lhcn consensus of expectations is assumed the focus of inquiry is directed away from sociological role conflict. Behaviors not in accord with the consensual expectations assumed by the scientific observer may well be treated as deviant. Investigation of conflict is confined to the difference between personality predispositions and role expectations. Parallel with the increased interest in the more generic con- cept ef role, students have become concerned with the theory and em- pirical testing of kindred concepts. Among those receiving special attention have been role conflict and role conflict resolution. Sue early students of role conflict carried out their theoreti- cal and empirical analyses within the context of a very broad defi- nition of the area of role conflict. 1 wide range of behaviors and types of conflict was subsumed under the rubric of role conflict. a situation in which 9.“! individual was confronted with a dillema of choice was regarded as role conflict .h it a later date students began to delimit the area of role conflict 3 by focussing on the conflict which is engendered by an individual's occupancy of bio or more positions. he positions are such that the fullfillmsnt of the expectations of one position necessarily results in the non-fullfillment of expectations associated with the other position. Such conflicts have been traditionally labeled as role-role conflicts or interrole conflicts. Until recently the main concern of role conflict students has been interrole conflict. hang notable recent exceptions to this emphasis on interrole conflict are Gross and his associates who focussed upon the conflicts engendered by an individual's occupancy of a single position.5 In this intrarcle type of conflict the individual is exposed to con- flicting expectations held by various individuals and groups who define his role. For example it may be the case that teachers and the school board hold expectations for the school superintendent that are mutually incompatible with reference to a particular situation, e.g., teacher salary increases. Ehrlich isolated potential intrarole conflict situations of state troOpere. The analysis carried out herein is based upon the potential conflict situations introduced by lhrlich.6 Although the foregoing overview has been cursory, it has only been intended to point to (l) the limited attempts to put role theory to the empirical test, (2) the increasing definitional precision which is necessary for both adequate theory construction and its subsequent empirical testing, (3) the existence of the assumption of consensus, which has blocked fruitful advances in the area of role conflict, and (h) the recent interest in role conflict, especially the conflict which results from an individual's occupancy of a single position. 1; Analysis of‘Diverso Operationalisations of Role Conflict Variables With the increasing precision of boy concepts in role theory there has been an attendant increase in the delineation and specification of variables deemed crucial for the prediction of role conflict resolution. Of particular concern are two concepts which have been utilised either explicitly or implicitly in various role conflict schemata, namely, legitimacy and sanctions.7 Parsons restricts role conflict to ”conflicting sets of legiti- dsed role expectations." For Parsons legitimised role expectations are expectations which hive been institutionalised.8 Getsels and Cuba refer to legitimacy as "mutual acceptanoo' by ego and alter of expec- tations in a given situdtion.'9 Prom this definitional bdckground Gross and his cohorts treat legitimacy as follows: “A logitinate expectation is one which the in- cumbent of a focal position fools others have a right to hold. In illegitimate expectation is one which he does not feel others have a right to hold."10 It 1. expected that an man—mun will be pre- disposed to conform to an expectation which others have a right to hold and predisposed not to conform to an expectation which others do not have a right to hold. In other words the individual rejects his responsibility to conform to an illegitimate expectation. It is further contended that action not in accordance with a legitimate expectation results in negative internal sanctions. legitimacy was Operationalisod in the following manner: "Do you think the indi- vidual or group named has a right to expect you to do this?"11 5 In a study ained at the validation of Gross's theory of role conflict resolution, liller and Shull worked with samples from four populations of business and labor leaders. For purposes of their analysis legitimacy was cperationalized as. 'Lssuning that you are the training director, do you think it was right and reasonable for the president to expect you to present a successful progran under the conditions described above‘l'la In an empirical investigation of role conflict lhrlich opera- tionalised legitimacy following the lead of Iarion Levy. levy states that responsibility is the accountability of an individual to another individual or group for his own acts and/or the acts of others.” lhrlioh's sample was asked the following question: “Are you account- able to these persons for what you actually do‘l'm It 1. entirely probable that the three Operationalisations of legitimacy would evohe three distinct sets of responses. i'o clarify this point a concrete situational example will be related. The following situation was reported to the writer by Ehrlich, who passed three months as a participant observer at various State Police posts in Michigan. at a certain post troopers were obligated to wash their patrol cars at the end of each patrol. this washing was nandatory and took place whether the car needed it or not. This practice was not departmental policy but was initiated by the Post Commander. If the troopers at this post were responding to this potential conflict situation in terms of the three operationalisations of legitimacy how would they react? It is first specified that we are concerned with the Post Comnander's expectation that tr00pers wash 6 their patrol cars at the end of each day. Then we ask following Gross and his associates, 'Do you think the Post Comander has the right to expect you to do this!" The designation ”has the right‘ refers to the authority vested in th. position of Post Commander. by virtue of this authority the Post Commander has the right to expect the tr00per to wash his patrol car. However, if "has the right to hold this expectation“ is included 'I. are not only referring to the authority of the position of Post Commander but also to the specific content of the expectation which he holds. It is neither the expectation as such iiich is legitinate, nor is it the position of Post Commander. Iegitinacy assessments evoked by this question are a dual function of both the position of the holder of the expectation and the content of the expectation. For the cane situation let us as]: the troopers their views on legitimacy as Operationalised by Miller and Shull. 'Do you think it was right and reasonable for the Post Commander to expect you to do this?" The respondents' attention is now directed to the specific content of the expectation. l'he emphasis of this cperationalization is on the 1051c of the expectation. The question does not ask if the Post Connander ”has the right” to hold this expectation but if it i; “right and reasonable“ for the Post Commander to hold this expectation. Cndoubtably the distribution of responses would differ if the two questions were put to this single sample. Finally, if we were to ask the trooPers in Ehrlich's words, "Are you accountable to the ‘ Post Commander for washing your car”, the. responses would undoubtably result from the troopers focussing on the 7 holder of the expectation, not the content of the expectation. The three operationalisations of legitimacy may he visualised as lying on an expectation-audience continuum. At one end of the con- tinuun we find Iiller and Shull's definition which emphasises the con- tent of the expectation. Toward the nidpoint of the continuu lies the construction of Gross and his associates, which takes into con- sideration both the content of the expectation and the authority of the audience or individual holding the expectation. it the other end of the continuum we find Ehrlich's defirdtion, which focusses prinarily upon the authority of the audience holding the expecta'ld on. The concept of sanctions, an old friend in sociology, has not been neglected in the area of role conflict. Among others Gross and his associates systematically integrate the concept of sanctions into their theory of role conflict resolution. Gross and his associates treat sanctions as "a role behavior the prinary significance of which is gratificational-deprivational.” It is assuned that an individual will act so as to ninilise the negative sanctions that nay result from not acting in accerdance with a par- ticular expectation. ”Sanction“ was operationalised in the following nnners The respondent was asked ”to indicate how those who expected bin to conform to expectation A, and then those who expected hin to conforn to expectation B, would react if he did not do what they ex- pected of hin."15 The concept of sanctions was utilised by Ehrlich in his analysis of role conflict, however, under the rubrie of “obligation." Ehrlich defines obligation in these terns: ”A role expectation will be said 8 to be obligatory if and only if failure to perfons the role expected imposes negative sanctions upon the actor and the performance of the role expected does not incur negative sanctions."16 Obligation was then operationalised by two questions: (1) "would they insist or demand that you do as they expect you to doi", and (2) “What would they do if you didn't do as they expect you told" . The second question above corresponds withathe operationali- sation of sanctions by Gross and his associates. However, this question had to be eliminated from Bhrlich's role conflict schedule because the troopers in his sample simply could not respond to it coherently. to are left then with the introduction of another role conflict variable, vis., obligation or the insistence of an audience or individual to have its expectation performed. Killer and Shull operationalised sanctions by a question which elicited (l) perception of audience reaction for failure to conform to the expectation and (2) perception of what would happen if the individual followed the expectation but failed to fullfill it. It is apparent that the concept of sanctions, along with legiti- nacy, has been operationally defined in diverse nanners. Already at this early stage in empirical role conflict analysis semantic diffi- culties appear. Indeed, diverse applications of key concepts are desireable at this time. However, the confusion which arises from grouping diverse operationalisations under the same nominal defi- mtion may, at some point in time, outweigh the positive contri- butions offered by the aforementioned analyses of role conflict. Students of role conflict met be onware of the diverse applications 9 of key concepts to accurately assess the results of role conflict analyses. They must also possess this awareness in order to en- hance their potential contributions in the area of role conflict. It is felt that pointing to semantic difficulties would have the additional function of serving as a frame of reference for the evaluation of certain sections of this study along with the work of past and future students of role conflict. Second and perhaps most inportant, the foregoing will serve as an introduction to many of the key concepts used in the following chapters. l. 2. 3. Li. 5e 6. 7. 10 FOOTNOTES Lionel J. Nieman and Jcl. Hughes, "The Problem of the Concept of Role - A Re-Survey of the literature," Social Forces, Vol. In (1951-1952). pp. Ila-1119. See for example John T. and Jeanne E. Gullahorn, "A Model for Role Conflict Analysis" (Paper read at the annual meeting of the American Sociological Association, 1959), and Neal J. Gross, Ward S. Mason and Alexander I. McEachern, Explorations in Role Analysis (New York: John Eloy and Sons Inc., 1958). Gross, Mason and McEachern, op. cit. , devote an entire chapter to criticism of the postulate of consensus. See for example the statements in Alfred R. Lindesmith and Anselm L. Strauss, Social Ps cholo (New York: The Dryden Press, 1949), pp. 33 -3 9, and F. Stuart Chopin, "Role Expec- tations and incompatibilities,” in J.B. Gittler (editor), Review of 3001010 (New York: John Paley and Sons Inc., 957 , p. 7. For empirical analyses carried out within the framework of a broad definition of the area of role conflict see Jerome Laulicht, "Role Conflict, The Pattern Variable Theory and Scalogram Analysis," Social Forces, Vol. IXXIII (19511-1955), pp. 250-2511, and Ro_bert' 'S‘."o‘rt‘ TM Study of 11.1. Conflicts as Related to Class Level," Journal of Abnormal and Social Psychc1ogy, Vol. mm (1952), M For an early empirical investigation of intrarcle conflict see Melvin Seaman, "Role Conflict and Ambivalence in leader- chip," American Sociological Review, Vol. mm (1955), pp. 373-330- Howard J. Ehrlich, "The Analysis of Role Conflicts in a Com- plex Organisation: The Police” (Unpublished Ph.D. dissertation, Michigan State University, 1999). For analyses employing both sanctions and legitimacy as crucial variables in the theory of role conflict resolution see Gross, Mason and McEachern, op. cit., pp. 21114 ff., and Howard J. Ehrlich, "Legitimacy, Obligation, and Ambivalence: An mpirical Evalu- ation of Some Neglected Aspects of Role Conflict" (Unpublished paper, The Ohio Stat. University, 1960). An empirical analysis utilising the concept of sanctions has been carried out by J.P. Sutcliffc and M. Habemn, ”Factors Influencing Choice in Role Conflict Situations ," American Sociological Review, Vol. III 8. 9. 10. 11. 12. 13. 11:. 15. 16. '17. 11 (1956), pp. 695-703. For early usage of the concept of legiti- macy in the area of role conflict see Samuel A. Stouffer, "An Analysis of Conflicti Social Norms,” American Sociological Review, Vol. m (191.19 , pp. 707-717. Talcott Parsons, The Social System (Glencoe: The Free Press, 1951)s P0 2800 J.W. Getsols and 13.6. Cuba, "Role, Role Conflict, and Effec- tiveness,” American SocioloLical Review, Vol. XIX (1951;), p. 165, quoted in Gross, Mason and McEachern, op. cit., p. 21.17. Gross, Mason and McEachern, op. cit., p. 2148. Ibide, Po 296e Delbert C. Miller and Fremont A. Shull, "Role Conflict Behavior in Administration: A Study in the Validation of a Theory of Role Conflict Resolution” (Paper read at the annual meeting of the American Sociological Association, 1960), p. 10. Marion Levy, The Structure of Society (Princeton, New Jersey: Princeton University Pressfl952), p. 1.168. Howard J. Ehrlich, "The Analysis of Role Conflicts in a Com- plex Organisation: The Police" (Unpublished Ph.D. dissertation, Michigan Stat. University, 1959), p. 75. Cross, Mason and McEachern, op. cit., p. 296. Ehrlich, op. cit., pp. 7-8. Ibide. p. 75o II THE PROBE] Knowledge of the empirical aspects of role conflict and role conflict resolution remains in a pristine state despite the success iiich Gross and his associates achieved in their prediction of role ecnflict resolution. There is both the need to replicate Gross's work and to explore the possibilities which his schema suggests. There is the additional need to expand and delineate variables which may be crucial in the determination of role conflict resolution. The primary objective of this study is the determination of the empirical significance of selected role conflict variables. The estab- lishment of such empirical significance implies continued usage of the variables under consideration, subject to continued refinement. On the other hand the determination of non-significance suggests the in- adequacies of the selected role conflict variables. The determination .of non-significance points to the reformulation of the variables or their withdrawal from subsequent role conflict analyses. [we specifically the problem of this inquiry is that of deter- mining the relationship and strength of the relationship between the variables of legitimacy, obligation, sanctions access, sanctions exercise 'and ego's preference on the one hand and ego's reported probable behavior on the other. Perhaps the most appropriate method for such determination is to strive for the greatest research flexi- bility possible. This thesis then will take the form of an exploratory study which seems to offer such flexibility. 12 ' 13 Because of limiting factors, this analysis is not exhaustive of the legical research possibilities. Some variables and combinations of variables seemed to warrant extensive consideration: sme did not. Research Setting and Procedures1 As part of a larger study of the State Police Department, every member of the June, 1960 recruit school class (n=58) was ahinisterod a role conflict questionnaire.2 The administration took place during the recruits' first week of training. To enter recruit school an individual must (1) be from 21 to 29 years of age, (2) pass a written civil service examination, and (3) pass a rigorous physical fitness test administered by members d the State Police Department. During the eigrt week training period the recruits learn various physical an! mental skills necessary for successful performance as a state trooper. Th rigor of the training is evidenced by the fact that only 55 of the original 58 recruits successfully completed the eight week program. Tb role conflict questionnaire is built around five potential role conflict situations that are apprOpriate to the position d State Police trooper. The potential conflicts are those engendered by the individual's occupancy of the single position of state trooper. Briefly, the conflict situations concern (1) whether the trooper should spend his time mainly on safety and traffic, spend his time mainly on complaint and criminal investigation or spend equal time on both: (2) whether the trooper should be a policeman twenty-four hours 11. a dqy or leave his job behind when.not on duty: (3) whether the trooper should report another trooper he thought was not properly perfosming his duty or keep quiet about such negligence: (h) whether the trOOper shoull be a “model citisen' and set a community example or whether he should be simply a “good citizen" not any different from anyone else: (5) whether the trooper should follow-the rules and regulations of the State Police organisation.to the letter or overlook than occasionallyc3 For each of the five potential conflict“eituations the recruit was asked to check which.cf three expectations he perceived each of nine audience groups to hold.h Only in situation 1 (safety-traffic, criminal) was the recruit faced with the selection of three expecta- tions. In.the remaining four conflict situations the recruit was offered three alternatives. However, the third alternative or expec- tation "C”, when chosen, constituted a perception of a particular audience .. holding “no expectation.” After his selection of expectations the recruit was asked about the legitimacy, access to sanctions, exercise of sanctions and obli- gation.that he perceived to be associated with each of the audiences and their expectations. Finally the recruit was asked to record which of the expectations he would personally m each of the audiences to hold. At the end of the two part questionnaire the recruit was asked, for each situation, what, he as a general rule, probably would do once he got out on the job. The alternatives from.which he chose corres- ponded to the particular expectations associated with the five different conflict situations. 15 The Selected Role Conflict Variables The role conflict variables under observation were (1) per- ceived expectation of audience, (2) perceived legitimacy of audience’s expectation, (3) perception of audience's access to sanctions, (h) perceived obligateriness with which audience holds expectation, (5) perception of the probability of audience exercising sanctions if their expectation is not followed, and (6) expectation which the recruit prefers each audience to hold. The dependent variable was the recruit's reported probable behavior (as ascertained from his responses to the questions on the last two pages of the role conflict questionnaire: Manifest Behavior Section). The operational definition of legitimagz used in our question- naire paralleled that of Gross and his associates. The recruits were asked, I"lhat right do you think this category has to hold this view!" The recruit w... expected to take both th. content of the expectation and the authority of the audience holding the expec- tation into account in making his assessment of legitimacy. The five point response category accompanying this question ranged from “has every right” to ”has no right." ’ Next, following ahrlich'. lead, the rigor with which expec- tations are held seemed to constitute an area for exploration. How rigidly or flexibly are expectations held? Obligation has been operationaliscd in our questionnaire as, "Hew strongly do you think persons in this category generally insistuthat troopers act according to this view?” The response category ranged from "absolutely insist" 16 to "do not really care.“ - Sanctions was broken down into two distinct questions. First, the recruits were asked, "In what position are the persons in this category to apply pressure to try to make troopers do as these per- sons expect?" The recruits were then asked to check one point along a five pointl‘seale which varied from "in the strongest position" to "in no position.“ This variable was designated as sanctions access. The second question concerning sanctions asked, "If troopers didn't hot the way persons in this category who hold this view ex- pect, what is the likelihood that such persons would actually de sonething to try to get troopers to act according to their view?" This variable was labeled sanctions exercise. Although not per--~ fectly equivalent, this question corresponded to Gross's Opera- tionalisation of sanctions. The response category ranged from ”every likelihood“ to "no likelihood.” . Finally the recruits were asked,“‘lhich of the three views presented at the top of this page is the one which you would most like persons in this category to have about troopers?" This con- stituted an important addition to the analysis of role conflict in that a personality factor was introduced. However vague the question night be, it did allow assessments of the relationship of a purely personal dimension with reoruits' reported probable behavior. This variable was labeled ggo's preference: On the last two pages of the role conflict questionnaire (Mam.- fest Behavior Section) the question was asked for each conflict situation, “much one of the following comes closest to describing 17 what, as a general rule, you will do once you get on the Job?" The question was followed by three statements of behavioral alternatives which coincided with the three statements provided to ascertain a rccruit's perceived expectations. The rocruits' responses to this question were referred to as gported probable behavior, and they constituted our dependent variable. l. 2. 3e h. 18 FOOTNOTES The research presented in this thesis constitutes a facet of a larger project: ”Role Image of the State Police", project M- 2957 supported by a grant from the National Institute of Mental Health. Principal investigator of the project is John C. Howell. The role conflict questionnaire is found in Appendix two. The five potential conflict situations were originally developed by Howard J. Ehrlich and incorporated into his role conflict schedule. This schedule was later modified, enlarged and worked into the role conflict questionnaire by Dr. Howell. For a noro comprehensive explanation of the five conflict situations see Howard J. Ehrlich, "The Analysis of Role Conflicts in a Complex Organisation: The Police" (Unpublished Ph.D. Mssertation, Michigan State University, 1959), pp. 26-49. The audiences are adopted in modified form from Dr. Ehrlioh's role conflict schedule. The nine audiences employed in this study are: Wives and/or Families of troopers, Troopers in general, Post Commanders, The General Public, Headquarters and District Command Officers, The Press, Personal Friends of troopers, Politicians, and Local and County Police. III RELATIONSHIPS OF THE ROLE CONFLICT VARIABLES AND REPORTED PROBABLE BEHAVIOR This chapter focusses upon the variables of legitimacy, obliga- tion, sanctions access and sanctions exercise and their empirical relationships with the recruits' reported probable behavior. The approach to the empirical testing of those role conflict variables presented in this chapter is different from previous analyses. For this reason a review of the basic method previously utilised to as- certain the efficacy of role conflict variables, that of Gross and associates, seems to be a necessary inclusion. Gross's theory of role conflict resolution contains three elements: (1) the individual's perception of legitimacy, il- legitimaoy of expectations, (2) the individual's perception of the sanctions that result from nonconformity to an audience's ex- pectation, and (3) the individual's orientation to legitimacy and sanction”:l From the logical possibilities of the combination of sanctions and legitimacy Gross posits sixteen types of role conflicts. These types range from expectation A and B both being viewed as legiti- mate and accompanied by negative sanctions for nonconformity to both expectations A and B being viewed as illegitimte and not accompanied by negative sanctions for non-perfonaance of the ex- pectation. Three types of individuals are then determined. One type responds primarily in terms of legitimacy; another type re- sponds primarily in terms of sanctions; the third type takes both 19 20 sanctions and legitimacy into account when.determining his course of behavior. Gross further breaks down his theory into three models of role conflict resolution. The basic difference between.the models and the theory is that, with the models, no knowledge of the indivi- dual's orientation.tc sanctions, legitimacy or both is necessary for behavioral predictions. The first model is labeled the "legiti- maoy model." It attempts to predict behavior simply from a knowledge of the individual's perception of audience-expectation legitimacy. The second, the "sanctions model," attempts to predict behavior from the knowledge of an individual's assessment of sanctions that result from nonconformity to an expectation. The third, the ”legitimacy- sanotions model," predicts behavior from.the knowledge of both the individual's perception of legitimacy and sanctions. Utilising these models Gross'was able to predict role conflict resolutionuwith a high degree of accuracy. In.plaoe of:models for the prediction of reported probable behavior the chi square test has been used in.this thesis to de- termine the relationship between‘the selected role conflict variables and reported probable behavior. In addition the coefficient of cone tingenoy has been.emplqyed to indicate the relative strength of the relationships between reported probable behavior and the various role conflict variables and combinations of variables. In.this analysis the value of chi square is reduced because the same expectation.nay be perceived to be held by audiences which are perceived to hold high access to sanctions, legitimate expectations 21 and so on, and simultaneously by audiences perceived as holding low access to sanctions, illegitimate expectations etc.2 Furthermore two audiences assessed as holding high access to sanctions etc., may be perceived to hold incompatible expectations.3 Although this diffi- cultwaas not mentioned by Gross and his associates, it must have confounded his predictions to some extent. It is felt that despite the different techniques, the results of this inwestigation.should be roughly comparable to those obtained by Gross and his associates. lith.these considerations in mind the intestigation of the re- lationship between obligation and reported probable behavior'nay proceed. Table 1 indicates that a highly significant relationship exists between obligation and reported probable behavior in every situation except IV, in which the relationship fails to attain statistical significanceJ" However, the values of the coefficient of contingency, ranging from .063 in situation IV to .237 in.situ- ation 11, point to a rather tenuous association. The strength of the relationship between obligation and reported probable behavior is moderately low. This relationship is not constant from situa- tion.to situation as evidenced by the flucuation in the contingency coefficients. In general then, the obligatoriness with which ex- pectations are perceived to be held manifests a moderately low as- sooiation with the reports of probable behavior.5 In regard to legitimacy and reported probable behavior Table 2 shows that their relationship is significant in four of the five conflict situations. In the remaining situation.the relationship, although in the expected direction, fails to attain statistical TABLEI RELATIONSHIP OF OBLIGATION AND REPORTED PROBABLE BEHAVIOR IN THE FIVE POTENTIAL ROLE CONFLICT SITUATIONS cm LEVEL or *CONTINGENCY SITUATION SQUARE SIGNIFICANCE COEFFICIENT N I 12.25 p<.001 .me 522 II 26.32 p<.oOI .237 1.143 III 17.9} 1.2.001 ' .207 hog IV 1.91 ms. .063 14.90 v 15.1.1 p<.OOl .179 1.52 df=1 *Interpretation of values of the coefficient of contingency is relative to the number of cells in a contingency table. The maximum value that C can attain in a 2x2 contingency table is .707. Sec Quinn IeNemar, PsthOIOgical Statistics (New York: John liley and Sons Inc., 19149), ppe 1" 2e RELATIONSHIP OF LEGITIMACY AND REPCRTED PROBABLE BEHAVIOR IN 23 TABLEZ THE FIVE POTENTIAL ROIE CONFLICT SITUATIONS CHI IEVEL 0E *CONTINGENCY SITUATIOI SQUARE SIGNIFICANCE COEFFICIENT N I 3.37 .10>p>.05 .075 522 II 12.8.15 p<~°01 .313 MB III 17.111 p<.001 .203 1.05 IV 5.73 p<-05 .109 190 v 19.o8 p<.OOl .200 1.62 t maximum value of C is .707 21; significance. The contingency coefficients are, in general, law, and flucuate from situation to situation. Interestingly the two situations in which the lowest association is found between legiti- macy and reported probable behavior are situations I and IV. The highest association between the variables is found in situation II. This pattern coincides with that manifested between obligation and reported probable behavior. The legitimacy which audiences and their expectations are perceived to have manifests a moderately low relationship with reported probable behavior. Table 3 reveals that the relationship between sanctions access and reported probable behavior is statistically significant at varying levels in four of the five conflict situations. Again the contingency coefficients are not as high as might be expected. The highest contingency coefficient, found in situation II, is only .167. Again too, the situational pattern of high association in situation II and low association in situations I and IV is evi- dent. As has been the case with the relationships of the previous variables tested and reported probable behavior, sanctions access manifests a significant but moderately weak association with re- ported probable behavior. Finally, rth. 1. indicates that the relationship between sanctions exercise and reported probable behavior is significant in four of the five conflict situations. The pattern of high and low association in situations II and I and IV respectively is again evident. The contingency coefficients generally point to a rather weak association between sanctions exercise and reported RELATIONSHIP OF SANCTIONS ACCESS AND REPORTED PROBABLE BEHAVIOR IN THE FIVE POTENTIAL ROLE CONFLICT SITUATIONS 25 TABLE3 CHI LEVEL (1“ tCONTINCENCY SITUATION SQUARE SIGNIFICANCE COEFFICIENT N I nod.- “’05 0089 522 II 12.92 p<.001 .167 1.1.3 III 5-55 p 80. .2: 84 E Sm. So.va at: H: RN. Sci «98 HH 9:. 300a 3.3 H a u a mu mAMTMMHUmMKm mon9024w o a me o a ma mamémmooq Eméosfiflofi monaozsm o a «a 20:4 .85 mmMJonsdeAmo WZOHadbhHm Bqumzoo maom AdHBZMBom MPHh Mme zH MOH>dmmm mumdmomm Qmamommm 924 mmHommNm monaozp>.05 .118 522 II 811.60 p<.001 .1100 1115 III 21.52 p<.001 .225 hos Iv 114.511 p<.OO5 .167 1190 v 21.50 p<.001 .210 1.62 «:5 *The maximum value of C for a Exh contingency table lies between .707 and .866. The precise value of maximum.c cannot be computed when the number of rows and columns in a contingency table are not equal. Quinn.McNemar, Ps cholo ical Statistic: (New'York: 71—78% Sons Inc., 19h9 , p. . John Wiley and Sec 56 . .aom .a Add." .35 Reason xoom Sansone: 373» I33 33:91.3 choomfieungoam .Momfloooa .0 egg eon .ugoo mo mace—Bu pockets A»: means». honemnflboo scum merged codewouumeoo Mo noeaasaaoc vogue 253s hopes.“ noavoeaaoo ens .noeflcafioo no eoeomba no.“ eevooauoo noon stun apnedoamooo honownavnoo one Th. Tee Tee meg. mom. floo.va om.H~ mmm. moo.va es.o~ 4am. Hoo.va mo.mH a om: mum. moo.va :m.:H meo. ...n we. Had. mo.va ma.m pH mo: mom. Hoo.va mm.Hm 4mm. Hoo.va :m.:a man. Hoo.vm :a.a~ HHH 93 3a. 300a 8.6 men. . 86.x 8.8 m3. Sofa m4. 3 3 «am we”. mo.AAAoH. me.a can. mo.va any: on“. mo.nanoa. am.m H a m a we m a mu m a we 202433 mam.mmHommxm mamtmmHummxm monsozsm mamvuoszHaHemu monaozem-aossHchmu monedaHm soHAkzoo WHOM aazmaom m2; E 2H “BU—”bag 30E anamommm n24 Qflszao mmHomE monaog 924 Baanmfl .mmHomme monsozdm .HogusmA ho mAHMmonadAmm Mme mo Mg mugs 37 behavior. The differences between the relationship of legitimacy and reported probable behavior on the one hand and the combination of legitimacy and sanctions exercise on the other are extremely slight. In two of the five conflict situations legitimacy is more closely related to reported probable behavior than are legitimacy and sanctions exercise combined. Since the combination of legiti- macy and sanctions exercise has resulted in only a slight increase in relationship strength with reported probable behavior, it is concluded that this procedure may be unwarranted in subsequent role conflict analyses, of course contingent upon improved statistical techniques and refinement of the variables. In summary it has been ampirically determined that: (l) Legitimacy, obligation, sanctions access and sanctions exercise all manifest a significant but moderately‘weak relationship with reported probable behavior. (2) The role conflict variables under consideration are interrelated to a.moderately high degree. lhen an audience and/er expectation is perceived to be legitimate, there is a strong tendency for it to be concomitantly perceived as obligatory, having high access to sanctions and having a great likelihood of exercising sanctions. (5) The procedure of pairing two variables, vis., legitimacy and sanctions exercise did not result in a substantial relationship increase with 38 reported probable behavior over that which obtains between legitimacy and sanctions exercise singly and reported probable behavior. Comparison With Previous Role Conflict Investigations A brief comparison will be carried out of the results of this thesis with those of other empirical studies of role conflict. Un- fortunately the statistical techniques utilised in this thesis represent a departure from those used in previous investigations of role conflict. However, it is felt that the results lend them- selves to rough comparisons. It has been determined that our role conflict ‘variablos mani- fest in general equivalent relationships with reported probable behavior. This finding differs from those of Gross and his as- sociates. They received their highest predictive accuracy utilising the “sanctions model." Compare for example the proportions of corrtct predictions given below. With the “legitimacy model" .8’4, .27, .65 and .66 correct predictions of role conflict resolution were attained. Employing the "sanctions model” Gross and his as- sociates correctly predicted .87, .79, .75 and- .87 role conflict resolutions.7 As was the case with the combination of legitimacy and sanctions exorcise investigated heroin, Gross and his cohorts did not improve predictions with their “legitimacy-sanctions model.” Sanctions was found to be the best predictor of role conflict rose; lotion by Gross and his associates. This finding was not replicated in this thesis. 39 The results of Ehrlich's investigation of role conflict are more in accordance with those of this thesis. This might be ex- pected considering that both studies utilised the same potential role conflict situations. With legitimacy Ehrlich correctly pre- dicted .55. .72, .00, .72 and .27 role conflict resolutions in the five potential conflict situations. Employing obligation Ehrlich correctly predicted .50, .65, .00, .66 and .2h role conflict reso- lutions. It is obvious that neither variable is "better" than the other in terms of correct predictions of role conflict resolution. It is also to be noted that the proportion of correct predictions is not nearly as high as that obtained by Gross and his associates. Ehrlich also combined legitimacy and obligation in an attempt to increase predictions. Hewever, this procedure did not produce the desired results. The combination of obligation and legitimacy re- sulted in.exactly the same proportions of correct role conflict resolution.predictions as those obtained with the single variable of legitimacy.8 Taking both sanctions and legitimacy'intc account, Miller and Shull achieved overall predictive accuracy of .71 from seven samples of business and labor leaders.9 In summary, no intesti- gators have achieved the high degree of correct predictions of role conflict resolution that'wero obtained by Gross and his as- sociates. 1. 2. ho FOOTNOTES For a complete explanation of this theory see Neal J. Groas, Hard S. Mason and.Alexander W3 McEachern, Explorations in Role Analysis (New York: John Wiley and Sons Inc., 1958), pp. 281-318. It is to be noted that Gross and his associates analyzed only those persons who were determined as having role conflict - role conflict being defined as the perception of contradictory expectations. Those per- sons who perceived each audience to hold similar expectations for their behavior in a particular situation (role congruency) were ex- cluded from.their analysis. In this study no distinction is made between persons having role conflict and role congruency in the sense that both are included in the empirical analyses of the re- lationship between role conflict variables and reported probable behavior. In a very real sense the empirical focus of this thesis is not role conflict resolution. The five point response categories were out after the first two points. In other words if the recruit checks one of these first two points his response is considered to constitute a perception of obligation, legitimacy, high access to sanctions and great like- lihood of exercise of sanctions. The remaining three points are regarded as non-obligatory, illegitimate perceptions etc. In addition, whenever a recruit reported his probable behavior as ”have no general practice on this matter" the value of chi square was reduced. It is incorrect to assume that the perception of ex- pectation C or "no expectation" corresponds with the selection of reported probable behavior 0 or "have no general practice on this matter." If a recruit perceives an audience to hold "no expecta- tion" for his behavior, it is reasonable to assume that he is rela- tively free to follow any of the three alternative behaviors listed in the Manifest Behavior Section. This discussion does not apply to situation I in that expectations.a, B and C all constitute be- havioral prescriptions, which correspond with the three alternatives listed for situation I in the Manifest Behavior Section. See the role conflict questionnaire in Appendix Two. The 2x2 contingency table for any one situation in Table l was tabu- lated as follows: If a respondent perceived four of the nine audiences as holding an obligatory expectation that was similar to his reported probable behavior, cell 1 of the contingency table ‘would receive four checks. If this same respondent perceived the remaining five audiences to hold a non-obligatory expectation not in accordance with his reported probable behavior five checks were placed in.cell h. (See Appendix One for the contingency table for- mats.) Thus, the re3ponses of each recruit were assigned to 5. 7. 8. L1 particular cells of the contingency table nine times for any one situation. While this procedure may inflate the value of chi square because of dependency of responses, it is felt that the factors contributing to the deflation of chi square outweigh this inflationary factor. The differing N's in each situation are derived from.the number of times 58 recruits respond to questions concerning legitimacy, obligation etc. N for any one situation becomes 9(number of audiences) x 58(number of recruits)-522. However, only in situation I will there always be an N of 522. In situations other than I the recruits were instructed not to respond to questions concerning legitimacy etc., when they checked expectation C or "no expectation.” The raw data from which the statistical tests were computed are given in tabular form.in.Appendix One. The data do not allow an empirical explanation of the pattern of low relationships between all the role conflict variables and reported probable behavior in situations I and IV and high relationships in the remaining situations. The following explanations are only conjectural. The situations may have had differential meaning and salience for the recruits. Perhaps the recruits were not concerned with the distinction between being a "model citizen" or a "good citizen." Situations I and IV then may not represent pervasive and important conflicts for the recruits. The third expectation in situation I might have negated the exclusive aspect of expectations.a and B. In other words the additional ex- pectation offered a compromise selection which might have been con- ducive to this situation not being perceived as an important, pervasive conflict. This explanation is tenable in the light of "compromise" often being cited as a means for eluding role conflict.. v The investigation of the relationship between two role conflict vari- ables and reported probable behavior necessitates the use of a 2xh contingency table. With an increase in the number of cells in a con- tingenoy table the maximum value which C can attain rises. Gross, Mason and McEachern, op. cit., p. 3114. The proportions cited are derived from.the four ro e conflict situations employed in this study. Howard J. Ehrlich, "The Analysis of Role Conflicts in a Complex Organization: The Police" (Unpublished Ph.D. dissertation, Michigan State University, 1959), p. 57. Delbert C. Miller and Fremont A. Shull, "Role Conflict Behavior in .Administration: .A Study in the validation of a Theory of Role Con- flict Resolution" (Paper read at the annual meeting of the.American Sociological.Association, 1960), p. 18. IV INTRODUCTION OF EGO'S PREFERENCE AS A ROLE CONFLICT VARIABLE In.this section.the relationship between ego's preference and reported probable behavior'will be investigated. The investigation of what expectations recruits prefer audiences to hold and their reported probable behavior constitutes a departure from the previous investigations of legitimacy, obligation and the like. In the pre- vious analyses personal dimensions were disregarded. It is felt that the introduction of personal factors into the analysis of role conflict constitutes a significant contribution in the direction of a.mcre comprehensive knowledge of role conflict and role conflict resolution. .After all, the concept of role has been repeatedly lauded for its inter-disciplinary integrational possibilities. Yet, in the hands of investigators identified with various disciplines the concept has often remained rather esoteric to the student and his particular discipline. It seems apparent that personal dimensions must be taken into account in any analysis of role conflict. Gross and his associates determined school superintendents' personal orientations to legiti- macy, sanctions or both in order to enhance their predictive ef- ficacy. However, this can hardly be construed as an investigation of the relationship of role conflict resolution and personal dimen- sicns.1 Table 1b shows the results of the investigation of the re- lationship of what expectation recruits prefer audiences to hold 142 h} and reported probable behavior. The differences in the chi square values for situations I and III on the one hand and II, IV and V on the other present an erratic pattern. In the latter situations the chi square values are extremely high, while in situations I and III the relationship 1.11. to attain statistical significance.2 It is difficult to compare the relative relationship strength between egc's preference and reported probable behavior and the previously tested variables and reported probable behavior. From the chi square values alone the relationship between egc's preference and reported probable behavior appears to be greater than that be- tween the other role conflict variables and reported probable behavior in three of the five situations. However, the overall stability of of the relationship between egc's preference and reported probable behavior is not as great as that, for instance, beheen legitimacy and reported probable behavior. Such instability, of course, would reduce the predictive utility of this variable. Because of the erratic pattern of significance manifested by the relationship of ego's preference and reported probable behavior, it is tentatively concluded that the strength of this relationship is, from an overall standpoint, roughly cqmnensurate with the re- laticnship betwoen the other role conflict variables and reported probable behavior. The results of the investigation of ego's preference and re- ported probable behavior must be interpreted with care. The role conflict questionnaire was administered during the respondents' first week in recruit school. For that particular point in time TABLElh RELATIONSHIP OF EGO'S PREFERENCE AND REPORTED PROBABLE BEHAVIQQ IN THE FIVE POTENTIAL ROLE CONFLICT SITUATIONS in ; 4— fl *CHI LEVEI.CF SITUATION SQUARE SIGNIFICANCE N I 2.92 e10>p>e05 522 II 139.86 p<.001 h05 III 1 e0? 11. I e 373 IV 71.16 p<.001 502 v 107.80 p<.001 wig df=l *The chi square values were derived from.a two cell contingency table. Expected frequencies were obtained by dividing each situation l'by 2. See Morris Zelditch, Jr.,.A Basic Course in S (lbw'Iork: Henry Holt and Company, 1959), p ociolo ical Statistics fl? p. 2 1‘2 0 h5 it might be expected that personal preferences would be a stronger determinant of role performance than would be the case at a later period in the respondents' careers. The plausibility of the above point is attested to in a study by Ehrlich, Rinehart and Howell. In this study the recruits who successfully completed the June, 1960 recruit school and a six month cn-the-job probationary period (n: 55) were re-adninistered the role conflict questionnaire. Theresults of this study re- vealed that the variables of legitimacy, sanctions access and the like, were more effective predictors of role conflict resolution than egc's preference in all five potential conflict situations.3 Despite the qualifications placed upon the interpretation of the relationship between egc's preference and reported probable behavior, and despite the tentative nature of the conclusions drawn concerning the strength of this relationship, the results do point to the subsequent inclusion of personal dimensions in role conflict schemata. Simultaneous Combination 0f Social.And Personal Dimensions The previously reported investigations have indicated that the variables of legitimacy, sanctions access, sanctions exercise, obligation and egc's preference are all significantly related to reported probable behavior. However, the relationship between these variables and reported probable behavior is not as strong as might be desired, and in the case of egc's preference not as stable as might be desired. The above considerations immediateby he suggest the possibility that the conceptual framework'withinrwhich the empirical investigations have been carried out is inadequate. In the previous chapter factors in the social situation, which some students have deemed crucial to the understanding and prediction of role conflict and role conflict resolution, have received the greater part of our attention. Specifically, dimensions of audiences and the content of their expectations were the focus. Those audiences and/or expectations viewed as legitimate, obligatory, backed by high access to and exercise of sanctions have been.thecretica1ly regarded as a cast into which individuals are automatically melded. lhile the socially patterned aspects of role conflict and role conflict reso- lution.have been.enphasised, the function of personal predispositions has been grossly neglected. Such a neglect led to the inclusion of an investigation of the relationship of personal preferences and re- ported probable behavior. In this analysis the social dimensions 'were disregarded. It is felt that both investigations did not result in the de- sired close relationship with reported probable behavior because each 'was, in effect, segmentising aspects of the total forces influencing individuals' decisions in the conflict situations. Each empirical investigation.was carried out without benefit of a comprehensive cons ccptual framework. It is contended that only by the simultaneous integration of both personal and social dimensions'will the under- standing of role conflict and role conflict resolution be adequately approached. The area of role conflict has not been completely bereft of h? attempts at the inclusion of both personal and social dimensions into conceptual frameworks. Gullahorn points out that role rep- resents "a point of interaction between the personality system and the social system.” He further contends that personality influences role behavior as well as being influenced by it. In essence, the influence of both personal and social dimensions upon the determina- tion of modes of behavior is attested to and integrated into Gulla- horn's model for role conflict analysis.h According to Getsols and Guba a social system.entails two major classes of phenomena: (1) Organizations with certain statuses and roles that will fulfill the goals of the system.and (2) indi- viduals with their personalities and need dispositions. Two di- mensions of activity are further specified: (a) The nomethetic dimension of activity, uhich is comprised of the organisation, status and role and (b) the idiOgraphio dimension of activity, which is comprised of the individual, personality and need dispo- sitions. Wa.given act is conceived.as deriving simultaneously from both nomethetio and idiographic dimensions.'5 In general, this conceptual framework can be adopted as a guide for further empirical investigation of the role conflict variables. Legitimacy, sanctions access, sanctions exercise and obliga- tion, being dimensions of audiences and the content of their expectations, can be considered as aspects of the nomothetic di- mension of activity. Ego's preference, which is a function of personality and need dispositions, can be considered as an aspect of the idiographic dimension of activity. It is felt that the 1+8 simultaneous inclusion of elements from.both the idicgraphic and ncmothotic dimensions of activity in the empirical investigation of role conflict variables will enhance the relationship with re- ported probable behavior. Specifically, it is hypothesized that the combination of legitimacy, for example, and cgo's preference will result in a relationship with reported probable behavior that is stronger than that between either of these variables alone and re- ported probable behavior. Table 15 indicates that the combination of ego's preference and legitimacy is significantly related to reported probable be- havior in each of the potential role conflict situations. With the exception of situation I, the contingency ceefficients are all moderately high. The important question of the relative strength of the relationship between the variables presently under considera- tion and reported probable behavior remains to be answered. Comparing the contingency coefficients in Table 16, it can be positively asserted that the combination of ego's preference and legitimacy has resulted in an enhanced relationship with reported probable behavior over that of any single variable or combination of variables and reported probable behavior. Unfortunately the relative strength of the relationship between the variables pre- sently under consideration and ego's preference and reported probable behavior cannot be directly compared. However, the re- lationship between this combination of variables and reported probable behavior is more stable than that between ego's preference and reported probable behavior. Because of the erratic nature ef 149 mm 15 RELATIONSHIP OF THE COMBINATION Q“ EGO'S PREFERENCE-EGITIMACY AND RE- PORTED PROBABLE BEHAVIOR IN THE FIVE POTENTIAL ROLE CONFLICT SITUATIONS CHI LEVEL or ccoNTINCENCI SITUATION SQUARE SIGNIFICANCE COEFFICIENT N I 11..03 p<.005 .161 522 II 152.11; p<.001 .538 373 III 28.23 p<.001 .279 333 N 60.38 p<.001 .333 has v 110.69 p<.001 .151. 1.26 $3 *The maximum value of C for a 211+ contingency table lies between .707 and .866. Sofia om.~oa .cansv hencwswvnco HHco es» s c» cansowfimme won ea hanemnapnoc Mo pueaowmmcce amass .encewasmaoc vocaae you ecvccaaco neon can honcwndpnco mo anenoamueoo age. So. Va «mo. $6: mom. Soda 3.8 4%. floods 8.2 > 8o.Va 3.2. can. 80.x 3.8 mum. mecca 4?: HS. no. vo mum B .3. 34 «mm. Soon no.8 mom. Soda «93. man. Gods in.“ HHH So.vo 8&2 Ra. Soda edema 3m. So.va 8:6 «on. 80.!“ m4. 3 E 50 853:. «ad 08. moo.va 3...: m3. moire“. me.» was. moKnAS. and H a «a b a «a b a «a b. a «a 2894 Inn—“Hm mmm mmmlmozfimammm Imozamhmmm m .oamee yommlhogeHma “H I: mmmlfimHomg monaozalwodfiHanW-H mQMIModEaHofi monHfiPun HUHQEAOO "whom aezmhom HPHm Mme zH Mogdmmm admomm anamommm Bad wodEHBHcH—H 924 gamma? mama .MoganmA and HmHommNm monaoim .HodaaHmmA ho mmHmmonamm E mo Mg flu 3mg 51 the relationship between egc's preference and reported probable behavior, it is concluded that the relationship between ego's preference - legitimacy and reported probable behavior is greater than that manifested by any other variable or variables and re- ported prebable behavior. There appears to be ample grounds, both theoretical and one pirical, to conclude that the inclusion of both personal and social dimensions represents a significant contribution to the under- standing of role conflict and role conflict resolution. 1. 2. 3. 5. 52 FOOTNOTES The method used by Gross and his associates to determine school superintendents' personal orientations to legitimacy, sanctions or both is one which does not lend itself to replication, even if it were desireable to do so. See Neal J. Gross, ward S. Mason and.Alexander‘W; McEachcrn, Explorations in Role Analysis (New York: John.Wiley and Sons Inc., 1958), pp.*289-295. The coefficient of contingency is not applicable to a two cell contingency table. Howard J. Ehrlich, James W3 Rinehart and John C. Howell, "The Study of Role Conflict: Explorations in methodolOgy" (Paper to be read at the annual meeting of the American SociOIOgical .Association, 1961). Jehn T. and Jeanne E. Gullahorn, ”A.Mbdel for Role Conflict Analysis" (Paper read at the annual meeting of the American SociolOgical Association, 1959), pp. hff. J.W} Getsols and E.G. Guba, "Social Behavior and the.Admini- strative Process" (Revised version of paper read at the annual meeting of the American Sociological Association, 1956). SUMMARY Five role conflict variables have been investigated. Two of the five, ego's preference and sanctions access, have been intro- duced for the first time. It has been empirically determined that the role conflict variables of legitimacy, obligation, sanctions access and sanctions exercise are significantly related to recruits' reported probable behavior in the majority of potential role cone flict situations. Hewever, the relationships between the afore- mentioned role conflict variables and reported probable behavior are moderately weak, as evidenced by the reported values of the coefficient of contingency. The role conflict variables of legitimacy, obligation, sanc- tions access and sanctions exercise were examined to determine if any one'was associated with reported probable behavior to a greater extent than were the others. Utilizing the coefficient of contin- gency as a test, it was found that the strength of the associations between each of the four role conflict variables and reported prob- able behavior was, in general, the same. No role conflict variable was dominant over any other in regard to their relationship with reported probable behavior. I Since none of the four variables was dominant in regard to its relationship with reported probable behavior, it was expected that the variables would manifest a high degree of interrelation- ship. .An analysis aimed at the determination of the dependence or 53 5h independence of the role conflict variables revealed that they were definitely dependent. A moderately high relationship was manifested between each pair of variables tested. The closest relationship was that between sanctions access and sanctions exercise. This finding was not surprising in that both variables were cperationalized to tap different aspects of the same underlying dimension. Although the four role conflict variables were determined to be highly interrelated, there remained a siseable proportion of the variance that was unexplained. Because the association between the variables was not perfect, it was felt that the procedure of combining two variables might increase the magnitude of the association with reported probable behavior over that which either variable manifested singly with reported probable behavior. V The variables of legitimacy and sanctions exercise were combined in an attempt to determine their association with reported probable behavior. The results of this investigation revealed that the re- lationship with reported probable behavior had not been substantially increased. The single variables of legitimacy and sanctions exercise were related to reported probable behavior to approximately the same degree as was the combination of legitimacy and sanctions exercise. a personal dimension, ego's preference, was introduced in this study as a role conflict variable. ‘The analysis of ego's preference and reported probable behavior revealed that these variables mani- fested an unstable relationship. In three of the five potential role conflict situations the relationship was found to be moderately high, while in the remaining two situations the relationship was found to 55 be moderately low. It was concluded that ego's preference was, in general, no more highly related to reported probable behavior than were the other role conflict variables and reported probable be- havior. Finally, empirical evidence was offered for the desireability of the simultaneous inclusion of both personal and social dimensions into role conflict schemata. The variables of ego's preference and legitimacy were combined. The results of this combination led to the conclusion that a significant relationship increase with reported probable behavior had been effected over and above that which had been previously manifested between personal and social role conflict variables singly and reported probable behavior. Conclusions And Suggestions For Future Research The following conclusions are drawn from.the empirical findings of 1:111. study: (1) That the role conflict variables of legitimacy, obligation, sanctions access and sanctions exercise are not inde- pendent, (2) that none of these variables is dominant in regard to its relationship with reported probable behavior and (3) that a combination of two variables did not increase the relationship with reported probable behavior. It is concluded that these findings strongly suggest that subsequent role conflict schemata consider the possibility of using only one of the four role conflict variables mentioned above. This conclusion is of course contingent upon sub- sequent refinement of the role conflict variables and improved statistical techniques for their handling. Which variable should 56 be chosen would of course depend upon the particular concerns of the investigator, the nature of his sample and the type of infor- mation.he is seeking. 0n the basis of this study the selection of legitimacy might constitute a wise choice, in as much as legiti- macy as operationalised herein evokes both the content of the expectation and the authority of the audience holding the expec- tation. .L fourth conclusion is that the study's exploration of per- sonal dimensions strongly suggests their subsequent inclusion in role conflict schemata. However, the results of the investigation of the combination of egois preference and legitimacy led to the further conclusion that the combination of personal and social variables was more closely related to reported probable behavior than‘was either singly. It cannot be cveremphasised that the separate analysis of personal and social dimensions only covers a segment of the total forces influencing individuals' behavioral decisions in role conflict situations. Only by the simultaneous inclusion and investigation of both personal and social dimensions 'will subsequent theoretical and empirical studies of role conflict approach the necessary requirements for explanation and prediction of behavior in role conflict situations. In regard to future research, there is a need to validate and replicate the theory and models of role conflict resolution as introduced by Gross and his associates. If the theory of Gross and his associates is valid, investigators should be able to achieve the same degree of predictive accuracy that was first attained. 57 The results presented in this thesis do not create the degree of excitation as that engendered by the analysis of Gross and his associates, nor do the results of other investigations of role conflict give the impression that "all that is to be done has been done." Second, the implications of the concept of sanctions can be more exhaustively probed. In this and previous analyses of role conflict onLy negative sanctions or the position of audiences to invoke negative sanctions for nonconformity to an expectation.has been explored. It is suggested that knowledge of individuals' perceptions of the positive sanctions that result from.conformity to an expectation.might be highly related to behavioral decisions in role conflict situations. Third, it goes without saying that both personal and social dimensions should be included in future role conflict schemata. Finally, the determination of factors which reduce the re- lationship between the role conflict variables and reported probable behavior offers an interesting direction for research. For example, such concepts as visibility and commitment might be relevant within this context. In general the area of role conflict is one which offers a multiplicity of research possibilities, both theoretical and empirical. The concept of role is one which is potentially capable of providing a necessary link to the integration of di- verse disciplines. Only by continued theoretical and empirical 'work can the full potential of the concept of role be realized. SELECTED BIBLIOGRAPHY Ehrlich, Howard J. "The Analysis of Role Conflicts in a Complex Organisation: The Police" (Unpublished Ph.D. dissertation, Michigan State University, 1959). Ehrlich, Howard J., Rinehart, James W. and Howell, John C. '.The Study of Role Conflict: Explorations in Methodologyu (Paper read at the annual meeting of the American Sociological Association, 1961). Getsols, J.W. and Guba, E.G. "Social Behavior and the Administration Process" (Revised version of paper read at the annual meeting of the American Sociological Association, 1956). Gross, Neal J., Mason, Ward S. and McEachern, Alexander W. E lora- tions in Role Analysis (New York: John Wiley and Sons Inc. , 1953). Gullahorn, John I. and Jeanne s. "A.Model for Role Conflict Analysis” (Paper read at the annual meeting of the American Sociological As- sociation, 1959). Levy, Marion. The Structure of Society (Princeton, New Jersey: Princeton University Press, 1952). Miller, Delbert C. and Shull, Fremont A. "Role Conflict Behavior in Administration: A Study in the Validation of a Theory of Role Conflict Resolution" (Paper read at the annual meeting of the American Sociological Association, 1960). lineman, Lionel J. and Hughes, J.W. "The Problem of the Concept of Role - A Re-Survey of the Literature," Social Forces, Vol. XXX (19514-952). PP. 1141'1h90 Parsons, Talcott. The Social System (Glencoe, Illinois: The Free Passe, 1951). Seaman, Melvin. "Role Conflict and Ambivalence in Leadership," American Sociological Review, Vol. XVIII (1953), pp. 573-380. Stouffer, Samuel A. ”An Analysis of Conflicting Social Norms," American Sociological Review, vol. xIV'(l9h9), pp. 707a717. Sutcliffe, J.P. and Haberman, M. "Factors Influencing Choice in Role Conflict Situations,” American SocioIOgical Review, Vol. HI (1956). ppo 695~703- 58 .APPENDIX ONE To facilitate the presentation of the raw data given in this section, the following abbreviations will be employed: H - high L - legitimacy L - low SA - sanctions access PE - perceived expectation SE - sanctions exercise RPB - reported probable behavior' EP - ego's preference 0 - obligation TABLE 1 page 22 SITUATION I II III Iv v RPB-PB RPB-PE same different H0 125 72 188 67 15h 155 1h9 155 182 81 L0 155 170 9h 9h 29 87 125 85 102 97 TABLE 2 page 25 SITUATION I II III Iv v RPB-PE RPB-PE same different HL 167 126 21h 69 155 151 205 1&1 198 88 LL 112 117 68 92 5o 91 69 77 86 9o 99 60, TABLE 5 page 25 SITUATION II III Iv v RPB-PE RPB-PE sane different HSA 151 111 192 79 115 11.5 1711 1511 188 89 1811 127 153 93 79 118 99 99 83 96 89 TABLE 14 page 26 SITUATION II III Iv v RPB-m RPB-PE same different H33 137 1141 176 109 117 128 m1 151 166 117 1.88 98 1116 56 102 1.6 111. 121 97 77 102 TABIE 6 page 51 SITUATION II III Iv v nsa ISA HST. 205 50 217 15 222 21 257 27 229 111 183 57 250 51. 157 56 126 68 158 50 169 61 TABLE 7 page 31 SITUATION I II III Iv 7 HL IL 216 116 228 1.2 250 50 270 58 2112 55 8h 176 6h 109 50 95 7h 108 1.6 159 TABLE 8 page 32 SITUATION I II III Iv V BL LL Us: 197 58 198 511 215 28 255 27 216 27 LST 105 181: 95 116 66 96 109 119 72 11.7 TABLE 9 page 32 SITUATION I II III Iv v BSA ISA 155 115 212 M. 231 59 235 55 22h 39 108 216 59 128 26 89 65 155 51: 11.5 62 TABLE 10 page 33 SITUATION I II III Iv v HSE LSE H0 150 117 195 61 225 62 211 70 211 51 1.0 85 2110 57 150 18 100 52 157 51 169 TABLE ll page 33 W SITUATION I II III Iv v HO 1.0 H1. 168 152 252 60 214.8 52 21.7 97 251 57 LL 28 19!. 55 118 112 85 55 111 94 1110 63 TABLE 12 page 35 SITUATION mm mm LIAcHSE LdeLSE RPB-PE same 118 511 25 811 159 65 16 147 RPB-PE ' different 80 £15 18 100 58 55 18 69 SITUATION III Iv v 108 25 9 22 156 67 6 65 150 117 19 68 108 no 19 7h 99 £12 21 56 66 25 10 79 TABIE 111 page Ma. SITUATION I II III IV V RPB-EP RPB-DP some different 281 2141 522 85 177 196 5'16 156 555 1111 ‘A TABIE 15 page 119 SITUATION HLacPE- MAPS- HIAPE- EP dif- LLacPE- EP dif- EP same ferent EP same ferent RPB-PE ' same 1111: 23 87 2h 193 9 56 3 RPB-PE . different 106 25 66 117 33 2O 11 118 SITUATION III Iv v 152 o 25 L. 181 20 511 15 187 11 65 15 102 11 59 2O 92 £15 26 119 1:5 52 27 5o ‘ APPENDIX TWO Only a sample of the two-part role conflict questionnaire is included in this section. The remaining pages of the questionnaire are identical to those included herein with one exception; the cats- gories or audiences labeled at the top of each page change after every fifth potential role conflict situation (rules and regulations situation). The nine audiences and the five conflict situations result in a 115 page questionnaire. 65 I 4‘ .. $335300 on» a." cane ocean» scum cannot: >5 non one .2333 coop. . on Ill." .53558 on» 5 03325 new c.3288 on 93.33 2331.... .50330 Hence. s on .H finch may no you so» coco on .323 adamnoum no» solasiu flaw-lemma u. Mm. Jag 93.830va ou umOnQ—d moEoo mnwgzom 05 mo moo £0.33 .v I \I. I I I ‘JIIII F'- 'IIIKAIII'I C 'lll F, I81 IOIF. LII Ins! EOIJHLF “(I'LH'uI‘ dug Infill!“ :90: ”0.32 encuidudo canto-flee. II -An! . 1.. I'-‘l N ‘ ' I - 0... II- 1‘" -O'I- "“ “ 'AI: ..'43‘ '0‘. -1"! 7&3‘6 a. mm. .0033 madnduuuov on unouoao amp—00 mcgddom 2.3 m0 moo £0.23» .N Ho>oz .v adoumm Ii :.m mosauoeom .N :wuwo .H «ebonm covenanca amen o>m£ no» can“ Hmuocom on» on mcouumooxm ome no» Haa3.noumo.303 .An“ on» no mono .nuon co haamnwo asap as tcomm : .m .coaummaumo>ca HMGHEHMU one ucfimameoo sua3.hacflme mafia he comma III .N .xuo3 oammmuu one avenue sufi3.chHmE eBay 35 macaw ..H , when can no you no» 00:0 00 Haaz.aanmnoum no» .oanu flammwwm m.mm..us£3 moanauomoo ou unemoao meson moaSOHHOM one no coo #0253 .H .oE«u many an on HHfi3_mHAunoum peas :0» umg3 .nmo 50% neon no .osamufia on an» unouumenv omega cook 50» cos: .Homoouu e we «mom m on nocmamme one one Hoonom udouuou Amacfim no» :0:3.oedu on» ounlousunm on» once xooa no» page ouanvou mcoaumonv méfizodaom 0&9 1411‘] .MHQOONH AV Md. .50? 2.90”.“ BDOQ4 EOHBWHDO UZHZHAVzfl HEP "H802 0.30 >=eon «3.3 000 0c ~32: $33 $33 you on «on 083% >832 inseam 3330094.. Amuse ofi 3 9.632— oo> «ea—3 00 «none? 3300 mac on» #0038 «33> 38. o» 9.50.2000.» «0d 0.33003 35 335 >3euocem >u0m3mo n3» 5 mnemuem 3:2» 90> at 393.30 30$ .~ 5* 02¢ A; .HnHHUNm m0Om< 70v OHMUHIU 50% 93 H$802 .0 .m .< H28 283 5.20.303 309m >u0m3mU flamenco 5 mnemooufi 05 3 23¢qu >n 30A >Zeueoem 305 3323.3 3 $0.33 090 05 on 3 9,0309 50> 06 £033 .95“? 05 HO 4 £039.: 03» :0 £050 on» no >m3 0:0 .uomoouu on» no onuoc «009nm .U .>:msowmmouo mcoflmgmou use and; on» £00225 3 nomoob on» «009nm .m 3333 one 3 moofiegmou vow m3?" 05 30:8 0» hen—00.3 05 309nm .< .muemoonu «done >u0ueueo e3» 5 odonuom >9 30A Ndmuenonw $05 >3enoum 3 25:02 50> A033 33> 000 may .93 50> amen 0a .0513. 50> «.05 eufinveu uaoflneov 03» «0.3“ 23. 6.23003 «none 933 >05 >u0meumo anyone 05 GM 38932 3033 more? 0.0.2.? 0.3 330m Aémzmo ZH mmmmoomh I, . i .3 . _ .“ P L .r ‘Jll §finugnlnrhiflxvvmier < ul.<. Ill‘lwrlffit‘fl \fln..!ill -I.-I! vlu /\ .«hltnil‘ nut-l..ln. uP.a.rtm.I-v|«i 4.3Iuiulwytllwu!l\l.,“.’ul. . - ...n .0. Ohio >dH‘-“ 9-1-nfl “Co: a“! 8-1-:1 U‘“':¥|llly “um-J'FFA. 0°: 0 ll Hana. “ON“.HHHIIII kflddvfldz \rdmfuohldm LPHOUHddoAaOr/V II -0 .l. ‘ ‘ 55:5 mama. Hmm .MHQOOMH d. 95 50> EOE BDOQ< H554 mZOHBwHDO OZHZH<2mm HEB "H902 05.00 505055 35005 505 05 505055 505055 505055 505 on 555 505055 >882 393.3% 3055500555 50000 055 05 0550305 50> 5.053 05 5000050 30505 050 055 050058 55805.... 0555 05 955550000 500 05050055 5055 505055 >205050m >50m0500 0555 55 0500505 055555 50> 05 >5m50550 303 .N .0 .m .< ”050 050050 «05050055 55050 >5050500 5.050500 55 05050055. 055 55 0500505 >5 305 >5505050m 5005 3505055 05 50553 050 055 05 05 05.0205 50> 05 50553 .2550 055 50 .5 .5505 50 05555 50550 55050 05 5050055 055 5005025 .U 505503500255 505555550 55.0 5550555500 5553 3.55.055 05555 055 55050 05 5050055 05 5005025 .m .5503 0555055 550 30500 5553 355055 05555 03 05050 05 5050055 055 5005025 .5. .05050055 55050 >50m0500 055... 55 0500505 >5 305 >205050m 50055 3505055 05 0.05505 50> 50553 30? 050 055 9500 50> 5005 00 .0903. 50> 5055 05555005 0505500550 035 505mm 055. .05050055 55050 00.05 >055 >50m0500 0>050 055 55 0500505 50553 03.05, 00555 050 3050.5 AEHZHO ZH wmmmOOMH 2' n...|’ I I In a JJPI.‘ idil.‘ manHvHiEv—FDG anvZuusz( 7.... -u _ ntflltv+vrcam ; 1“: 5 Uphill HMHUNJ — -_\ \ t / / 00055 059 r 05050059 50 055555 0050555 5050500 055 05005550 50500505 0505500 50555055 050 50550550005 mammoomfi NEH 50 BUW5NW Mudflmmzmw 305mm mmHMOUmH¢U mmfi 2H mzommmm Bdflfi m>mHQmm DOV OD B¢m3 "050550050 035 5553 >55505 05000 0550550550050 055305505 059 .50550055 >50 5550.055 00 000055 .50550055 >50>0 05 mmuumulmmm;xdmu 50050 05 0550 05 000055 .0000 055 05 05 0>05505 50> 5053 05 5000050 00500 50553 050 055 3055 500 50> 555 ~00005 530 550> 555 >5500X0 5553.0503050 055 50 0505 ~008550500 .503050 0555 50 55055 55 m.0 00055 .505505 055 55050 00555005 050 505505 530 550>_05 5000050 00500 50553 503050 055 550 5055 05 00500 050 505 .0503050 =m5053= 50 :55055: 05 050 05058 .5050 00 5005 0 505 05 0550550550050 055305505 055 umZOHBUmMHQ "fi— ‘1 mmflHmOUWBANU "5.0.35.5. 2H mzommmm a .50 E55 HEB .5502 Hawk DON On— 303 .mmmooma 4 m4 Ozgmmm mmmmmmw 055. \ 05050059 . 50 00555505 50 ; 005.53 00.5...”05 mugs 5555500 in ’ 0:0 5000.5 5050500 W 55.0505m0595 [’11 «4' ‘ v\. *1! .IT I ‘ I u.“ n “In“ A a ‘ .II J“ in! q.- ire-H. fill \ /’ nWIHGflHOO.HpH. WO|/ ‘— & 3-”) "l... ‘iIQII‘. I... l'llt -Il"a|n' "l... ‘IC‘IIa I‘li‘. lull 0-.-.-. .‘II.’ 0..--.qzlll 0 ‘II' -II|'O|III...|5 I‘I N .. a. undo .330." 3?: mac.» «0 93.30 93.30 «on on 3.5 3.3mm 330 33> ~. 3 .5 .3323 003 .30.» 3 030035 253 .09» £033 33> one 0» 30A 300.? romuamu 02» 5 0000.3.“ 005 9.503 .30.» 0M M0030 30a £09093 0 0.0 magmmmm .D .m .4 “000 £026 «muomoouu 0:090 23: 3 .Comoumu 030 5 0000.33 2.3 0x2 «005 3903 :0.» #033 2.0 on» 3 omdm 03» m0 no» on» ad 6300092— 0303 00.23 05 H0 £023 #2393 a 0.0 933025 000: .33: 38:05 no: 3 :85 025233 oz 350% 300:".— o< 005.01 >uo>m «33> .305 0» 950.3000 «00 0» 0.3.30.3 now 3 33 3 90300500 00 #20300 3.903 occuuom A030 «.05 00043332 05 3 «0:3 £0098 33> 03» 30A 053 .Cououuo out 3 uncouoa >03 05 000 2:30 0.3.30.3 3 $230.3 0 00 $53on 003309 033009 00330“ aoflmmom .3332“ .309 .33 000» «0098.30 00 a m 5 0 an d 5 05 3 «00098 0000.39 0004» 00 00 30900.3 33E 0» >3 00 93000.2" 39mm 3 333000 0.30 5 0.330% on» 0.3 003309 0043 3 #2393 d an muflmomm in: in: in: Ema 3% But 00 0.03 33$ :3 0500 cam £03230 003 3193 .33 «33> 030 303 0» «an romouuo 34.3 #030 :0.» av 3m? 00:3 .uomoonu a no uqflnonm .n I91 1" Jr 1" C 4‘. cmuuo -- (G 30:08 “0.902 II. 0 mamumm .m mmEHUOEOm . N 0300 III. a mm>onm vmumodcca push m>mn so» was“ Hmuwcmm 050 cu maoaummoxm mwa 00% HHH3.cmumo.3O£ .30“ may no mono .Hmuume mfisu co moduomum kumcmm o: m>mm .m .mmzasu Sosa uzonm amasv mmmx .N .mmausu moqaom ma: m:aE90mnmm maummoum 000 mp3 usmzonu H Eog3 Homoouu Hosuocm vacuum :i.-.H finch any so 90¢ 90» mono ow Had3 mflnmnoum 50% ~0Hsu amumnum Im.mm.~umg3 maflnfiuommo cu ummmoao mmeoo mfldzoaaom on» no mno sauna .m H0302 .0 Mawumm IIIII. m mmfiaumaom .IIIII.~ 0300 III. a wm>onm cmumuavca yuan m>mn so» was“ Hmumcmm 0:» on macaumooxm 0x05 50» Haaz.cmumo.3on .non on» no mono .Hmuume many so muauomum Hmnmcmm on m>mm .m .5030 000 nmg3 now he o>moq .N Ii .11 J .3 \ J"! 13.11 1 . .. ‘ .v... n .. o u-Iuv.l. ..... 2! v.'.- . . ...... . . o. i .o .. . .I . . ... . ..'!-I‘I‘...-. J . z . f. 2‘ . . . ll. . A . A . .. n..\.. I...‘..OIII‘I.I I-‘h.~ln-.I-.~u . u - .t . .y. .~ - .Icl . . p. x ’ .u \ lav . ,n . nus-nuluulnl 0“: v.l‘. - . .. .0 133. .. V ~ ~ - .... A. .0 :- . . . . .. o... . v .u v to n. .. Q-I.a .. ‘ ,..u-u. . » l Io. .0 n u u \ v. ..a.‘!..‘.‘ ..|,.u .. .1 a . .l . . I.x.. . ....Il-n.n .. ..- . v 0 ul v-y no.v.v . . . . ... o n . $5”... ...u. 1.1-... It}! ~ .n I1... ..-. . -.,. .u. a. u . In-Il .vl ‘23:... ...Iclo. a o. . :vnr . Ft l . .nll..-v.|:.p-vuqn.lu. ..-.vu«v .1 u..... .o 2.. l.....- v I 0.....uvnu..v u. a. . . . .v . - ult- . n ! ...I- n: a n a... . . . .o .‘ . .. ‘. .. . . o n . . . . . . I u . .. . . u A . .. I. . 2 .. ... - . . ... .2. ...u . . u.» .. u . . .J... . .. u .. a... v ..o.:...~.. .. .3. o-In.v . ,u I.. ~|II.I I.- . ..I . V . .‘ . .a Into 0 unnlnu .oo no a . .. . I. . . O. .u.- .~. I. U. u . a . .o. I .. I... . u . 0 :I .l I "ITITm“W\IflfijflfifliflfiflwflfgfifflES 312