27
The Effect of Variations in Relatedness Need Satisfaction on Relatedness Desires Author(s): Clayton P. Alderfer, Robert E. Kaplan and Ken K. Smith Source: Administrative Science Quarterly, Vol. 19, No. 4 (Dec., 1974), pp. 507-532 Published by: Sage Publications, Inc. on behalf of the Johnson Graduate School of Management, Cornell University Stable URL: http://www.jstor.org/stable/2391808 . Accessed: 18/06/2014 09:09 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. . Sage Publications, Inc. and Johnson Graduate School of Management, Cornell University are collaborating with JSTOR to digitize, preserve and extend access to Administrative Science Quarterly. http://www.jstor.org This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AM All use subject to JSTOR Terms and Conditions

The Effect of Variations in Relatedness Need Satisfaction on Relatedness Desires

Embed Size (px)

Citation preview

The Effect of Variations in Relatedness Need Satisfaction on Relatedness DesiresAuthor(s): Clayton P. Alderfer, Robert E. Kaplan and Ken K. SmithSource: Administrative Science Quarterly, Vol. 19, No. 4 (Dec., 1974), pp. 507-532Published by: Sage Publications, Inc. on behalf of the Johnson Graduate School of Management,Cornell UniversityStable URL: http://www.jstor.org/stable/2391808 .

Accessed: 18/06/2014 09:09

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

Sage Publications, Inc. and Johnson Graduate School of Management, Cornell University are collaboratingwith JSTOR to digitize, preserve and extend access to Administrative Science Quarterly.

http://www.jstor.org

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

The Effect of Varia- tions in Relatedness Need Satisfaction on Relatedness Desires

Clayton P. Alderfer, Robert E. Kaplan, and Ken K. Smith

The authors wish to thank John Holbrook, Martin Greller, Thomas Barbieri, Jack Hautolauma, Richard Hackman, and Chris Argyris for their help as experimenters and/or readers of an earlier draft of this article. Although naming the men who served as subjects in the study and the organization where the study took place would violate a promise of confidentiality, they would also like to thank the former for their participation and the latter for its financial aid. Primary support for the study came from the Office of Naval Research (Contract Number N001 4-67-A-0097- 001 7). whose help is gratefully acknowledged.

Hypotheses pertaining to relatedness satisfactions and desires were derived from ERG theory and tested in an experimental social psychology laboratory. Twenty-one adult male middle and upper middle managers participated in role playing activi- ties based upon actual work experiences described by the men in interviews beforehand. Three degrees of mutuality and relatedness satisfaction were created in the role playing ex- periment. The managers' desires for relatedness varied as a function of the experimental conditions and supported the hypothesized curvilinear relationship between relatedness satisfaction and desires. These results proved to be consistent with previous results obtained by field correlational methods and thus provide a link between the two methodologies for the study of human needs in organizational settings. The study also presents a new behavior coding system for scoring mutuality in interpersonal behavior and demonstrates that high degrees of relatedness satisfaction can be produced in a laboratory setting.

The quality of the relationship between an individual member and an organization to which he belongs depends, in part, on the degree that the person's needs are satisfied by partici- pation in organizational activities. Some human need theories (Maslow, 1 954; Alderfer, 1 972) further propose that need satisfaction affects human desires. A person's desires may determine the type of satisfactions he seeks from organiza- tional life. Taken together, these propositions imply a dy- namic cycle of changing individual satisfactions and desires as a function of person-organization interaction (Alderfer, 1 972).

This article has two major purposes: (1 ) to present the results of a laboratory experiment designed to test derivations from existence, relatedness, and growth (ERG) theory which, prior to this investigation, had been studied only by field correla- tional methods and (2) to introduce a new approach to laboratory methodology which employs adult subjects who participate in role playing activities closely modeled after their real life organizational experiences, without employing the usual laboratory deceptions.

ERG theory is a conceptual and empirical system for under- standing, explaining, and predicting the satisfaction and desire properties of human needs (Alderfer, 1972). Each of the basic needs in ERG theory is defined in terms of a target toward which efforts at gratification are aimed and in terms of a process through which, and only through which, satis- faction can be obtained. For relatedness needs, the primary topic of this study, the targets are significant others who may include both individuals and groups. The behavioral process necessary for satisfying relatedness needs is called mutuality.

Field correlational studies of relatedness satisfaction and desire have produced results consistent with postulating a curvilinear function between the two variables. When related- ness satisfaction was quite low, desires were high and, as satisfaction increased, desires decreased. In certain settings, however-a manufacturing plant using behavioral science methods in management, a college fraternity with relationship- oriented leadership, and a human relations training labora- tory-the desire curve turned upward again for the highest levels of satisfaction (Alderfer, 1972). To reproduce these findings under the more controlled conditions of a laboratory experiment, it would be necessary to create not only low and

507/ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

moderate conditions of relatedness satisfaction, but also unusually high degrees of relatedness satisfaction.

4.5.

(4.1 7)

4.0

a 3.5 - 0 U,

a)

a 30 by 3.0 - \ (2.70)

2.5

(2.31)

2.0

(-8.47) (4.37) (10.46) Low Moderate High

Relatedness Satisfaction

Figure. Relatedness desires as a function of relatedness satisfaction

In recent years, there has been increasing criticism of certain characteristics of laboratory experiments in social psychology. Much of the knowledge generated by laboratory social psychology is based on studies of college freshmen and sophomores (Festinger and Katz, 1 954). Investigators have become increasingly aware of the relatively low social and situational power of many subjects (Argyris, 1 970; Kelman, 1968, 1972). The processes studied in many laboratory settings have also been criticized for their mechanistic tenden- cies to reproduce only those features which show an indi- vidual's defensive and simple qualities rather than his expan- sive and complex characteristics (Maslow, 1 966; Argyris, 1969). In short, much of laboratory social psychology has been built upon an exploitative relationship between experi- menter and subject and, not accidentally, has investigated primarily those variables which can be observed under such circumstances. To create genuinely high levels of relatedness satisfaction in a social psychology laboratory setting, it is necessary to alter substantially many of the usual methods for developing a psychological contract between experimenter and subject.

MUTUALITY AND RELATEDNESS NEEDS

The degree of mutuality in a relationship is defined as the extent to which all relevant ideas and feelings are given and received by the parties in the relationship. The highest level of mutuality occurs when both agreements and disagreements about ideas are understood by the parties forming the relation- ship and when both positive feelings, such as liking, support, and trust, and negative feelings, such as fear, anxiety, and mistrust, are expressed and accepted by the parties who form

508/ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

Relatedness Need Satisfaction

the relationship. Less than full mutuality occurs to the degree that relevant ideas and feelings are withheld or ignored in an exchange, to the extent that positive or negative reactions are systematically excluded from discussion, and to the degree that the exchange is primarily one in which one party largely does the giving, while the other does the receiving. As here defined, mutuality is a behavioral process capable of being observed by someone, such as a researcher, who is not directly participating in the exchange, as well as being per- ceived by the parties in the on-going relationship. Previous research on relatedness needs relied completely on question- naire measures, but the present study also includes the development of a behavioral coding system for measuring the degree of mutuality observed in an exchange between two people. Relatedness satisfaction is operationalized by the questionnaire responses of the person whose needs are being studied. The behavioral coding system allows an investigator to correlate self reports of relatedness satisfaction with inde- pendent observations of those mutuality behaviors deemed necessary for need satisfaction according to the theoretical construct. The behavioral process identified as mutuality draws signifi- cantly from formulations proposed by Rogers (1 959) and Argyris (1 962). Assuming a minimal level of mutual willingness to be in contact and to receive communications, we may say that the greater the communicated congruence of experience, awareness, and behavior on the part of one individual, the more the ensuing relationship will involve a tendency toward reciprocal communication with the same qualities, mutually accurate understanding of the communications, improved psychological adjustment and functioning in both parties, and mutual satisfaction in the relationship. Conversely, the greater the communicated incongruence of experience, awareness, and behavior, the more the ensuing relationship will involve further communication with the same quality, disintegration of accurate understanding, lessened psychological adjustment in both parties, and mutual dissatisfaction in the relationship (Rogers, 1 959: 240).

Authentic relationships are, therefore, those relationships in which an individual enhances his sense of self- and other awareness and ac- ceptance in such a way that others can do the same.

Human authenticity is an interpersonal phenomenon. An individual cannot be authentic independent of the relationships he has with others. His feelings of authenticity will depend as much upon the capacity of others to create authentic relationships as . . . upon his own (Argyris, 1962: 21).

Building upon the Rogers' and Argyris' concepts,-the mutu- ality formulation specifies crucial characteristics of the com- munication process among parties who have significant relations with each other. The idea of mutuality includes more than just the process of communication; it also is concerned with the content communicated in human relationships, a feature which Rogers and Argyris did not explicitly address. In this sense, the concept of mutuality draws upon the theory of object relations as developed by such investigators as Klein (1 955) and Fairbairn (1 952). The idea of mutuality explicitly assumes that ambivalence-that is, both positive and negative feelings-is a significant part of all human relationships and suggests how this property might be optimally dealt with by the parties in their relationship.

Relatedness satisfaction and desires pertain not only to the process and content of communication, but also to the ob- jects, namely, individuals and groups, among whom exchange

509/ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

is taking place. Derived in part from Sigmund Freud's (1 963: 106) notion of transference and Anna Freud's (1 946: 46, 66, and 1 28) delineation of the process of projection, the concept of relatedness needs addresses the question of how the objects of satisfactions and desires may vary among significant others. Previous field correlational research on relatedness needs has generally found that desires pertaining to a person's relationships with one significant other tend to transfer to additional human objects (Alderfer, 1972: 113-131). The present research examines how desires about human objects vary as a function of the mutuality of the exchange and how desires for the self may be projected onto others in an interpersonal exchange.

The concepts of mutuality and relatedness needs differ sub- stantially from motives for social approval (Eisenberger, 1 970), need for affiliation (Murray, 1 938), and need for interpersonal safety, belongingness, and esteem (Maslow, 1 954). Each of these addresses a person's wish to be highly regarded by one's fellows, which is certainly an element in the present constructs, but none of these other motives deal directly with the presence of negative opinion and affect, except implicitly as it is detrimental to the more desirable positive state.

Little attention is paid by theorists of positive feelings to the behavioral processes by which communication takes place, while this is the central feature in the relatedness-mutuality framework. Although mutuality as a behavioral process draws substantially on Rogers' (1 959) work on communication processes, its basic assumption of ambivalence in human emo- tions denies the utility of Rogers' concept of unconditional positive regard. Significant human relationships, including those between a therapist and a client, do not exist without mixed feelings among the parties. Neither the possibility nor the desirability of having the prevalence of positive feelings outweigh negative feelings is denied here. The present study will examine the presence of both positive and negative feelings as they are observed to occur in interpersonal ex- change and as they are perceived by the parties in the exchange.

A number of predictions derived from the portion of ERG theory dealing with relatedness needs will thus be tested: (1) there is a tendency for relatedness desires to vary curvi- linearly with relatedness satisfaction and mutuality, such that the higher desires will occur at the lower and higher degrees of satisfaction and mutuality and (2) desire functions of mutuality apply not only to the self but, through projection in this study, to other persons as well. Methodologically, this study was conducted in a laboratory setting permitting causal inference in order to develop a valid behavioral coding system to measure mutuality and to develop a high degree of mutuality between the investigators and the subjects so that the often exploitive atmosphere of the social psychology laboratory would be minimized.

METHOD Relating to Subjects Selection of subjects. Considerable effort was devoted to allowing opportunities for the participants in this study to do so on the basis of free and informed choice and to creating experimental conditions which would be both real for them in terms of their work experiences and rich with opportunities for their own learning. The research program began with a

51 O/ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

Relatedness Need Satisfaction

series of conversations between the investigators and the plant management of a New England manufacturing plant of a national chemical corporation. The researchers explained that their goals were to develop an experimental seminar in which the participants could learn how they dealt with difficult managerial interpersonal problems and the researchers could collect research data.

Because an activity of this kind had not been conducted before, two steps were outlined to provide for termination in advance of the actual experimental sessions if the proposal proved unworkable. The managers were told that the project would not continue if it became apparent that its objectives were not being attained. This also ensured that participation in the study was voluntary.

The first step after discussions with plant management was to have a meeting of potential participants to explain the idea of the experimental seminar and answer questions. Twenty- seven managers attended this session, after they had received a memorandum from the plant manager with a letter from one of the investigators attached. The plant manager intro- duced the first two authors, who ran the meeting. Shortly before the session ended, by prior agreement, the plant manager left the room, explaining that he wanted to be sure that the men had an opportunity to raise questions with the researchers. The meeting continued for approximately ten minutes after the plant manager left; the managers did use the opportunity to raise some issues. During this time, for example, the researchers' assertions that the sessions would not be used to evaluate the men's competence as managers were questioned. At the close of the meeting, the researchers circulated a paper on which those who wished to participate in the next phase could sign. Twenty-eight names appeared on the list. One man had apparently left word for a friend to decide for him.

The second exploratory step consisted of one-hour semi- structured interviews with all of the men who signed up for this phase. These interviews fulfilled four functions. (1 ) Since role playing episodes were to be used to create varying degrees of relatedness satisfaction, data concerning real work experiences of the managers were required. In the interviews, managers were asked to describe some human relations problems that they had to handle. (2) The interviews also provided an opportunity to get to know the men on a face- to-face basis, to begin to establish trust between them and the investigators, and to ask them how they decided to participate in the seminar. (3) Recognizing that there are considerable individual differences in the ability to play roles, the interview was used to give each participant a brief role playing experience. After the man described a particular ex- perience he had had at work, he was asked to actually be- have as he did, while the investigator took on the role of the other person. A few of the men turned out to be unable to play the role; they repeated their previous descriptions of what had happened rather than actually performing the action. (4) After the role playing preview, the interviewer asked each manager whether he wanted to participate and those unable to enter their own roles said no.

The understanding reached with plant management pro- vided that, if the interviews revealed that the project was infeasible, feedback of the data gathered in the interviews would be provided to all who had participated in a group session. This information was to be treated confidentially and

511 /ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

would focus on some themes cutting across all interviews. As it turned out, 23 men were able to role play and continued to want to be involved. When the time came to actually conduct the study, 2 of the 23 had left the organization, so a total of 21 managers participated in the experimental sessions.

Design

Each subject played roles in low (L), moderate (M), and high (H) mutuality conditions to form a repeated measures design. There are six different ways that the three conditions can be ordered to give a completely counterbalanced format. The distribution of subjects to each of the orders was as follows: MLH MHL LMH LHM HLM HML

4 3 3 4 3 4

Experimental Procedures

While the interviews took place at the plant, the role playing sessions occurred at the university. When a manager arrived, he was greeted by the investigator who had interviewed him. The two men proceeded to a laboratory room which the interviewer explained would serve as the manager's office during the role playing sessions. In the room, the investigator reviewed the purpose of the study and described the se- quence of events that would take place during the manager's time in his office. Before each of three role playing episodes, the manager would receive a written account of his situation in relation to a subordinate who would be played by a graduate student. The manager was given about five minutes alone to read the description of his role, after which the investigator returned to answer any questions he might have. When questions were answered, the investigator left the room, telling the manager the name of the subordinate who would soon be knocking on the door. He also explained to the manager that the investigator would be observing the role playing through a one-way glass in order to be able to talk with the manager about how he handled the various situations after all three had taken place. There was a tape recorder on the manager's desk to record the interaction and be available for listening by the manager and investigator after each episode. The investigator also reappeared after each session to give the manager a questionnaire designed to assess relatedness satisfaction and desires. When he had completed the questionnaire, the manager was given the role description for the subsequent episode. Each manager faced three different subordinate role players and each experimenter played each role approximately the same number of times. The distribution of role playing subordinates by episodes was

Low Moderate High Player A 6 5 6 Player B 6 7 6 Player C 4 4 4 Player D 5 5 5

At the completion of the three episodes, the investigator returned with a final brief questionnaire asking the manager to reflect on all three episodes in addition to the last regular questionnaire.

After the questionnaire had been completed, the investigator discussed with the manager what had happened during the interactions. In the role of consultant, the investigator took primarily a nondirective stance, inviting the manager to identify those dimensions about which he wanted to learn. Sometimes managers preferred to question the consultant

51 2/ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

Relatedness Need Satisfaction

who would describe common patterns in the managers' behavior which seemed to recur across role playing episodes. If the manager did not initiate it himself, the consultant suggested that they might want to listen to portions of the taped interactions. Most-though not all-managers wanted to do this and the role play of most interest was the low mutuality condition. It was usually possible for a manager to see how he significantly misheard what the subordinate was saying in the episode. Usually, this consultation period lasted about an hour and proved to be a satisfying experience for the manager. Before the manager left, the investigator asked him to be sure not to tell other managers what happened during the role playing so that each man might come to the sessions without knowing the role playing details.

About seven weeks after the last role playing session, a meet- ing at the company was held for all of the managers who participated in the study. At this time, the logic and theory behind the study were explained and the four role players were introduced to the managers. This lively session lasted about three hours and resulted in a number of managers expressing a variety of disbeliefs about the study. Some wanted to check whether the role players had actually spread themselves among the different roles and others questioned the results of the study. This feedback session also provided an opportunity to explore some unexpected results from the study.

Role Descriptions

The primary method for creating the interpersonal episodes was through writing role descriptions from both the manager's and subordinate's perspectives. Prior to each episode, both the manager and his experimenter-subordinate read his own version of the role which presented a conglomeration of events portrayed by the managers in their individual interviews. The managers were told that their subordinates had slightly different versions of the events than they did. The experi- menters carefully studied the managers' roles as well as their own in order to prepare for creating the desired conditions.

While controlling for role player and for the order of roles, the design of this experiment confounds content of role with degree of relatedness satisfaction. Technically, this means that the results obtained may be valid only for the specific roles utilized. The decision to proceed with this confounding was made strictly for economic and logistical reasons. It took a considerable amount of time for the role players to learn the three roles they performed and there were only 21 managers available from the plant. The interviews produced more than 20 possible situations that might have been used to create different role playing episodes. The three episodes that were used were the first three that were attempted. Though not demonstrated empirically, the three roles used could have been pushed toward any of the three conditions with appropriate practice and an indefinitely large number of role playing episodes could be created using the procedures developed in this study. A detailed description of the role playing procedures is contained in the appendix to this article.

Behavior Measures

The conceptual definition of mutuality specifies that the key components of this complex variable are receiving ideas, receiving feelings, giving ideas, and giving feelings. A useful operational definition of the concept should include em-

51 3/ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

pirical indicators of these four components if the measure is to be maximally close to the theoretical construct. The pro- cedure developed coded each unit of behavior for degrees of the four components and, in addition, rated the quality of affect being expressed, if any. Thus, each unit of behavior received at least four scores and a fifth value was added when the coder could identify feelings being expressed. It takes empathic individuals to utilize this system reliably, because the operational criterion of relevant ideas and feelings is whether they can be observed in the exchange. A detailed account of the behavior coding system is also contained in the appendix.

Questionnaire Measures At the close of each role playing episode, the investigator entered the experimental room after the subordinate had left and asked the manager to complete a brief questionnaire. This instrument consisted of three pages which, in turn, dealt with the manager's reactions to (1 ) entering the role, (2) expressing his desires about the relationship just ended, and (3) stating his perceptions of the relatedness satisfaction he had obtained from the episode. Each of the pages had a different objective in the study and the items on each page were presented in a different format.

Entering the role. On entering the role, the manager was asked to check, on vertically arranged 4-point scales, his perception of (1 ) the realism of the subordinate's behavior, (2) his own naturalness in entering his own role, (3) his degree of involvement in his own role, and (4) his degree of satisfaction with the role playing experience. These items were designed to be a general manipulation check and an opportunity for the manager to drain off some of his strongest feelings about the episode before he was asked some more differentiated questions on the following pages. Relationship desires. The second page consisted of 1 3 items describing characteristics of a relationship. Five of these items were included as fillers, while the rest made up two a priori scales intended to measure relatedness desires for two human objects, self and other. The filler items pertained to friendship, influence, and mutual cooperation, while the a priori desire scale items dealt with trust, leveling, psychological distance, and respect. The items were arranged so that self and other items appeared alternately and the managers were asked to say how much more or less of.each quality they wanted by using a 7-point scale going from -3-very much less- through 0-no more, no less-to +3-very much more- applied to each of the 1 3 items. Summed scales were formed by adding the values given to trust, respect, and leveling and then subtracting the value for psychological distance. Three such scales were then utilized as dependent measures in the study: one with desire for self as the stem, one with desire for other as the stem, and one adding the self and other scales. Relatedness need satisfaction. The third page of the ques- tionnaire contained 10 items in 6-point agree-disagree format to measure the manager's perception of how much his re- latedness needs were being satisfied. Items in this set were very similar and, in some cases, identical to those that had been used in field correlational studies and were closely tied to the conceptual definition of relatedness satisfaction. Four items, counterbalanced in terms of whether agreement or disagreement indicated satisfaction, formed an a priori scale called subordinate receives me. The purpose of this scale was

51 4/ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

Relatedness Need Satisfaction

to ascertain the degree to which the manager perceived that he was being received as he wished by his subordinate role player. Two items, counterbalanced in terms of whether agreement or disagreement indicated satisfaction, formed an a priori scale called give of self, which was to ascertain how free the manager felt to give his own point of view in the discussion. A third subscale consisting of 2 counterbalanced items measured how much the manager thought he received the subordinate. In addition to these three scales relating to the giving and receiving process as it would be perceived by the manager, there were also 2 items to allow the manager to re- port his positive and negative feelings as a result of the episode. Person desires. At the close of the three role playing episodes, each manager was asked to express his desires about the three subordinates he faced by means of a hypothetical hiring situation. He was asked which of the three men he would hire if he could only choose one and which one he would not hire if he could choose two.

RESULTS Reliability and Validity of Measures Reliability of behavior measures. To estimate the reliability of the behavior measures, sample interactions were taken from each of the three experimental conditions. A total of 1 38 units were coded independently by the first and third authors; 41 units were from low conditions, 46 from moderate condi- tions, and 51 from high conditions. The sample of units in- cluded all of the role players and four different managers, one of whom was specifically selected because of the variability in his behavior and the other three who were chosen randomly. Usually, the first 10 units, plus or minus, of a given episode were coded. The percentage agreement on the definition of units was 98. The results shown in Table 1 indicate that the range of Spearman-Brown estimates was from .80 to .89. Once this reliability study had been completed, the same two coders scored all the behavior in all the role playing episodes. Since one of the coders was a role player, he did not score his own episodes; otherwise, the various conditions, managers, and role players were evenly divided between the two coders.

Reliability of questionnaire measures. Internal consistency estimates of the relatedness satisfaction scales were also computed and ranged in value from .94 to .29. Managers seemed to be the most internally consistent in how they perceived their subordinates receiving them and the least internally consistent in how they themselves received their subordinates. Of the three types of measures-behavior, satisfaction, and desire-the desire scales had the lowest Spearman-Brown reliabilities. These coefficients were also computed on an internal consistency basis and were lower in value than all of the satisfaction scales except receive of subordinate. One place where the satisfaction and desire reliability patterns are similar is in the comparison between the self-oriented and other-oriented subscales. In both in- stances, the managers were more internally consistent in reporting perceptions of what was happening to them, in terms of satisfaction and desire, than they were in reporting what was happening to their subordinates.

The reliabilities found in the present study are quite similar to the results obtained in field correlational and quasi-experi- mental studies where comparable measures were employed (Alderfer and Lodahl, 1 971; Alderfer, 1 967; Alderfer, 1 972). The reliability of the mutuality behavior subscales and total

51 5/ASO

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

scale tends to be equal to or greater than the precursors of this scale, namely the openness and flight scales used by Alderfer and Lodahl (1 971). The satisfaction and desire scales reliabilities are comparable to or slightly higher than the results obtained in field settings where no direct experimental manipulations were undertaken. Also similar to the field results is the finding that the desire reliabilities tend to run below the satisfaction estimates (Alderfer, 1967, 1972).

Table 1

Spearman-Brown Reliability Estimates of Behavior and Perception Measures

Mutuality behavior Relatedness satisfaction Relatedness desires n=138 n=63 n=63

Receiving ideas .82 Subordinate receive me .94 Self .66 Receiving feelings .89 Give of self .81 Other .44 Giving ideas .89 Receive of subordinate .29 Giving feelings .85 Total relatedness satisfaction .68 Total mutuality behavior .87 Feelings .80

Also of interest is the interrelationship among the subscale of the mutuality behavior and relatedness satisfaction. The re- sults shown in Table 2 indicate that for the behavior measures, there was a tendency for both ideas scales to be highly cor- related with each other and for both feelings scales to be highly related to each other. Generally, however, the inter- correlation of all the subscales of mutuality behavior tended to be high, thereby justifying combining them into a single summated scale to assess the complex behavioral variable. The subscales of the relatedness satisfaction measures also tend to interrelate with each other, except for the give of self and the receive of subordinate scales. Since these two scales have the lowest reliabilities of the three measures, one assumes that this low subscale correlation is best explained by the low reliabilities.

Table 2

Subscale Intercorrelations for Behavior and Satisfaction Measurest

Behavior Satisfaction (n=21, 3 replications) (n=21, 3 replications)

RF GI GF GoS RS Receive ideas (RI) .60O0 .8400 .430 Subordinate receive me .5800 .52** Receive feelings (RF) .57000 .7200 Give of self (GoS) .1 8* Give ideas (GI) .54000 Receive subordinate (RS) Give feelings (GF)

p<.05.

p<.01.

p<.001.

t The design of this experiment potentially provided 63 (21 managers x 3 conditions) data points where both behavior and satisfaction measures were available, but failure of the tape recorder twice reduced this to 61. Because of the repeated measures on the managers, the 61 data points were not strictly independent. But it was possible to form 3 sets of 21-or 20-independent observations and then compute correlations and sig- nificance levels by combining the results of the 3 independent calculations. Stratifying for role player, 7 episodes were selected randomly from the low condition, 7 from the moderate condition, and 7 from the high condition. Thus, each set of 21-20-points contains data from 21-20-different individual managers. The correlations reported in Tables 2 and 3 are based on the average of 3 such coefficients; computed by Z transformation, and the significance levels are the results of combining the p values of the separate correlations (Winer, 1 962: 43).

51 6/ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

Relatedness Need Satisfaction

The fact that there were subscales of mutuality and of re- latedness satisfaction provides an opportunity to observe not only the validity of the overall association between the be- havior and questionnaire measures, but also the linkages between the various subscales of both behavior and self reports. The results in Table 3 indicate where convergence was expected between subscale measures and that in nearly every instance, significant results were obtained. The receive feelings subscale was not significantly correlated with any perceptual measure. In addition to the predicted relationships, other patterns emerge from the findings. The behavior scale most highly and consistently related to the perceptual measures was the simple rating of the feelings being expressed. Positive and negative feelings were measured by the manager's per- ceptions and coded behaviorally. This behavior was the most highly correlated with each of the relatedness satisfaction subscales and with overall relatedness satisfaction. Another notable pattern is that in every case, ideas behavior, whether receiving or giving, correlates more highly with the perception measures than the feelings behavior, whether receiving or giving. Similarly, giving behavior, whether ideas or feeling, correlates more highly with the perceptual measures than receiving behavior, whether ideas or feelings. These two effects seem to combine and result in giving ideas being the behavior subscale most highly correlated with the perceptions measure of overall relatedness satisfaction.

Table 3

Validity Coefficients for Mutuality Behaviors and Relatedness Satisfaction

Relatedness satisfactiont (n=21, 3 replications)4

Total Positive Negative Subordinate Give of Receive relatedness

Behaviors feelings feelings receive self self subordinate satisfaction

Receive ideas .21 -.22 .32 .25 .340 .370 Receive feelings .02 -.05 .12 .21 .23 .18 Give ideas .360 -.350 .470 .340 .39g .510* Give feelings .21 -.1 5 .29 .23 .39 .370 Total mutuality .1 9 -.21 .34 .29 .42g .4100 Feelings .7100 -.570 .750 .50 .510Ag .740

p<.05.

p<.01.

p<.001.

t Subscale and total scale validity coefficients in italics.

See fn., Table 2.

Manipulation Checks Entry into the roles. Three measures were used to assess the way that the managers were able psychologically to enter the roles presented to them in the experiment. On two of these variables-the degree of realism perceived by the manager and his involvement in the role-the data showed no sig- nificant differences between the conditions and were uni-

51 7/ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

formly of high values. A 4-point scale was employed in these measures and the lowest value of any mean for a condition was 3.48 on these two measures. The third variable for checking role entry-perceived naturalness of the role-did differ among the conditions. The moderate and high condi- tions seemed significantly more natural to the managers than the low condition; there was a monotonically increasing function between mutuality level and naturalness of role entry.

Another question related to role entry pertains to whether there were differences among the role players in the degree to which they were able to create high involvement, realness, and naturalness for the managers. The data shown in Table 4 indicate that there were no differences among the role playing subordinates on any of the three role entry dimensions.

Table 4

Condition and Role Player Effects on Role Entry Variables

Condition Role Player

low moderate high df Fi P W2 A B C D df Ft p

Realism 3.71 3.67 3.62 2/40 .14 n.s. Realism 3.71 3.42 3.83 3.80 3/59 1.91 n.s.

Involve- Involve- ment 3.48 3.62 3.57 2/40 .38 n.s. ment 3.53 3.58 3.58 3.53 3/59 .03 n.s.

Natural- Natural- ness 2.95 3.43 3.71 2/40 9.48 .001 17 ness 3.29 3.47 3.33 3.33 3/59 .18 n.s.

One-way analysis of variance with repeated measures.

t One-way analysis of variance without repeated measures.

Behavioral differences among conditions. The mutuality sub- scales, the total mutuality measure, and the feelings scale provide the bases for determining the ways that the behavior of the manager and his role playing subordinate differed among the three experimental conditions. Results shown in Table 5 indicate that there were significant behavioral differ- ences among the conditions on all measures. Examination of the means of each condition on the different measures reveals that not every measure simply increased as the conditions changed from low to moderate to high.

Receiving ideas, for example, increased significantly from the low to the moderate condition, but it did not change sig- nificantly from the moderate to high condition. Receiving feelings did not change from the low to moderate condition, but it did significantly increase from the moderate to the high condition. Giving ideas increased monotonically from the low to the moderate condition and from the moderate to the high condition. Giving feelings decreased from the low to the moderate condition and then increased from the moderate to the high condition. The combined effect of these subscale differences was virtually no behavioral difference between the low and moderate conditions on total mutuality, although there was a substantial increase in total mutuality from the moderate to high conditions. From the low to moderate conditions, increases in giving and receiving ideas were compensated for by a decrease in the giving of feelings, while from the moderate to high conditions, there were in- creases in receiving feelings, giving ideas, and giving feelings. Expressed feelings became increasingly more positive across all three conditions.

51 8/ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

Relatedness Need Satisfaction

Table 5

Condition Effects on Behavior Measures

Behavior Low Moderate High df Fi p W2

Receiving ideas 1.70t 1.81 T 1.85 2/58 4.73 .05 .11

Receiving feelings? 1.39t 1.32t 1.53 2/58 3.80 .05 .08

Giving ideas 1.88t 2.03t 2.16 2/58 11.94 .001 .26

Giving feelings? 1.77t 1.63t 2.07 2/58 41.11 .001 .57

Total mutuality? 6.68t 6.73t 7.52 2/58 11.70 .001 .26

Feelings 1.92t 2.38t 2.93 2/58 58.95 .001 .65

Analysis of variance without repeated measures due to missing data.

t Indicates that adjacent means differ at p<.05 or less.

Indicates that adjacent means do not differ significantly.

? Departure from linearity significant, p<.OS or less.

Observing only the mean changes in feeling scores across the three conditions might lead one to the conclusion that the conditions differed in emotionality mainly in that the expressed feelings became more positive as condition level increased. The changes in emotional expression were more subtle than this, as examination of Table 6 shows. The ex- plicitness of all emotional expression-both negative and positive-increased as a function of condition. Twenty-two percent of the very negative feelings were stated explicitly in the low condition, 33 percent in the moderate condition, and 62 percent in the high condition. Eleven percent of the mixed feelings were stated explicitly in the low condition, 10 percent in the moderate condition, and 47 percent in the high condition. Six percent of the very positive feelings were stated explicitly in the low condition, 7 percent in the moderate condition, and 25 percent in the high condition.

Table 6

Interrelation among Condition, Quality of Emotion, and Explicitness of Emotional Expression"

Expressed emotions Give Very Very

Condition feelings negative Negative Mixed Positive positive Totals

Low Implicit 212 593 33 73 0 911 Explicit 62 40 4 5 1 112

Moderate Implicit 38 373 63 95 1 570 Explicit 19 26 7 7 1 60

High Implicit 30 147 48 200 7 432 Explicit 50 82 43 66 24 265 Totals 411 1,261 198 446 34

X2 (d.f.= 22) =1114.54, p<.001, for the 3-way interaction among condition, give feelings, and expressed emotions (see Lewis, 1 962).

A further question pertains to how much the behavioral differences among conditions were due to the role players' behavior in comparison to the managers'. Table 7 shows that the experimenters generally tended to be more explicit about their feelings than the subjects, regardless of condition, but the subjects tended to demonstrate the same relative pattern of explicitness across conditions as the experimenters. Par- ticularly, there was no difference in explicitness of feelings between the low and moderate conditions for either party,

51 9/ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

but there was a decided increase for both parties from the moderate to high condition. Moreover, the relative increase in explicitness of feelings from the moderate to high condi- tion was greater for the subjects-i 1 percent to 67 percent- than for the experimenters-i 5 percent to 56 percent. Thus, although the experimenters' behavior was leading the estab- lishment of the differences among conditions, the behavioral variance evolved from the actions of both parties, not just the experimenters'. Table 7

Interaction among Condition, Explicitness of Emotion, and Person-Role Player or Subject*

Give Role Condition feelings player Subject Totals

Low Implicit 536 375 911 Explicit 79 33 112

Moderate Implicit 320 250 570 Explicit 42 1 8 60

High Implicit 196 236 432 Explicit 1 57 108 265 Totals 1,330 1,020

X2 (d.f.=7)=184.28p<.001 forthe 3-way interaction among conditions, give feelings, and person (see Lewis, 1962).

Perceived differences among the conditions. The general satisfaction scale, the relatedness satisfaction subscales, overall relatedness satisfaction, positive feelings, and negative feelings all differed significantly across the three experimental conditions. Like the behavior measures, however, not all of the scales discriminate among all the conditions. General satisfaction did differ significantly between the low and moderate conditions and between the moderate and high conditions, but the perception of giving freely of one's own views differed significantly only between the low and moder- ate conditions. The managers' perceptions of receiving the subordinate's message differed significantly between low and moderate conditions and between moderate and high condi- tions. The perception of being received by the subordinate also differed significantly between both pairs of conditions. As would be expected from the behavior of the three sub- scales, overall relatedness satisfaction also differed significantly between the low and moderate conditions and between the moderate and high conditions. Positive feelings increased monotonically across the three conditions in significant ways, while negative feelings only decreased significantly from the low to the moderate condition.

In general, both the behavioral and the perceptual data con- firm that it was possible to create experimentally three sig- nificantly different levels of mutuality and relatedness satis- faction as those terms have been defined conceptually and operationally.

Four quite different experimenters were employed in the study. Two had considerable training and experience with experi- ential learning and social intervention, while the other two were less experienced. The men differed substantially in terms of their overt personality characteristics. Several re- spected organizational behavior colleagues doubted whether some of the experimenters had the interpersonal competence to produce the variations in their own behavior necessary to create the three different experimental conditions. The group

5 20/ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

Relatedness Need Satisfaction

Table 8

Condition Effects on Perceptual Measures

Low Moderate High df F* p W2

General satisfaction 2.08t 3.1 7t 3.74 2/40 24.71 .001 .41 Give of self -.80t 2.1 4t 3.00 2/40 1 7.45 .001 .34 Receive subordinate -.71 t 1.1 4t 2.52 2/40 1 3.71 .001 .36 Received by subordinate -6.76t 1.62t 5.05 2/40 55.52 .001 .65 Overall relatedness satisfaction -8.47t 4.37t 10.46 2/40 88.81 .001 .75 Positive feelings 2.05t 4.29t 5.1 4 2/40 48.67 .001 .63 Negative feelings 4.05t 1.81 t 1.43 2/40 38.90 .001 .57

Repeated measures 1 -way analysis of variance. t Indicates that adjacent means differ at p<.05 or less.

Indicates that adjacent means do not differ significantly.

of experimenters, in preparation for the study, worked hard to learn their roles and, in particular, to respond to the various approaches the managers employed. Despite their efforts and commitment, it was expected at the end of the study that there would be differences among the experimenters or, at the very least, interactions between experimenters and roles. The interpretation of an interaction between experimenter and role would be that the men differed among themselves in their capacity to behave in the ways expected in a particular role. The data shown in Table 9 indicate that there were no interactions between experimenters and the roles and on the three major measures, there was only one main effect differ- ence among the experimenters. There was a barely significant difference among the four experimenters in the overall level of mutuality they created across all conditions and the rank order difference among the experimenters on this variable was not what would have been expected on an a priori basis.

Table 9

Role Player Effects on Independent Variables

Role player Role player X condition interactions A B C D df F p df F p

Mean mutuality behavior 7.1 9 7.02 6.55 7.00 3/49 2.82 .05 Mutuality behavior 6/49 .38 n.s. Mean general satisfaction 2.65 3.42 2.92 3.00 3/51 2.72 n.s. General satisfaction 6/51 1.19 n.s. Mean relatedness satisfaction 9.62 11.18 9.00 8.80 3/51 1.42 n.s. Relatedness satisfaction 6/51 .22 n.s.

d.f.=49 rather than d.f.= 51, because the tape recordings of 2 role playing episodes were lost.

Tests of Hypotheses

Data for testing the hypotheses of this study are shown in Table 9. There was a significant difference across conditions in the strength of relatedness desires that were expressed. Tests of linear and quadratic comparisons were performed on the combined measure and both comparisons were sig- nificant (Flinear-5.28, d.f.=1/40, p< .05; Fquadratic=4.10,

521 /ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

d.f.=1/40,p<.05). Within the self and other measures, main effect, linear, and quadratic tests were also performed. Both measures were significantly different across conditions (Fself = 4.80, d.f. = 2/40, p < .05; Fother= 4.42, d.f. = 2/40, p<.05). On the self measure, only the linear comparison was significant, but the quadratic was not (Fself Iinear=6.70, d.f.=1 /40, p < .01). On the other measure, only the quadratic comparison was significant, while the linear was not (Fother

quadratic= 5.63, d.f. = 1 /40, p< .05).

These results support a number of key features of ERG theory's conceptualization of relatedness needs (Alderfer, 1972: 144-1 50). Consistent support for a curvilinear relation- ship between relatedness satisfaction and desire was obtained. Since this major result was demonstrated by experimental methods, more confidence may be placed in the theory than when it was based on correlational data alone. The effects of satisfaction on desire were shown to apply to more than one human object. Separate effects on the self and other desire measures showed that the experimental manipulation affected the managers' acknowledged desires for themselves and their projected desires for the role players. Despite the similar pattern of results for the self and other measures, the sig- nificant difference between these two conditions supports the notion that the managers were indeed responding to two different human objects.

Table 10

Condition and Person Effects on Desire Measures

Condition means Analysis of variance Conditions (A = managers, B = conditions, C = measures)

Person low moderate high Person means Effect S.S. d.f. M.S. F p

Desired for self 5.20 3.05 3.20 3.81 Conditions 80.82 2 40.41 4.72 .05

Desired for other 3.14 1.57 2.1 9 2.30 Persons 71.63 1 71.63 1 5.66 .001 (self and other)

Condition means: 4.17 2.31 2.70 AB 342.51 40 8.56 (total desire)

AC 91.53 20 4.57

These data were analyzed by a 3-way analysis of variance: 21 managers, by 3 conditions, by 2 measures. Managers were treated as random effects, while conditions and measures were treated as fixed effects (Winer, 1 962: 1 72 and 254).

Because the theory about relatedness needs is concerned both with behavioral processes and human objects, a second- ary measure of desire for the human object was also em- ployed in the study. The managers were asked at the end of the role playing episodes to indicate who among the three men who played their subordinates they would hire and reject. The data shown in Table 11 indicate a monotonically de- creasing tendency to reject the person as conditions change from low to high and a monotonically increasing tendency to hire the person as conditions changed from low to high satisfaction. After indicating his choice to hire and reject, the manager was also asked to explain why he made the choice he did and these reasons were coded for interpersonal and noninterpersonal reasons. A given answer could be coded for both interpersonal and noninterpersonal reasons. Sixteen of the managers gave interpersonal reasons for their hire choice and 1 7 gave interpersonal reasons for their reject decisions. Five gave noninterpersonal reasons for their hire choice and 8 gave noninterpersonal reasons for their reject choice. It was

522/ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

Relatedness Need Satisfaction

apparent that the managers were making their choices pri- marily for interpersonal reasons. One way to interpret these object choices along with the process desires shown in Table 1 0 is that subjects under the low satisfaction condition ex- press a wish to avoid undesirable human objects while, under the high satisfaction condition, they express a wish to approach desirable people. Table 11

Hire and Fire Patterns by Conditions

Condition Low Moderate High

Reject 13 5 3 Hire 2 6 13

X2 (d.f. = 2) = 1 4.60, p <.001.

DISCUSSION AND CONCLUSION

Perhaps one of the more important features of the present research is that it produced in the laboratory setting, with extensive experimental control, results which support similar findings obtained by field correlational methods. These results were obtained through the participation of adult middle class subjects without deception. The research closest to the current formulation was reviewed by Eisenberger (1 970) who con- cluded that there was sufficient evidence to postulate a deprivation-satiation function for social approval. While mutuality and relatedness satisfaction are both conceptually and operationally more complex than social approval, the difference between the low and moderate conditions in this study includes a substantial difference in the degree of posi- tive affect exchanged and between these two conditions, desire decreases in a way that is consistent with Eisenberger's (1 970) review. Finding a generalized linear component across all three conditions is also consistent with Eisenberger's analysis. Where the present concepts and findings depart most markedly from other work is in the creation of the highest level of mutuality and in showing that the desire curves turn upward in this condition. Although the highest level of mutuality has a more positive flavor than either the moderate or low conditions, it also differs significantly from these other states in the explicitness of how all affect is ex- pressed. Thus, it would be an oversimplification to assert that the upward turning desire curves stem only from the in- creasingly positive affect in the highest condition of mutuality and relatedness satisfaction.

To further complete the link between the laboratory results of this investigation and the field correlational results of other studies, product-moment correlations between satisfaction and desire were also computed in the manner described in Table 2. The linear correlation between relatedness satisfaction and desire was -.42 (p < .001 ). Relatedness satisfaction was transformed to develop a quadratic term by subtracting the mean relatedness satisfaction of the moderate condition from each individual score and then squaring this difference score to be correlated with desire. The resultant correlation between relatedness satisfaction squared and desire was +.30 (p<.01). The multiple correlation using both the linear and quadratic terms was .53 (p <.001), but the analysis testing whether the quadratic term added significantly to the

523/ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

variance accounted for by the linear term alone proved insig- nificant. These results produced by correlational methods were consistent with the findings emerging from the analysis of variance, except for the failure of the quadratic term to add significantly to the variance accounted for by the linear term. The analysis of variance model utilizes orthogonal compari- sons in generating tests for the linear and quadratic com- ponents, while the correlational approach statistically controls for the correlation between the linear and quadratic terms. The size of the correlation between the linear and quadratic terms is thus partially a function of the transformation used. The size employed in this study resulted in an average correla- tion of -.83 between the linear and quadratic terms across the three data sets and other transformations were not tried in search of the one that would minimize the association between the linear and quadratic terms.

If it is not simply the increasing positiveness of the interaction as one moves from the low to high conditions that accounts for the upturn in the desire curves, what is it? It is the explicit dealing with all emotions, not just positive feelings. Three different bits of evidence tend to be consistent with this speculation.

First, the results shown in Table 5 indicate that both receiving feelings and giving feelings vary curvilinearly across the conditions in a way that parallels the pattern of the desire measures. The minimum points of receiving feelings, giving feelings, and desires all occur in the moderate relatedness satisfaction condition. Second, an individual level exploration of the impact of given and received emotionality on evoking desires was performed by correlating the two feelings scores with the desire scores by the procedure described in Table 2. One of six possible product-moment correlations -3 data sets by 2 behavior measure by 1 desire measure-was sig- nificant and the other five were not different from zero. In one data set receiving feelings was positively correlated with strength of relatedness desires (r= + .38, p < .05).

Third, the matter was explored with the managers during their feedback session, which happened before the behavior coding had been carried out. Analyzing the data, it was possible to identify three ways managers responded to the desire meas- ures across the conditions: one group showed the expected curvilinear relationship upturn, a second showed desires monotonically decreasing across all three conditions, and a third group showed no identifiable pattern. In the feedback session, it was explained that the study was designed to see which law was more valid and it was determined that different ones seemed to behave according to different laws. It was suggested that three groups be formed according to who seemed to follow the same pattern of reporting desires without initially telling the men which law their group had followed. Each group would be charged with attempting to discover commonalities among their members. They agreed to this proposal and formed subgroups for half-hour discussions to search similarities among themselves.

On returning, the curvilinear group described itself in its words as uniformly a volatile group in which its members differed from each other in many ways, but were similar in the one characteristic that each man's feelings were easily aroused. Members of the monotonically decreasing group reported that they never got to discussing themselves, but spent much of their time trying to figure out which theory they fit. The group decided that it was the curvilinear theory

524/ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

Relatedness Need Satisfaction

and was wrong. Members of the erratic group, that is, those whose responses fit neither the linear or curvilinear model, noted that they all approached the cases very analytically and pointed out that some of the details of each case were sub- stantively wrong. They noted that each of them tried quite hard to control the experimental situation. As a group, they predicted that they fit the monotonicaly decreasing function, which also was wrong. The curvilinear group seemed at ease with the experimental situation and additional exploration during the feedback session, the linearly decreasing group desirous of pleasing the experimenter, and the erratic group impelled to fight the experimental procedures. None of the three bases for inferring that the presence of observable emotionality heightens relatedness desires is unequivocal in itself but, taken together, the curvilinearity of emotional explicitness across the three conditions, the managers' unstructured self reports, and the one significant correlation between explicitly receiving feelings and desires all strengthen the case for this explanation. The experimental procedures employed in this study differ in important ways from what is done in many social psychology experiments. Rather than working with college sophomores, adults were used who included the highest ranking members of their social system. The company paid $1 5 per man toward the study's expenses. Special efforts were made to obtain informed consent from each subject. Each person who participated received feedback on his own behavior through counseling and tape listening-if he chose-and on the study as a whole through a group meeting devoted to explaining all facets of the investigation. The time given to feedback for each individual approximately matched the time used to collect data from the managers. At no time was deception employed. Although unfamiliar with Kelman's (1 972) state- ment about the rights of subjects in experimentation when the study was designed and executed, it is clear that the methods were very consistent with his proposals-despite the fact that little of the approach was based on ethical considerations. To create high levels of mutuality experi- mentally, it was necessary to build an experimental social system which itself demonstrated high mutuality. Two noteworthy outcomes of this study, which have implica- tions for interpersonal behavior in organizations and the design of role-playing experiments, are the potency of the experimenters' behavior and the managers' reaction to it. First, 62 percent of the fired subordinates were from the low mutuality condition, and 62 percent of the hired subordinates were from the high mutuality condition. This finding speaks powerfully to how readily professional managers may convert behavioral processes of which they are a partial cause to personal evaluations in which they use their authority to determine another person's destiny. Second, the experimenters were remarkably successful in evoking all degrees of mutuality from the managers whose own education was devoid of sophisticated knowledge of interpersonal behavior. By them- selves behaving according to specified degrees of mutuality, the experimenters influenced the managers to demonstrate similar behavior. This result suggests that experimenters and managers alike can and do determine the nature of their own interpersonal worlds by how they themselves act. While there are now many programs to teach managers about inter- personal style and its relation to their work, there is less awareness of the relevance of this kind of education for the scientific work of researchers. Perhaps it is time for social

525/ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

psychology experimenters to consider enlarging the tools of their scientific trade to include more complex interpersonal styles.

This study showed that it is possible to bring significant features of organizational behavior into the social psychology laboratory for controlled study. Parts of the ERG theory of human needs are now based on empirical data from both field and laboratory settings. This study was designed from a theory of person-organization interaction which viewed the laboratory as an organization and the subject as a member. Confirmation of the need theory propositions through this methodology implies support for the broader dynamic organi- zation model.

Clayton P. Alderfer is associate professor in the School of Organizations and Management at Yale University. Robert E. Kaplan is assistant professor in the School of Management at Case Western Reserve University. Ken K. Smith is a graduate student in the Department of Administrative Sciences at Yale University.

REFERENCES

Alderfer, Clayton P. 1 967 "Convergent and dis-

criminant validation of satisfaction and desire measures by interviews and questionnaires." Journal of Applied Psy- chology, 51: 509-520.

1 972 Existence, Relatedness, and Growth: Human Needs in Organizational Settings. Glencoe, Ill.: Free Press.

Alderfer, Clayton P., and Thomas M. Lodahl 1 971 "A quasi-experiment on the

use of experiential methods in the classroom." Journal of Applied Behavioral Science, 7: 43-70.

Argyris, Chris 1 962 Interpersonal Competence

and Organizational Effec- tiveness. Homewood, Ill.: Dorsey.

1 969 "The incompleteness of social psychological theory." American Psychologist, 24: 893-908.

1 970 Intervention Theory and Method. Reading, Mass.: Addison -Wesley.

Eisenberger, R. 1 970 'Is there a deprivation-

satiation function for social approval?" Psychological Bulletin, 74: 255-275.

Fairbairn, W. R. D. 1952 An Object Relations Theory

of Personality. New York: Basic Books.

Festinger, L., and D. Katz 1 954 Research methods in the

behavioral sciences. New York: Holt, Rinehart, and Winston.

Freud, Anna 1 946 The Ego and the Mecha-

nisms of Defense. New York: International Universities Press.

Freud, Sigmund 1 963 Therapy and Technique.

New York: Collier.

Kelman, H. C. 1968 A Time to Speak. San

Francisco: Jossey-Bass. 1 972 "The rights of the subject

in social research." American Psychologist, 27: 989-1 01 6.

Klein, M. (ed.) 1 955 New Directions in Psycho-

Analysis. London: Tavistock.

Lewis, B. N. 1 962 "On the analysis of inter-

action in multi-dimensional contingency tables." Journal of the Royal Statistical Society, 25: 88-117.

Maslow, A. H. 1 954 Motivation and Personality.

New York: Harper. 1 966 The Psychology of Science.

New York: Harper and Row.

Murray, H. A. 1 938 Explorations in Personality.

New York: Science Editions.

Rogers, C. R. 1 959 "A theory of therapy,

personality, and interpersonal relationships, as developed in the client-centered frame- work." In S. Koch (ed.), Psychology: A Study of a Science, 3: 184-256. New York: McGraw-Hill.

Winer, B. J. 1 962 Statistical Principles in

Experimental Design. New York: McGraw-Hill.

APPENDIX

A. ROLE PLAYING INSTRUCTIONS

Low Mutuality (Case of the Underpaid Manager)

Manager's position: You have asked one of your subordinates to come to your office for his performance review. Two months late, the review has been held up by your boss who has been slow to approve your proposed rating. You think that the reason for his slowness was his reluctance to agree to lowering the man's rating from A-1 to B-2. After sitting on the review for seven weeks, your boss returned it to you with his approval, noting that he attended a meeting in which this man fell

526/ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

Relatedness Need Satisfaction

asleep. He also commented, "We can't afford to lose this man; he's too talented." You guess that your boss was not in favor of the lowered evaluation, but decided to go along with it after he saw the man fall asleep. Your relationship with your boss is more or less cool and distant. You know he would be very upset if you even intimated that he held up the review process. You will give the man a 3 percent raise as a result of his performance review and the general economic squeeze being felt throughout the company. Subordinate's (experimenter's) position: Your boss has asked you to come to his office for your biannual performance review. The review is more than two months late. So far your previous reviews in the company have been outstanding; all of your ratings have been A-1. The cutback in operations has hurt your financial remuneration, however. Despite good reviews, your last two increases have been minimal.

During the last six months, sickness in your family has caused you to seek out and find a second job at night. This has helped a little on the financial side, but has brought its own set of problems, because you have to be away from home more than usual at a time when your presence would be very helpful. You are feeling decidedly angry at the company for not paying you better when your performance has been outstanding.

Lately, you have been feeling tired and fatigued at work and have actually begun the process of looking for another position. You have decided that if your next review does not lead to a substantial raise, you will leave just as soon as you can obtain a better job elsewhere. You intend to ask your boss in very concrete terms what to expect in the future.

Moderate Mutuality (Case of the Broken Toilet)

Manager's position: You are feeling a lot of pressure from sales and from top management at the plant and corporate level to complete a particu- larly large batch of product. During the last several days, you have gotten the impression that the employees in one area have been performing below their average level and, on top of that, seem to be complaining about all sorts of things, including a toilet that has not been fixed for several days. One of the reasons why the toilet has remained unfixed is because top management has set priorities to keep the maintenance activities all on production facilities in order to get the current batch completed on time. The man in charge of this area is coming to see you, at your request, to discuss these matters. Your general impression of this man is positive. His performance is generally above average, but you sometimes feel as though he pays too much attention to how people feel about things and not enough to getting the job done.

Top management has told you that corporate group is paying special attention to the current batch and will respond very favorably to your meeting the customer's deadline. Contract negotiations are not far away and you have been alerted by industrial relations to be careful in order to avoid any needless grievances.

Subordinate's (experimenter's) position: A broken toilet in your area has not been fixed for several days and your people have been coming to you to complain. Most recently, you have learned that there might be a movement to file a grievance. You have spoken both to your boss and to the maintenance people. From both you have heard that they are sympathetic to the problem, will do everything they can, but are pressed by other higher priority demands.

Now that you have heard about the potential grievance, you feel that you want to let your boss know and see one more time if something can not be done to get the problem fixed. You increasingly sense that the employees want action. At the same time, you have the impression that they may be looking for an excuse to make trouble. On the other hand, you understand that your boss is feeling pressure from sales for deliveries, but you also have the impression that he is not fully leveling with you. You feel that you could make a stronger case with your people if you believed that your boss told you the complete story of what he is facing.

527/ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

High Mutuality (Case of the Exiabor Manager)

Manager's position: You have just had a new man reassigned to you as a result of the cutback. This manager had formerly been a union official with whom you had had a severe conflict. In getting this new man, you had a choice of three men. You selected this man because his technical competence was far superior to the others you might have chosen. At the same time, however, you were concerned about your past relationship with him.

The conflict had been over whether a company employee had stolen some tools. You felt sure that he did but you did not quite have all the evi- dence to document the case. The union official admitted that he knew more than you did, but would not share his information. One time you got very angry at him and he began to wave an umbrella at you. Nothing more came of this event, but the man was not fired because of insufficient evidence. He has subsequently left the company on his own initiative.

Now you want very much to start the new working relationship off on a new foot and feel it necessary to discuss these former events. You believe that you were right in trying to get the missing information from the union leader, but now recognize how he might have seen you as trying to intimidate him and resisted for those reasons. You have asked to meet with him to discuss how you two will work together.

Subordinate's (experimenter's) position: You are now a manager, although you had formerly been a high-ranking union officer. While you were a union officer, you prevented the man you now work for from firing an employee who had taken some company tools to use at his home. The manager's position at the time had been that the man had stolen the tools. You had been unsure about the employee's intent, but felt he needed help against arbitrary action from management. Some of the tools in question were eventually returned, while others never reappeared.

As a result of the economic cutback in operations, you have been re- assigned to a reporting relationship with the man with whom you had formerly been in conflict when you were a union official. For many different reasons, you would like to get the new relationship off to a good start. While you do not feel that you were wrong in the old conflict, you do recognize that your union role may have prevented you from fully seeing the manager's point of view. You are prepared to try to have an open discussion with your new supervisor about this past event in order to clear the air and start things off in a constructive vein.

Behavioral Heuristics for Experimenters

Despite the predisposing effect of the role descriptions, there was no expectation from the investigators that these instructions alone would create the desired behavioral outcomes. To aid the experimenters in their objective of influencing the interaction to proceed in the direction needed by the study, a set of guidelines was developed. There were some issues that pertained to the total subordinate role and others that applied only to the specific episodes.

General guidelines.

(1 ) Remember that it is up to you to influence the direction of the interaction. At the same time, however, recognize that you may not be able to affect every manager. Use at least 10 minutes.

(2) As a general rule, you will want to create some tension in order to engage the man you face. The manager's tendency will probably be to minimize the tension and to look good.

(3) No matter what happens, stay in role. The investigators will be observing events and will intervene if it seems necessary.

(4) If the interchange unfolds in such a way that you feel in need of some additional facts, invent them on the spot. Use your history, personal expe- riences, and immediate feelings. When in doubt, be yourself within the role.

(5) Study both the part you have to play and the manager's role and information. You have a valuable advantage in knowing his role, while he will not know yours. Use this information to create tension or avoid being seduced away from the primary direction of the role playing.

(6) In preparing for the role, try to think of ways the manager might behave in his role, then plan how you might respond in order to move things in the intended direction.

528/ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

Relatedness Need Satisfaction

(7) Plan to learn a few things yourself in the first session or two. This way you will probably be able to stay loose enough to respond flexibly to the behaviors offered by the managers.

On the specific roles.

High mutuality. (1 ) Be wary of the manager's early moves to smooth things over. To counter this, you have two basic strategies: (a) confronting him about his past behavior and (b) owning up to the irrationalities in your own past behavior. You also have two targets for discussion: (a) your roles as union leader and manager and (b) your and his personal choices and emotions.

Moderate mutuality. (2) You want this interaction to show some signs of both positive and negative effects and end without a clearcut resolution. To develop this affect-and effect-consider the following. (a) Being insistent about your needs in the situation and accommodating to your boss's position. (b) Both you and your boss lose if there is a grievance. Remind him of this if he seems to need it. Choose which part to remind him of, depend- ing on whether, in your judgment, the interaction is too positive or too negative. Try to end the interaction with a positive and negative ex- change in close proximity. For example, you agree to do more to stop the grievance and move production and you are not sure it will work, unless your boss comes through with getting the priorities changed.

Low mutuality. (3) Gradually, but persistently, you want this interaction to deteriorate. Discuss both performance and potential aspects of the review. You may choose to accept the poorer performance, but never the potential de- crease. Hammer at your boss for not telling you on a day-to-day basis what the problems were as well as for being two months late in giving you the review. Do not get stuck on just the money aspect, although be sure to use it. Deal with the company's obligations to help you grow and learn and to be considerate of your personal needs. Be alert to incon- gruities or evasiveness in the manager's present behavior. You might ask him whether you are being made to suffer because he and his boss have a poor relationship.

B. BEHAVIOR CODING SYSTEM

Definition of a unit.

A unit of interaction is defined as all of the words spoken by one person, A, from the time he begins to speak until a second person, B, starts to talk. B may start talking either because A has finished talking or because A has been interrupted. Generally, a unit consists of a number of sen- tences, expressing at least one complete thought and perhaps more, bounded at either end by the words of speakers other than the one whose unit is immediately being coded. This particular definition of a unit is arbitrary to a certain degree, being particularly useful for dyadic and small group interaction. It would not be particularly useful in coding speeches or other kinds of monologues. This general definition of a unit is supported by a number of subsidiary guidelines for aiding the scorer in deciding how to code a unit.

(1 ) When it is hard to decide whether a unit of behavior has happened, (a) score a unit if the words, however brief, express a complete thought, even if they do not form a grammatically complete sentence. Example: "Hello" is scored. (b) Do not score a unit if some words are spoken which do not alter the speech of the main talker. Example: "Ugh" is not scored. (c) Score a one word interruption, even though it does not express a complete thought if it does provoke a response. Example: "But . . ." followed by, "Don't interrupt me," would be scored.

(2) When a long unit of behavior seems to deserve two different scores, select the value that seems to reflect the longest or most potent portion of the interaction.

(3) If there is a seeming discrepancy between two levels of inference in coding the behavior, select the value which least departs from the literal meaning of the words.

Each unit of interaction is coded before the next unit is attended to. For each unit, the central focus of coding pertains to what is being ex-

5291ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

pressed in that unit and in the unit immediately preceding it. A coder does not trace back over the coding scheme and recode an earlier unit because an event subsequent to it seems to change the meaning. A coder does begin coding any unit with the receiving scores-ideas and feel- ings-and then proceeds to code the giving scores-ideas and feelings. The first unit in any sequence does not have any receiving scores because it has no behavior preceding it.

It is assumed that each speaker responds in some way to the giving behavior of the preceding speaker. In following a speaker, the subsequent speaker gives some indication, which could be zero, how much he re- ceived of what the other person said. The scoring system does not attempt to speculate about how much the subsequent speaker actually receives; it does attempt to describe observable behavior on the part of the second person which indicates how much he received of what the first person said.

Receiving Ideas

This behavioral component of mutuality is rated on a three-point scale designed to reflect how much of the preceding speaker's ideas are overtly received by the subsequent speaker.

One is scored when the subsequent speaker either totally ignores or completely misunderstands what the preceding speaker's ideas were. Example: Speaker A, "I think we should stop worrying about whether we get the current batch of supplies out for the moment and decide to fix the toilet, because if we don't, we're certain to have a grievance." Speaker B, "I don't think the workers are behaving very well." Speaker B's score on receiving ideas for that unit would be 1, because there was no indication in his words that he in any way addressed himself to the issues raised by Speaker A. Two is scored when there is some indication that a portion of the pre- vious unit was received, although the remainder of the ideas might be ignored or misunderstood. Example: Speaker B, after Speaker A above, "I don't think we can put the amenities higher than the current batch, because management has commitments it must meet." Speaker B's score on receiving ideas for this unit would be two, because there were signs in his comments that he addressed himself to some of the issues raised by Speaker A, but there were also some ideas unresponded to, such as the potential for a grievance. Three is scored when the subsequent speaker responds to essentially all of what the preceding speaker said and uses it as a springboard to his own comments. Example: Speaker B, after Speaker A above, "I don't believe the workers will file a grievance yet because they have too much to lose themselves. Our incentive scheme almost guarantees that won't happen. We haven't had any disturbance for 1 8 months since we de- veloped that new system. At the moment our top priority is the batch. We'll attend to the toilet after that." Speaker B's score on receiving ideas would be three, because he responded to all of the major components of the preceding speaker's words.

In summary, score receiving ideas I if none of the preceding ideas' content is received, 3 if all of the preceding ideas are received, and 2 if some of the preceding ideas are received.

Receiving Feelings This behavioral component of mutuality is rated on a three-point scale to reflect how much of the preceding speaker's feelings are overtly received by the subsequent speaker. One is scored when the subsequent speaker totally ignores both verbally and with his own tone of voice whatever feelings were conveyed by the preceding speaker. Example: Speaker A, "What you have just said makes me feel slightly hostile, because I don't understand why I'm being treated this way." Speaker B, "I wonder what will happen when the boss comes back." Speaker B's receiving feelings score for that unit would be 1, because he in no way responded to the feelings expressed by Speaker A. Two is scored when the subsequent speaker indicates implicitly that he has received some of the affect given by the preceding speaker. This might be done by indicating a reaction to the affect, such as defensive-

5 30/ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

Relatedness Need Satisfaction

ness, or by a change in voice tone. Example: Speaker B, responding after Speaker A above, "I am confused about why you are reacting that way."' Speaker B's receiving feelings score for that unit would be 2, because he responded in part to the obvious feelings expressed by Speaker A.

Three is scored when there is an explicitly stated recognition by the subsequent speaker of the feelings expressed by the preceding speaker. Example: Speaker B, responding to Speaker A above, "I can see that you are annoyed, but I don't see why. I have explained why I think it is necessary." Speaker B's receiving feelings score for that unit would be 3, because he responded explicitly to the feelings expressed by Speaker A, even though he did so somewhat defensively. In summary, score receiving feelings 1 if the expressed feelings are ignored, 3 if they are explicitly and fully acknowledged, and 2 if they are implicitly and or incompletely acknowledged.

Both receiving behavior-ideas and feelings-depend in part on the giving behavior of the preceding speaker and questions arise with regard to this interdependency. Two further rules were developed to deal with this contingency:

(1 ) The automatic receiving ideas or receiving feelings score for a sub- sequent speaker is 1, if the preceding speaker's giving ideas or giving feelings score was 1.

(2) A subsequent speaker can receive a score of 3 on either receiving behavior, even if the giving behavior was only 2, provided that he behaviorally indicates that he received all of the ideas or feelings that were stated.

Giving Ideas This behavioral component of mutuality is rated on a 3-point scale to reflect how fully and completely, as compared to unilaterally and guardedly, the speaker expresses his ideas.

One is scored when the speaker states something without background, explanation, or reason. Example: "I can't accept that level of perform- ance." Two is scored when the statement or message contains an ac- count of its relevance, background, meaning, implications, or an invitation to explore these things, but it is also clear that more needs to be said by the speaker to give full contextual meaning to the statement. Example: "I don't understand why you said what you did since it seems to con- flict with what you said a few minutes ago." Three is scored when the statement or message is accompanied by a full and unambiguous state- ment of the issues, background, and implications and the scorer con- siders that the speaker has been as explicit as he can about his own views and the reasons for them. Example: "What you just said now, that my behavior was unacceptable, seems to conflict with what you said a few minutes ago when you stated that you could understand why I acted as I did in my role as union steward."

Giving Feelings This behavioral component of mutuality is also rated on a 3-point scale to reflect the extent to which an explicit statement of feelings is made by the speaker. One is scored when there is no indication of any feeling expressed in the unit. Example: "It is 3 o'clock." Two is scored when feelings can be observed in the speaker's words, tone, or frequency of speech without these emotions being acknowledged or accepted by the speaker. Ex- ample: "You've been telling me that over and over again, and I simply do not believe you." Three is scored when feelings are expressed and at the same time acknowledged, accepted, and owned by the speaker. These feelings need not, or they may, be relevant to the people or events in the immediate situation. Example: "I am very happy that you have decided to accept my offer."

When the giving feelings scores were 2 or 3, the coding system also provided for a fifth category, feelings, to score the nature of the feelings being expressed. This variable ranged from 1 to 5 as follows:

One was very negative feelings. Example: "'You sonofabitch!" Two was mildly negative feelings. Example: ''I am kind of upset with what just

531 /ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions

happened." Three was mixed feelings, some positive and some negative. Example: "I don't like the idea of having to discharge too many men, but I see the necessity for it and will do my best for their sake and the company's." Four was mildly positive feelings. Example: "That was a good idea." Five was very positive feelings. Example: "I am really pleased and excited by what you just did." In total, the mutuality behavior coding system consists of four com- ponents applied to every unit of interaction, a fifth category called feelings which is scored whenever feelings are expressed, and rules for defining a unit of interaction. The nature of this system is such that it is the interpersonal behavior that is scored, not the individual's behaving, because the system of operational definitions requires that the scorer refer to the preceding unit of interaction in scoring any subsequent behavior. Thus, the interdependency which is assumed to characterize interpersonal relations is a part of this scoring system. Nevertheless, separate judgments are made about each of the components. There were independent operational rules for each giving and receiving behavior. Feelings scores did require feelings to be expressed, but their score did not depend on the degree of implicit-explicitness.

532/ASQ

This content downloaded from 62.122.79.21 on Wed, 18 Jun 2014 09:09:30 AMAll use subject to JSTOR Terms and Conditions