Accentuate the Positive:
Positive Sentiments and Status in Task Groups*
Alison J. Bianchi
Donna A. Lancianese
The University of Iowa
* Direct correspondence to Alison J. Bianchi, Department of Sociology, W118 Seashore Hall, The University of Iowa, Iowa City, Iowa 52242-1401; firstname.lastname@example.org.
We explore the capacity of positive sentiments, those enduring affective states one achieves when one likes another, to impact status structures. Do positive sentiments combine with existing aspects of interaction to create status hiearchies and potentially change the social order, or do they moderate the effects of extant structure by dampening the magnitude of status differences? Using the theoretical framework of Status Characteristics Theory (SCT) and the Camilleri-Berger model of decision-making, we designed an experiment to adjudicate between these two potentialities. Participation in the study consisted of 168 students. Results found support for the notion that positive sentiments moderate the effects of structural factors on indicators of social status; interestingly, this moderation effect varies by gender. We discuss the theoretical implications of these findings, and new directions for research concerning sentiment and status processes.
What is a group process, and what are status and sentiment processes? A group process is a pattern of behavior that can be observed in groups with similar characteristics. For example, the social status, or social prestige, that some group members have tends to afford them more influence than others. Group members expect those with higher prestige to be at the “top” of the group or in a leadership position; group members presume that those with lower prestige are somehow less than higher status persons, and are thus allowed less influence. The development of this informal rank ordering in a group, termed a status hierarchy, is a group process that emerges during interactions in groups whose members are working together to complete a task that they value. (We classify the kind of group that we study because there are many diverse types, and group processes are distinctive depending on the type of group. For example, consider the way one would behave in a group of friends versus a group of professional colleagues. Influencing friends is one kind of group process; influencing colleagues is clearly a different matter.) The type of group process during which a status hierarchy emerges and organizes group members’ behavior is called a status process.
In a similar type of group, a sentiment process is the development of a pattern of influence based on whether or not group members like or dislike one another. Researchers have argued that those who are liked tend to have more influence than those who are disliked.
Both of these processes have serious sociological implications: status hierarchies that are enacted in collective, goal-oriented groups mirror the macro-level social inequalities of the society in which these groups are embedded. For example, in the United States today, men have higher salaries than women, on average. And, on average, men have more social prestige than women, and thus tend to be the informal leaders of groups more than women (it is no accident that most jury forepersons are male!). As you can see, the enactment of a status hierarchy based on gender (or other markers of social status) in small groups is one mechanism that helps maintain society-wide disparities due to social status.
As for sentiment processes, symbolic interactionists have shown us that we tend to like individuals who are most like ourselves, such as those of the same race/ethnicity and age. Consequently, we have a tendency, on average, to create social networks that exclude others that are different from ourselves, perhaps subjecting them to discrimination. Understanding status and sentiment processes could lead to intervention policies that eliminate their deleterious effects.
Many researchers have examined status and sentiment processes in isolation; few have examined them together, which is how they occur in “the real world”. We decided to explore both status and sentiment processes in small groups, but with a twist. Our central research question was: do sentiment processes operate the same way as status processes? Or, put in other words, does sentiment equal status in collective, goal-oriented groups? This question may seem strange, given how we just described sentiment and status processes, but think about the question this way: do you consider the people that you like as having more prestige and value than those who you have no feelings for? Would you allow those that you like to influence you more than those whom you have no opinion about because you think that they are more competent? These are the types of questions that our study investigates.
To explore these questions, we have to understand how the status process works in detail for collective, goal-oriented groups. Luckily, sociological social psychologists have developed and extensively tested a theory that describes just how they work called Status Characteristics Theory (or SCT) (Berger et al. 1977). SCT describes how, during collective, goal-oriented group interaction, information that conveys differences in social status creates status hierarchies that, in turn, direct how group members behave. Group members who are perceived to be on the top of the status hierarchy will exert more influence than group members perceived to be closer to the bottom of the status hierarchy. So, our research question then becomes: does liking someone give them more social prestige in the “liker’s” eyes? And consequently, will that liked person have more influence?
SCT is a highly technical theory, and therefore some technical knowledge and jargon are needed to understand it fully. Accordingly, we now put forth a very, very abbreviated version of SCT, so that one can understand why we set up our test of positive sentiments’ capacity to convey social status as we did. (We studied only “positive sentiments” or liking each other in small groups because the first author has already studied negative sentiments – see Bianchi ).
The elements of interaction with information that convey social status are called status characteristics. A status characteristic is any recognized social distinction that has attached to it widely shared beliefs about at least two categories, or states, of the distinction. For example, gender is a status characteristic, and its states are “male” and “female.” Those possessing one category are more valued socially and perceived as generally more competent than others possessing the complementary category (Berger, Cohen, and Zelditch 1966; Correll and Ridgeway 2003). Different states of a status characteristic convey social advantages and disadvantages upon actors who have the traits and attributes associated with the respective state (Webster and Hysom 1998). In groups whose members are working together on tasks that they perceive have successful or unsuccessful outcomes, status characteristics become associated with performance expectations, which are out-of-awareness hunches about who has the ability to get the task done properly (Ridgeway and Walker 1995). So, social status and prestige become equated with competence in groups working to get a job done.
Consciously, we know that if one possesses social prestige it does not necessarily mean that one is more competent. However, in situations during which we want to get that task done well, and there are no formal titles designating leaders and followers, we nonconsciously (almost like we are on auto-pilot) fall back on visual cues, expressions, and the like as information that connotes competence at the task. Unfortunately, we associate visual cues, such as perceiving others’ gender, race/ethnicity, and social class, with cultural stereotypes about competence. We do not mean to do this ? it just happens automatically. This knowledge alone should prompt us to better understand these processes, since once we understand them, they can be eliminated!
A status hierarchy is enacted if a group of actors who possess different states of status characteristics work on a collective task that they value. Then, status generalization, the process by which differences in actors’ states of status characteristics form performance expectations, and how these performance expectations determine influence, ensues.
Briefly, the status generalization process is as follows: some actors in task groups have constellations of status characteristics that give them higher status and more social advantage in comparison to other group members. Because they are perceived to be higher in status, these group members are given more opportunities to control the direction of the group task. High status members do not exercise power in the group because their ability to persuade other actors is not achieved through coercion. Instead, when status differences are activated and made salient to group members, differential performance expectations are evoked. As beliefs about task performance are linked to social status, higher status individuals will then be perceived to be generally more competent to complete the group task successfully. Differences in performance expectations then translate into deferential behavior on the part of low-status group members. Low-status group members accept the influence of high-status group members more than high-status actors accept low-status actors’ influence. This difference in acceptance of influence is behaviorally displayed as group members complete the group task. Hence, an observable influence hierarchy can be measured that indicates the status differences between actors with unequal performance expectations.
Ridgeway and Johnson (1990) theorized that group members perceive persuasive discourse by liked actors as a sign of competence, as their resolutions of disagreements may be listened to more, considered more, and more positively evaluated than disliked actors’ attempts at influence. This differentiated pattern of influence behaviors could indicate an elevation in status for liked actors. Fisek and Berger (1998) then posited that positive sentiments could act as status elements that combine with other status elements, such as status characteristics, to form activated expectations that affect influence behaviors. They refer to this potentiality as the “sentiments as constitutive of expectations” model. However, they also suggest that if positive sentiments do not fit this model, then sentiments may “transform, filter or moderate” (Fisek and Berger 1998:26) the effects of differential performance expectations on influence behaviors.
To begin finding answers to our research question, the first step was to create a laboratory test that isolated the effects of positive sentiments on influence, the indicator of social status, and determine if positive sentiments operated as a status characteristic.
So, how did we design our study? Our dependent or outcome variable, influence, would be measured in dyads of subjects and partners interacting over a computer network in a group processes laboratory. (For experimental control, our “partners’” responses were computer generated, so subjects were actually not interacting with another person, but were led to believe they were. We, of course, debriefed subjects after the fact and let them in on the deception.) We used a measure of influence that is an “industry standard” (Moore 1968; Berger, Cohen, and Zelditch 1972): subjects are paired together as a team and make choices between alternatives in three steps: first, the subject and “partner” are individually presented with a decision to select between two alternatives. Next, the initial (computer-generated) choice of the partner is revealed to the subject. Taking the partner’s initial choice into consideration, the subject then makes a final choice between the alternatives. The final choice of the partner is not divulged, but subjects are told that all final choices will be judged as "correct" or "incorrect," and these results form an overall team score. Influence is measured by how often a subject changes his or her initial choice when making his or her final choice, given that his or her initial choice differed from the partner’s. In other words, if subjects stay with their initial choices when making their final choices, then their partners, who did not agree with the subjects’ initial choices, did not sway the subjects. To measure the rate of stay responses (referred to as a P(s) score), subjects are typically presented with 20 initial choice disagreements from partners, and are asked to make final choices after taking their partners' choices into consideration. The P(s) score is, therefore, a measure of the rate at which a subject rejects influence. Its range is from 0 to 1, and the higher the P(s) score, then the more status the subject has in relation to his or her partner (a person who rejects influence from another person has more social status). Finally, it matters not what the choices are about (so we will not go into the minute details about the group task). What matters is that initial and final choices are made so that that we may measure how much influence a person rejects.
For our independent variable, we administered a “Basic Values Orientation” survey to subjects (the survey had no scientific purpose, it just allowed us a way to “rate” how much the subjects and partners would like each other). After they completed it, either we told the subjects that their partners’ and their similar responses meant that they would like each other if they were to meet during a social situation or we gave the subjects no information about their partners. Remember, the partners are actually computer-generated responses, so this information was a deception, too. In this way, we varied sentiments to compare those who had positive or no sentiment relations with their partners.
For a robust design (i.e., the effects of positive sentiments on influence should be tested in as many different situations as possible), we varied one situational feature – the responsibility for the team score. Some subjects were told that their final choice only, and not their partners’, counted as the team score, and so they were the Decision Makers for the team. Other subjects were told that their final choice was just for the record, or that they were Advisors to their partners. If positive sentiments act as status characteristics, it should not matter if the subjects were Decision Makers or Advisors. (We used the Camilleri/Berger (1967) game theory model for this idea.)
We also controlled for gender, since sentiment processes could differ by gender, and so we told the male subjects that their partners were male and the female subjects that their partners were female. And, we told every subject that they were good at the task that they were about to do (that’s what the [+,+] means in Figure 1).
So, our final experimental design was a 2 x 2 x 2 factorial design that crosses gender with sentiments and decision-making responsibility. This means that we had eight groups of subjects; each subject in his or her group got the same information about who their partner was (male or female), whether the partner and the subject would like each other or not know if they would like each other, and if the subject was going to be the Decision Maker or the Advisor for his or her dyad (the partner would thus be the Advisor or Decision Maker, respectively). We calculated the average P(s) score for each group. Note that we also label the average P(s) scores with subscripts 1 through 8 for convenience.
Since one of the factors/independent variables is positive sentiments, if we saw a main effect in average P(s) across the relevant conditions, then sentiment would be acting as a status characteristic. This would mean that despite level of responsibility and gender, average P(s) scores had to be lower in all positive sentiment experimental groups/conditions (assuming that one would reject influence less from someone one likes) compared to the corresponding no sentiment groups/conditions.
We administered a postsession questionnaire to all subjects to capture their impressions about themselves and their partners in this situation. We had 168 subjects participate in our study, and all but 9 believed our deceptions. Here are our results (in Figure 1):
Figure 1: Average Stay Responses (P(s) Scores) for Experimental Conditions
As you can see, there was no main effect for positive sentiments for either men (P(s)1 is greater than P(s)2) or women (P(s)5 and P(s)7 are equal), so positive sentiments do not operate as status characteristics in these situations. However, what we did find, interestingly, is a three–way interaction effect for gender by sentiments and by level of responsibility! This means that men and women manage their positive sentiments differently in same-gender groups.
To examine this difference, note the differences in the Decision Maker conditions for men versus women in Figure 1. As Decision Makers have full responsibility for the team score, men defer to those that they like (i.e., reject influence less) as compared to those about whom they have no sentiment information (average P(s)4 Why do female Decision Makers defer less to partners that they like, as compared to their male counterparts? When asked how responsible they felt when being assigned to the role of the Decision Maker, the female subjects felt significantly less responsible than the male subjects. When asked how responsible they felt during their initial decisions, the female subjects again felt significantly less responsible than the male subjects. These results suggest that, unlike males, females never owned the feeling of responsibility that comes with the role of the Decision Maker.
We argue that responsibility is one aspect of a leadership position, and again, results demonstrate that women often feel as though they are not legitimate candidates for roles involving leadership and authority (Meeker and Weitzel-O’Neill 1977; Ridgeway 1988). To compensate for this perceived lack of legitimacy, women use group-oriented, positive socioemotional behaviors to increase the legitimacy of their influence attempts (Meeker and Weitzel-O’Neill 1977; Ridgeway 1982). The female subjects of this experiment did not feel as though they had the legitimacy of being the responsible ones, in comparison to males, and thus used the good-will of positive sentiments to gain influence. The male subjects did not feel the need to do this. This finding demonstrates an important difference in the way men and women interact with others. Perhaps women in leadership positions feel that not accepting influence from those that they like, and like them in return, is a way to appear responsible. Men already feel that they are responsible, and so they accept influence without concern about appearing to be irresponsible.
In conclusion, sentiment processes in this study did not turn out to be like status processes – they have a mind of their own, so to speak. However, interesting sociological insights were found, as we demonstrated that women and men have different styles of interaction regarding positive sentiments – a finding that is all our own!
We hope to continue our research into positive sentiments, as one study certainly does not answer all of the questions about status and sentiment processes. We may invite friendship pairs into the laboratory and do similar tests, as friends have much more long-lasting and enduring social ties than acquaintances who meet in the lab. We will also continue to examine these processes by gender, as clearly men and women demonstrate differing patterns of positive sentiment expressions and usages. It might also be interesting to see if patterns of influence behaviors emerge due to positive sentiments in groups with all African American, Asian American, European American, Native American, Hispanic, or biracial group members. The possibilities are nearly endless, or at least enough to keep us both employed for a long time (we hope).
REFERENCESBerger, Joseph, Bernard P. Cohen, and Morris Zelditch, Jr. 1966. “Status haracteristics and Expectation States.” Pp. 26-49 in Sociological Theories in Progress, Volume 1, edited by Joseph Berger, Morris Zelditch, Jr., and Bo Anderson. Boston, MA: Houghton Mifflin.
Berger, Joseph, Bernard P. Cohen and Morris Zelditch, Jr. 1972. “Status Characteristics and Social Interaction.” American Sociological Review 37:241-55.
Berger, Joseph, M. Hamit Fisek, Robert Z. Norman, and Morris Zelditch, Jr. 1977. Status Characteristics and Social Interaction: An Expectation States Approach. New York: Elsevier Press.
Bianchi, Alison J. 2005. “Rejecting Others’ Influence: Negative Sentiment and Status in Task Groups.” Sociological Perspectives 47(4):339-55.
Camilleri, Santo F. and Joseph Berger. 1967. “Decision-Making and Social Influence: A Model and an Experimental Test.” Sociometry 30:367-78.
Correll, Shelley and Cecilia L. Ridgeway. 2003. “Expectation States Theory.” Pp. 29-51 in Handbook of Social Psychology, edited by John Delamater. New York: Kluwer Academic/Plenum.
Fisek, M. Hamit and Joseph Berger. 1998. “Sentiment and Task Performance Expectations.” Advances in Group Processes 15:23-39.
Meeker, Barbara and Patricia Weitzel-O'Neill. 1977. “Sex Roles and Interpersonal Behavior in Task-Oriented Groups.” American Sociological Review 42:91-105.
Moore, James C., Jr. 1968. “Status and Influence in Small Group Interactions.” Sociometry 31:47-63.
Ridgeway, Cecilia L. 1982. “Status in Groups: The Importance of Motivation.” American Sociological Review 47(1):76-88.
Ridgeway, Cecilia L. 1988. “Gender Differences in Task Groups: A Status and Legitimacy Account.” Pp. 188-206 in Status Generalization: New Theory and Research, edited by Murray Webster, Jr. and Martha Foschi. Stanford, CA: Stanford University Press.
Ridgeway, Cecilia L. and Cathryn Johnson. 1990. “What is the Relationship between Socioemotional Behavior and Status in Task Groups?” American Journal of Sociology 95:5:1189-1212.
Ridgeway, Cecilia L. and Henry A. Walker. 1995. “Status Structures.” Pp. 281-310 in Sociological Perspectives on Social Psychology, edited by Karen S. Cook, Gary A. Fine, and James S. House. Boston, MA: Allyn and Bacon.
Webster, Murray Jr. and Stuart J. Hysom. 1998. “Creating Status Characteristics.” American Sociological Review 63(3):351-78.