THIS CHAPTER WILL DISCUSS:

1. How cohesiveness is a result of all the forces that attract people to groups.

2. How some of these forces, including liking, identification with the group, and psychological needs, affect maintenance-based cohesiveness.

3. How other forces, such as group and personal goals and attraction to group activities, affect task-based cohesiveness.

4. How cohesiveness affects diverse group variables, such as group process and productivity.

5. What groups can to do increase cohesiveness.

INTRODUCTION

"Cohesiveness" is a term that describes one of the maintenance variables of groups. What does it mean to say that a group is cohesive? Like many concepts, cohesiveness is easy to recognize but difficult to define. People intuitively understand the term. However, it is difficult to define it precisely to everyone's satisfaction.

Cohesive groups have several positive qualities that we all can recognize and agree upon. For example, cohesive groups have a general perception of "we-ness." There is a feeling of friendship and loyalty among group members. The group also has high morale. However, it is problematic to use these qualities to describe the concept of cohesiveness. For example, "morale" is a troublesome term in itself that is difficult to define. Also, a list of qualities will not serve as a definition.

How, then, can we define cohesiveness? Scientists have suggested several definitions for the term. In this book, however, we will adopt only one of these suggestions.

The definition of cohesiveness that we will employ comes from the "group dynamics" school of researchers. This school flourished in the 1950s and was a descendant of the work of the great psychologist Kurt Lewin. We have already described some of his work in Chapter 2, and we will continue to refer to him throughout the book. Lewin suggested a definition of cohesiveness that focuses our attention on the individual within the group. He believed that the term depends on how the individual member perceives his or her relationship with a particular group.

Field Theory

Lewin's hypotheses about groups are part of a more general proposal he called "field theory." Field theory examines the relationship between a person's goals and his or her behaviors in pursuit of those goals. For example, Ed wants to graduate from college with honors. How does he work toward that goal and why does he act the way he does in order to reach it?

Lewin proposed a model in order to represent this relationship between goals and behaviors. The model includes all of the factors that affect a person at a given time. These factors are primarily psychological, such as the person's goals and impressions of the current situation. For example, a factor for Ed might be that sometimes it is more fun to play basketball than to study. Ed must deal with the conflicting forces of something that he considers fun at the moment and something that he believes is a good long-term goal. In addition, Lewin's model includes those biological and physical factors that significantly influence a person's psychological state. For instance, Ed might feel so sick one day that he cannot study as much as he thinks he should.

Lewin called the totality of these factors the person's life-space. Figure 3.1 represents a possible life-space.

FIGURE 3.1

Factors exist at the centers of "regions" within a person's life-space. The factors include the person's possible goals, as well as other influences. The areas within the diagram illustrate these regions. The relative locations of the areas represent the relationship that exists between the person (labeled "P" on the model) and the various factors in his or her life (labeled "G"). Regions can have the additional quality of "valences." These are attractive or repulsive characteristics. The characteristics have "forces." In the diagram the arrows portray the forces at work. They affect the person's movement around his or her life-space.

A region that has a positive valence has the label "+" on the model. It "produces" forces that induce the person toward its center. In contrast, the negatively valenced region creates forces that impede the person's progress toward its center. It has a "-" label on it. For example, the diagram could illustrate a day when Ed found that a basketball game was more attractive to him than the idea of studying for an exam. He might have thought that he already knew the exam material. Whatever the case, the option to study became the "G-" in the model. Ed wanted to avoid doing it. The "G+" represented the basketball game. As you can clearly see, all the arrows point toward the "G+"; Ed's tendency will be to play basketball that day. (Lewin's model is summarized well by Levinger, 1957.)

Field Theory and Cohesiveness

A field theorist believes that a "group" has a special place in the life-space model. It exists at the center of its own region in each group member's life-space. For instance, a group called "The Study Group" might consist of Ed, Joan, Mike, and Kim. All four people have unique life-space situations. For each of them, "The Study Group" has a specific place at the center of one of the areas in his or her life-space. The area has a valence that depends on the member's goals and the member's perception concerning whether or not the group can fulfill those goals.

If a group member perceives that the group can fulfill his or her goals, the group becomes attractive. It has a positive valence. There can be various degrees of this attraction. However, to the extent that the group has any positive valence at all, forces in the member's life-space will drive the member toward the group's region. These forces increase the member's attraction toward the group. For example, Ed may want to study mathematics in a group setting. Further, Ed wants to talk with Kim, because he is attracted to her. He also wants to get to know more students who are serious about their studies. He sees "The Study Group" as something that can fulfill those goals. All of his various goals, combined with his perception that the group can fulfill them in some way, act as forces that increase Ed's attraction to "The Study Group."

The overall cohesiveness of a group is the sum of these positive forces in each member's life-space. Cohesiveness is "the resultant of all forces acting on all the members to remain in the group" (Cartwright, 1968, p. 91). This is the definition that we will be using throughout this book.

For instance, if Joan and Mike have goals that are similar to Ed's, they probably also have strong forces working on them that attract them to "The Study Group." In that case it would be a very cohesive group. However, it may be that Kim is not very interested in studying in a group because she already knows mathematics very well. She would prefer to study alone. Or perhaps she already has a group of friends that she enjoys very much. She does not want to find a group of similar people, as Ed does. In these cases "The Study Group" will not be so cohesive as it would have been if all members had strong forces working on them to remain in the group.

Back in Chapter 1 we made distinctions among six perspectives that theorists take toward groups. We called one of these the "motivational" approach. Motivational viewpoint theorists see group behavior as a result of factors that prompt individual behavior, such as needs or goals. Lewin's idea of forces moving people around their life-space is an example of the motivational approach.

Member Commitment

It is important that we do not become confused about the difference between a group's overall amount of cohesiveness and each member's attraction to the group. After all, it makes no sense to talk about individual cohesiveness. To make this distinction clear, we will distinguish between group cohesiveness and member commitment to the group. Moreland and Levine (1982) have discussed the idea of member commitment to the group in great depth. In their view, a member is committed to the group to the extent that the member is getting what they want from the group more or less than they would get from other groups. If Mike feels that he is getting more from "The Study Group" than from any other comparable group, he will remain a member. If Kim feels she could get more from another group, she will leave.

Moreland and Levine describe how people consider both the present, past, and future when deciding if they want to remain a member of a group. In other words, one weighs whether one is presently getting more from the group than from other groups, if one got more from the group than other groups in the past, and if one expects to get more from the group than from others in the future when making their decision. The theorists feel that in general, present experiences are the most important in determining a person's commitment to a group, because what is happening at the present is normally what is most salient in people's minds. Similarly, past experiences are usually more important than future expectations just because they are real experiences the person had.

There are however exceptions to these generalities. When a person has just joined a group, there is no past experience, so the person must decide based on present experiences and expectations about the future. Analogously, if a person is ready to leave a group, there is no future, so the decision is based on present and past experiences only.

Although Moreland and Levine do not consider this, it follows from their analysis that every member makes these same considerations and, as a result, each member has a certain degree of commitment to the group. Thus group cohesiveness depends on how committed each member is to the group as a whole.

COHESIVENESS AS AN OUTPUT VARIABLE

Cohesiveness is a particularly interesting variable to study because it has a dual role in group processes. Clearly, it stands as one of the major output variables in group process. A group can become cohesive or noncohesive over time. In addition, however, cohesiveness acts as an input variable that affects later group activity. Members of a cohesive group act differently together than members of a noncohesive organization. Hence, cohesiveness has a dual role as both an input and an output variable.

We will first discuss the role of cohesiveness as an output variable. In doing this, we will make a distinction between two different aspects of cohesiveness. The distinction has been made by Tziner (1982) and is consistent with the task versus maintenance distinction that we use throughout this book. One aspect of cohesiveness is based on group members' liking for one another and on their desire to be in the group. We will call this aspect "maintenance-based" cohesiveness. For example, several people may form a group because they are attracted to one another. One reason Joan might want to be in "The Study Group" is because she likes the other members. One reason Mike might want to be in "The Study Group" is because he just enjoys the experience of being with other people.

The other aspect of cohesiveness is based on the extent to which the group helps its members reach important goals or participate in desired activities. We will call this aspect "task-based" cohesiveness. Sometimes many members of the group share the same goal. For example, Ed, Joan, and Mike find the idea of studying in a group attractive because they think it will help them understand mathematics better. The group satisfies their desire for this understanding. In addition, a member might join because the group is a means for satisfying "personal goals." In other words, the person attempts to use the group to satisfy goals that other members may not share. For instance, Ed wants to meet Kim because he finds her attractive. That is not a group goal, but Ed thinks that joining the group will still satisfy this desire.

In Chapter 1 we made the claim that cohesiveness is a maintenance variable. Readers should not confuse that claim with what we are saying now. We are saying that there are two aspects of cohesiveness. One aspect is caused by task-based factors and the other aspect is caused by maintenance-based factors. This does not change the fact that cohesiveness itself is a maintenance variable.

Maintenance-Based Cohesiveness

As we have noted, a group's cohesiveness is measured by the sum total of the forces attracting its members to it. Many scholars have considered the amount of liking among the members to be the most significant of these forces. If the members do not enjoy being with one another, it probably would be very difficult for them to be attracted to the group. In fact, according to Lott and Lott (1965), one would not lose too much by defining "cohesiveness" simply as the extent to which each group member likes each other member. Researchers have generally acted as if they agreed with this claim. In a large majority of cases, scientists have actually used "liking" among members as the sole measurement of group cohesiveness. They have similarly used liking as the main method of manipulating cohesiveness. Hence, in practice, most scientists have acted as if cohesiveness and liking are the same in a group. Although we do not agree with this claim, we do recognize the central importance of liking in determining group cohesiveness. For this reason, our discussion regarding liking will be extensive.

Liking

Many researchers have examined the processes that may lead one person to find another person likable. One of these researchers, Newcomb (1960), created a method for classifying the reasons that can lead to one person liking or disliking another person. In his system, there are three general "reasons" that Person A may like or dislike Person B.

Admiration. First, Person A may perceive certain qualities in Person B that Person A likes. Person A may also see things that he or she dislikes. These qualities affect whether A likes B. They are specific "reasons" for liking, and Newcomb grouped them together. He labeled this category the degree of admiration that exists between the two people.

Reciprocation. Second, Person A may like Person B because A believes, in turn, that B likes A. This circular effect can also work to make people dislike each other. Newcomb called this reason reciprocation.

Similarity. Third, Person A may like or dislike Person B according to how Person A thinks Person B feels about the topic under discussion. If A thinks they agree, A might like B. In contrast, if Person A thinks they disagree, A may dislike B. It is a question of similarity of feelings about X.

Before we begin to examine these reasons for liking, we must briefly explain the idea of contact. As we all know, Person A cannot come to like Person B unless they meet each other in the first place.

Contact

People come to like only those people they have the opportunity to meet. This claim is intuitively obvious. Even so, we should not ignore its importance. There have been many studies examining the importance of contact as it relates to liking, and they have had some interesting results.

One thing we know is that the odds that people will come into contact with each other is a function of the extent to which they are located near each other. This physical nearness is significant. Many studies have shown that friendships at school or work are more likely to form among people who sit near each other, as opposed to those who sit far away from each other. This is also true among people who live near one another.

These findings are clearly relevant to liking in small groups. People in groups are unquestionably in contact with one another. Therefore the potential is there for group members to come to like one another, increasing the odds of a cohesive group.

Initial meetings and likability. However, contact with other members in a decision-making group is often involuntary. It might seem that increased contact, when it is involuntary, would also increase the odds that people will come to dislike one another. One could hypothesize that group members would resent being forced together. However, scientists have found that dislike is relatively rare among new acquaintances. Indeed, it appears that people expect to like other people they meet. There seems to be a bias within people toward mutually liking one another until they see a good reason not to do so. One could say that people find one another "likable" until proven otherwise.

A study by Darley and Berscheid (1967) examined this idea. Their study led some women to anticipate that they would take part in a dyadic discussion. The topic would be dating behavior. The researchers then showed the women information about their "fictional" partners. They also showed them data concerning an additional fictional participant in the experiment. These two sets of information were identical. The study then asked the women to evaluate the fictional participants. The women expressed more liking for, as well as a greater desire to work with, their future "partner" than for the other fictional "woman," despite the fact that the data about the two fictional women had been identical.

It appears that the anticipation of meeting a person leads to the expectation of liking that person. This works at least until further information about the person comes to light. The study implies that, when someone assigns us to a group, we generally come to the first meeting expecting to like the other members.

Admiration

As we said earlier, one reason why Person A may like Person B is that Person A perceives certain qualities in Person B that Person A likes. Analogously, Person A may dislike Person B because A see qualities in B that A dislikes. We call the perceptions we have about another person's qualities our personality impression of that other person.

Personality impressions are probably the biggest factor in long-term liking. These impressions are the mental pictures we have of people. These mental pictures consist of descriptions of what kind of people they are, what they look like, what kinds of things they do, and so on. Apparently, our beliefs about people's permanent characteristics--the kind of people they are--have great impact on whether we like them.

The beliefs that we hold about a person exist as a network of traits. These traits are mostly adjectives that are meant to describe a person's personality and character. There are many trait terms that we use to describe personality. For example, we might call a person "nice," "clumsy," "attractive," or "weak."

It is important to realize that traits are abstractions. They are generalized terms for a set of behaviors, but they are not behaviors themselves. For example, you cannot see a "clumsy." What you can see is an action that you can classify as "clumsy." The classification is a decision of sorts. It is not inherent in the behavior (Reeder & Brewer, 1979). After all, conduct that is "clumsy" in most instances may be "smart" in some circumstances. For example, dropping a glass may seem to be a clumsy behavior, but it is a smart action if the glass is hot enough to burn you badly.

We use the behaviors that we see people to perform in order to form our impressions of them. At a party, a person might perform a series of behaviors that we consider clumsy. For example, the person drops a glass, spills popcorn, and dances on his or her partner's feet.

We might notice this behavior rather than others because it is particularly salient. By "salient" behavior, we mean actions that are particularly attention-grabbing. A lot of research has shown that the impression formation begins when someone performs salient behavior that we notice and begin to think about.

After noticing this behavior, we then judge whether performing behaviors of that type is an inherent part of the person's personality. We could instead decide that the behavior is a momentary lapse caused by circumstance. For example, when we see the person dropping the glass and spilling the popcorn, we might think that the person is naturally clumsy. In contrast, we might believe that the person could be drunk or nervous. In other words, we must decide whether the person's character or the circumstance is responsible for the clumsiness.

Theorists have put a lot of thought into determining the conditions under which observers see a person's character as responsible for his or her actions. Jones and Davis (1965) examined what happened when observers saw the behavior only once. In this case, onlookers assign responsibility for the action to the extent that the behavior is unexpected, unusual, or unexplainable by circumstance. For example, one day people see Ted complaining about standing in a line at the bookstore. If Ted is late for an appointment or if the line seems unusually long to everybody in it, observers may say that the circumstance is responsible for Ted being impatient. If the line is short and to be expected, Ted's complaints are unusual. In that case, onlookers might say that Ted has the personality trait of being an impatient person.

Kelley (1967) described examples of behavior that occurred in different situations. Observers had more than one example that they could use to form a judgment. In such a circumstance, observers assign responsibility for the behavior to the extent that the action is similar across situations, consistent with different people, and different from the norm. For example, onlookers see Ted in many situations, and he is consistently complaining about having to wait. He complains to many people and in inappropriate situations. Other people complain far less about waiting than Ted. This series of behaviors may lead the observers to consider Ted's character to be that of an impatient person.

Once we judge that the person's character is responsible for the behavior, we attribute the relevant trait to the person. For example, we attribute the trait of impatience to Ted. Thus, we call the work of Jones, Kelley, and their associates attribution theory.

Let us return to the example of the clumsy person at the party. If the circumstances apply, we could decide that the person has the character trait of clumsiness. Having made this attribution, we tentatively assign other traits to this person to the extent that clumsiness implies the new traits. For example, we could believe that clumsiness is associated with stupidity. In that case, we would also expect to see "stupid" actions from this person.

Researchers call this tendency the "halo effect" when it applies to "good" attributes and the "horns effect" when it involves "bad" character traits. For example, in one study, researchers showed participants photographs and then asked them to make judgments about the person photographed. The participants saw more attractive people as being more intelligent and more exciting to be with than less attractive people. The "good" characteristic "attractive" led to the assignment of other "good" characteristics to the person. The halo effect is not all-powerful, however. If you later recognize "smart" behavior, for example, from the person you judged "clumsy," you can change your impression of the person. However, the person you thought clumsy must behave clearly in a way that is "smart," for it is difficult to alter initial impressions.

We gradually build up a unique impression of a person as time passes. This impression comes from our own observations of the person and from our interpretations of what we hear about him or her from other people (Hewes et al., 1985). As we have shown, various factors and situations come into play as this "impression formation" process takes place.

Why is it that we use traits as descriptors for people? Why do we call people "dull" or "energetic"? We seem to want to describe a person's character. Why? There appear to be two reasons.

First, we feel a need to be able to explain and predict other people's behavior. Traits help us explain and predict. They can give us an intuitively satisfying explanation for behaviors. For instance, Jacob dropped a glass because he is "clumsy." We do not need to worry about a further explanation. In addition, traits are a device we can use to predict later behavior. Because Jacob is clumsy, we better not invite him into a china shop.

Second, we use traits because they serve as a basis for evaluating a person. This is crucial to our interest in liking and admiration. For example, we may think badly of clumsiness. We believe that Jacob is clumsy. It follows that we have a reason to think badly of Jacob. Of course, we assign quite a few attributes to a person. The way that we evaluate one trait may contradict our evaluation of another trait. For example, we might also think that Jacob is "funny." We evaluate that trait as being a good attribute. Because we like "funny" people, we may think well of Jacob because he is funny. The question is, Which trait will we value the most? Will we end up liking Jacob or disliking him?

Researchers have tried to predict our overall "liking" of a person who has several traits that we evaluate differently. Quite a lot of work has gone into developing algebraic formulas to answer those questions. The best-known model holds that whether we like a person depends on the average of the evaluations of the individual traits. Each evaluation is weighted according to the importance of the trait (Anderson, 1974). This model serves fairly well, but a few cases elude it. Sometimes combinations of traits lead to unique situations that do not occur when each attribute occurs separately.

For example, if someone tells you that Jill is "dumb" and "blonde," the two traits together may create a rather negative picture in your mind. Without meeting Jill, you may think her to be "flighty" or "irresponsible." However, if the person only says that Jill is "blonde," you might think Jill to be "friendly" and "exciting" because she is blonde. Similarly, the combination of "dishonest" and "generous" is another instance in which a unique combination changes things. These traits may lead to a picture of a Robin Hood-type character with a particularly positive overall evaluation.

Impressions and evaluations of the personalities of people are very important in a small group. They influence the manner in which group members interact. Certain impressions have considerable significance in the small group setting. One type is the attribution of intelligence and competence. Suls and Miller (1978) conducted a study concerning this. They had their participants take an exam that tested their "ability" at social psychology. They then rated them as either "very good," "good," "average," or"poor" at social psychology skills. The researchers asked the participants to choose the type of person with whom they would wish to form study groups. Not surprisingly, 93.1 and 89.4 percent of the participants in two respective studies wanted partners who were rated either "very good" or "good."

As we can see, this study points out that personality impressions play a major role in whether we admire a person. This factor can help determine whether "admiration" will affect liking among small-group members.

Reciprocation

As we have discussed, one of the most important factors in the degree to which Cisco, for example, likes Kristin, is the degree to which Kristin likes Cisco. Researchers call this process "reciprocation." There are many possible reasons that Cisco will tend to reciprocate Kristin's liking. The approval that Cisco gets from Kristin, as a sole factor, should cause Cisco to return Kristin's liking. In addition, Cisco can look forward to cooperation and support from Kristin. This adds to Cisco's reciprocation.

We can see how this could work in a small group. For example, the group leader may give compliments or criticisms to a group member. These might give the member an impression of the extent to which the leader likes or dislikes the member. The impression will, in turn, influence the member to reciprocate in a similar way. Indeed, many studies have used criticisms and compliments in order to support the tendency for degree of liking to be reciprocated. In fact, one study showed compliments and criticisms to be directly related to group cohesiveness. Dittes (1959) asked groups of participants to discuss a problem concerning juvenile delinquency, and interrupted the discussion on three occasions to allow members to rate how desirable they found one another as group members. After the discussion, each participant was shown what was said to be the other members' ratings of her or him. Actually, the participants were shown phony ratings that implied they were either accepted or rejected by the other members. After a second group discussion, the members were asked how much they wanted to remain in the group. Participants who thought they had been accepted were more attracted to the group than participants who thought they had been rejected.

However, we will not delve into these specific findings in detail. Instead, we will describe two instances in which the reciprocation effect does not occur.

Incongruency. The incongruency exception occurs when Person A believes that Person B's praise or blame is inconsistent with Person A's own self-appraisal. For example, Cisco may think poorly of himself. However, he believes that Kristin likes him. In such a situation, Cisco may believe that Kristin's judgment is faulty because he believes that she is wrong to like him. He will then dislike her as well as her opinion.

Deutsch and Solomon (1959) performed an experiment that highlighted the incongruency problem. They had their participants perform as "members" of two "teams." They then gave their participants phony evaluations of their performances, rating them either good or bad. Each participant next wrote a note to one of his or her "teammates." The researchers collected the real notes but gave fake notes back to the participants. These phony notes indicated that the "teammate" either did or did not want the participant on their "team" in the future. The participants next made written evaluations of the "teammates" who they thought had written notes to them.

Results of the study showed that participants thought a great deal about their own performance rating as they read their fake notes. Each participant kept in mind his or her "grade" in the experiment. Participants who believed that they had performed well liked their "teammate" the most if the note writer wanted the participant to stay on the "team." In contrast, these "good" performers liked the "teammate" least if he or she wanted the participant off the "team."

This process did not take place if the participants believed that they had not done well. These "bad" performers reported that their liking for the "teammate" was intermediate. It was unaffected by the "teammate's" evaluation of them. Neither liking nor disliking was reciprocated. It appears that when somebody knows that he or she has performed a task badly, the process of reciprocation is unaffected. The tendency to return verbal approval may be offset by the knowledge that the approver is wrong.

However, interpretation of this study does lead to some problems. Berscheid and Walster (1978) point out some of them. If Person B praises Person A after a bad performance, Person A may come to think that Person B is either stupid or insincere. These feelings will affect the reciprocation process in a different way from the incongruency exception. The possibility that Person B is insincere leads us to the second exception regarding reciprocation.

Ingratiation. The ingratiation exception occurs when Person A comes to believe that Person B's praise is due to some ulterior motive for gaining Person A's favor. For example, Cisco believes that Kristin is praising him in order to get later rewards from him. The ingratiation exception is most likely to occur if Cisco has reason to believe that Kristin's words are false. This is most likely to happen when Cisco is confident about the degree to which he possesses the characteristic on which Kristin's praise is based.

For example, it is useless to compliment people on their looks if they know that they are good-looking. They have heard it before and probably do not want to hear it again. It is equally useless to compliment people on their looks if they know for certain that they are not good-looking. They will know that the compliment is a lie.

There are perhaps times when a person can succeed in using false praise to be ingratiating. The most successful time would be when the person being praised does not know their qualities well. For example, perhaps a person does not know whether he or she is really good-looking (Jones, 1964). In such a case, ingratiation may create liking. The person does not perceive that the compliment is false. He or she does not suspect any ulterior motive and liking can occur through reciprocation.

Similarity

Newcomb included "similarity" within his typology system. He believed it was a cause for liking. The idea refers to the pleasure that one feels when interacting with someone who has similar beliefs and opinions as oneself. Attitudinal agreement, along with reciprocity of liking, is particularly important in the establishment of casual friendships. As such, it is extremely important in the small-group setting. Small groups often involve casual relationships.

Newcomb (1960, 1961) explored this issue in a classic study of the acquaintance process. For the study, Newcomb obtained the use of a boarding house near the University of Michigan campus. On two occasions he invited 17 male transfer students to live there for a semester free of charge. In exchange, Newcomb required them to participate in four to five hours of research a week. He performed various studies with the groups. Some of the experiments explored the patterns of the friendships that formed and the reasons for the friendships.

Early in the semester, functional distance and reciprocation of liking were the major determinants in liking. This was expected. Students were attracted to students who were close to them and who seemed to like them in return. However, attitude similarity became a critical factor as the semester progressed. As the students got to know one another better, they learned how they all really felt. Opinions came out into the open. As this happened, the students started to like others who shared their attitudes. They also lessened in their liking for those who did not agree with them about topics.

The relationship between the extent to which people believe they agree with one another and the extent to which they like ne another is extremely strong. Byrne (1971) and his coworkers conducted a series of studies that showed how strong that relationship can be. Byrne's research centered on a basic experimental procedure. For his studies, participants rated their attitudes toward each of 26 political, social, and religious issues. Several weeks later, the experimenters gave the participants a set of ratings for these same issues. They told them that the ratings were those of a stranger they were to meet. The researchers then asked

the participants to rate their liking for the unknown person and their willingness to work with him or her on a project. The participants also rated the stranger on other characteristics, such as intelligence and character.

What the participants did not know was that there was no actual "stranger." Instead, the researchers had given them a list based upon the participants' own opinions. The experimenters wrote down opinions that agreed or disagreed with the attitudes of the participants in a predetermined way. The lists ranged from total agreement, through proportions of agreement, to total disagreement. The study found that the degree of liking of the unknown person was strongly related to the proportion of statements on which the participant and the "stranger" agreed. In addition, greater accord led to a halo effect. Participants also rated an "agreeing" stranger as more intelligent, well informed, moral, and well adjusted than a "disagreeing" person.

In his later studies, Byrne found this relationship to hold for people of all ages and socioeconomic groups. It also was true for people from different countries and even for hospitalized schizophrenics. In all the cases, perceived agreement led to a liking of the "stranger." However, as with any such relationship, there are exceptions.

One exception is that, if the topic of agreement or disagreement is not important, it has no effect on liking. This is particularly true if the people interacting talk about other topics that they consider important. For example, Cisco and Kristin might disagree about their favorite colors, but they agree about the person they should elect president. The topic of their favorite colors is not important, and it does not figure into whether they like each other. The topic of the presidential candidate is important to them, however, and because they agree about that, they probably will like each other.

Another exception that Byrne found was that agreement with a person with undesirable qualities has no effect on liking. For example, you may agree with a drug addict who says that nuclear disarmament is good. However, if you do not like the life-style of a person dependent on drugs, you will probably not become friends with the drug addict even though you agree with his or her opinion.

Similarity in opinions has been directly related to group cohesiveness. Festinger (1954) reported a study he performed with Gerard, Hymovitch, Kelley, and Raven in which groups studied a labor dispute and then evaluated the union's decisions in the dispute. Afterwards, each member was asked how much they thought other group member agreed with them about the dispute and how much they attracted they were to the group. Those members who thought they agreed with the other members were more attracted to the group than hose members who thought they disagreed.

Conclusions. Liking is the most important element in maintenance-based group cohesiveness. It is a powerful force that influences whether people want to be with one another. As such, it has an impact on whether a group is drawn together cohesively. Research into the topic of liking has revealed a variety of factors that affect the processes of how and why one person finds another person likable. All of these factors come into play in a group setting. There are other forces, however, that affect maintenance-based cohesiveness.

Identification with the Group

Sometimes, our membership in a group becomes a very important part of our self-identity. For example, Karintha is a member of a basketball team called the {Sharks}. When she thinks about herself, one of the first things that comes to Karintha's mind is "I am a Shark." In this situation, we can say that Karintha identifies with the Sharks. According to Hogg (1992), when we identify with a group we look favorably upon the other members of the group, even those we do not find particularly likable. This is particularly true when circumstances make us think about our group membership. Nadine is also a member of the Sharks. Karintha does not find Nadine especially likable. However, if Karintha hears an outsider criticizing Nadine's performance as a Shark, Karintha is likely to jump to Nadine's defense. According to Hogg, although Karintha does not like Nadine, Karintha identifies with Nadine because they are members of the same team. Hogg called Karintha's feelings toward Nadine social attraction to distinguish it from regular liking.

Hogg claimed that when we identify ourselves with a group, we form in our heads a prototype of the "ideal" member of the group. For example, Karintha might see the "ideal" Shark as skilled, hard-working, and supportive. Hogg believed that we become socially attractive to other members of a group we identify ourselves with to the extent that we believe these other members approximate the "ideal." Recall our recent discussion of personality impressions. We can imagine that if Karintha's impression of Nadine is as a skilled, hard-working, and supportive team member, Karintha will be socially attracted to Nadine. In contrast, if Karintha sees Cecilee, another Shark, as far from the ideal, Karintha will not be socially attracted to Cecilee and would be less likely to defend Cecilee from outside criticism as Karintha defended Nadine.

Hogg and associates (Hogg & Hardie, 1991; Hogg, Cooper-Shaw, & Holzworth, 1993) have done some studies to try to distinguish social attraction from normal liking. They have used both real pre-existing groups (a sports team, an educational organization) and experimental groups in these studies. In these studies, they have asked group members to list the three group members they thought were closest to the group prototype, the three members they liked the most, and the three members they most wanted to be in their group. As Hogg expected, they generally found that there was more overlap between the members seen as "prototypical" and the members they wanted to be in their group than there was between the members seen as prototypical and the member they liked. This effect was strongest for those members who particularly identified with the group.

Hogg felt that liking and social attraction have separate effects on group cohesiveness. It is possible that each is a more important factor than the other in determining group cohesiveness in different circumstances. It would follow from Hogg's claims that when group members identify strongly with their group, cohesiveness would depend more on how much social attraction group members have for one another than on how much they like one another. In contrast, when group members do not identify with their group very much, cohesiveness would depend more on mutual liking. Research to evaluate these ideas needs to be performed before we can have confidence in them.

Psychological Needs

There has always been speculation that people may be attracted to a group in order to satisfy some deep-seated psychological need. As we discussed in Chapter 1, McClelland (1961) hypothesized that some people have a particular high need for affiliation with others. It follows that people high on this need would be likely to want to be a member of a group. Thus the extent to which group members have a need for affiliation would be a factor in the amount of maintenance-based cohesiveness in their group.

Similarly, McClelland also believed that some people have a particular high need for power. One can imagine that a person who has a strong desire to dominate others may join a group in order to take a leadership role. Being leader would satisfy his or her need to be in charge. The group will continue to be attractive only as long as the person can successfully dominate the other members.

A group member's desire to satisfy such psychological needs may indirectly affect cohesiveness through its direct effect on group process. Imagine that a group contains a member who has a very strong desire to dominate. Picture what his or her communication would be like. As you can envision, this communication would probably have a great effect on the other members' attraction to the group. What exactly would this communication be like? Would the group member who desires to be in charge try to make the other group members feel worthless? Or would he or she feel secure as leader and try to make the group attractive to the other members?

Communication and status needs. A study by Kelley (1951) suggests some answers to these questions. He theorized that people desire to have high status in their groups. His idea was that most people have the psychological need to have status. Kelley further hypothesized that this desire affects the type of communication that group members exchange. He divided eight-member groups into two four-member subgroups. The researcher then told the subgroups that one of them would receive a series of patterns of rectangles. He gave them instructions that they were to write messages to the second subgroup that would help the second subgroup reproduce the patterns. The researcher also said that group members could write notes to one another within their subgroups. The messages did not have to be limited to technical information. The members of each subgroup then went into different rooms. The researcher said he would give them their assignments and take their messages back and forth for them.

The participants thought the experiment would be as the researcher had described it to them. In actuality, he never delivered the real messages. Instead, the experimenter collected all the participants' notes and gave out phony ones he had written himself. Further, he assigned both groups the task of reproducing patterns. There was no "sending" subgroup. Most important, Kelley told participants about the status and the permanence of their jobs. He assigned each participant a job, as a message translator or a pattern reproducer. He told them that the former was a high-status job and that the latter was a low-status task. He also told each member whether the job was subject to change (mobile) or permanent (nonmobile).

Kelley then analyzed the messages that the participants sent one another. The notes sent to the fictional "sending" subgroup revealed differences among the group members. The researcher determined that the high-status/nonmobile and the low status/mobile participants sent more "cohesiveness-building" messages than other members. Their notes contained overtures to friendship, encouragement, and praise and other such statements. Kelley interpreted these results under the assumption that the participants wanted to have high status if they felt they could attain it. The high-status/nonmobile participants knew that they would retain their status, and so were comfortable sending cohesiveness-building messages to the other subgroup. The low-status/mobile participants probably thought that sending these messages might help them gain high status. In contrast, Kelley believed that the high-status/mobile participants probably felt their status threatened and so refrained from encouraging the other members. Finally, the low-status/nonmobile participants were resigned to their fate and saw no point in sending cohesiveness-building messages. These latter two group members sent far fewer messages that were designed to build cohesiveness.

Interestingly, within the subgroups these patterns changed. In their own groups, the mobile members sent more "cohesiveness-building" messages than the nonmobile members did. This was true no matter what their status was. Unfortunately, it is impossible to determine the effect of the members' notes on one another. Because no member ever received the actual notes, the study could not examine whether the notes affected the receiver's attraction to the subgroup. However, the study did reveal that group members communicated differently, based on their status and ability to change that status. These different types of communication could affect group cohesiveness.

Evaluation needs. There is also some evidence that people want to be with others in order to learn more about themselves. Festinger's (1954) theory of social comparison proposes that people have a desire to evaluate their own abilities. We will be explaining this theory more fully in Chapter 6, "Conformity and Deviance," but Festinger's idea has applications for this section also.

Singer and Shockley (1965) conducted an experiment to test Festinger's idea. They asked 38 participants to complete a phony task. The participants then received a number representing their "score" on the task. The researchers gave 24 of them an interpretation of the meaning of the score. Out of these 24, only 2 chose to wait with other participants while the researchers set up a second part of the experiment. In contrast, 6 of the 14 who had not received an interpretation of their scores wanted to wait with other participants. They could then talk among themselves. It appeared that the people who did not know how well they had done felt the desire to associate with others who had done the task.

Conclusion. As we have discussed, group membership in itself may be attractive to some humans because membership can satisfy various psychological needs. In particular, needs for affiliation, power and status, and knowledge about themselves may draw people into groups. Therefore, needs such as these can be a factor in maintenance-based cohesiveness.

Task-Based Cohesiveness

As we mentioned earlier, task-based cohesiveness can be the result of group members' desires to achieve either group or personal goals or the members' attraction to group activities. Studies have been performed in order to understand the impact of these factors.

Group Goals

A group might attract a person if the group has an attractive goal. A classic study reported in Sherif and Sherif (1953) showed this tendency at work. The study took place at a summer camp for 12-year-old boys. Experimenters first established intergroup hostility between the boys. The researchers then strove to evaluate various methods for reducing this hostility.

The study divided the boys into two groups, called the "Bull Dogs" and the "Red Devils." Each group lived in a different part of the camp and performed its necessary activities, such as food preparation, separately. To create intergroup hostility, the researchers arranged games, contests, and circumstances in such a way that one group interfered with the activities of the other. These methods succeeded in arousing a great deal of animosity. Name-calling and fights broke out.

The experimenters next tried to reduce this hostility by first merely arranging for contact between the groups. This had no effect upon hostility. They then arranged tasks that required intergroup cooperation. The researchers secretly broke the entire camp's water system and then arranged for the groups to fix it together. They also had the whole camp raise money to go to the movies. In addition, the experimenters chose a camp-wide baseball team to play a team from another camp. These cooperative tasks eventually led to some reduction in hostility. Further, there was some identification with the camp as a whole.

This study demonstrates that, even within this atmosphere of hostility, group goals had the power to affect cohesiveness. The campers had the desire to reach camp goals. This led, first, to attraction to their own group and, next, to some attraction to the larger group of the whole camp.

Personal Goals

A member may also be attracted to a group because it provides the means for satisfying goals unrelated to any of the group's expressed purposes. For example, Fred might decide that he wants to be his college's class president. He has heard that five out of the last six class presidents were members of a certain fraternity. The fraternity's members are popular and get a lot of exposure on campus. Fred finds the idea of membership in the fraternity very attractive because membership would help him win the election for class president.

Deutsch (1959) performed an experiment based on this idea. He reasoned that the chance to receive a personal prize when the group is successful would increase a person's attraction to a group. He further believed that the greater chance of success, the greater the attraction would be. Deutsch worked with Air Force cadets in his study. The cadets received personal rewards of money or a three-day pass if their group performed the task successfully.

Researchers believe that three factors affect a member's perception of the chance of his or her group succeeding at a task. They are

1. the probability that the group can win the prize,

2. the motivation of the other group members to succeed, and

3. the group's ability to perform the task.

Deutsch's study used these three factors to create various test situations.

Deutsch formed three-person groups. The experiment consisted of the following sequence:

1. The researcher told the cadets that they would have either a 90- or a 10-percent chance of winning the prize. The experimenter thus manipulated the factor regarding perceived probability of winning.

2. The participants wrote brief notes to one another about their feelings concerning working in the group. The experimenter collected these notes.

3. The cadets took a phony "group intelligence" test. This supposedly predicted their ability to perform their assigned tasks.

4. The researcher gave the cadets phony notes, supposedly written by other members. These expressed either desire or dislike for working in the group. The experimenter thus manipulated the factor regarding the perception of other members' motivation.

5. The participants received the results of the phony group intelligence test. The researcher thus manipulated the factor for perceived group ability.

6. The participants performed five tasks of varying types.

7. The cadets rated their feeling about the group and their performance.

The results of the study supported Deutsch's idea. The participants were attracted to the group if they thought the group could help them win their personal goal. High perceived group motivation and ability, and in some cases perceived probability of winning, led participants to want to stay in their group more often than not. In addition, these members felt more team spirit, felt more obligated to the group, believed the group helped their performance, and gave a higher rating to the group's performance than subjects who felt that they were in less capable groups.

There is a corollary within the idea that a group is more attractive if it helps a person meet a personal goal. The corollary is that a group will be more cohesive if its members' personal goals coincide. Deutsch (1949, 1973) performed an earlier study that highlighted the way this process works. He assigned groups of his students various tasks. Next, he told the groups that their work would be graded either cooperatively or competitively. In the cooperative groups, each member of the group would receive the same grade, based on the entire group's performance. For the competitive groups, Deutsch said he would compare the individual performances of each member and grade them against one another.

In comparison with the competitive groups, the cooperative groups showed more friendliness and helpfulness in group discussion. They were also more satisfied with the group as a whole and with each member's contributions to the organization. In addition, cooperative group members felt a greater desire to win one another's respect, as well as feeling more of an obligation to help one another.

It is probably not that surprising to find that members are attracted to groups that help them achieve their goals. The findings of a study by Shaw and Gilchrist (1955) may be more surprising. They indicate that there is a limit to this effect. In the study, one pair of confederates, who were collaborating with the experimenters, and one pair of participants worked at a series of four tasks. These were rigged so that the confederates would continually "succeed." In contrast, the participants would always "fail." After each task, the researchers asked the participants with whom they wished to perform the subsequent assignment. After the first task, 36 out of 48 wanted to work with one of the "successful" confederates. However, this number dropped to 25 after the second assignment, and to 21 for the last two tasks. There are at least two possible interpretations of this finding. One interpretation could be that people who constantly fail may come to prefer working with other consistent failures. Another explanation might be that people may still wish to succeed but come to prefer working with people to whom they are accustomed.

Attraction to Group Activities

People also join groups if the groups do the sorts of things that the prospective members like to do. Cohesiveness increases when people do things that they enjoy in groups. Thibaut (1950) conducted an experiment to test if a group's activity in itself can affect cohesiveness.

Thibaut's study divided preexisting groups of boys from summer camps and placement houses into two teams. The teams played a series of four games. The games required the two teams to perform different activities. In all four cases, the researcher gave the same team the more pleasant of the two tasks. For example, in one game the "privileged" team got to throw beanbags at a target. In contrast, the "underprivileged" team had to hold up the target and retrieve the beanbags. The boys were already in groups, so the experiment tested the effect of the activity itself upon cohesiveness.

Thibaut asked the boys with which members of the entire group they preferred to play. He asked them twice, both before and after the games. If they chose more members of their own team after playing the game than before, one could conclude that the cohesiveness of the team had improved. Results showed that the "privileged" team increased its cohesiveness markedly. The "underprivileged" team also became more cohesive, but to a lesser extent. The group activity had affected cohesiveness.

Toward A General Theory of Cohesiveness as Output

Putting everything together, we have a start at a general theory of the factors affecting the amount of cohesiveness in a group. Factors such as liking for other members, identification with the group, and the satisfaction of psychological needs serve as input variables affecting member commitment and maintenance-based cohesiveness as output variables. Similarly, the opportunity to reach group and personal goals and participate in attractive activities serve as input variables affecting member commitment and task-based cohesiveness as output variables.

The problem is that very little research and theorizing has been performed about the role of process as the intermediary between the input and output factors. Clearly there must be certain types of communication that lead to greater or lesser amounts of cohesiveness. There are a few findings that suggest what these types might be. In Chapter 2 we discussed the research Bales and his associates performed using groups discussing "human relations" problems. Bales (1953) compared the most and the least satisfied among sixteen of these groups, and found that the most satisfied of the groups agreed with one another more, disagreed with one another less, and showed less tension and antagonism than the least satisfied of the groups. Much more work needs to be done about this issue before we can make any definitive claims about the role of communication as a factor in group cohesiveness. Until then, we cannot propose a true input-process-output theory accounting for cohesiveness as an output variable.

Nonetheless, we do have some general ideas about how a group can produce cohesiveness as an output variable. In addition, cohesiveness can also affect the group as an input variable. We will discuss this process next.

COHESIVENESS AS AN INPUT VARIABLE

Thus far we have conceived of cohesiveness as a result of several factors. As such it is an output variable. There is good reason for this. When a group first begins, there is little or no cohesiveness. However, once a group has been together for some time, some level of cohesiveness is established. This level of cohesiveness will then become a factor in subsequent group discussion. Hence, it will become an input variable. As such a variable, it will be mediated through group discussion. When this takes place, it can affect all of the output variables we have discussed in this book. Cohesiveness can even indirectly affect itself.

We have distinguished between maintenance-based and task-based cohesiveness. As we shall see, when serving as input variables, these two aspects of cohesiveness have different effects on group process and output.

Cohesiveness, Communication, and Social Influence

Lott and Lott's Study

It stands to reason that members of groups that are cohesive are likely to talk with one another more than members of noncohesive groups. It also stands to reason that group cohesiveness would affect the process of "social influence." We will describe "social influence" further in Chapter 7. For now, we can simply say that "social influence" refers to the ways in which group members influence one another's beliefs, attitudes, and behaviors. Cohesiveness should affect attempts at social influence during group discussions, and these attempts should alter group members' attitudes after the discussion is over.

Lott and Lott (1961) presented evidence that seems consistent with these claims. They asked 15 ongoing campus group of 6 to 10 members to discuss whether or not students are too contented, lazy, and self-centered. They measured how much communication occurred. They also asked group members' opinions about this issue both before and after the discussion, to see if the discussion led to the members changing these opinions. Finally, they asked group members how much they liked one another in order to measure group cohesiveness. They found that high liking among the members led to greater amounts of talk about the issue under consideration. In turn, there was a significant group influence on member opinions concerning the issue. Members' opinions became closer to one another after the discussion in high liking groups.

Unfortunately, Lott and Lott's study underestimates the complexity of the relationship between cohesiveness, communication, and social influence. As mentioned earlier in this chapter, Lott and Lott in another article claimed that liking among members is a sufficient measure of group cohesiveness. Unfortunately, this claim overlooks the distinction between maintenance- and task-based cohesiveness.

Back's Study

A classic study by Back (1951) illustrates the importance of this distinction. Back wanted to see if various factors that cause cohesiveness would affect social influence differently. He used three different manipulations to induce cohesiveness in his dyads, or two-member groups. In the "liking" manipulation, he told the participants that they would either definitely get along with their dyad partner, that they would probably get along, or that they might irritate one another. This first manipulation was clearly an attempt to encourage high versus low versus no maintenance-based cohesiveness. In the "task" manipulation, he either told the participants that the best individual task performer would receive five dollars for their performance or just encouraged them to do the task well. This second manipulation was clearly an attempt to encourage high versus low task-based maintenance. In the "prestige" manipulation, he told the participants that, based on their classwork, they should be the best group of all or just a good performing group. It is not clear whether this is a task or maintenance-based manipulation.

In the actual study, Back gave his participants a series of three photographs. He then told them to write a preliminary draft of a story explaining and connecting the photographs. After completing the story, the subjects returned the photographs and then met in groups of two, or dyads. The researcher told the participants that they had written stories about the same pictures and that they were to exchange information so that each could do a better job on subsequent, final stories explaining the photographs. In actuality, each member of a dyad had seen slightly different pictures. This meant that some disagreement between the dyad members was likely.

Back analyzed each dyad's interaction to see how much the members communicated and how they attempted to influence each other. He then compared each participant's preliminary and final stories to see whether the partners had actually influenced each other. The findings depended on the type of cohesiveness manipulation. Let us look at these results for different variables in turn.

Amount of communication. Back measured amount of communication according to how many seconds the dyads spent talking. The results can be found in Table 3.1.

Table 3.1 Type of Manipulation
Cohesiveness Liking Task Prestige
High 449 321.5 362.5
Low 412.5 415.5 307
None 330

Not surprisingly, the more maintenance-based cohesion in the group, the more the group members talked to one another. Even the dyads told that they would "probably" like one another talked for a relatively long amount of time. This is consistent with the findings of Lott and Lott's (1961) study. However, low task-maintenance dyads talked more than high. These participants seemed to view each other merely as tools for achieving good task performance, and the high task-maintenance dyads seemed to feel they did not need to talk much to perform the task well. The hi and low "prestige" groups seemed to act like the "liking" groups in this regard, although they talked a lot less as a whole.

Social influence. Overall, highly cohesive dyads of all three types engaged in more attempted influence than low-cohesive dyads. The attempts at influence involved stating their own positions, arguing, and reasoning with each other. Similarly, the high-cohesive dyads had more verbal reactions to these influence attempts than did the other dyads. The cohesive partners would agree or disagree with each other, counterargue, and reject each other's stories. Both attempted influence and verbal reactions to these attempts were highest for the "task" manipulation groups, providing more evidence that members of these dyads used each other as tools for achieving good task performance. Finally, there was more actual influence of one another's stories in the high cohesive dyads than the low for the "liking" and "task" manipulations, although cohesiveness had no effect on actual influence in the "prestige" dyads.

Implications. Back's study leads to some interesting further implications involving the effect of cohesiveness on interaction. It would seem to be intuitively reasonable to predict that the more cohesive a group is, the more "friendly" its conversation will be. One could also expect that such a group will have greater cohesiveness in the future. These predictions do contain some truth. However, Back's results clearly show that they are too simplistic. First, this expected finding did not occur in the "task" groups. Second, the high cohesive groups were more argumentative than the less cohesive groups. Their discussions seemed to arouse a great deal of conflict.

Why would this be so? It may be that highly cohesive groups are not afraid to argue. They know that they are strongly "glued together," and that some conflict will not tear them apart. In contrast, a noncohesive group may be afraid that a fight will split them apart, and so be hesitant to argue.

In fact, because of their feeling of being "glued together," a cohesive group should be more likely to engage in any type of unpleasant communication. A study by Pepitone and Reichling (1955) is relevant to this point. Members of dyads were either told they would or would not like one another. Then, a person introduced as "the experimenter" insulted both participants. While subsequently performing a task, the dyads composed of members who had been told they would like one another spent more time expressing hostility about the experimenter and their task to one another than the dyads composed of member who had been told they would not like one another.

What we should expect is for communication among members of cohesive groups to be alternatively more friendly and more argumentative and unfriendly than that in noncohesive groups. In fact, in Chapter 8, "Group Process," we will be describing work by Robert F. Bales that provides an explanation for why this should occur.

Cohesiveness and Satisfaction

Exline (1957) performed an interesting study regarding the role of cohesiveness as an input variable. His experiment is problematical, but it does shed more light on the question of how cohesiveness affects a group. Exline manipulated maintenance-based cohesiveness by telling his participants that, based phony personality measures the subjects had filled out earlier, they would be "congenial" or not be congenial. Exline next had his participants take part in a "role-play" exercise. This was an improvised dramatic situation in which each participant had a secret role to play. In comparison with the less congenial groups, the more congenial groups had members who stated that they had more liking for one another. They also had a greater desire to work together. In addition, they were more satisfied with their performance and better able to guess one another's assigned role in the exercise. The findings appeared to reveal that cohesiveness has a positive effect upon group member satisfaction.

There are problems with this study, however. Exline did not analyze the group's exercises. Therefore, we cannot get a clear picture of the connection that may have existed between the input of cohesiveness and the output of satisfaction. On the one hand, it could be that the more congenial groups actually did do a better job of performing the exercise than the less "congenial" groups. On the other hand, it may be that the more congenial groups did not actually perform better. Perhaps they were more satisfied only because of their cohesiveness. Exline's results do suggest that there is a connection between cohesiveness and satisfaction. However, his study is an example of the ambiguities that result when researchers ignore group process. We need to know more before we can make definite statements.

Cohesiveness and Task Productivity

Our next concern regarding cohesiveness as an input variable involves the relationship between it and task productivity. For many years, this relationship was considered to be complex and somewhat unclear. Dozens of studies have been performed since the early 1950s, with contradictory findings. Sometimes cohesive groups were more productive than noncohesive groups, sometimes they were less productive, and sometimes cohesiveness did not seem to affect productivity at all.

Recently, it has become evident that considering the distinction between task- and maintenance-based cohesiveness clarifies much of this confusion. A study by Zaccaro and Lowe (1988) reflects this new understanding. The researchers formed four-person groups to perform a productivity task. Before the task, the experimenters carried out manipulations to induce either high or low maintenance-based and task-based cohesiveness. First, half the groups participated in an exercise designed to generate maintenance-based cohesiveness. The group broke up temporarily into dyads who talked to one another, and then all four group members took turns introducing their dyadic partner to the two other members. The other half of the groups did not perform this exercise, leading to lower maintenance-based cohesiveness. Second, half of the high and half of the low maintenance-based cohesive groups were told about the importance of good performance on their upcoming task and offered extra credit if they were the best performing group. This manipulation was designed to enhance task-based cohesiveness. The other groups were not told anything, resulting in lower task-based maintenance. As a consequence of these manipulations, the groups were either high or low on task-based cohesiveness and either high or low on maintenance-based cohesiveness. Finally, the group performed the productivity task, which involved folding as many sheets of paper into tent-shaped forms as they could in fifteen minutes. During the performance, the researchers measured how much communication occurred. The researchers also asked the participants how committed they were to performing their task well.

The results showed that groups high on task-based cohesiveness were more committed to task performance and more productive than groups low on task-based cohesiveness. In contrast, groups high and low on maintenance-based cohesiveness were equal in productivity. Interestingly, the groups that were high in maintenance-based cohesiveness were also highly committed to the task. However, consistently with the Lott and Lott (1961) and Back (1951), they were also extremely talkative, and the more talkative groups were less productive. This is just one of several studies that have found that talking hurts performance in productivity tasks. Thus the higher talking canceled out the effects of high motivation in the high maintenance-based cohesive groups.

A review of literature by Mullen & Copper (1994) shows that, in general, cohesive groups are more productive than noncohesive groups, particularly when group size is relatively small. However, this relationship is much stronger for task-based cohesiveness than it is for maintenance-based cohesiveness. Clearly, when groups are cohesive because their members care about their task, they will usually be more productive than groups that are not cohesive because their members do not care about their task. This relationship does not hold true when group cohesiveness is determined by liking or disliking. However, groups that are cohesive because of liking may still care about their task. Would this make them productive?

A study by Schachter, Ellertson, McBride, and Gregory (1951) addressed this issue.

They performed a study similar to the study by Kelley that we described earlier. However, in Schachter et al.'s experiment, the role of cohesiveness changed from an output to an input variable. In the study, the researchers asked three-women groups to work on producing cardboard checkerboards. The study divided the task into three parts. The task involved cutting the cardboard, pasting it on heavier stock, and painting the checkerboard pattern on it through a stencil. The experimenters said they would assign each participant in the group one of these tasks. The experimenters said that the participants would work in separate rooms, and they could exchange notes via messengers. Researchers also told the participants either that they definitely would or that they might not like their "coworkers." This was a successful manipulation of cohesiveness, based on later ratings for liking.

In actuality, the participants all did the same job. Each ended up cutting cardboard. The experimenters also intercepted their notes and replaced them with a standardized series of messages that they gave the participants every four minutes.

The first four notes were identical for all the participants. However, starting at 16 minutes into the study, the notes started to request that the women either speed up or slow down the cardboard cutting. These notes served as a manipulation to see how the group could influence each member. The researchers measured productivity by counting the number of checkerboards cut in eight-minute segments. The results of the study are shown in Table 3.2.

Table 3.2 Segment (in minutes)
Induction Cohesive 8-16 16-24 24-32
Faster High 5.31 8.23 11.23
Faster Low 6.16 9.08 11.25
Slower High 6.31 5.31 4.15
Slower Low 6.42 5.84 6.00

It appears that members in groups high in maintenance-based cohesiveness were susceptible to influence from the other group members. As a consequence, they aimed toward the group's apparent goal. They either speeded up or slowed down when the other members asked them. In other words, people in cohesive groups tended to follow the group's actions even if those actions hurt the task output. It was more important to go along with the group than to work productively if the group gave the message to slow down. In contrast, members of groups low in maintenance-based did follow the suggestion to speed up, but they did not slow down when other members asked them. They did not always do what the rest of the whole group wanted.

The implication of these findings is that groups high on maintenance-based cohesiveness are susceptible to influence from their group. If their group cares about task performance, they will be productive, and if their group does not care about the task, they will not be productive. In contrast, groups low on maintenance-based cohesiveness are less susceptible to influence from their group, so that whether the group cares about task performance will have little effect.

Cohesiveness, Decision Accuracy and Quality

The relationship of cohesiveness with decision accuracy and quality is not entirely clear. After reviewing past research, Mullen, Anthony, Salas, and Driskell (1994) concluded that task-based cohesiveness led to better decisions, whereas maintenance-based cohesiveness to worse decisions. If true, this finding would be similar to that for cohesiveness and productivity. As in that case, a group that is cohesive because its members care about the task would work harder and thus make better decisions than a group whose members do not care about the task. In contrast, groups that are cohesive because their members like one another may place more emphasis on getting along well with one another than in making a good decision, and as a result make worse decisions than groups whose members do not like one another. In fact, there is reason to believe that, under some conditions, getting along well becomes so important to highly cohesive groups that their members become hesitant to freely exchange ideas. In this circumstance, groups can make disastrously bad decisions. This condition is known as groupthink. We will discuss groupthink in detail in Chapter 12, "Decision Theory."

The problem with this general conclusion is that Mullen et al. did not distinguish between accuracy and quality tasks in their review. It is not clear that their conclusion is true for both types of tasks. For example, Zaccaro and McCoy (1988) performed a study in which they manipulated task- and maintenance-based cohesiveness the same way as in the Zaccaro and Lowe (1988) discussed earlier. Afterward, the groups performed a survival game accuracy task such as the ones we discussed in Chapter 2. Accuracy was highest for groups high in both maintenance- and task-based cohesiveness than in groups low in either or both types of cohesiveness. This finding suggests that accuracy tasks may be performed better by groups whose members like one another and care about their tasks. If so, then Mullen et al.'s generalization is false.

Toward A General Theory of Cohesiveness as Input

We are in far better shape to propose a general theory of cohesiveness as an input variable than we were for cohesiveness as an output variable, because we know more about the role of communication as process in this case. When cohesiveness is due to task factors, members of highly cohesive groups are more dedicated to their task than members of less cohesive groups. Communication is largely task-oriented and, as a consequence, task performance is better for highly cohesive groups. In contrast, when cohesiveness is due to maintenance factors, members of highly cohesive groups may or may not be dedicated to their task. If they are dedicated to their task, they will act like high task-cohesive groups and perform well. If they are not dedicated to their task, their communication will be largely maintenance-oriented, and they will perform more poorly on their task than noncohesive groups.

We see that, as an input variable, cohesiveness has an effect on the social influence process and on task performance. As such, it can be a powerful tool. Increased cohesiveness can lead to more successful groups. Can we influence whether a group is cohesive? In the next section we provide some recommendations.

PUTTING THEORY INTO PRACTICE

There are definite advantages for a group to have members that are attracted to it. Cohesiveness can improve the functioning of the group. Members of highly cohesive groups tend to trust and have confidence in one another. This leads to an atmosphere that allows the expression of conflicting points of view. Such free expression is necessary for the group to make its best possible decision. Cohesiveness also provides the potential for a group to work at its fastest and most productive levels. Finally, cohesive groups have members who enjoy being together. Interaction in a friendly group creates good feelings. The entire cohesive group experience brings satisfaction. These are valuable goals in and of themselves. We should not forget them even in the face of pressing task requirements.

There is however a bad side to cohesiveness. Cohesiveness that is solely a result of group members liking one another can lead to poor group performance. If the group is more concerned with maintaining good relationships or having a good time with one another, task work will suffer. Therefore, no matter what factors have led to cohesiveness, the group must be motivated to perform its task. If it is, the group should work well together.

As you can see, the cohesive group has a great many good qualities. Unfortunately, one cannot force a group to be cohesive. For example, members may honestly dislike one another, or its members not care about its task. There is little that can be done about this.

However, in most cases, a group can perform specific actions in order to increase its own cohesiveness. The following recommendations may be helpful:

1. The group can increase the amount of communication among its members. The success of this strategy, however, depends on the content of the communication. The strategy will backfire if the extra talk consists mostly of disagreement or expressions of dislike. However, keep in mind that once a group achieves cohesiveness, there is a danger of too much concentration on agreement and good feeling among members. The group must attain a proper balance.

2. The group can elevate the attractiveness of group interaction. For example, encourage some humor during group decision making. This will help counteract the anxieties and pressures brought on by attention to the task. A little fun at the beginning, and particularly at the end, of a meeting is good. In addition, have members perform the less interesting tasks together, as suggested in Chapter 2, even at the expense of productivity.

3. The group can heighten the perceived value of being a member of the group in three ways:

a. The group can accentuate the importance of its goals. Making it clear that the group goals are important to the individual members can increase the members' commitment to the group and enhance their effort.

b. The group can stress the interdependence among members as they perform tasks. Give each member leadership roles to perform (see Chapter 11, "Communication-Based Approaches to Leadership"). Praise and blame the group as a whole for good and bad performance. Do not focus on the individual members.

c. The group can emphasize its identity. Be sure everyone is aware of the group' s history. Talk about the common experiences that members have shared. Meet at the same place and time. Give the group a name. Perform social activities together. If you do not mind being a little extreme, invite competition with other groups. You can also enforce barriers, such as initiation rites, that people must cross in order to become group members.

These recommendations can help a group become more cohesive. The members of a highly cohesive group are well adapted to cooperate with one another and to compete with other groups. This ability to compete is an important aspect of cohesiveness. In the next chapter we will shift our attention from the dynamics within a group to the dynamics of conflict between groups.

SUMMARY

In this chapter, we agree with the theorists who have defined cohesiveness as the result of all the forces that attract members to their groups. As such a result, it is an output variable. Groups are more cohesive when their members are committed to the group. There are two general types of forces that lead to member commitment and group cohesiveness. These are maintenance-based factors and task-based factors. Because of this distinction, we can talk about maintenance- and task-based cohesiveness separately.

The most important factor in maintenance-based cohesiveness is the amount of liking among group members. For liking to occur, people must first meet one another. After this initial contact, people may like each other for several reasons. They may come to like each other because of characteristics they admire in their impressions of one another's personality. In addition, a person may like someone because he or she appears to like the person or because the person agrees with him or her on other important issues.

A second factor in maintenance-based cohesiveness is member identification with the group. When this occurs, member come to respect those other group members who they feel reflect the ideals of the group. A third factor in maintenance-based cohesiveness are members' psychological needs. These include needs for power, affiliation, status, and evaluation of one's beliefs.

Task-based maintenance is a result of group members' attraction to the group's goals and activities. These factors are shared by group members. In addition, there can be personal goals unshared by other group members that may attract people to groups.

Cohesiveness also serves as an input variable. As such, it affects group process and output. The effect of cohesiveness as an input variable can depend on whether it is task- or maintenance-based. In comparison with less cohesive groups, members of cohesive groups attempt to influence one another more often, and they find it more comfortable to disagree with one another. However, whereas groups high in maintenance-based cohesiveness tend to be more talkative overall than groups low in maintenance-based cohesiveness, groups high in task-based cohesiveness often talk less than groups low in task-based maintenance. In addition, whereas groups high in task-based cohesiveness perform their tasks better than members low in task-based cohesiveness, the same is true for maintenance-based cohesiveness only when members are motivated to perform their task. When group members are not interested in their task performance, maintenance-based cohesive groups are poorer at their task than noncohesive groups.