Sie sind auf Seite 1von 17

Group Dynamics: Theory, Research, and Practice 2005, Vol. 9, No.

1, 1531

Copyright 2005 by the Educational Publishing Foundation 1089-2699/05/$12.00 DOI: 10.1037/1089-2699.9.1.15

Group Decision Making and Perceived Decision Success: The Role of Communication Medium
Sylvia G. Roch
University at Albany, State University of New York

Roya Ayman
Illinois Institute of Technology

The authors investigated the role of communication medium on the relationship between team member perceptions and decision success. Seventy-three 3-person groups participated in a consensus intellective task either face-to-face (FTF) or via computer-mediated communication (CMC). The participants also assessed their groups decision success and team member competencies. CMC group members success perceptions signicantly predicted their groups performance, but FTF group members perceptions did not. Furthermore, only CMC group members judgments regarding their groups problem-solving ability signicantly predicted their decision success. Last, judgments of decision success mediated the relationship between perceptions of members problem-solving ability and decision success only for CMC group members. Implications are discussed.

Two trends have revolutionized the workplace: the increased use of work teams and computers. Organizations are becoming increasingly team based, and work that was once accomplished by individuals is now delegated to teams (e.g., Goldstein, 1995). At the same time, employees are increasingly using computers to communicate with each other (e.g., Howard, 1995; Ilgen & Pulakos, 1999). This increased reliance on computer-mediated communication has practical advantages, especially in the context of teams. For example, it frees team members from geographical and time constraints (Hollingshead & McGrath, 1995). Where once team members were constrained to meet in a single geographical location, they can now meet from the comfort of their ofces, free from geographical limitations. It is now feasible that a team will never meet in person but have members scattered throughout the world. E-mail, chat rooms, and teleconferencing

Sylvia G. Roch, Department of Psychology, University at Albany, State University of New York; Roya Ayman, Institute of Psychology, Illinois Institute of Technology. This research was supported by Illinois Institute of Technology Research Grant 2330. We thank Susan Adams and Sandra Nelson for their assistance in data collection. Correspondence concerning this article should be addressed to Sylvia G. Roch, Department of Psychology, University at Albany, State University of New York, 1400 Washington Avenue, Albany, NY 12222. E-mail: roch@albany.edu 15

have allowed a level of communication between geographically dispersed teams that was not possible 20 years ago. Furthermore, because of the potential for teams to include members from diverse geographical locations, many of the teams will consist of virtual strangers who are assembled for the time it takes to make a decision or complete a task (e.g., Hinds & Weisband, 2003). Furthermore, teams, regardless of whether they are face-to-face (FTF) or computer-mediated (CMC), are often not aware of whether their decision is the correct one. In most cases outside the laboratory, the correct answer is not available; if it were available, there would have been no need to convene a team to make a decision. It is only after the decision is implemented that teams may receive a sense of whether they made the correct decision. Thus, it is valuable for teams to have some sense of whether their decision is a good one before it is implemented. The groupthink literature (see Turner & Pratkanis, 1998, for a review) is full of examples of teams that made decisions later shown to be unfortunate, to say the least, such as those responsible for the Bay of Pigs and the launch of the space shuttle Challenger. Thus, it is valuable for teams to have some sense of whether they have made a decision that will be successful before it is implemented with undesired consequences. For example, in a work context, if a team believes that its decision has

16

ROCH AND AYMAN

a low probability of success, even though it may have been the best decision in the given circumstance, it may be worthwhile for the manager to reframe the situation or to provide additional resources until the team believes that its decision has a high probability of success. This, of course, assumes that the teams estimate of its probability of success is realistic. Whether a teams estimate of success is realistic may be partly determined by how well the team members evaluate the expertise available to the group. Empirical research has shown that group performance often is related not only to the expertise in the group but also to the ability of the team to recognize expertise (e.g., Einhorn, Hogarth, & Klempner, 1977; Libby, Trotman, & Zimmer, 1987; Littlepage & Silbiger, 1992). Littlepage and Silbiger (1992), however, pointed out that even though much research has focused on the groups ability to judge individual member expertise, not much research has investigated a groups judgments of its nal decision. If a group cannot adequately evaluate the quality of a decision it has reached, they noted, it runs the risk of implementing a faulty decision (Littlepage & Silbiger, 1992, p. 346). Thus, the present study focuses on how well group members perceptions of decision success predict decision success. Furthermore, it may be that in some contexts, estimates of success more closely mirror actual success. Littlepage and Silbiger (1992) found that larger groups were better at recognizing group expertise and that this ability, in conjunction with a greater amount of expertise, led to the increased performance of larger groups. However, the authors discuss the need to nd other factors that inuence the accuracy of group evaluations of their decisions. One such factor may be communication medium. Specically, the present study focuses on whether the predictive validity of success perceptions differs according to communication medium, FTF versus CMC, and on the role of team members perceptions of their groups ability in determining both perceived and actual decision success in both types of medium.

Communication Medium Research


The research investigating the effect of technological innovations on team functioning has not kept pace with the advances. A recent meta-

analysis (Baltes, Dickson, Sherman, Bauer, & LaGanke, 2002) investigating the impact of CMC on group decision making included only 22 published and 5 unpublished studies. The Baltes et al. meta-analysis used McGraths (1984) task circumplex, which divided group performance into four quadrants. One quadrant consists of choosing tasks, and this quadrant is further subdivided into intellective and decision-making tasks, based on the nature of the task. According to McGrath (1984), decisionmaking tasks do not have correct answers. In contrast, intellective tasks do have correct answers. McGrath further divided intellective tasks into eureka tasks (truth wins), non-eureka logic tasks, and non-eureka consensus tasks. In the eureka type of task, the correct answer is so compelling that the group immediately recognizes the answer. In the non-eureka tasks, the correct answer tends to be less compelling, based either on logic or on the consensus of experts. The results of the Baltes et al. (2002) metaanalysis showed that in intellective tasks, CMC groups make less successful decisions than FTF groups, but the effect was not large (r .24) and was based on only 16 studies. According to Cohen (1988), a correlation of .10 is a small effect, and a correlation of .30 is a medium effect. Furthermore, in an exhaustive narrative review of the literature investigating the effect of communication technology on work teams, Hollingshead and McGrath (1995) found only 8 studies that compared FTF with CMC teams performance in intellective tasks. They found that 5 studies reported worse performance for CMC versus FTF groups, 1 study reported better performance for CMC groups, and 2 studies reported no difference. Nevertheless, the Baltes et al. meta-analysis and the narrative review of the literature by Hollingshead and McGrath both arrive at the same cautionary conclusion: CMC may result in less successful group decisions relative to teams that meet FTF when completing intellective tasks. It appears, however, that this is not a large effect. The reason, however, why CMC groups at times perform worse on intellective tasks is not well understood. Baltes et al. (2002) based their hypothesis that CMC groups would perform worse than FTF groups on Straus and McGraths (1994) assertion that communication will most likely affect outcomes when tasks

PERCEPTIONS AND DECISION SUCCESS

17

require coordination, persuasion, consensus, and both the perception and expression of emotion. Baltes et al. argued that intellective and decision-making tasks have these properties, and these properties seem to be dependent on social cues. Therefore, because CMC lacks social cues, CMC groups will perform worse. It may be indeed that Baltes et al. (2002) are correct and that the lack of social cues underlies any performance difference between FTF and CMC groups. It may be that one cannot successfully argue for ones position and read nonverbal cues to determine members true positions without the social cues inherent in FTF communication, resulting in decreased performance. But the converse can also be argued: The lack of social cues may help to increase performance because social cues often convey information unrelated to the quality of the decision, such as information associated with gender, race, and age. Moreover, it may be that in CMC, persuasion is based on logic or facts rather than on nonverbal dominance.

Elaboration Likelihood Model


Petty and Cacioppos (1981, 1986) elaboration likelihood model (ELM) provides an explanation of why social cues can reduce decision quality. Petty and Cacioppos ELM model proposes that people can be persuaded either by carefully evaluating message content (central processing) or by peripheral cues, often independent of the content of the message, such as attractiveness, length of the message, appearance of the speaker, and so forth. Because the only information available to CMC group members is the typed word on a screen, they most likely process information centrally by evaluating the content of the message. Of course, there still are some peripheral cues associated with the written word, such as the writing style and the length of the argument. But because CMC groups are focused on the message and not the messenger, they most likely are not inuenced by these peripheral cues to a great degree. In contrast, FTF group members may be persuaded both by the content of the message (central processing) and by peripheral cues such as personality, gender, attractiveness, and speaking style, because the person making the argument is salient. As a result of the salience of the peripheral cues associated with FTF interaction,

FTF group members may consider these peripheral cues to a greater extent when making their decisions. As mentioned earlier, however, FTF groups do have a slight tendency to outperform CMC groups on intellective tasks (Baltes et al., 2002; Hollingshead & McGrath, 1995). This nding may be a function of the amount of information exchanged in the two types of groups. FTF groups may at times arrive at a more successful decision simply because they have the information necessary to reach the correct decision. Hollingshead and McGrath (1995) observed that the computer does not simply reduce the participation of loquacious group members or simply increase the participation of quiet group members; it does not democratize group discussion, as some researchers have implied. Rather, it reduces participation for all participants (p. 74). In the Hollingshead and McGrath (1995) review, 10 studies found less participation, that is, fewer verbal acts, for CMC groups in comparison to FTF groups, and only 2 studies found no difference. No study found more participation in CMC groups. For example, Straus and McGrath (1994) found that CMC groups had less total communication than FTF groups; Siegel, Dubrovsky, Kiesler, and McGuire (1986) found that CMC groups made fewer remarks; and Smith and Vanecek (1988) found that CMC groups shared less information. Other, more recent studies have also found that CMC group members exchange less information (Hedlund, Ilgen, & Hollenbeck, 1998; Hollingshead, 1996a, 1996b). In fact, Hollingshead (1996a, 1996b) refers to the lower frequency of communication in CMC groups as information suppression. Thus, the nding that CMC groups communicate less and, therefore, share less information than FTF groups is a very robust nding. This reduction in the total amount of information available may be especially important in intellective tasks, for which there is a correct answer. With decreased information, the information needed to determine the correct answer might not be available. Hedlund et al. (1998) also hypothesized that information exchange is critical when completing intellective tasks and that the lower frequency of communication in CMC groups inhibits effective decision making. They found support for this hypothesis.

18

ROCH AND AYMAN

It may be, however, that because CMC groups are less focused on peripheral cues, they also evaluate the available information more accurately. In other words, because CMC limits communication to the written word, CMC group members may be better judges of whether their group has indeed reached a successful decision. Thus, it may be that CMC group members are better judges of the success of their decisions than FTF group members because their attention is focused more on the logical or factual basis of the presented arguments.

Perceptions of Decision Effectiveness


Previous research has not directly investigated the relationship between perceptions of decision success and decision success. It has, however, investigated a related construct, decision condence. Perceived decision success addresses the question How successful is your decision?, but condence addresses the question How sure are you that you made a successful decision? Condence is dened as the strength of a persons belief that a specic statement is the best or most accurate response (Peterson & Pitz, 1988). These two concepts are closely related, and thus the literature on decision condence is relevant to the current study. As mentioned earlier, previous research has investigated the relationship between decision condence and success, and this relationship has been investigated on both the individual and the group level. Research focusing on the individual level has shown that ones condence that one has made the correct decision is often unrelated or only weakly related to decision success (e.g., Paese & Sniezek, 1991; Ryback, 1967; Sniezek, Paese, & Switzer, 1990). It appears, however, that decision condence in the group decision is a better predictor of group decision success than condence in ones decision is of individual decision success. Sniezek and Henry (1990) found that in an intellective task consisting of estimating the price of cars, individual condence levels regarding initial individual decisions were not signicantly related to decision accuracy but condence in the group consensus decision was signicantly, but not strongly, correlated with group decision accuracy. On the basis of these ndings, Sniezek and Henry (1989, 1990) proposed a two-stage

model for consensus group judgment. The rst stage, revision, consists of social interaction. During this stage, group members may voluntarily modify their individual judgments. The second stage, weighing, occurs either when the group has no more time to consider new ideas or when the group members no longer propose new ideas. At this stage, the group needs to determine a consensus rating that represents the often diverse individual judgments. According to Sniezek and Henry (1989, 1990), this consensus decision does not consist of an equally weighted average of the individual opinions: Some individuals opinions are given more weight than others according to the perceived quality of the contributions. Other research has also supported the importance of recognizing the expertise available to the group in determining group performance (e.g., Einhorn et al., 1977; Littlepage & Silbiger, 1992). This research provides additional insight for why CMC group members may be better judges of their groups decision. It may be that because during group discussion CMC group members are focused more on the logical and factual basis of the arguments (central processing) than on the personality behind the statements (peripheral processing), they may also do a better job of recognizing team member ability and knowledge. Hedlund et al. (1998) proposed a similar argument, albeit in a different context. They argued that the richness of FTF communication makes it more difcult for team leaders to appropriately weigh team members recommendations because of the increased salience of demographic characteristics, increased number of messages, and increased emotional content of the messages. They showed that in an intellective task, CMC group leaders showed greater hierarchical sensitivity (appropriately weighing group members recommendations) than FTF group leaders and that hierarchical sensitivity signicantly predicted decision success. It should be noted, however, that in their study, FTF group leaders did make more successful decisions, which they attributed to FTF groups superior team informity (available needed information; p. 32) and staff validity (validity of group members recommendations). Even though this was a structured task with distributed expertise and a designated leader, similar processes may occur in leaderless groups. CMC group members may be better judges of whether

PERCEPTIONS AND DECISION SUCCESS

19

the group has adequate information to make a successful decision than FTF group members, similar to Hedlund et al.s (1998) hierarchical sensitivity.

Hypotheses
In summary, CMC group members should be better judges of their decision success than FTF group members because their communication is limited to the written word, which should focus their attention on the quality of the arguments of their fellow team members and not on the social cues associated with various group members. Thus, in terms of Petty and Cacioppos (1981, 1986) ELM model, CMC groups most likely use mainly central processing and focus on the arguments made by the group members because they are not as distracted by peripheral cues. Thus, they should be able to more accurately evaluate the information available to the group, similar to the results of Hedlund et al. (1998). As mentioned earlier, even though Hedlund et al. examined group decision-making in the context of a structured task with distributed expertise and a designated leader, their results are relevant. They found that CMC group leaders showed greater hierarchical sensitivity (appropriately weighing group members recommendations) than FTF group leaders and that hierarchical sensitivity signicantly predicted decision accuracy. Even though Hedlund et al. (1998) did not explain their results in terms of ELM theory, they speculated that the CMC group leaders showed more hierarchical sensitivity because they were less inuenced by demographic characteristics, number of messages, and emotional content of the messages. Similar processes are likely to occur in a leaderless group. Because the arguments are more salient as a result of the decreased number and salience of peripheral cues in CMC groups versus FTF groups, CMC group members should also be better judges of their own and team members competence, expressed in terms of task knowledge, team ability, and problem-solving ability. For FTF group members, team competence perceptions may not only be a function of the information provided and the quality of the arguments but could also be inuenced by peripheral cues associated with the person presenting the argument to a

much greater degree than for CMC group members. Competence judgments that are less biased by peripheral cues should also allow CMC group members to determine whether they have the ability to make a successful decision. Thus, for CMC group members, perceptions of group competence should be directly related to perceived decision success, and perceived success should be related to decision success. In other words, for CMC group members, perceived success should mediate the relationship between perceived group competence and decision success. In FTF groups, perceived success should not mediate the relationship between perceived group competence and group decision success, because there should not be a signicant relationship between perceived competence and decision success. Given the nature of the communication in FTF groups, both perceived competence and perceived success should be based on interpersonal factors unrelated to decision success to a much greater degree than for CMC group members. Hypothesis 1: CMC group members success perceptions will be better predictors of the success of their decision than FTF group members success perceptions. Hypothesis 2: Only CMC group members competence judgments will predict their decision success. Hypothesis 3: Perceived success will mediate the relationship between perceived group competence and decision success only for CMC group members.

Method Participants
Data were collected from 228 undergraduate students who participated in 77 groups, with each group consisting of 2 men and 1 woman, from a private Midwestern technical university and a public Midwestern university. Gender was kept constant to control for potential gender effects. Three groups (9 participants) were eliminated because of technical difculties, resulting in 74 groups, 40 FTF groups and 34 CMC group. Furthermore, 8 participants were eliminated from the individual analyses because they

20

ROCH AND AYMAN

had spoken English for less than 3 years. Completing the measures required a high level of English comprehension, and limited experience with the language made the requisite English ability highly unlikely. Participants received extra credit for participating.

Materials and Measures


Computer chat room. A computer chat room created for this study was used to record the conversation and the time elapsed for the CMC groups. Decision task and success scores. The Lost at Sea task (Nemiroff & Pasmore, 1975) requires participants, taking the role of survivors of a recently burned down yacht in the South Pacic, to rank 15 items in order of importance for survival. Decision success was determined from the absolute difference between the groups ranking and the ranking determined from the consensus of experts (Nemiroff & Pasmore, 1975). Smaller scores indicated greater ranking success. McGrath (1984) identied survival tasks, such as this one, as non-eureka consensus intellective tasks. Perceived member competence. Participants were asked to assess the competence of both themselves and their two fellow group members on three itemsrepresenting member knowledge, member problem-solving skill, and member team skill on a 5-point scale (1 poor, 5 excellent) after the completion of the survival task. Each participant provided a rating for each group member on each of the three competencies. The perceptions regarding each type of competence were aggregated (averaged) within competency to determine a members perception of the combined knowledge, problem-solving skill, or team skill available to the group when reaching the consensus decision. The three types of competencies were not combined because they are conceptually distinct. Perceived decision success. Participants were asked, How successful do you think your group was at arriving at a high-quality decision? and were asked to individually provide a percentage. Participants were instructed at the beginning of the experiment that their rankings of the items in the survival task would be compared with expert rankings and that their rankings would be judged by how closely they resembled the expert rankings. Thus, in this case,

perceived success was determined by how closely the participants thought their groups ranking of the items matched expert rankings. Questionnaire. Along with unrelated items, the questionnaire contained items assessing demographics and perceived interests of group members in maintaining interpersonal relationships and successfully completing the task.

Procedure
This study was part of a larger study investigating group dynamics that included several variables associated with topics not related to the present study, such as leader emergence, group atmosphere, satisfaction, and member familiarity. Participants, recruited within undergraduate psychology courses, were randomly chosen to participate in one of the two conditions, FTF or CMC, in groups of three. Upon arrival, informed consent was obtained, and each participant was randomly assigned a letter (X, Y, or Z) for his or her identication on all materials. Next, participants were given the rst questionnaire packet. This packet contained questions unrelated to the study, such as whether they were familiar with the other group members. It should be noted that familiarity was evenly dispersed among communication conditions. Next, the Lost at Sea task was distributed. Initially, participants worked on the task individually without communicating. Upon completion of the Lost at Sea individual worksheets, participants were given a group worksheet and informed that they had 30 min to complete the task as a group. They were allowed to keep their individual worksheets. At this point, they were informed that the two most effective groups (dened as the groups most closely matching the expert rankings, and in cases of ties, also the fastest) would win $30 per group member. The groups then completed the task together either face-to-face in a small room or via computer in a large computer lab, where it was not possible to see the other members. The CMC groups received instructions on how to use the chat room before proceeding with the group discussion, and the FTF groups were videotaped. All group members, both FTF and CMC, were informed that they had 30 min to complete the task and received a reminder to choose a designated recorder 5 min into the task. The designated recorder entered the nal rankings on

PERCEPTIONS AND DECISION SUCCESS

21

the computer for the CMC groups or handed in the nal copy of the rankings for the FTF groups. Once groups indicated that they had reached consensus, the Lost at Sea materials were collected, and group members were asked to sign the group worksheet signifying their agreement with the group decision. If the group had not completed the task after 25 min, they were given a 5-min warning to indicate that they should complete the task within the next 5 min. Last, participants completed a second questionnaire that contained questions regarding perceived member competence, perceived interests, perceived success, and demographics, in that order. It should be noted, however, that questions unrelated to the current experiment separated the questions of interest.

Results
Level of analysis is an issue whenever analyzing group data, especially in studies such as this one in which some variables are measured at the group level (decision success) and other variables are measured at the individual level (perceived success and perceived member competence). Thus, we rst determined whether individual perceptions of perceived success and member competence, dened in terms of team ability, problem-solving ability, and knowledge, could be analyzed at the group level. The level of within-group agreement was assessed two ways: by calculating rwg (James, Demaree, & Wolf, 1984) and by examining the intraclass correlation, ICC(2). According to Klein et al. (2000), rwg assesses variability in one group at a time. In contrast, the intraclass correlation assesses variability across the entire sample. We examined both of these indicators. For the perceived success measure, 22% of the FTF groups and 18% of the CMC groups had a rwg of less than the recommended minimum level of .70 (e.g., Klein et al., 2000). When examining the intraclass correlations, only that for the CMC groups was above the recommended minimum level of .70 (.76 for the CMC groups and .21 for the FTF groups). Thus, CMC group members showed adequate agreement for their success perceptions to be aggregated to the group level, but this was not the case for FTF group members. In terms of group competence, however, both FTF and CMC groups showed relatively low

levels of within-group agreement for both rwg and the intraclass correlation. For FTF groups, rwg was less than .70 for 18% of the groups regarding team ability, 45% regarding knowledge, and 29% regarding problem-solving ability. Similarly, for CMC groups, rwg was less than .70 for 12% of the groups regarding team ability, 44% regarding knowledge, and 26% regarding problem-solving ability. Furthermore, none of the intraclass correlations reached the .70 level: For FTF groups the intraclass correlations were .50 for team ability, .53 for knowledge, and .38 for problem-solving ability, and for CMC groups the intraclass correlations were .62 for team ability, .54 for knowledge, and .43 for problem-solving ability. These results indicate too much variability at the individual level to aggregate the data to the group level. Thus, we decided to analyze the perception data on the individual level. The means, standard deviations, and correlations for the study variables can be found in Table 1.

Communication Medium, Perceived Decision Success, and Decision Success


First, before investigating the hypotheses, we explored whether the communication conditions performed equally well on the task. We conducted this analysis on the group level, because it does not include individual perception data. An independent-samples t test revealed no signicant difference between the groups, t(72) 1.25, ns. Although the difference did not reach statistical signicance, it should be noted that the FTF groups did provide more successful rankings (M 58.65, SD 13.53) than the CMC groups (M 62.97, SD 16.16), indicating a small effect (d 0.29). Lower numbers denote higher levels of decision success. Second, as specied in our rst hypothesis, we determined whether our CMC group members could predict their decision success better than FTF group members using a hierarchical regression with the rst step consisting of communication medium and perceived decision success predicting decision success and the second step consisting of the addition of the interaction between communication medium and perceived decision success. The independent variables were at the individual level, and following the recommendation of Cohen, Cohen,

22

ROCH AND AYMAN

Table 1 Means, Standard Deviations, and Intercorrelations for the Study Variables
Variable M SD 1 Face-to-face 1. 2. 3. 4. 5. Decision success Perceived success Team ability Knowledge Problem solving 58.65 84.07 4.19 3.42 3.99 13.53 15.47 0.72 0.87 0.70 .04 .03 .01 .06 Computer mediated 1. 2. 3. 4. 5. Decision success Perceived success Team ability Knowledge Problem solving .01. 62.97 79.30 4.03 3.42 3.91 16.16 18.51 0.58 0.84 0.54 .30** .01 .05 .17 2 3 4 5

.27** .13 .24**

.45** .70**

.52**

.36** .29** .53**

.32** .62**

.44**

** p

West, and Aiken (2003), we centered perceived decision success across the sample before entering it into the regression equation. Both steps were signicant. In Step 1, perceived success was a signicant predictor of decision success ( .18), and in Step 2, the interaction between perceived success and communication condition was signicant ( .19). (See Table 2.) To further analyze the moderating effect of communication medium on success perceptions, we conducted simple slope analyses. Results showed that perceived success was not a signicant predictor of decision success for FTF group members ( .04, ns) but was a signicant predictor for CMC group members ( .30, p .01). It should be noted that increased success is associated with smaller valTable 2 Communication Condition and Perceived Decision Success Predicting Decision Success
Variable Step 1 Communication condition Perceived success Step 2 Communication condition Perceived success Condition Perceived Success Note. R2 for Step 2. * p .05. .06 ( p B SE B .13* .18* .13* .04 .19* .05)

3.90 2.02 0.15 0.06 3.90 2.00 0.03 0.09 0.23 0.12 .02 ( p

.01) for Step 1; R2

ues, and increased perceived success is associated with larger numbers, hence the negative relationship. (See Figure 1.) In Figure 1, low perceived success is dened as one standard deviation below the mean for perceived success (65% perceived success), and high perceived success is dened as one standard deviation above the mean (99% perceived success). As can be seen in Figure 1, the line representing the slope between perceived and decision success is almost at for FTF groups, indicating almost no relationship between how well FTF group members thought their group performed and decision success. In contrast, the signicant negative slope for CMC groups indicates that their success perceptions predicted their decision success. Furthermore, Figure 1 also shows that at high levels of perceived success, both communication mediums have similar success scores. In fact, further simple slopes analyses show that the difference between communication mediums is not signicant at high levels of perceived success but is signicant at low levels of perceived success. CMC group members appear to be much better judges of when their group is not doing well than FTF group members. Thus, in support of Hypothesis 1, CMC group members success perceptions predicted decision success and FTF group members success perceptions did not. Furthermore, because CMC group members demonstrated adequate agreement regarding perceived success, we also investigated the re-

PERCEPTIONS AND DECISION SUCCESS

23

Figure 1. Perceived success and decision success by communication medium. CMC computer-mediated communication; FTF face-to-face.

lationship between perceived success and decision success for the CMC condition on the group level. We created a composite index for perceived success by averaging the members perceived success scores and then correlated this index with decision success. Similar to the results from the individual-level analyses, CMC group success perceptions were signicantly related to decision success (r .33, p .05). Even though FTF group members did not demonstrate adequate agreement regarding perceived success, we nonetheless also correlated their aggregated perceived success index with decision success: The relationship was not signicant (r .06, ns). It should be noted that for all group-level analyses, we used one-tailed signicance tests because we had strong directional hypothesesthat is, we expected the same results as with the individual-level analysesand the power to nd signicant effects in the group-level analyses was much less than that in the individual-level analyses.

Communication Medium, Perceived Competence, and Decision Success


Next, we determined whether CMC group members competence judgments signicantly predicted decision success. We used hierarchical regression with the rst step consisting of communication medium and the three types of

competencies predicting decision success and the second step consisting of the addition of the interaction between communication medium and each type of competence. Again, the independent variables were at the individual level, and the types of perceived competencies were centered across the sample. In Step 1, both communication medium ( .15) and per.21) sigceived problem-solving ability ( nicantly predicted decision success. In Step 2, the interaction between perceived problemsolving ability and communication medium approached signicance ( .19). It should be noted that neither of the other two interactions involving the other competencies was signicant. (See Table 3.) To further analyze the interaction between communication medium and perceived problem-solving ability in predicting decision success, we conducted simple slope analyses. Results showed that problem solving was not a signicant predictor of decision success for FTF group members ( .09, ns) but was a signicant predictor for CMC group members ( .33, p .05). To present a more complete picture, these relationships were also examined on the group level, and the results mirror those found on the individual level. Simple slopes analyses at the group level showed that problem solving was not a signicant predictor .11) of decision success for FTF groups (

24

ROCH AND AYMAN

Table 3 Communication Condition and Perceived Competence Predicting Decision Success


Variable Step 1 Communication condition Team ability Knowledge Problem solving Step 2 Communication condition Team ability Knowledge Problem solving Condition Team Ability Condition Knowledge Condition Problem Solving Note. R2 p .10. .05 for Step 1 ( p * p .05. B 4.55 2.05 1.43 4.77 4.53 0.44 0.42 1.81 4.05 2.31 7.79 SE B 2.03 2.09 1.35 2.26 2.03 2.69 1.85 2.87 4.27 2.72 4.68 .15* .09 .08 .21* .15* .02 .02 .08 .11 .09 .19

Mediation Analysis
Next, we conducted a mediation analysis, as recommended by Baron and Kenny (1986), to explore whether the perception of decision success mediates the relationship between perceived member competence and decision success for the CMC group members (Hypothesis 3). This analysis was also conducted on the individual level, and all continuous variables were centered. Baron and Kenny recommend examining a series of three regression equations. The rst equation, the regression of perceived success on perceived competence (team ability, knowledge, and problem-solving ability), showed that perceived problem-solving ability was the only signicant predictor of per.47, p .01; see ceived decision success ( Table 4, Regression Equation 1). Second, the regression of decision success on perceived competence (team ability, knowledge, and problem-solving ability) also showed that only perceived problem-solving ability signicantly .33, p .05; predicted decision success ( see Table 4, Regression Equation 2). Thus, perceptions of problem-solving ability were related to both perceived decision success and decision success for CMC group members.

.05); R2

.02 for Step 2.

but was a signicant predictor for CMC groups ( .52). Thus, in support of Hypothesis 2, only CMC group members competence judgments, specifically problem-solving ability, predicted their decision success. This also establishes that CMC group members made relatively accurate assessments of the problem-solving ability available to the group.

Table 4 Summary of Mediation Analysis for Computer-Mediated Communication Group Members on the Individual Level
Variable Regression Equation 1: Regressing perceived success on perceived member competence Team ability Knowledge Problem-solving ability Regression Equation 2: Regressing decision success on perceived member competence Team ability Knowledge Problem-solving ability Regression Equation 3: Regressing decision success on perceived member competence and perceived success Team ability Knowledge Problem-solving ability Perceived success Note. R2 .28 ( p .01) for Regression Equation 1; R2 R2 .07 ( p .01) for Regression Equation 3. * p .05. ** p .01. B SE B

1.36 1.43 16.11

3.62 2.17 4.04

.04 .07 .47**

4.49 2.72 9.59

3.53 2.12 3.94

.16 .15 .33*

4.85 3.11 5.27 0.27

3.41 2.05 4.12 0.10

.18 .17 .18 .32**

.06 for Regression Equation 2;

PERCEPTIONS AND DECISION SUCCESS

25

Third, the regression of decision success on both perceived member competence and perceived success revealed that the three types of perceived competence, including problem-solving ability, were not signicant predictors, but perceived success was a signicant predictor ( .32, p .01; see Table 4, Regression Equation 3). These results indicate that perceived success mediates the relationship between perceived problem-solving ability and decision success for CMC group members, providing support for Hypothesis 3. We conducted a similar three-step procedure for the FTF groups. None of the requirements for mediation were met (see Table 5). Last, to present a complete picture of the results, we also conducted the analyses on the group level (see Tables 6 and 7). The results do not differ from those on the individual level, except that for CMC groups, neither problem solving nor perceived accuracy predicted decision success when considered simultaneously, even though both signicantly predicted decision success when considered individually. Again, it should also be noted that the analyses at the group level had much less power to nd signicant results than the analyses at the individual level.

Ancillary Analyses
We conducted two sets of analyses to provide additional support for Hypothesis 1, predicting that CMC group members success perceptions should be better predictors of decision success than FTF group members perceptions. The rst set of analyses was conducted to rule out purely statistical explanations. Specically, we wanted to rule out the possibility that our nding that CMC group members predicted decision success better than FTF group members was a statistical artifact due to a combination of a ceiling effect and range restriction for FTF group members. The second set of analyses was conducted to provide support for our explanatory mechanism underlying Hypothesis 1. Perceived success. Members of FTF groups thought that their groups were signicantly more successful than CMC group members (M 84%, SD 15.47, for FTF; M 79%, SD 18.51, for CMC), t(209) 2.03, p .05. On the other hand, even though CMC group members success estimates had a larger standard deviation than FTF group members estimates, Levenes test for equality of variance showed that the variances of the FTF and CMC group members estimates were not signi-

Table 5 Summary of Mediation Analysis for Face-to-Face Group Members on the Individual Level
Variable Regression Equation 1: Regressing perceived success on perceived member competence Team ability Knowledge Problem-solving ability Regression Equation 2: Regressing decision success on perceived member competence Team ability Knowledge Problem-solving ability Regression Equation 3: Regressing decision success on perceived member competence and perceived success Team ability Knowledge Problem-solving ability Perceived success Note. R2 .08 ( p .05) for Regression Equation 1; R2 R2 .00 for Regression Equation 3. B SE B

4.19 0.17 2.48

2.83 1.94 3.01

.19 .01 .11

0.44 0.42 1.81

2.54 1.75 2.70

.02 .03 .09

0.53 0.41 1.76 0.02

2.58 1.75 2.72 0.09

.03 .03 .09 .02

.01 for Regression Equation 2;

26

ROCH AND AYMAN

Table 6 Summary of Mediation Analysis for Computer-Mediated Communication Groups on the Group Level
Variable Regression Equation 1: Regressing perceived success on perceived member competence Team ability Knowledge Problem-solving ability Regression Equation 2: Regressing decision success on perceived member competence Team ability Knowledge Problem-solving ability Regression Equation 3: Regressing decision success on perceived member competence and perceived success Team ability Knowledge Problem-solving ability Perceived success Note. R2 .42 ( p .01) for Regression Equation 1; R2 R2 .04 for Regression Equation 3. * p .05. ** p .01. B SE B

3.40 1.97 32.01

7.50 4.38 8.56

.09 .07 .72**

14.87 2.57 23.69

9.42 5.50 10.75

.38 .09 .52*

13.95 2.04 15.02 0.27

9.39 5.49 12.93 0.23

.36 .07 .33 .26

.14 for Regression Equation 2;

cantly different. Thus, although FTF group members thought that their groups were more successful, this was not due to a ceiling effect. Furthermore, range restriction is not an issue: Both types of groups showed similar variability in their success estimates.

Interests. The postexperimental questionnaire contained two questions asking to what extent participants expressed interest in maintaining interpersonal relationships and to what extent they expressed interest in solving the problem accurately on an 8-point scale (1

Table 7 Summary of Mediation Analysis for Face-to-Face Groups on the Group Level
Variable Regression Equation 1: Regressing perceived success on perceived member competence Team ability Knowledge Problem-solving ability Regression Equation 2: Regressing decision success on perceived member competence Team ability Knowledge Problem-solving ability Regression Equation 3: Regressing decision success on perceived member competence and perceived success Team ability Knowledge Problem-solving ability Perceived success Note. R2 .16 for Regression Equation 1; R2 for Regression Equation 3. B SE B

6.85 0.00 4.57

4.47 3.59 5.80

.29 .00 .15

1.35 1.91 4.48

6.80 5.45 8.82

.04 .06 .11

1.95 1.91 4.08 0.09

7.10 5.52 9.00 0.26

.06 .06 .10 .06 .00

.00 for Regression Equation 2; R2

PERCEPTIONS AND DECISION SUCCESS

27

none, 8 a great deal). It should be noted that participants answered these two general questions three times, once with their own interest in mind and then regarding their perceptions of the interests of their two group members. All three perceptions were averaged to determine an individuals general perception of his or her groups interest. FTF group members reported signicantly more interest in maintaining interpersonal relationships (M 6.66, SD 1.43) than CMC group members (M 6.07, SD 1.38), t(209) 3.05, p .05. On the other hand, there was no signicant difference between FTF and CMC group members regarding the amount of interest in solving the problem, t(209) 0.78, ns. In summary, the results show that CMC group members estimates of perceived success were signicantly related to decision success, and their perceptions of problem-solving ability were also related to decision success. This was not the case for FTF group members. Only for CMC groups did perceived decision success mediate the relationship between perceived problem-solving ability and decision success. Last, FTF group members were signicantly more interested in maintaining interpersonal relationships than CMC group members.

Discussion
The differences between CMC and FTF groups extend beyond potential performance differences. In the current study, CMC group members perceptions of decision success signicantly predicted their decision success, regardless of whether the analyses were conducted on the individual or the group level. In contrast, almost no relationship existed between FTF group members perceived decision success and decision success. CMC group members appear to have been more aware of when their group had not reached a successful decision than FTF group members. Overall, FTF group members believed that their groups were more successful than CMC group members, even though the performance difference between the two types of groups was not signicant. The lack of a relationship between FTF group members perceived success and decision success, however, cannot be attributed to a ceiling effect. FTF group members reported a mean of 84% estimated success, with

a standard deviation of 15.47. Thus, it was not the case that all members of FTF groups believed they were part of successful groups whereas CMC group members thought their groups were not successful (79% perceived success rate for CMC group members). Furthermore, the variances of the FTF and CMC group members estimates of success were not significantly different. Therefore, it also was not the case that CMC members signicantly predicted their groups decision success and FTF members did not simply because FTF member estimates showed range restriction whereas CMC member estimates did not. The predictive differences between the two groups cannot be attributed to statistical artifacts. Another indication that these predictive differences cannot be attributed to statistical artifacts but represent real differences between the two types of groups can be found by examining the intraclass correlations for perceived success. As described, we rst conducted the analyses involving member perceptions on the individual level and not the group level because the level of agreement was below the suggested minimum level for most of the variables. There was, however, one exception: For CMC groups, the intraclass correlation for perceived success was above the .70 minimum (.76), indicating that CMC group members showed an adequate level of agreement regarding whether their group was successful, in contrast to FTF groups, which had an intraclass correlation of .21. This indicates that CMC group members had similar perceptions of the success of their group decision whereas FTF group members did not. One reason why members of CMC groups could predict their groups decision success and demonstrated greater agreement in comparison to FTF group members may be that, as mentioned earlier, communicating via the computer puts the focus on the message and not the messenger. In other words, in terms of Petty and Cacioppos ELM model (Petty & Cacioppo, 1981, 1986), because the only information available to CMC group members was the typed word on a computer screen, they most likely processed the information centrally by evaluating the message content and were distracted by few peripheral cues. On the other hand, members of FTF groups were most likely inuenced both by the content of the message (central processing) and by peripheral cues such as the

28

ROCH AND AYMAN

personality, gender, attractiveness, and speaking style, resulting in less accurate estimates of whether their group was successful and less agreement among group members. It appears that each FTF group member was affected by a different set of peripheral cues and to different degrees, depending on his or her personal views, resulting in relatively low group member agreement of perceived group success. Hedlund et al. (1998) also speculated that the social cues inherent in FTF communication may be distracting, even though they concentrated on group leaders ability to differentiate between members who made good recommendations and those who did not, and did not investigate perceptions of decision success. Nonetheless, they proposed and found support for a similar process in which CMC group leaders did a better job of weighing individual recommendations than FTF group leaders. Thus, even though social cues may provide interpersonal information, they appear to distract individuals from concentrating on the message instead of the messenger. Tangential evidence indicating that CMC group members focused more on the message than the messenger in comparison to FTF group members can be found in the postexperimental questionnaire. The results from the questions regarding interest in maintaining interpersonal relationships and completing the task showed that FTF group members were signicantly more interested in relationship maintenance than CMC group members, even though both were equally interested in completing the task. Thus, FTF group members most likely focused more on interpersonal factors necessary to maintain working relationships within the group. Because CMC group members could evaluate the content of members messages without the distraction of peripheral cues associated with FTF communication and relationship maintenance, they could also make realistic judgments of the combined problem-solving skill in the group. The results showed that CMC group members perceptions of problem-solving skill signicantly predicted decision success, but this was not the case for FTF group members. Furthermore, the mediation analysis shows that perceptions of problem-solving ability played an indirect role in predicting decision success for CMC group members. This indi-

cates that CMC group members perceptions of their problem-solving ability were related to their perceptions of decision success and that these perceptions of decision success, in turn, were related to actual decision success. However, it should be noted that evidence for mediation was much stronger in the individual-level analyses than in the group-level analyses, most likely because of the difference in power between the individual- and group-level analyses. These relationships were not evident for FTF group members: Their perceptions of problemsolving ability were not related to decision success, and also their perceptions of success were not related to decision success. Nonetheless, even though FTF group members perceptions of their team and of their groups success were not related to decision success, FTF groups did make more successful decisions, albeit not signicantly more in this case. This is not surprising, given that the Baltes et al. (2002) meta-analysis showed that the effect representing the superiority of FTF groups in intellective tasks is not large, and the Hollingshead and McGrath (1995) narrative review found only slightly more than half of the FTF groups performed better than CMC groups. Thus, even though overall it does appear that CMC groups perform worse than FTF groups on intellective tasks, this may not always be the case. For example, Baltes et al. found that CMC groups given unlimited amounts of time performed as well as FTF groups. Thus, even though we imposed a time limit, it was a generous time limit, which allowed the CMC groups to perform almost as well as the FTF groups. It should be noted, however, that our perceived success measure differs from both group efcacy and group potency. Group potency and efcacy assess whether group members believe that the group can succeed in the certain situation (e.g., Guzzo, Yost, Campbell, & Shea, 1993; Lester, Meglino, & Korsgaard, 2002). Group potency and group efcacy are closely related to the concept of self-efcacy but elevated to the group level (Guzzo & Shea, 1992). Furthermore, group potency is a generalized belief about group capabilities across tasks and contexts, but group efcacy is task specic (Gully, Joshi, Incalcaterra, & Beaubien, 2002; Lee, Tinsley, & Bobko, 2002). In the present study, we asked not whether the groups can

PERCEPTIONS AND DECISION SUCCESS

29

succeed on a specic task (group efcacy) but to what degree they have succeeded (perceived decision success). It may be that group members think that their group could have succeeded but that for various reasons their group did not perform well. In the current study we were interested in how well group members believe the group performed and not how well they believe the group could perform. Last, one concern shared by all laboratory research is limited external validity. We acknowledge that the present results have unknown generalizability to other situations and populations; however, as Mook (1983) reminded us, the purpose of research is not always to generalize to other populations. To highlight this point, Mook used the example of the classic study by Harlow and Harlow (1962) in which the researchers investigated whether young monkeys would prefer a terry-cloth-covered wire mother to a wire mother with food. The ndings of this study were never designed to generalize to the general population of monkeys, or even to human children, but were intended rather to test the authors theory of attachment against the more traditional drive-reduction theories of attachment popular during that era. In a similar vein, the purpose of this study was to conduct a more in-depth analysis into the differences between FTF and CMC groups and to determine whether members of CMC groups can be better judges of their groups performance, at least on intellective tasks. We chose to concentrate on intellective tasks because it is especially important to understand decision making in cases where there are correct answers. Further research is needed to determine the generalizability of the results, but as Dipboye (1990) reminded us, generalizability is an issue for all research, both laboratory and eld.

members. This nding has implications for both theory and practice. On the theoretical level, the present results demonstrate the need to go beyond the traditional outcome variables of time, satisfaction, and decision. Research should also explore potential differences in group processes used to reach group decisions and the implications of these process differences. The nding that CMC group members can predict their groups success more accurately and with greater agreement also has important practical implications. These results imply that it may be benecial for managers to simply ask CMC group members whether they have arrived at a successful decision, especially when the decisions have correct answers. As indicated in the introduction, if group members indicate that they do not believe their decision is the right one, managers have the opportunity to give the group more information, time, and resources to avoid a potentially costly mistake. On the other hand, the results suggest that managers should not give FTF group members estimates of decision success the same weight. Also, it is very likely that when asking FTF group members whether they have made a successful decision, managers will receive different estimates of success from the various group members. Thus, it appears that simply avoiding using CMC groups because they may make a less effective decision than FTF groups is simplistic and short sighted. It appears that CMC group members are better judges of their performance than FTF group members, and further research is needed to uncover qualitative differences between FTF and CMC groups that may help to alleviate potential performance differences between the two types of groups. In todays global community, understanding the differences between CMC and FTF groups is becoming increasingly important, because at times FTF meetings will not be an option.

Conclusion
The present research adds to our understanding of the differences between FTF and CMC groups by investigating differences in perceptions related to outcomes. The results show that CMC group members are signicantly better judges of whether they made a successful decision and agree on the success of their groups decision to a greater extent than FTF group

References
Baltes, B. B., Dickson, M. W., Sherman, M. P., Bauer, C. C., & LaGanke, J. S. (2002). Computermediated communication and group decision making: A meta-analysis. Organizational Behavior and Human Decision Processes, 87, 156 179. Baron, R. M., & Kenny, D. A. (1986). The moderatormediator variable distinction in social psycho-

30

ROCH AND AYMAN

logical research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 11731182. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum. Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/correction analysis for the behavioral sciences (3rd ed.). Mahwah, NJ: Erlbaum. Dipboye, R. L. (1990). Laboratory vs. eld research in industrial and organizational psychology. In C. L. Cooper & I. T. Robertson (Eds.), International review of industrial and organizational psychology (Vol. 5, pp. 134). Oxford, England: Wiley. Einhorn, H. J., Hogarth, R. M., & Klempner, E. (1977). Quality of group judgment. Psychological Bulletin, 84, 158 172. Goldstein, I. L. (1995). Foreword. In R. A. Guzzo, E. Salas, & Associates (Eds.), Team effectiveness and decision-making in organizations (pp. xixiv). San Francisco: Jossey-Bass. Gully, S. M., Joshi, A., Incalcaterra, K. A., & Beaubien, J. M. (2002). A meta-analysis of team efcacy, potency, and performance: Interdependence and level of analysis as moderators of observed relationships. Journal of Applied Psychology, 87, 819 832. Guzzo, R. A., & Shea, G. P. (1992). Group performance and intergroup relations in organizations. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (2nd ed., pp. 269 313). Palo Alto, CA: Consulting Psychologists Press. Guzzo, R. A., Yost, P. R., Campbell, R. J., & Shea, G. P. (1993). Potency in groups: Articulating a construct. British Journal of Social Psychology, 32, 87106. Harlow, H., & Harlow, M. (1962). Social deprivation in monkeys. Scientic American, 207, 136 146. Hedlund, J., Ilgen, D. R., & Hollenbeck, J. R. (1998). Decision accuracy in computer-mediated versus face to face decision-making teams. Organizational Behavior and Human Decision Processes, 76, 30 47. Hinds, P. J., & Weisband, S. P. (2003). Knowledge sharing and shared understanding in virtual teams. In C. B. Gibson & S. G. Cohen (Eds.), Virtual teams that work (pp. 2136). San Francisco: Jossey-Bass. Hollingshead, A. B. (1996a). Information suppression and status persistence in group decision making: The effects of communication media. Human Communication Research, 23, 193219. Hollingshead, A. B. (1996b). The rank-order effect in group decision making. Organizational Behavior and Human Decision Making, 68, 181193.

Hollingshead, A. B., & McGrath, J. E. (1995). Computer-assisted groups: A critical review of the empirical research. In R. A. Guzzo, E. Salas, & Associates (Eds.), Team effectiveness and decisionmaking in organizations (pp. 46 78). San Francisco: Jossey-Bass. Howard, A. (1995). The changing nature of work. San Francisco: Jossey-Bass. Ilgen, D. R., & Pulakos, E. D. (Eds.). (1999). The changing nature of performance: Implications for stafng, motivation, and development. San Francisco: Jossey-Bass. James, L. R., Demaree, R. G., & Wolf, G. (1984). Estimating within-group interrater reliability with and without response bias. Journal of Applied Psychology, 69, 8598. Klein, K. J., Bliese, P. D., Kozlowski, S. W. J., Dansereau, F., Gavin, M., Grifn, M. A., et al. (2000). Multilevel analytical techniques. In K. J. Klein & S. W. J. Kozlowski (Eds.), Multilevel theory, research, and methods in organizations: Foundations, extensions, and new directions. San Francisco: Jossey-Bass. Lee, C., Tinsley, C., & Bobko, P. (2002). An investigation of the antecedents and consequences of group level condence. Journal of Applied Social Psychology, 32, 1628 1652. Lester, S. W., Meglino, B. M., & Korsgaard, M. A. (2002). The antecedents and consequences of group potency: A longitudinal investigation of newly formed work groups. Academy of Management Journal, 45, 352368. Libby, R., Trotman, K. T., & Zimmer, I. (1987). Member variation, recognition of expertise, and group performance. Journal of Applied Psychology, 72, 81 87. Littlepage, G. E., & Silbiger, H. (1992). Recognition of expertise in decision-making groups: Effects of group size and participation patterns. Small Group Research, 23, 344 355. McGrath, J. E. (1984). Groups, interaction and performance. Englewood Cliffs, NJ: Prentice Hall. Mook, D. G. (1983). In defense of external invalidity. American Psychologist, 38, 379 387. Nemiroff, P. M., & Pasmore, W. A. (1975). Lost at Sea: A consensus-seeking task. In J. W. Pfeiffer & J. E. Jones (Eds.), The 1975 annual handbook for group facilitators (pp. 28 34). San Diego, CA: Pfeiffer. Paese, P. W., & Sniezek, J. A. (1991). Inuences on the appropriateness of condence in judgment: Practice, effort, information, and decision-making. Organizational Behavior and Human Decision Processes, 48, 100 130. Peterson, D. K., & Pitz, G. F. (1988). Condence, uncertainty, and the use of information. Journal of Experimental Psychology: Learning, Memory, and Cognition, 14, 8592.

PERCEPTIONS AND DECISION SUCCESS

31

Petty, R. E., & Cacioppo, J. T. (1981). Attitudes and persuasion: Classic and contemporary approaches. Dubuque, IA: William C. Brown Communications. Petty, R. E., & Cacioppo, J. T. (1986). Communication and persuasion: Central and peripheral routes to attitude change. New York: SpringerVerlag. Ryback, D. (1967). Condence and accuracy as a function of experience in judgment making in the absence of systematic feedback. Perceptual and Motor Skills, 34, 331334. Siegel, J., Dubrovsky, V., Kiesler, S., & McGuire, T. W. (1986). Group processes in computer-mediated communication. Organizational Behavior and Human Decision Processes, 37, 157187. Smith, J., & Vanecek, M. (1988). Computer conferencing and task-oriented decisions: Implications for group decision support. Information and Management, 14, 123132. Sniezek, J. A., & Henry, R. A. (1989). Accuracy and condence in group judgment. Organizational Behavior and Human Decision Processes, 43, 128.

Sniezek, J. A., & Henry, R. A. (1990). Revision, weighting, and commitment in consensus group judgment. Organizational Behavior and Human Decision Processes, 45, 66 84. Sniezek, J. A., Paese, P. W., & Switzer, F. S., III. (1990). The effect of choosing on condence and choice. Organizational Behavior and Human Decision Processes, 46, 264 282. Straus, S. G., & McGrath, J. E. (1994). Does medium matter? The interaction of task type and technology on group performance and member reactions. Journal of Applied Psychology, 79, 8797. Turner, M. E., & Pratkanis, A. R. (1998). Twentyve years of groupthink theory and research: Lessons from the evaluation of a theory. Organizational Behavior and Human Decision Processes, 73, 105115. Received April 18, 2003 Revision received October 14, 2004 Accepted October 14, 2004

Das könnte Ihnen auch gefallen