Content uploaded by Yi-Ching Liu
Author content
All content in this area was uploaded by Yi-Ching Liu on Dec 16, 2019
Content may be subject to copyright.
http://sgr.sagepub.com/
Small Group Research
http://sgr.sagepub.com/content/45/2/154
The online version of this article can be found at:
DOI: 10.1177/1046496414522716
2014 45: 154Small Group Research
Yi-Ching Liu and Poppy Lauretta McLeod
Accountability in Group Problem-Solving
Individual Preference for Procedural Order and Process
Published by:
http://www.sagepublications.com
can be found at:Small Group ResearchAdditional services and information for
http://sgr.sagepub.com/cgi/alertsEmail Alerts:
http://sgr.sagepub.com/subscriptionsSubscriptions:
http://www.sagepub.com/journalsReprints.navReprints:
http://www.sagepub.com/journalsPermissions.navPermissions:
http://sgr.sagepub.com/content/45/2/154.refs.htmlCitations:
What is This?
- Mar 24, 2014Version of Record >>
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
Small Group Research
2014, Vol. 45(2) 154 –175
© The Author(s) 2014
Reprints and permissions:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/1046496414522716
sgr.sagepub.com
Article
Individual Preference for
Procedural Order and
Process Accountability in
Group Problem-Solving
Yi-Ching Liu1 and Poppy Lauretta McLeod1
Abstract
An experiment compared interaction process and task performance of
groups composed of members with high or low preference for procedural
order, assigned to an individual or group process accountability condition.
Procedural order preference did not affect the use of procedures nor
did procedures affect task performance. Competing hypotheses on the
effects of process accountability were tested: One set of hypotheses was
based on higher motivation among high order preference groups because
accountability amplifies their natural work predispositions; the other set was
based on higher motivation among low order preference groups because
accountability induces cognitive dissonance in them. Results supported
the amplifying hypotheses. High order preference groups exchanged more
information, perceived their discussions as more orderly, and produced
more feasible task solutions. Compared with group process accountability,
individual accountability led to less information exchange in both high and low
order preference groups, but low order preference groups under individual
accountability produced the least feasible task solutions of all groups.
Implications for the match between individual work habit preferences and
task context demands are discussed.
1Cornell University, Ithaca, NY, USA
Corresponding Author:
Yi-Ching Liu, Department of Communication, Cornell University, 336 Kennedy Hall, Ithaca,
NY 14853, USA.
Email: yl839@cornell.edu
522716SGRXXX10.1177/1046496414522716Small Group ResearchLiu and McLeod
research-article2014
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
Liu and McLeod 155
Keywords
group procedural order preference, group procedure, motivated information
processing, epistemic motivation, process accountability
Differences in the work habits of group members have crucial effects on task
coordination. Some members may prefer meetings that follow a structured
agenda, have systematic decision-making procedures, and use consistent
decision criteria while others may feel uncomfortable under such structures
and would prefer more flexible procedures. Early work on group decision-
making processes has shown that such preferences for procedural order are
relatively stable characteristics (Putnam, 1976). More recent work suggests
that different circumstances, such as level of motivation, affect the expres-
sion of different work habit preferences and how the interaction between con-
text and preferences affects task performance (e.g., De Dreu, Nijstad, & van
Knippenberg, 2008). Although procedures are vital to strong task perfor-
mance (Davis, Tindale, Nagao, Hinsz, & Robertson, 1984; Poole, Holmes,
Watson, & DeSanctis, 1993), Pavitt (1993) has suggested that procedures
must be combined with high motivation to yield high-quality decisions.
Building upon recent research on motivation within decision-making groups
(e.g., De Dreu et al., 2008; Scholten, van Knippenberg, Nijstad, & De Dreu,
2007) the current study examines how placing groups under different condi-
tions of process accountability affects the impact of individual members’
work habit preferences on interaction patterns and on task performance
quality.
Procedural Order Preference
The importance of procedural order in group decision-making process was first
identified by Bormann (1975) and defined as “the degree of systematic order
inherent in the type of procedural behaviors” (Putnam, 1979, p. 194) that occur
in small group discussion. Putnam suggested that individuals join group dis-
cussions with preferences for work habits and procedural order. This group
procedural order preference (GPOP) might be thought of as a predisposition
toward or expectation about what constitutes appropriate group behaviors, and
although it is a relatively stable attitude, it may oscillate due to changes in
group experience (Putnam, 1976). Based on these ideas, Putnam (1979) devel-
oped the Group Procedural Order Questionnaire (GPOQ) to understand indi-
vidual preference for structured procedures in groups. She established the
validity of this instrument by demonstrating positive relationships with mea-
sures for order, cognitive structure, and self-management, negative relationship
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
156 Small Group Research 45(2)
with tolerance of ambiguity, and group flexibility, while empirically distin-
guishing it from several other characteristics (e.g., task-orientation and leader-
ship). GPOP herein is perceived as a unique concept that depicts an individual’s
preference for work habits related to use of procedures in groups (Putnam,
1976).
High procedural order preference (HPO) individuals are identified by four
characteristics: “(a) preference for planned and sequential patterns for orga-
nizing task activities; (b) concern for time-management; (c) emphasis on
regular and predictable procedures; and (d) emphasis on clarifying group
procedures and reminding members to adhere to the task.” Low procedural
order preference (LPO) individuals, on the contrary, are characterized by the
following four characteristics: “(a) the use of chain-association; (b) flexibil-
ity in establishing and changing plans; (c) oblivious to time constraints; and
(d) emphasis on a balance between the task and the socio-emotional needs of
the group’s members” (Putnam, 1979, pp. 195-197).
When working in groups, HPO individuals would be expected to follow
procedures that produce organized and orderly discussions, whereas LPO
individuals would produce less organized discussions. Consistent with these
expectations, Putnam (1979) reported that in groups composed of HPO indi-
viduals behaviors such as summarizing, referring to the agenda, clarifying
procedures, and so forth were seen whereas discussions in LPO groups more
often displayed interruptions, multiple side conversations, oscillation
between task and socio-emotional issues, and chain associations. Burke and
Aytes (2001) found changes over time in the effects of procedural order pref-
erence on group interaction. In their study, HPO groups initially displayed
more discussion structuring behaviors than did LPO groups, but over the
course of four meetings the amount of structuring behaviors in HPO groups
decreased, while it increased within the LPO groups. Their interpretation was
that because of the low natural inclination toward establishing procedures in
LPO groups, the realization of the importance of having a structure occurred
relatively late in these groups and a flurry of organizing activities were seen
in the later stages of their project. On the contrary, HPO groups’ natural incli-
nation was to put organizing structures in place right away, but once estab-
lished there was less need for continued explicit structuring behaviors.
The first hypothesis we propose is a replication of the previous research
showing that the procedural order preference of group members will affect
the pattern of procedural behaviors seen in their discussions. Specifically, we
expect that HPO groups will use agendas, summarize the discussion, focus on
the task, use clear signaling to change the topic, and discuss goals, division of
labor, and procedures more often than will LPO groups; LPO groups will
change topics without clear signaling, continue discussion after topic change
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
Liu and McLeod 157
without signaling, and digress more often than will HPO groups (Hypothesis
1 [H1]).
Procedural Order Preference and Task Performance
There is limited empirical evidence about the effect of GPOP on task perfor-
mance. Hirokawa, Ice, and Cook (1988) found that the performance of HPO
groups was enhanced by following a high structure agenda whereas the per-
formance of LPO groups was equivalent in high versus low structure con-
texts; moreover, in the low structure context, LPO groups outperformed HPO
groups. Comparable results from a computer-mediated communication con-
text, reported in Wheeler, Mennecke, and Scudder’s (1993, see Table 2),
showed that LPO groups outperformed HPO groups in general, but the differ-
ence was about four times greater when no process structures were imposed.
In contrast to the benefits of procedures for HPO groups seen in Hirokawa et
al.’s study, the HPO groups in the Wheeler et al. study showed the lowest
performance scores of all when given a set of procedures to follow.
Taken together, these findings suggest that the LPO preference benefits
from less restrictive contexts, which Hirokawa et al. (1988) argued was an
indication of flexibility and adaptability; the implications for the HPO prefer-
ence are less clear. Moreover, because these researchers examined the effects
of imposed procedures, it was not possible for them to observe how group
member procedural order preferences would shape the processes they would
follow. It was of interest in the current article to stimulate implicitly HPO and
LPO individuals’ motivation to use procedures in group discussion process so
that the procedures they freely chose could be observed and then related to
task outcomes. Research has suggested that such motivation might be stimu-
lated by holding groups accountable for their decision-making process (e.g.,
De Dreu et al., 2008; Scholten et al., 2007). Building on H1, we would expect
that using relatively unstructured procedures would benefit LPO groups more
than HPO groups, and despite the ambiguity of previous findings we never-
theless argue that it is theoretically reasonable to hypothesize that using
structured procedures would benefit HPO groups more than LPO groups
(Hypothesis 2 [H2]).
Epistemic Motivation and Process Accountability
It is clear that group member motivation in general is important for task per-
formance, but perhaps more predictive of performance is the match between
specific motivation and task demands. Problem-solving tasks such as the one
we use in this study require motivation related to acquisition and processing
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
158 Small Group Research 45(2)
of information, which can be labeled as epistemic motivation (De Dreu et al.,
2008; Scholten et al., 2007). Epistemic motivation tends to rise when people
recognize that their knowledge is not sufficient to address the problem, thus
increasing the likelihood they will engage in systematic search and process-
ing of information (Chaiken & Trope, 1999). When group members have
high motivation to develop an accurate and deep understanding of the task
circumstances, they rely more on systematic than on heuristic information
processing, and group discussions are characterized more by information-
driven than by preference-driven interaction (De Dreu, Koole, & Oldersma,
1999; Scholten et al., 2007). Information-driven interaction generally leads
to higher quality problem-solving (Cruz, Henningsen, & Williams, 2000; De
Dreu et al., 2008; van Swol, Savadori, & Sniezek, 2003).
One way of raising epistemic motivation is to hold people accountable for
their decision-making process (Kerschreiter, Schulz-Hardt, Mojzisch, &
Frey, 2008; Scholten et al., 2007). For example, Scholten et al. (2007) found
that group information sharing and decision quality in a hidden profile task
increased when members were told they would be later interviewed about
their groups’ decision-making process. It has also been shown that account-
ability increases the difficulty and time needed for groups to reach consensus
(e.g., van Hiel & Schittekatte, 1998), which has been argued to be important
for the avoidance of groupthink (Kroon, Hart, & van Kreveld, 1991; Kroon,
van Kreveld, & Rabbie, 1992).
Within a group context, the question arises of whether collective or indi-
vidual accountability is more effective. The limited empirical evidence sug-
gests that individual accountability is more powerful, primarily because in
groups, the effects of accountability are diffused across members (Kroon et
al., 1991; O’Connor, 1997), though in follow-up research, Kroon et al. (1992)
expected but did not find differences. In the current study, we contribute addi-
tional data relative to the question of the difference between group and indi-
vidual accountability by testing the hypothesis that the effects of individual
accountability will be stronger than will the effects of group accountability
(Hypothesis 3 [H3]).
Individual Differences
Pavitt (1993) suggested that traits such as preferences for procedural order may
explain differences in how people respond to group discussion procedures.
Individuals with high preferences for order might be expected to respond more
positively to procedures that raise process accountability than would individuals
with low preferences for order. Accountability increases the complexity of deci-
sion-making strategies (Kroon et al., 1992), which would correspond with a
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
Liu and McLeod 159
preference for formal procedures. Also, accountability increases self-awareness
(De Dreu, Koole, & Steinel, 2000; van Hiel & Schittekatte, 1998) and may
heighten for group members the salience of how closely their preferences match
demands of the task context. Therefore, when groups are held accountable for
their problem-solving process HPO individuals might perceive this task demand
to be a closer match to their natural inclinations than might LPO individuals.
Accountability would thus have an enhancing effect on HPO groups, and a detri-
mental effect on LPO groups.
We hypothesize that accountability may amplify the effect of procedural
order preference such that HPO groups will exchange more information
(Hypothesis 4a [H4a]), have higher perception of epistemic motivation, pro-
cess accountability, and ordered process (Hypothesis 4b [H4b]), and we argue
that this higher attention to process will lead them to achieve higher quality
task performance (Hypothesis 4c [H4c]) than will LPO groups. In addition,
building on the prediction of H3, we expect the difference between HPO and
LPO groups to be greater under the individual than the group accountability
condition (Hypothesis 5 [H5]).
Suggested by the results reported in Burke and Aytes (2001), a competing set
of hypotheses based on a cognitive dissonance argument can be proposed. The
heightened awareness of the distance between their preferences and the task
demands could induce dissonance within LPO individuals, thus producing stron-
ger effects of process accountability on LPO than on HPO groups. LPO groups
may then be hyper-vigilant in their use of procedures and attention to process in
an effort to reduce the dissonance. HPO groups, on the other hand, will not expe-
rience dissonance because the task demands are close to their typical way of
working, and therefore these groups would not be expected to put forth any spe-
cial effort. Based on this reasoning we propose that LPO groups will exchange
more information (Hypothesis 6a [H6a]), have higher perception of epistemic
motivation, process accountability, and ordered process (Hypothesis 6b [H6b])
and we argue that this higher attention to process will lead them to achieve higher
quality task performance (Hypothesis 6c [H6c]) than will HPO groups. Finally,
building on the prediction of H3, we expect the difference between LPO and
HPO groups to be greater under the individual than the group accountability con-
dition (Hypothesis 7 [H7]).
Method
Design
This study was a 2 × 2 factorial laboratory experiment with GPOP (LPO/HPO
groups) and process accountability (individual/group) as the independent
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
160 Small Group Research 45(2)
variables. The assignment of people into HPO or LPO groups was decided prior
to the experiment, and these two kinds of groups were then randomly assigned to
either the individual or group process accountability condition.
Participants
To solicit a relatively representative sample of participants other than univer-
sity students for the main study, we hired a virtual world market research
company,1 to recruit 585 users of a 3D virtual environment from all over the
world to participate in our online pre-task survey. The goal of pre-task survey
was to obtain a large sample in which the range of participants’ scores on the
scale of individual preference for procedural order was wide enough to com-
pose HPO and LPO groups. 399 participants agreed to participate in the main
experiment. Participants ranged in age from 19 to 70 (M = 35 years). Of the
participants, 70% were female. Approximately 73% of the participants were
from North America, 15% were from Europe, and the remaining participants
were distributed across the other continents.
The pre-task survey consisted of Putnam’s (1979) GPOQ. The reliability and
validity of this 34-item Likert-type scale have been established, and the scale has
been used in previous studies to identify an individual’s preference for proce-
dures in group meetings (Burke & Aytes, 2001; Hirokawa et al., 1988; Wheeler
et al., 1993). The reliability obtained in our study based on 585 participants was
Cronbach’s alpha = .86. Following Putnam (1979, 1983), the participants’ raw
scores were transformed for the convenience of interpretation into standard T
scores that center around a mean of 50 and have a standard deviation of 10. The
standardized scores range from 21.88 to 75.25. Following Putnam (1979, 1983),
participants who received scores at least one standard deviation above (n = 72)
and below (n = 72) the mean on the scale were identified as HPO and LPO can-
didates, respectively, and were invited to sign up for the main study, which was
conducted in a lab setting within the 3D virtual environment, Second Life. Due
to attrition, another invitation was sent out to the participants who scored a 0.75
standard deviation both above and below the mean. In the final sample, 92 par-
ticipants were organized into 14 LPO groups and 12 HPO groups with three to
five participants in each group. A t test confirmed there was a significant differ-
ence in the mean of group members’ scores on the GPOQ between these two
types of groups, MHPO = 62.87, MLPO = 34.72, t(24) = −27.43, p < .001.
Procedure
Participants, using their own avatars, were provided the link to the experi-
mental site in Second Life, in which the researcher placed a table, chairs, and
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
Liu and McLeod 161
white boards with task information posted on them (see Figure 1). After they
completed informed consent forms, the experimenter, also through an avatar,
provided the task instructions through the public chat window. All task com-
munication was based on text chat.
Manipulation of process accountability. Following procedures used in Scholten
et al. (2007) groups in the individual process accountability condition were
told that there would be an individual interview following the group discus-
sion asking them to explain how their group reached their solutions to the
case, and how they used information to make their judgments. Participants
were told that they should therefore pay attention not only to the task infor-
mation, but also to how they worked, and were encouraged to take notes for
the interview to remember how they made their decisions. Participants in the
group process accountability condition were told that there would be a group
interview following the group discussion, and all the remaining instructions
were the same as in the individual process accountability manipulation. No
interviews were actually conducted, and this mild deception was debriefed.
Five HPO and nine LPO groups were randomly assigned to the group
Figure 1. The experimental site in Second Life.
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
162 Small Group Research 45(2)
accountability condition, and seven HPO and five LPO groups were ran-
domly assigned to the individual accountability condition.
Task. The task, adapted from Wheeler and Mennecke (1992), presents com-
plex information about a high school that is experiencing a number of prob-
lems related to budgets, enrollments, teaching quality, and so forth. The
groups were asked to identify what they thought was the main problem facing
the school, and to prepare a written set of recommendations for addressing
the problem. We assessed performance quality based on which problems
groups identified and the kinds of recommendations they produced using
Wheeler and Mennecke’s scoring protocol. The participants were given 40
min to complete the task. Once they were finished, the participants were
given a link to the post-task survey and then they were debriefed and pro-
vided with the agreed compensation (US$10).
Measurements
Group procedural order. Participants were asked to rate their discussions on a
seven-point semantic differential scale consisting of six bipolar adjectives
and phrases (organized vs. unorganized; time-conscious vs. not time-con-
scious; flexible vs. inflexible; cyclical vs. linear; task-focused vs. not task-
focused; followed no particular order vs. followed a definite order), developed
based on guidance from Putnam (1979). To check whether group composi-
tion produced the expected procedural patterns, participants were asked to
rate on a five-point scale (1 = strongly disagree; 5 = strongly agree) four
descriptive questions adapted from the GPOQ (Putnam, 1979): (a) “Our
group decided to follow a procedure for making a decision and suddenly they
seemed to have forgotten it”; (b) “My group members brainstormed wild
ideas for the fun of it”; (c) “Our group often digressed”; and (d) “In general,
people in my group followed an orderly process.” The four questions were
combined into an index and labeled perception of order (α = .63).
Group process accountability. Five questions (α = .82) used in previous studies
(e.g., Scholten et al., 2007) were used to assess if participants experienced
individual or group process accountability. On a scale of 1 to 5, participants
were asked, for example, to what extent they agreed with the following sen-
tence: “While working on the task with my group, I (my group) took into
consideration that I (we) would be interviewed after we finished the task.”
Epistemic motivation. Participants were asked to rate their motivation to think
and process the information systematically by three questions (α = .74),
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
Liu and McLeod 163
which were used by De Dreu et al. (1999, p. 352): (a) “I tried to take into
consideration all possible perspectives,” (b) “I tried to make judgments and
decisions as thoroughly as possible,” and (c) “I thought deeply before making
a decision.”
Decision quality. A group’s decision quality was evaluated by two coders
using Wheeler and Mennecke’s (1992) scoring manual. This manual consists
of more than 300 possible solutions with two sets of scores that were vali-
dated by a team of 13 raters in their study. Each group solution was given a
problem-solving score for how well the solution solved the problems in the
case, and a feasibility score for the extent to which a solution is feasible
within the constraints of the case. The two coders, one of the authors and an
undergraduate research assistant unaware of treatment condition, both coded
the 26 groups independently by finding the best match between the groups’
decisions and the potential answers with the scores listed in the manual (α =
.77). Disagreements between the two coders were discussed and the recon-
ciled scores were used in the analyses.
Coding of group interaction. A codebook adapted from Putnam’s (1983) coding
system of group procedure order was applied to analyze each group’s proce-
dural activity. The adapted coding scheme consists of 11 categories, which
includes the original 10 categories and two newly created ones. We found
that Putnam’s original category topic change without heading could not ade-
quately capture a topic change that is still related to the general theme of
discussion, so we created two new categories: relevant topic change without
heading and the category irrelevant topic change without heading (see appen-
dix for the definitions of each category).
The transcripts of two groups that were to be excluded from the sample
(one was a dyad and the other contained one member who had not provided
the appropriate consent information) were first used for the coding training,
and another five groups’ transcripts were added for the training. Two coders,
one of the authors and an undergraduate research assistant, practiced for
approximately 9 hr before attaining inter-coder reliability of Krippendorff’s
alpha (Hayes & Krippendorff, 2007) = .8. They then coded the transcripts,
which were masked as to experimental condition.
Following Putnam’s (1983) directions for using the coding system,2 the cod-
ing unit is “a complete thought unit on a particular topic until that topic changes
or until a procedural message is initiated” (p. 4). Discrepancies were settled
through discussion and the reconciled codes were used for the final data analy-
sis. The overall inter-coder reliability was k = .86 among the 26 transcripts and
the category-by-category reliabilities are listed as follows: agenda = .94;
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
164 Small Group Research 45(2)
goals = .92; summaries = .91; division of labor = .92; procedural direction =
.89; topic change with heading = .89; continue discussion after topic change
with heading = .84; relevant topic change without heading = .75; irrelevant
topic change without heading = .76; continue discussion after topic change
without heading = .84; digression = .93. Guetzkow’s (1950) unitizing reliabil-
ity was .99 (U = .01). Two dependent variables were then created by combining
the counts in categories. That is, (a) counts in the categories of agenda, goal,
summarizing division of labor, procedural directions, and changing topics with
general headings were combined to constitute HPO activities and (b) counts in
the categories of relevant/irrelevant topic change without heading, continue
discussion after change without heading, and digression were aggregated and
defined as LPO activities.
Finally, we examined information exchange, measured by the ratio of the
information mentioned during discussion out of the total pieces of informa-
tion provided in the task (Lu, Yuan, & McLeod, 2012). One of the authors and
another undergraduate research assistant, unaware of treatment condition,
used a checklist that included every piece of information in the task to count
the number of pieces of information mentioned in each group’s discussion.
Discrepancies were reconciled through discussion and the final numbers
were used to calculate the proportion of information exchange out of the total
pieces of information. The inter-coder reliability was Krippendorff’s α = .73.
Results
Manipulation Check
To examine if the manipulations of both individual and group process
accountability stimulated epistemic motivation, we calculated the correlation
between the survey questions regarding perception of process accountability
and epistemic motivation and found that perception of process accountability
was significantly associated with epistemic motivation, r = .26, p = .02.
Group Procedural Activities
Because of the nested structure of the data, we first calculated intraclass cor-
relations (ICC) to estimate the effect of group membership on the dependent
variables, and all ICCs were low and non-significant. We therefore used uni-
variate statistics to test all hypotheses. Given the small final sample size (N =
26), we conducted power analyses using the statistical application G*Power
3 (Faul, Erdfelder, Lang, & Buchner, 2007). Using Cohen’s (1969) effect size
criteria, our power for a large effect size (d = .80) was .50, for a medium
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
Liu and McLeod 165
Table 1. Effects of Procedural Order Preference and Process Accountability on
Procedural Acts; Info-Exchange; Perception of Epistemic Motivation, Order, and
Process Accountability; and Task Performance.
Group composition
HPO groups LPO groups
Variable
IPA (n = 7)
M (SD)
GPA (n = 5)
M (SD)
IPA (n = 5)
M (SD)
GPA (n =9)
M (SD)
Proportion of high
procedural acts
0.44 (0.05) 0.42 (0.06) 0.48 (0.06) 0.48 (0.04)
Proportion of info-
exchange
0.33 (0.04) 0.46 (0.04) 0.29 (0.04) 0.33 (0.03)
Perception of epistemic
motivation
4.38 (0.48) 4.28 (0.46) 4.07 (0.66) 4.11 (0.43)
Perception of process
accountability
3.53 (1.12) 2.89 (1.02) 3.56 (0.85) 3.22 (0.94)
Perception of order 4.37 (0.60) 4.43 (0.45) 4.20 (0.71) 3.97 (0.54)
Semantic differential scales
Organized 6.32 (0.90) 5.95 (1.13) 5.67 (1.29) 4.78 (1.72)
Time-conscious 6.20 (1.04) 5.89 (1.29) 5.67 (1.84) 4.93 (1.62)
Inflexible 1.68 (1.07) 2.00 (1.10) 1.67 (0.72) 2.63 (1.45)
Linear 3.28 (1.95) 3.68 (2.03) 4.27 (1.49) 3.74 (1.51)
Task-focused 6.76 (0.44) 6.32 (1.06) 6.67 (0.62) 6.15 (1.10)
Followed a definite
order
4.68 (2.19) 4.95 (1.84) 4.87 (1.60) 4.26 (1.97)
Problem-solving score 68.10 (7.47) 65.87 (4.98) 64.12 (3.15) 68.25 (5.49)
Feasibility score 72.44 (4.65) 67.93 (16.76) 48.42 (21.85) 71.05 (8.52)
Note. HPO = high procedural order preference; LPO = low procedural order preference;
IPA = individual process accountability; GPA = group process accountability.
effect size (d = .50) was .23, and for a small effect size (d = .20) was .08. The
accepted criterion is that ≥.80 is statistically powerful (Cohen, 1969). We will
take the relatively low power of our tests into account in the discussion of the
results.
Table 1 presents the means and standard deviations of all dependent vari-
ables for the experimental conditions. It should be first noted that HPO
groups displayed a significantly higher total of procedural acts than did LPO
groups, MHPO = 133.20, MLPO = 93.41, F(1, 22) = 6.71, p = .02, ω2 = .18. The
proportion of HPO procedures was used to test the hypotheses.3 To test the
prediction of H1 that HPO groups would use more structured procedures
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
166 Small Group Research 45(2)
(i.e., HPO procedures) than would LPO groups, and that LPO groups would
display more unstructured procedures (i.e., LPO procedures) than would
HPO groups, we conducted a 2GroupComposition × 2ProcessAccountability ANOVA on
the proportion of HPO procedures. The group composition main effect indi-
cated that there was no significant difference between HPO and LPO groups.
H1 is not supported.
H2 predicted that LPO procedural acts would benefit LPO groups more
than HPO groups whereas HPO procedural acts would benefit HPO groups
more than LPO groups. This hypothesis involves the interaction effect
between group composition and the proportion of procedural acts. To test this
hypothesis, we created dummy variables for group composition and interac-
tions of group composition and both kinds of procedural acts, and regressed
the two performance measures, problem-solving score and feasibility score,
respectively, on these variables.4 The regression coefficients for the interac-
tion terms provide the hypothesis test. We found that the impact of LPO pro-
cedures on problem-solving scores was marginally stronger for HPO groups
than for LPO groups, t(24) = 1.78, p = .09, ω2 = .08, but no other differences
were found. These results offer no support for H2.
H3 queried whether there were differences between the individual and the
group accountability conditions in group members’ perception of process
accountability, epistemic motivation, information exchange, and group perfor-
mance. As predicted, significantly higher perception of process accountability
was found in the individual process accountability condition, F(1, 22) = 4.81, p =
.03, ω2 = .04, but opposite to expectations, we found significantly lower informa-
tion exchange in this condition, F(1, 22) = 5.22, p = .03, ω2 = .13. The means for
perceptions of epistemic motivation and the two task performance measures did
not significantly differ. H3 received only partial support.
The Impact of Process Accountability
H4a to H4c predicted that process accountability would amplify the effect of
procedural order preference whereas hypotheses H6a to H6c predicts oppo-
site effects based on cognitive dissonance. Overall, ANOVA results showed
more support for the amplifying effect than for the cognitive dissonance
effect. We found that compared with LPO groups, HPO groups exchanged
more information (H4a), F(1, 22) = 4.56, p = .04, ω2 = .11; had higher percep-
tion of epistemic motivation (H4b), F(1, 22) = 4.77, p = .03, ω2 = .04; and
perceived their discussions to be more orderly (H4b), F(1, 82) = 4.77, p = .03,
ω2 = .04. The groups did not differ in the perception of process accountability
(H4b). The results of participants’ responses to the six bipolar descriptions
provided further support for H4b: HPO individuals described their groups as
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
Liu and McLeod 167
significantly more organized, F(1, 82) = 9.74, p < .001, ω2 = .09, and as more
time-conscious, F(1, 82) = 5.51, p = .02, ω2 = .05, than did LPO individuals.
With respect to performance (H4c), HPO groups had marginally higher fea-
sibility scores than did LPO groups, F(1, 22) = 3.93, p = .06, ω2 = .08, but as
will be discussed below, this was qualified by a significant interaction effect
with level of process accountability. Scores on the problem-solving measure
did not significantly differ.
H5 examined whether the difference between HPO and LPO groups would
be greater under the individual process accountability condition than the
group accountability condition. Consistent with the hypothesis, the differ-
ence in the solution feasibility score between the HPO and LPO groups was
significantly larger under individual than group process accountability, F(1,
22) = 6.62, p = .02, ω2 = .16. No significant interactions were found for any
of the other measures; thus, H5 has limited support.
Discussion
This research examined the relationship between individual preference for
procedural order, group procedural activities, and group performance, and
further explored the impact of process accountability on group decision-mak-
ing process and decision quality. Consistent with the findings of Wheeler et
al. (1993) that HPO groups participated more than did LPO groups, we also
saw a higher total number of procedures in HPO groups, but the proportion
of high versus low structure procedures did not differ. We also did not find
much evidence that the effects of procedures differed between LPO and HPO
groups. The most important findings of this study relate to the central interest
in the interaction between procedural order preferences and process
accountability.
First, the hypothesized main effect of stronger effects under individual
than group process accountability generally held. Groups by and large
showed more indicators of process vigilance when the members were held
individually accountable. The findings for information exchange were oppo-
site to this prediction, however. Based on van Hiel and Schittekatte’s (1998)
findings that groups took longer to reach decisions under individual than col-
lective accountability, we speculate that groups in the individual process
accountability condition may have spent more time in reaching consensus on
the information discussed thus leaving them less time to exchange additional
information than did the groups in the group process accountability
condition.
Second, we found that individual process accountability amplified the
effects of procedural order preference on group members’ perceptions and
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
168 Small Group Research 45(2)
performance such that HPO groups exchanged more information, had higher
perception of epistemic motivation, and perception of order, and higher fea-
sibility scores. Moreover, the interaction showed that the difference in feasi-
bility score was higher under the individual than collective process
accountability. Examination of the means in Table 1 shows that the interac-
tion between procedural order preference and process accountability is
accounted for primarily by the very negative impact on LPO groups’ feasibil-
ity score under the individual process accountability condition. The scores in
the other three conditions essentially are not different from each other. This
pattern suggests that individual process accountability may have exerted
more pressure on LPO groups than on HPO groups, given that LPO groups
tend to be more flexible in their decision-making process. This mismatch
between preferences and task demands appears to have had a detrimental
effect rather than a rallying effect on these groups. Burke and Aytes (2001)
found, however, that over time LPO groups eventually responded to demands
for greater structure. Hirokawa et al. (1988) reported that LPO groups tended
to perform better under low structure conditions, but that because of their
greater flexibility the effect of structure was lower for them than for HPO
groups. The data from our study suggest a stronger effect of structure on LPO
groups. Taking into consideration the Burke and Aytes findings suggests that
the flexibility of LPO groups may manifest over time. A fruitful direction for
future research would be further examination of the temporal effects of pro-
cedural order preferences and process accountability. These findings also
point to the importance of additional investigation into the effects of match-
ing work habit preferences with task demands.
We were puzzled by why we found effects on task performance for the
feasibility but not for the problem-solving scores. We suspect that the reason
may be related to inherent differences in the properties of these two mea-
sures. Based on the way Wheeler and Mennecke (1992) described the scoring
it may be that the feasibility scores are evaluated by stricter criteria. That is,
to have high feasibility scores, solutions would have to meet several specific
constraints described in the task whereas problem-solving involve a more
global judgment of the extent to which the solutions would fix the problem.
The easier task of providing a solution that could solve the problem would
result in relatively little variance on that measure, compared with the more
difficult task of providing solutions that fit within specified constraints of
feasibility. In support of this explanation, we indeed saw more variance in the
feasibility scores, range = 71.50; SD = 15.26, than in the problem-solving
scores, range = 24.33; SD = 5.60.
Several limitations of this study affect the interpretation of the results and
point to possible directions for future research. First, we examined only
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
Liu and McLeod 169
homogeneous groups, whereas groups composed of people who all prefer
one type of procedural order is rarely the case in daily life. It would be impor-
tant and interesting to study how heterogeneous groups negotiate a single
approach to using procedures. It has been suggested that such differences in
work values could trigger group faultlines, which may cause more damage to
group processes and performances than faultlines resulting from demo-
graphic diversity (e.g., Thatcher & Patel, 2011).
Second is the small sample size. Although over 500 individuals partici-
pated in the first phase of the research, the need to ensure that we had groups
characterized by high and low procedural order preference severely reduced
the final sample. This is a difficulty inherent in research designs of this type
that depend on the distribution the population of some specific characteris-
tics. As indicated by the post hoc power analyses, the highest power we could
have achieved with this sample size was .50, well below the conventional
criterion of .80. Small samples also limit the ability to use some more
advanced statistical techniques, such as structural equation modeling, which
may allow the examination of some of the more dynamic aspect of the data,
for example, understanding how and to what extent individual difference in
preference for group procedure shapes the group decision-making process
and, in turn, affects group decision quality. Related to the issue of small sam-
ple size, we eliminated a no process accountability control condition to maxi-
mize the numbers within each of the other conditions. As a result, it must be
kept in mind that our results for process accountability need to be interpreted
as individual accountability relative to group-level accountability. Controlling
for the main effect of accountability would be an important extension to these
findings.
Despite these limitations, an implication can be drawn from this study that
individual preference for procedural order is an important consideration
when deciding how to structure group discussion and how to motivate mem-
bers of work teams. People who naturally are inclined toward structure seem
to be able to function well when asked to work under conditions where struc-
ture is emphasized, but putting such constraints on people disinclined toward
structure may be harmful, at least in the short run.
Appendix
Codebook for Group Procedural Order5
1. Agenda: Requests or suggests deadlines, agendas, or list of activities
a. Messages that concern the order or arrangement of the in-group
system activities.
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
170 Small Group Research 45(2)
b. Messages that mention a timetable, an agenda in use, or an
agenda item proposed for that purpose.
c. Signposts or summaries at the beginning of the meeting, which
serve the purpose of setting an agenda for the discussion.
d. Development of a list or rank order of alternatives—the mes-
sages that initiate this and organize this list.
2. Goals: Requests or makes statements about group goals or group
jurisdiction
a. Message that sets goals for the group—goals differ from agendas
in establishing objectives for what the group should be doing,
but not necessarily placing these issues on the agenda for group
interaction.
b. Group goals deal with the overall purpose or responsibilities of
the group.
c. Group goals also question the jurisdiction of the group. Any mes-
sages that discuss whether this is the responsibility of this par-
ticular group.
3. Summaries: Summarizes and integrates contribution (exclude sign-
post summaries)
a. Messages that provide an internal review of what has been dis-
cussed.
b. Responses that attempt to integrate one person’s comments with
a previous statement or observation.
4. Division of labor: Suggests or requests division of labor or implemen-
tation of a course of action
a. Messages that attempt to divide responsibilities among the
immediate group members.
b. Messages that address how the group should implement a deci-
sion or a course of action.
c. The primary focus is the group system and the duties or functions
within this particular group.
d. Discussion of how to carry out decisions made by the group. This
may involve procedures but not the internal group procedures as
much as the notion of where do we go from here. Remember to
make sure the group is trying to implement a decision or divide
its labors to accomplish the task.
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
Liu and McLeod 171
5. Request procedural direction: Requests or suggests procedural
direction.
a. Comments with metacommunication on the procedural habits of
the group.
b. Comments that question whether the group has strayed from the
topic, or whether the group has time to finish their agenda.
c. Messages that deal with clarifying and repeating the internal pro-
cedural matters of running the group meeting.
6. Topic changes with heading (abstract label): Changes task-related
topic of discussion by introducing an abstract label or general heading
that suggests a procedural activity.
a. Changes or switches the topic that was previously under discus-
sion and must facilitate the group’s attention.
b. Introduces a label or category that would subsume a number
of particulars or specific details. This label or category should
resemble what one would list on a topic outline for a speech or
a paper. The category is a general heading like refreshments to
cover food and drink at a party; place to cover park or bar; time
to cover day, week, and hour.
7. Continue discussion after topic change with heading: Continues task-
related discussion initiated by a general heading or by an agenda
category.
8. Relevant topic change without heading: Changes task-related discus-
sion by switching topics via jumping from specific detail to specific
detail. Content switches by moving from the details of one category
to the details of another category.
9. Irrelevant topic changes without heading: Changes topic by switch-
ing topics via jumping from specific detail of one category to details
of another irrelevant category. The change is not detected or discussed
by following comment(s).
10. Continue discussion after topic change without heading: Messages
that follow detailed topic change and do not initiate a procedural
activity or a digression.
11. Digression: Message that digresses from the group task to socio-emo-
tional issues. Contribution that strays from the here-and-now interac-
tion that have little bearing on the group. The digression does not
have to address social issues—it simply has to indicate a shift from
the immediate concerns of the group to other events, which have very
little to do with this particular group per se.
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
172 Small Group Research 45(2)
Authors’ Note
A previous version of this work was presented at the 2013 INGRoup Conference,
Atlanta, Georgia. Any opinions, findings, conclusions, or recommendations expressed
in this publication are those of the author(s) and do not necessarily reflect the view of
the U.S. Department of Agriculture.
Acknowledgment
The authors thank Linda Putnam for providing the coding manual for procedural mes-
sages in group interaction and Bradley C. Wheeler for providing the School of
Business Policy Task manual.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research,
authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research,
authorship, and/or publication of this article: This research was supported in part by
the Cornell Agricultural Experiment Station federal formula funds, Project No. NYC-
131413 received from Cooperative State Research, Education, and Extension Service,
U.S. Department of Agriculture.
Notes
1. See www.markettruths.com.
2. Please see Putnam (1983) for the concise introduction of the coding scheme. The
more specific coding instructions (including the definition of what constitutes a
coding unit) were obtained directly from Linda Putnam.
3. All procedures were coded into either the high procedural order preference
(HPO) or the low procedural order preference (LPO) category, so the proportion
of one gives complete information about the other category.
4. For a discussion of these procedures, see http://www.ats.ucla.edu/stat/spss/faq/
compreg2.htm
5. This codebook is adapted from Putnam (1983), and the coding manual with
examples and instructions were provided directly by Linda Putnam.
References
Bormann, E. G. (1975). Discussion and group methods: Theory and practice. New
York, NY: Harper and Row.
Burke, K., & Aytes, K. (2001, January). Preference for procedural ordering in dis-
tributed groups: How do media and repeated interaction affect perceptions and
procedural structuring? Proceedings of the 34th Hawaii International Conference
on System Sciences, Washington, DC.
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
Liu and McLeod 173
Chaiken, S., & Trope, Y. (1999). Dual-process theories in social psychology. New
York, NY: Guilford Press.
Cohen, J. (1969). Statistical power analysis for the behavioral sciences. New York,
NY: Academic Press.
Cruz, M. G., Henningsen, D. D., & Williams, M. L. M. (2000). The presence of
norms in the absence of groups? The impact of normative influence under
hidden-profile conditions. Human Communication Research, 26, 105-124.
doi:10.1111/j.1468-2958.2000.tb00752.x
Davis, J. H., Tindale, R. S., Nagao, D. H., Hinsz, V. B., & Robertson, B. (1984).
Order effects in multiple decisions by groups: A demonstration with mock juries
and trial procedures. Journal of Personality and Social Psychology, 47, 1003-
1012. doi:10.1037/0022-3514.47.5.1003
Dreu, C. K. W. De, Koole, S. L., & Steinel, W. (2000). Unfixing the fixed pie : A
motivated information-processing approach to integrative negotiation. Journal
of Personality and Social Psychology, 79, 975–987. doi:10.1037//0022-
35I4.79.6.975
De Dreu, C. K. W., Koole, S. L., & Oldersma, F. L. (1999). On the seizing and freez-
ing of negotiator inferences: Need for cognitive closure moderates the use of heu-
ristics in negotiation. Personality and Social Psychology Bulletin, 25, 348-362.
doi:10.1177/0146167299025003007
De Dreu, C. K. W., Nijstad, B. A., & van Knippenberg, D. (2008). Motivated informa-
tion processing in group judgment and decision making. Personality and Social
Psychology Review, 12, 22-49. doi:10.1177/1088868307304092
Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: A flexible
statistical power analysis program for the social, behavioral, and biomedical sci-
ences. Behavior Research Methods, 39, 175-191. doi:10.3758/BF03193146
Guetzkow, H. (1950). Unitizing and categorizing problems in coding qualitative data.
Journal of Clinical Psychology, 6, 47-58. doi:10.1002/1097-4679(195001)
Hayes, A. F., & Krippendorff, K. (2007). Answering the call for a standard reliabil-
ity measure for coding data. Communication Methods and Measures, 1, 77-89.
doi:10.1080/19312450709336664
Hirokawa, R. Y., Ice, R., & Cook, J. (1988). Preference for procedural order, discus-
sion structure and group decision performance. Communication Quarterly, 36,
217-226. doi:10.1080/01463378809369724
Kerschreiter, R., Schulz-Hardt, S., Mojzisch, A., & Frey, D. (2008). Biased informa-
tion search in homogeneous groups: Confidence as a moderator for the effect of
anticipated task requirements. Personality and Social Psychology Bulletin, 34,
679-691. doi:10.1177/0146167207313934
Kroon, M. B. R., Hart, P., & van Kreveld, D. (1991). Managing group decision
making processes: Individual versus collective accountability and group think.
International Journal of Conflict Management, 2, 91-115. doi:10.1108/eb022695
Kroon, M. B. R., van Kreveld, D., & Rabbie, J. M. (1992). Group versus individ-
ual decision making: Effects of accountability and gender on groupthink. Small
Group Research, 23, 427–458. doi:10.1177/1046496492234001
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
174 Small Group Research 45(2)
Lu, L., Yuan, Y. C., & McLeod, P. L. (2012). Twenty-five years of hidden profiles
in group decision making: A meta-analysis. Personality and Social Psychology
Review, 16, 54-75. doi:10.1177/1088868311417243
O’Connor, K. M. (1997). Groups and solos in context: The effects of accountability
on team negotiation. Organizational Behavior and Human Decision Processes,
72, 384-407. doi:10.1006/obhd.1997.2743
Pavitt, T. C. (1993). What (little) we know about formal group discussion proce-
dures: A review of relevant research. Small Group Research, 24, 217-235.
doi:10.1177/1046496493242004
Poole, M. S., Holmes, M. E., Watson, R., & DeSanctis, G. (1993). Group decision
support systems and group communication: A comparison of decision making
processes in computer-supported and nonsupported groups. Communication
Research, 20, 176-213. doi:10.1177/009365093020002002
Putnam, L. L. (1976). Construction and testing of a questionnaire to assess an indi-
vidual’s preference for procedural order in small task-oriented groups (Doctoral
dissertation). University of Minnesota. Available from https://www.lib.umn.edu/
Putnam, L. L. (1979). Preference for procedural order in task-oriented small groups.
Communication Monographs, 46, 193-218. doi:10.1080/03637757909376006
Putnam, L. L. (1983). Small group work climates: A lag-sequential analysis of group inter-
action. Small Group Research, 14, 465-494. doi:10.1177/104649648301400405
Scholten, L., van Knippenberg, D., Nijstad, B. A., & De Dreu, C. K. W. (2007).
Motivated information processing and group decision-making: Effects of pro-
cess accountability on information processing and decision quality. Journal of
Experimental Social Psychology, 43, 539-552. doi:10.1016/j.jesp.2006.05.010
Thatcher, S. M. B., & Patel, P. C. (2011). Group faultlines: A review, integra-
tion, and guide to future research. Journal of Management, 38, 969-1009.
doi:10.1177/0149206311426187
van Hiel, A., & Schittekatte, M. (1998). Information exchange in context: Effects
of gender composition of group, accountability, and intergroup perception on
group decision making. Journal of Applied Social Psychology, 28, 2049-2067.
doi:10.1111/j.1559-1816.1998.tb01360.x
van Swol, L. M., Savadori, L., & Sniezek, J. A. (2003). Factors that may affect the dif-
ficulty of uncovering hidden profiles. Group Processes & Intergroup Relations,
6, 285-304. doi:10.1177/13684302030063005
Wheeler, B. C., & Mennecke, B. (1992). The school of business policy task man-
ual (Working Paper No. 92-524, Unpublished manuscript). Indiana University
Graduate School of Business, Bloomington. Retrieved from https://scholarworks.
iu.edu/dspace/bitstream/handle/2022/13427/policy_task_manual_wheeler.
pdf?sequence=1
Wheeler, B. C., Mennecke, B., & Scudder, J. N. (1993). Restrictive group support
systems as a source of process structure for high and low procedural order for
groups. Small Group Research, 24, 504-522. doi:10.1177/1046496493244005
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
Liu and McLeod 175
Author Biographies
Yi-Ching Liu is a doctoral candidate in the Department of Communication at Cornell
University. Her research interests in the field of group communication include indi-
vidual differences and motivated information sharing, social influence in group inter-
action process, and the impact of communication technology.
Poppy Lauretta McLeod is an associate professor at Cornell University. She earned
her PhD in social psychology from Harvard University. Her research interests include
technology usage, information exchange, decision-making, social identity and social
influence within small groups, and intergroup communication.
at CORNELL UNIV on July 21, 2014sgr.sagepub.comDownloaded from
