Content uploaded by Mary Jo Hatch
Author content
All content in this area was uploaded by Mary Jo Hatch on Oct 29, 2015
Content may be subject to copyright.
Physical Barriers, Task Characteristics, and Interaction Activity in Research and Development
Firms
Author(s): Mary Jo Hatch
Source:
Administrative Science Quarterly,
Vol. 32, No. 3 (Sep., 1987), pp. 387-399
Published by: Johnson Graduate School of Management, Cornell University
Stable URL: http://www.jstor.org/stable/2392911
Accessed: 30/11/2010 15:24
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=cjohn.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Johnson Graduate School of Management, Cornell University is collaborating with JSTOR to digitize, preserve
and extend access to Administrative Science Quarterly.
http://www.jstor.org
Physical Barriers, Task
Characteristics, and
Interaction Activity
in Research and
Development Firms
Mary Jo Hatch
San Diego State University
? 1987 by Cornell University.
0001 -8392/3203-0387/$1 .00.
This research is based on a Ph.D. disser-
tation completed at Stanford University. I
would like to express my deep apprecia-
tion to the members of my committee,
Jeffrey Pfeffer (Chair),
Harold Leavitt, and
Jerry Porras. Gerald Salancik and three
anonymous ASQ reviewers provided sub-
stantial help with earlier drafts of this
paper. I would also like to thank Jai Ghor-
pade, Mark Butler, and Sanford Ehrlich.
A field study was used to examine the common belief that
barriers around offices are desirable because they reduce
interaction, thus allowing more time for accomplishing
tasks. If this were true, we would expect to find more in-
teraction in offices with fewer barriers. The opposite was
found in the field study conducted in two high-technology
firms. Partition height, number of partitions, and the use
of a door or a secretary were all positively associated with
one or more forms of interaction activity. A desk posi-
tioned away from the office entrance was the only barrier
found to be negatively associated with interaction. These
relationships were independent of variables representing
task characteristics, job type, work experience, demo-
graphic characteristics and sociability.'
Physical settings provide contexts for behavior. They are
thought to have influence through their ability to support the
range of activities that becomes associated with them and to
constrain other forms of activity (Barker, 1968; Wicker, 1979).
Several organizational theorists have proposed models in
which communication and social interaction are offered as the
primary work-related activities affected by physical structure
in organizations (Steele, 1973, 1986; Becker, 1981; Pfeffer,
1982; Sundstrom, 1986). This study explores the relation-
ships between several characteristics of physical structure
and interaction activity found to occur in two research and
development firms.
At least three aspects of interaction have been linked with
physical structure in organizations: choice of interaction
partners, type of interaction (e.g., face to face, formal versus
informal), and amount of interaction. Two dimensions of
physical structure have been investigated with respect to in-
teraction activity: (1) distance between potential communica-
tion partners and (2) degree of enclosure provided by physical
barriers.
Clear evidence of a relationship between distance and choice
of interaction partners has emerged from the literature (Gulla-
horn, 1952; Wells, 1965; Gerstberger and Allen, 1968;
Conrath, 1973; Allen and Fusfeld, 1974; Szilagyi and Holland,
1980). All of these researchers found an inverse relationship
between the distance separating employees and the likeli-
hood of communication occurring between them. However,
Conrath (1973) found evidence that suggests that this rela-
tionship may be specific to certain types of interaction. He
found distance to be negatively correlated with face-to-face
contact and telephone calls, but not with written communica-
tion between pairs of employees.
The relationship between interaction and enclosure is far less
clear than that between interaction and distance. While a
number of researchers have made statements to the effect
that a greater amount of interaction occurs in offices that im-
pose fewer physical barriers on their occupants (Mehrabian,
1976; Oldham and Brass, 1979; Becker, 1981; Sutton and
Rafaeli, 1987), empirical evidence regarding this relationship
is mixed.
In support of a negative relationship between barriers and in-
teraction, Gullahorn (1952) found that clerical workers whose
desks were separated by filing cabinets engaged in less
387/Administrative
Science Quarterly,
32 (1987): 387-399
conversation with each other than did workers who were not
separated. Similarly, Brookes and Kaplan (1972), Allen and
Gerstberger (1973), and Szilagyi and Holland (1980) all re-
ported that moves from closed offices (floor-to-ceiling walls
and doors) to open offices (unenclosed or partially
enclosed
by moveable partitions) were associated with increased op-
portunities to interact or improved ease of communication. A
positive relationship between enclosure and interaction was
suggested by Oldham and Brass (1979), who found that per-
ceptions of supervisor feedback and friendship opportunities
decreased with a move to open offices. Sundstrom, Herbert,
and Brown (1982) reported different relationships between
enclosure and interaction depending on the type of interac-
tion. They discovered that supervision and interdepartmental
contact increased while confidential conversations decreased
following a move from a closed to an open office. The Buffalo
Organization for Social and Technological Innovation (BOSTI,
1981) also reported enclosure to be positively associated with
confidential conversations.
The available evidence suggests that specifying types of in-
teraction may be critical to understanding the relationships
between physical structure and interaction activity. Also,
compared with distance, which has been studied, the issue of
enclosure by physical barriers appears to be the aspect of
physical structure in greatest need of further research. There-
fore, the relationships between physical barriers and type and
amount of interaction form the focal point of this study.
AN EXPLORATORY
MODEL OF INTERACTION ACTIVITY
Since theory concerning the relationships between interaction
activity and enclosure by physical barriers is limited and the
empirical evidence contradictory, an exploratory approach was
taken in this study. An initial set of physical structure and ac-
tivity variables was derived from previous research and was
extended and refined in a pilot study conducted in an organi-
zation. Also, because interaction has received considerable
attention from investigators working outside the domain of
physical structure research, some additional variables shown
to be associated with interaction (described below) were in-
cluded.
This study differs from previous research in three ways. First,
it focuses on different types of interaction activities than
have been studied in the past, and it uses a different method
for measuring these activities- self-reported-activity logs.
Second, it draws together a set of physical-barrier
measures
that have not been examined together before. Third,
this
study includes controls for some other obvious contributors
to interaction in organizations so as to better assert the inde-
pendent relationships of physical barriers with the activities
examined.
Interaction Activity
Interest in identifying categories of work activity has focused
primarily
on the role of the manager (e.g., Mahoney, Jerdee,
and Carroll, 1963, 1965; Mintzberg, 1973; Kurke
and Aldrich,
1983; Carroll and Gillen, 1987). No typology of office work
activity has been offered in the literature, although a few vari-
ables have been examined: confidential conversation (BOSTI,
388/ASQ, September 1987
Barriers and Interaction
1981; Sundstrom, Herbert, and Brown, 1982), face-to-face in-
teraction (Conrath, 1973; Wofford, Gerloff, and Cummins,
1977), and telecommunication (Conrath, 1973; Goddard,
1973; Galitz, 1980). However, none of these variables is
presently in common use.
Because type of interaction was to be a central focus of this
study, and because no particular
set of activities was indi-
cated in the literature, a pilot study was conducted to identify
the activities to be examined (see below for a discussion of
the pilot study). An exhaustive set of nine activities was se-
lected: work alone, work together, meetings, interruptions,
build relationships, socialize, phone or computer mail, breaks,
and personal time.
Physical Barriers
Physical enclosure is provided by barriers such as doors,
walls, and partitions (Archea, 1977). Many studies of enclo-
sure have not used direct measures of physical barriers, but
have relied on distinctions of open versus closed offices. A
few studies have operationalized enclosure using an index of
the numbers and types of partitions around workspaces
(BOSTI, 1981; Oldham and Rotchford, 1983). In this study,
number of partitions, height of partitions, and use of a door
are examined as separate measures.
Desk position and the placement of a secretary are also po-
tentially important barriers. Facing an office door, as opposed
to facing away from it, may imply a readiness for interaction
that increases its occurrence (Joiner, 1976). The use of a
secretary to screen callers may imply a lack of willingness to
interact, although a turned back or a secretary receiving calls
may also be a social rather than a physical barrier. Neverthe-
less, in the interest of completeness, in this study, desk posi-
tion and secretarial screening were both examined.
Other Contributors to Interaction Activity
Physical structure is believed to be embedded in a web of in-
fluences on behavior. In particular,
the tasks performed and
the social positions held within the organization are related to
work activities in systematic ways (e.g., Katz and Kahn, 1978;
Rousseau, 1978; Brass, 1985). Similarly, since no two people
approach the same task identically, individual differences are
expected in the distribution of employees' activities at work.
Therefore, physical barriers were examined along with vari-
ables representing task characteristics and individual
differ-
ences. Unmeasured aspects of jobs and firms were
controlled using dummy variables.
Interaction has been linked to at least three characteristics of
task structure: position level, task interdependence, and un-
certainty. These three were examined here, while other task
characteristics are left for future studies.
Task characteristics. Empirical
evidence of the importance of
sociotechnical variables in physical structure research was
provided by Oldham and Brass (1979) and Oldham and
Rotchford (1983). These studies found that job characteristics
moderate relationships between office design and satisfaction
and motivation. This study, however, focuses on interaction
389/ASQ, September 1-987
activity. Because task characteristics are at the same level of
analysis as are task activities, while job characteristics are
conceptually associated with the more aggregate level of
satisfaction and performance, task rather than job
characteristics were selected for study. Oldham and
Rotchford (1983) provided support for this choice with their
finding that job characteristics were unrelated to two
behaviors they examined but were related to their measures
of satisfaction.
Position level reflects an individual's social location within the
organizational hierarchy. As such, position level is associated
with authority, status, and power. These dimensions are rec-
ognized determinants of organizational behavior (e.g., Weber,
1947; Simon, 1957; Pfeffer, 1981). Mintzberg (1973) empiri-
cally linked position level with interaction activity for his
sample of managers, suggesting that position level is posi-
tively correlated with interaction activity.
Task interdependence is defined as the extent to which the
elements upon which work is performed or the work pro-
cesses themselves are interrelated (Scott, 1981). Task inter-
dependence tends to increase the amount of contact required
among workers and should, therefore, be positively related to
interaction activity (Thompson, 1967; Kiggundu, 1981).
Uncertainty. Simon (1957) suggested that individuals reach
their limited capacities for attention much sooner under con-
ditions of uncertainty than under conditions of certainty.
Building on Simon, Williamson (1975) suggested that in-
creasing the number of decision makers alleviates the load on
the capacity for attention of any given member of the pool.
Thus, uncertainty should increase the time individuals spend
interacting (i.e., pooling for attention capacity). Uncertainty
has been shown to be positively correlated with interaction
(Festinger, 1954) and also with task interdependence (Brass,
1985).
Individual differences. In order to determine the extent to
which the relationships found in this study are independent of
individual differences, measures of demographic variables
(age, education, and sex), and work experience (company and
job tenure) are included. Sociability is also included because
some individuals seem more likely than others to engage in
interaction, due to personal preferences, habits, or traits.
Sociability should be positively correlated with interaction
activity. While sociability is but one of a number of trait
variables that might be included, the examination of a more
complete representation was deemed to be beyond the
scope of the present study.
METHOD
Pilot Study
Work was observed and interviews were conducted with 12
volunteers who performed technical and professional work in
a high-technology firm (a different firm from that in the main
study reported below). The subjects responded to an informal
solicitation for support of the study by one of the firm's man-
agers (the author's contact). A typology was constructed from
interaction
activities mentioned in the literature
and from in-
390/ASQ, September 1987
Barriers and Interaction
terviewees' responses to the question: How do you spend
your time? Feedback from participants who used the typology
to log their activities was used to develop and refine the cat-
egories further. Three iterative phases produced the nine
types of activity used in the main study.
The participants' verbal feedback led to several changes in
the log. First, it was made clear that the category labels for
the interaction activities required careful definition to render
reliable interpretations across respondents who would use
the categories to report their activities. For instance, much
concern was expressed over a distinction between socializing
and working together: socializing did not sound work-related,
while working together seemed too task-specific. Many par-
ticipants felt that some of their non-task-specific interactions
were directly relevant to their work and could not be accu-
rately categorized. To alleviate this problem, a category la-
beled "build relationships" was created. Second, participants
felt that it was easier to remember to log their activities and
their entries were more reliable when all activities (not just
interactions) were included in the log. The categories of
working alone and breaks and personal time were added in
response to this suggestion. The addition of these categories
also appeared to set up a response habit that reduced the in-
trusiveness of the logging procedure. Third, it was observed
that using more than nine log categories required substantial
attention each time an activity was logged, while fewer items
quickly became familiar enough to respondents that a glance
was all that was needed to log an entry accurately. These ex-
periences led to the final design of the activity log used by
subjects in the main study.
Sample
A field study was conducted in two high-technology firms in
the San Francisco Bay Area. Both firms that participated in
the main study employed over 8,000 individuals and had been
in business for more than 25 years at the time of the study.
All phases of the study were completed by 99 participants
(87.6 percent response rate). This sample contained indi-
viduals assigned to both open (N = 28; 9 employees from
Firm 1, and 19 from Firm 2) and closed (N = 71; all from Firm
1) office spaces.
The research sites for the study were selected on the basis of
availability
and management's willingness to allow the study
to be conducted. Organizational concerns about security
made it impractical for the researcher personally to select in-
dividuals for the study. After the study was underway, the re-
searcher was allowed unconstrained access. Because initial
access was so limited, liaisons designated by the firms solic-
ited participants with a written invitation to become involved
in a research study. Five liaisons assisted in the recruitment
of study participants. Participation in the study was strictly
voluntary and all materials were coded to provide confidenti-
ality.
Measures
To avoid common-method bias, barriers, task characteristics
and individual
differences, and work activities
were assessed
using different
data-collection
techniques. The researcher
391/ASQ, September 1987
measured barriers using a checklist. Survey items measuring
task and individual-difference variables were administered in
a questionnaire using seven-point, Likert-type scales. Work
activity data were provided by the participants in self-reported
activity logs.
Activities. Participants logged their activities for one week
using preprinted forms listing the nine work activities se-
lected in the pilot study: work alone, work together,
meetings, interruptions, build relationships, socialize, phone
or computer mail, breaks, and personal time. Participants en-
tered the time they began a new activity, making log entries
under the appropriate category heading each time they
changed from one activity to another. The amount of time re-
ported in each column of the log was summed and divided by
the total time reported in all activities. This calculation pro-
duced measures of the percentage of time reported in each
type of activity. Proportions of time rather than absolute time
was used, because the total time reported across individuals
was highly variable, ranging from 1,532 to 3,546 minutes in
the sample studied. Using proportions of time allows statis-
tical control over unmeasured factors influencing the determi-
nation of total time spent at the work site.
Number of partitions. The number of walls or partitions
around the focal individual's primary
work station.
Partition height. The average height of the walls or partitions
around an individual's work station (to a maximum of 8 feet).
Door. When their offices had doors, participants were asked
to estimate the percentage of time their door was closed.
When offices had no door, a score of 0 was assigned.
Desk position. An individual facing away from the door was
coded 1; an individual
facing the door was coded 0.
Secretary. An individual
assigned a secretary who provided
screening was coded 1, and an individual who had no secre-
tarial screening was coded 0.
Position level. Position level data were collected via a single
item on the questionnaire describing the respondent's level of
responsibility: no supervisory duties (coded 1), supervise
others who are not supervisors (2), supervise subordinates
who are supervisors (3), supervise subordinates who are
managers who supervise other managers (4).
Job type. Three jobs were represented in the sample studied:
technicians (N = 41), staff scientists (N = 34), and engineers
(N = 24). Two dummy variables were created to control for
job type differences. Technicians were assigned a 1 on the
variable labeled "technicians," while all others were assigned
a 0. Engineers were coded 1, while technicians and staff sci-
entists were assigned a 0, on the variable labelled "engineers."
Uncertainty. Four items adapted from Hage and Aiken (1969)
were included in the questionnaire to measure routineness.
However, these items were highly intercorrelated with three
uncertainty items developed by Seashore et al. (1983). All
seven items loaded on one factor in a factor analysis. The
routineness items were therefore reverse scored and com-
bined with the uncertainty items. The combined scale mea-
sures the extent to which respondents deal with surprising or
unpredictable situations, deal with new problems, consider
392/ASQ, September 1987
Barriers and Interaction
their jobs to involve uncertainty, perform the same tasks day
to day, encounter the same situations and problems, are re-
quired to use a variety of procedures, and consider their work
to be routine (alpha = .82).
Task interdependence. Task interdependence was measured
using two items developed by Seashore et al. (1983). The
items measure how much cooperation is required by the job
and how often the respondent is required to meet or check
with other people in the organization (alpha = .76).
Sociability. A six-item scale, adapted from Eaves and Eysenck
(1975), measured the extent to which respondents prefer
reading to meeting people, like going out, prefer to have few
but special friends, are quiet around others, like talking to
people, and would be unhappy if they could not see lots of
people most of the time (alpha = .61).
Analysis
The unit of analysis in this research is the individual. Correla-
tions for the measures employed in the study were computed
to assess the relationships among physical barriers, task
characteristics, individual
differences, and the work activity
measures. In order to determine whether the relationships
between barriers and types of interaction were independent
of the task and individual
variables examined, each of the ac-
tivity measures was regressed on the barrier
measures, po-
sition level, task interdependence, uncertainty, and sociability,
forcing firm and job type (step 1) and demographic variables
(step 2) to enter in hierarchical fashion. Task, position, and
sociability variables (step 3) were forced to enter before the
barrier
measures in order to estimate conservatively the sig-
nificance of adding barriers to these models.
RESULTS
The zero-order correlations shown in Table 1 reveal that, in
this data set, task and barrier
measures are statistically inde-
pendent of each other, with the exception of a significant re-
lationship between level and secretary (r = .32).
Multicollinearity
was likely to occur due to high correlations
between door and partition height (r = .82) and door and
number of partitions (r = .41), so the door variable was not
included in the regression analyses reported below in tables
2 and 3.
Allowing firm, job, and demographic characteristics to enter
hierarchically into the regressions before the study variables
provides a conservative estimate of the association between
work activities and task and barrier
measures. The hierarchical
regressions shown in Table 2 reveal that the study variables
account for sizable portions of additional explained variance in
seven categories of self-reported work activity including:
working alone (.42 change in R2), working with others (.09),
attending meetings (.31), interruptions (.1 1), building relation-
ships (.18), on the phone (.32), and taking work breaks (.16).
The barrier
measures make unique contributions in every
case. The models for socializing and personal time did not
achieve significance and so are not included in the table.
It was expected that barriers would be negatively related to
activities involving
interaction.
As Table 3 shows, however,
the pattern
of results in the regression models indicates
that
393/ASQ, September 1987
Table 1
Correlations and Descriptive Statistics for Work Activities and Variables for Tasks,
Individual Differences, and Barriers (N = 99)
Variable Mean SD 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17
1. Work alone .52 .17
2. Work together .13 .09 -.49-
3. Meetings .08 .08 -.58- -.04
4. Interruptions .06 .04 -.35- .01 .05
5. Build
relationships .02 .02 - .34- .07 .14 .03
6. Socialize .03 .03 -.14 -.07 -.01 .18- .12
7. Phone .05 .08 -.51- -.08 .19- .15 .14 -.15
8. Breaks .11 .04 -.01 -.1 1 -.15 -.17- .02 .16 -.10
9. Personal .02 .03 -.13 .01 .06 - .06 - .07 - .02 - .05 -.16
10. Number of
partitions 3.68 .74 -.09 .05 .08 -.1 1 .02 .10 .03 -.03 .13
11. Partition
height 7.09 1.39 -.25- .16- .28- .14- -.07 .09 - .12 .00 .12 .38-
12. Door .44 .36 - .20- .25- .15 .05 .02 .02 - .15 .06 .08 .41- .82-
13. Desk position .87 .33 -.09 .10 -.05 -.27- .12 -.04 -.15 .05 - .09 .05 .00 .03
14. Secretary .05 .22 -.37- - .07 .30- .15 .01 .00 .44- .05 .02 - .21- .09 .10 - .20-
15. Position level 1.34 .48 -.42- .06 .35- .09 .39- .03 .25- .08 -.10 .00 .01 .02 .14 .32-
16. Uncertainty 34.03 6.37 -.33- .22- .31- .00 .12 -.06 .19- -.24- .08 -.06 -.01 .06 -.04 .09 .09
17. Task
interdependence 10.26 2.31 -.37- .17- .23- .22- .10 -.03 .23- -.04 -.06 -.21 -.01 .00 -.01 .10 .00 .41-
18. Sociability 21.97 4.72 - .30- .08 .20- .19 - .03 .09 .20- .12 - .16- .09 .13 .10 -.15 .21- .02 .22- .22-
p .10; Up .05; -p .01; -p i .001 (two-tailed test).
every barrier investigated, except for desk position, was posi-
tively related to time reported to be spent in at least one form
of interaction activity. In this study, individuals in offices with
higher partitions reported more time spent working together
and in meetings than did those with lower partitions. Indi-
viduals in offices with a greater number of partitions or walls
reported spending more time on the telephone. More time
spent on the phone was reported by individuals who had
secretaries to screen their calls than by those who did not.
Table 2
Hierarchical Regressions for Work Activities
Variables Overall
Dependent variable Step added R2 AR2 F P
Work alone 1 Firm, job type .02 .02 0.78 n.s.
2 Demographics .10 .08 1
.27 n.s.
3 Task, sociability .43 .33 5.35 .001
4 Barriers .52 .10 5.62 .001
Work together 1 Firm, job type .10 .10 3.31 .02
2 Demographics .23 .13 3.35 .002
3 Task, sociability .28 .05 2.83 .003
4 Barriers .32 .06 2.44 .005
Meetings 1 Firm, job type .01 .01 0.44 n.s.
2 Demographics .10 .09 1.26 n.s.
3 Task, sociability .31 .21 3.17 .001
4 Barriers .41 .10 3.52 .001
Interruptions 1 Firm, job type .13 .13 4.51 .005
2 Demographics .22 .09 3.17 .003
3 Task, sociability .26 .04 2.55 .006
4 Barriers .33 .07 2.51 .004
Build relationships 1 Firm, job type .06 .06 2.10 n.s.
2 Demographics .15 .09 1.91 .07
3 Task, sociability .29 .14 2.97 .002
4 Barriers .33 .04 2.55 .003
Phone 1 Firm, job type .11 .11 3.98 .01
2 Demographics .13 .02 1.66 n.s.
3 Task, sociability .32 .19 3.31 .001
4 Barriers .45 .13 4.13 .001
Breaks 1 Firm, job type .07 .07 2.28 .08
2 Demographics .11 .04 1.39 n.s.
3 Task, sociability .19 .08 1.62 n.s.
4 Barriers .27 .08 1.91 .03
394/ASQ, September 1987
Barriers and Interaction
Table 3
Regression of Work Activities on Variables for Tasks, Individual Differences, and Barriers (N = 99)*
Dependent Variables
Build
Independent Work Work Inter- relation-
Variables alone together Meetings ruptions ships Phone Breaks
Firm -.028 -.041 .054- -.012 .022-- .029 -.033-
(0.56) (1.20) (1.87) (0.71) (2.65) (1.17) (2.16)
Technician - .057 - .010 .030 .033-- .001 .047-- - .023-
(1.55) (0.41) (1.44) (2.75) (0.16) (2.61) (2.07)
Engineer - .104- .087-- .029 .031- - .005 .004 - .036-
(2.05) (2.52) (1.03) (1.90) (0.68) (0.16) (2.31)
Age .031- - .030- - .018- .000 .006- .015- .004
(1.64) (2.32) (1.75) (0.00) (2.03) (1.64) (0.65)
Sex .026 .011 -.026 -.007 .002 -.007 .000
(0.76) (0.48) (1.34) (0.66) (0.26) (0.41) (0.03)
Education - .002 .006 - .002 .005 .0003 - .007 - .002
(0.14) (0.75) (0.28) (1.25) (0.15) (1.24) (0.59)
Job tenure .0001 - .001 .001 .0005- - .0003- - .0003 .0001
(0.07) (1.38) (1.12) (1.81) (2.43) (0.074) (0.27)
Co m pa ny te n u re .0001 .001 - .0002 - .0001 .0001 - .0002 - .0003
(0.13) (1.33) (0.64) (0.31) (0.84) (0.75) (1.57)
Sociability - .002 - .001 - .0001 .001 .0002 .001 .001
(0.68) (0.26) (0.07) (1.42) (0.35) (0.076) (1.30)
Position level - .142-- .008 .068- .011 .022" .038- .005
(4.47) (0.38) (3.77) (1.10) (4.25) (2.42) (0.47)
Task interdependence -.019m .004 .005 .003 .0003 .005 .003
(2.71) (0.79) (1.11) (1.34) (0.24) (1.43) (1.26)
Uncertainty - .003 .002 .002 - .001 .0001 .002 - .002--
(1.41) (1.42) (1.53) (0.66) (0.18) (1.27) (2.54)
Number of partitions - .028 - .013 .009 - .011 .005 .036" - .004
(1.29) (0.88) (0.74) (1.52) (1.46) (3.38) (0.61)
Partition height - .041w .022- .027-- .008 - .002 - .009 - .013--
(2.67) (2.09) (3.11) (1.64) (0.68) (1.18) (2.78)
Desk .016 .020 .006 - .029- .007 - .003 - .002
(0.35) (0.64) (0.22) (1.98) (0.93) (0.11) (0.17)
Secretary - .138- - 0.14 .046 -.030 -.007 .117-- .005
(1.96) (0.29) (1.14) (1.30) (0.61) (3.39) (0.24)
Constant 1.401 -.046 -.333-- -.021 - .105-- -.267-- .290--
(6.71) (0.32) (2.81) (0.30) (3.05) (2.61) (4.56)
P2 (adjusted) .43 .19 .29 .20 .20 .35 .13
-p < .10; -p < .05; Up < .01; Up < .001.
* Unstandardized regression coefficients are reported. T-values appear in parentheses.
The only barrier
relationship that was found to be negative
was the association between desk position and interruptions.
Individuals who faced away from their office entrance re-
ported fewer interruptions than did those whose desks faced
the entrance.
Due to the likelihood of multicollinearity, separate regressions
substituting door for partitions were run. The door variable
contributed significantly to two models-working together
and building relationships. In both cases the association was
positive. The result for partitions was not significant in the
comparison model of time reported to be spent building rela-
tionships, which indicates that the door variable contributes
independently to the variance in this activity measure. The
significant result for working together, however, shifts be-
tween partition height and door.
The relationships found between position-level and work-ac-
tivity measures were in the directions reported
in previous
395/ASQ, September 1987
studies, but position level was not associated with all types of
activity examined in the study. Position level was positively
associated with time reported in meetings, building relation-
ships, and on the phone; negatively associated with working
alone; and not associated at all with working together or in-
terruptions.
Task interdependence and uncertainty were not significantly
related to time reported in any type of interaction examined.
In fact, task interdependence and uncertainty were conspic-
uous by their absence from the regression models of all types
of self-reported interaction activity. However, significant neg-
ative relationships were found between task interdependence
and time reported working alone and between uncertainty
and time reported taking breaks. These measures were also
positively related to each other, replicating Brass (1985).
The results indicate some interesting job-related differences
in time reported spent in some types of interaction activity.
Scientists reported spending significantly less time on inter-
ruptions than did engineers or technicians, while engineers
reported more time working together and technicians re-
ported more time on the phone. These results provide an ex-
tension to Sundstrom, Herbert, and Brown's (1982) finding
that professional-technical employees are particularly
sensi-
tive to intrusions.
DISCUSSION
Individuals and organizations facing decisions about physical
space undoubtedly make assumptions about how physical
space influences behavior. For instance, the assumption that
a lack of physical barriers increases interaction among office
workers can be found in much of the recent literature on
physical space (e.g., Oldham and Brass, 1979; Becker, 1981;
Sutton and Rafaeli, 1987). The research reported here indi-
cates that such assumptions may be erroneous.
This study provides evidence that interactions among profes-
sional-technical workers in research and development firms
may be greater for workers who are given enclosed work
spaces than for those lacking physical barriers. What is more,
the research presented indicates that different barrier mea-
sures are associated with different interaction measures.
However, caution must be exercised in generalizing from
these findings, since causality cannot be inferred on the basis
of cross-sectional data. The reported study is also limited in
scope. The relationships between barriers, task character-
istics, and interaction activities were only examined for three
jobs, all of which involved industrial research activity. Controls
for these jobs and the two firms sampled were included, but
these variables need far more research attention than was
possible in this initial, exploratory examination.
Enclosure by partitions (or walls) and a door was found to be
positively associated with the amount of time individuals re-
ported working with others. Partition height was also posi-
tively related to reported amounts of meeting time. These
results are consistent with the view that enclosure supports
interpersonal and group interaction. Similar observations have
been reported by Richards and Dobyns (1957) and Oldham
and Brass (1979). These previous studies found that reduc-
396/ASQ, September 1987
Barriers and Interaction
tions in physical enclosure decreased the amount of interac-
tion among clerical coworkers and newspaper employees'
perceptions of interaction opportunities, respectively. This
study extends these findings to research and development
firms and suggests that the particular
categories of interaction
positively associated with enclosure are working together and
meetings.
The basis for the relationships between interaction activity
and enclosure by partitions and doors requires further study.
These relationships could reflect any of a number of mecha-
nisms. For instance, physical enclosure may help a group to
define its boundary and to enhance its sense of identity, se-
curity, or cohesiveness. Similarly, the act of closing a door
may be a symbol of intimacy that stimulates interaction via
social norms that apply to such occasions. Any or all of these
factors could contribute to a tendency for employees to
spend more time together when they occupy enclosed office
spaces.
Time reported to be spent building relationships was posi-
tively correlated with the use of a door. One possible expla-
nation for this association is that relationship building is not
viewed as directly related to work, and individuals are un-
comfortable engaging in such activities in the presence of
others whom they fear might negatively evaluate their be-
havior. Another possibility is that this category contains re-
ports of confidential conversations (e.g., pertaining to coalition
formation, negotiation, or secretive action) that may require a
closed door if they are to take place at the work site at all
(BOSTI, 1981; Sundstrom, Herbert, and Brown, 1982). These
and other possibilities might be investigated in future studies.
The door-measure results have an interesting implication for
understanding open-door policies common in many American
business organizations. For most people, such policies repre-
sent the organization's desire for open communication and
strong interpersonal relationships. The positive association
between a closed door and time reported to be spent building
relationships may indicate that such policies are either coun-
terproductive or that they operate on the basis of their sym-
bolic expression of openness rather than their presumed
effects on interaction activity. That is, while an open door may
communicate a desire to build relationships, closing the door
while engaging in interaction may be necessary if relation-
ship-building activity is to occur. Future studies might usefully
investigate the symbolic versus the behavioral implications of
this aspect of physical structures.
A common organizational belief is that busy managers require
secretaries. According to the results of this study, secretarial
screening is positively associated with phone calls and nega-
tively associated with time reported working alone. This may
simply confirm the assumption that secretaries are assigned
according to need-busier people get secretaries. However,
the expectation that secretarial screening will be associated
with reduced interaction demands is not supported by the re-
sults. The relationship between a secretary and interaction
activities was independent of position level, job type, and task
characteristics.
That
is, among individuals
doing similar
work
at similar position levels, those without secretaries reported
397/ASQ, September 1987
spending more time working alone and less time on the
phone than did those with secretaries.
A desk positioned away from the office entrance was the only
physical barrier found to be negatively associated with re-
ported interruptions-the relationship most often assumed to
exist between physical barriers and behavior in offices. Since
desk position is more often under the control of the office
occupant than are partition height, a door, or the assignment
of a secretary, perhaps overgeneralization of this relationship
has occurred, resulting in the misperception that all physical
barriers reduce interaction.
CONCLUSION
Several types of interaction were found to be associated with
physical barriers, independent of their association with task
characteristics and individual differences. Bearing in mind the
limitations of this study, these results argue for accepting
Barker's (1968) view that physical settings support some
forms of activity and constrain others. Additional research is
needed to determine whether these relationships generalize
to other jobs and other firms. Research is also needed to as-
sess the causal relationships between physical barriers, inter-
action activities, and important outcomes such as
performance.
REFERENCES
Allen, Thomas J., and Alan R.
Fusfeld
1974 "Research laboratory architec-
ture and the structuring of
communications." Working
Paper 692-74, M.I.T. Sloan
School of Management.
Allen, Thomas J., and Peter G.
Gerstberger
1973 "A field experiment to improve
communications in a product
engineering department: The
nonterritorial
office." Human
Factors, 15: 487-498.
Archea, John
1977 "The place of architectural
factors in behavioral theories
of privacy." Journal of Social
Issues, 33: 116-137.
Barker, Roger G.
1968 Ecological Psychology: Con-
cepts and Methods for
Studying the Environment of
Human Behavior. Stanford,
CA: Stanford University Press.
Becker, Franklin D.
1981 Workspace: Creating Environ-
ments in Organizations. New
York: Praeger.
Buffalo Organization for Social and
Technological Innovation (BOSTI)
1981 The Impact of Office Environ-
ment on Productivity and
Quality of Working Life: Com-
prehensive Findings. Buffalo,
NY: Buffalo Organization for
Social and Technological Inno-
vation.
Brass, Daniel J.
1985 "Technology and the struc-
turing of jobs: Employee satis-
faction, performance, and
influence." Organizational Be-
havior and Human Decision
Processes, 35: 216-240.
Brookes, Malcolm J., and Archie
Kaplan
1972 "The office environment:
Space planning and affective
behavior." Human Factors, 14:
373-391.
Carroll, Stephen J., and Dennis J.
Gillen
1987 "Are the classical management
functions useful in describing
managerial work?" Academy
of Management Review, 12:
38-51.
Conrath, C. W.
1973 "Communication patterns, or-
ganizational structure, and
man: Some relationships."
Human Factors, 15: 459-470.
Eaves, Lindon, and Hans Eysenck
1975 "The nature of extraversion: A
genetical analysis." Journal of
Personality and Social Psy-
chology, 32: 102-112.
Festinger, Leon
1954 "A theory of social comparison
processes." Human Relations,
7: 117-140.
Galitz, W. D.
1980 Human Factors in Office Auto-
mation. Atlanta, GA: Life Of-
fice Management Association.
Gerstberger, Peter G., and Thomas
J. Allen
1968 "Criteria used by research and
development engineers in the
selection of an information
source." Journal of Applied
Psychology, 52: 272-279.
Goddard, J. G.
1973 Office Linkages and Location:
A Study of Communications
and Spatial Patterns in Central
London. Oxford: Pergamon.
Gullahorn, J. T.
1952 "Distance and friendship as
factors in the gross interaction
matrix." Sociometry, 15:
123-134.
Hage, J., and M. Aiken
1969 "Routine technology, social
structure and organizational
goals." Administrative Science
Quarterly, 14: 366-376.
Joiner, D.
1976 "Social ritual and architectural
space." In H. Proshansky, W.
Ittelson and L. RivIin
(eds.),
Environmental Psychology, 2d
ed.: 224-241. New York: Holt,
Rinehart & Winston.
Katz, Daniel, and Robert L. Kahn
1978 Social Psychology of Organiza-
tions, 2d ed. New York: Wiley.
Kiggundu, Moses N.
1981 "Task interdependence and
the theory of job design."
Academy of Management Re-
view, 6: 499-508.
398/ASQ, September 1987
Barriers and Interaction
Kurke, Lance B., and Howard E.
Aldrich
1983 "Mintzberg was right! A repli-
cation and extension of the
nature of managerial work."
Management Science, 29:
975-984.
Mahoney, T. A., T. H. Jerdee, and
S. J. Carroll, Jr.
1963 Development of Managerial
Performance: A Research Ap-
proach. Cincinnati: South-
Western.
1965 "The jobs of management."
Industrial Relations, 4:
97-110.
Mehrabian, Albert
1978 Public Places and Private
Spaces: The Psychology of
Work, Play and Living Environ-
ments. New York: Basic
Books.
Mintzberg, Henry
1973 The Nature of Managerial
Work. New York: Harper &
Row.
Oldham, Greg R., and Daniel J.
Brass
1979 "Employee reactions to an
open-plan office: A naturally
occurring quasi-experiment."
Administrative Science Quar-
terly, 24: 267-284.
Oldham, Greg R., and Nancy L.
Rotchford
1983 "Relationships between office
characteristics and employee
reactions: A study of the
physical environment." Ad-
ministrative Science Quarterly,
28: 542-556.
Pfeffer, Jeffrey
1981 Power in Organizations.
Marshfield, MA: Pitman.
1982 Organizations and Organization
Theory. Boston: Pitman.
Richards, C. B., and H. F. Dobyns
1957 "Topography and culture: The
case of the changing cage."
Human Organization, 16:
16-20.
Rousseau, Denise M.
1978 "Characteristics of depart-
ments, positions, and indi-
viduals: Contexts for attitudes
and behavior." Administrative
Science Quarterly, 23:
521 -540.
Scott, W. Richard
1981 Organizations: Rational, Nat-
ural, and Open Systems. En-
glewood Cliffs, NJ:
Prentice-Hall.
Seashore, Stanley E., Edward E.
Lawler, Philip H. Mirvis, and
Cortlandt Cammann
1983 Assessing Organizational
Change: A Guide to Methods,
Measures, and Practices. New
York: Wiley.
Simon, Herbert A.
1957 Administrative Behavior: A
Study of Decision-Making Pro-
cesses in Administrative Orga-
nization. New York: Macmillan.
Steele, Fritz
1973 Physical Settings and OrgaRiz
zation Development. Reading,
MA: Addison-Wesley.
1986 Making and Managing High-
Quality Workplaces: An Orga-
nizational Ecology. New York:
Teachers College Press.
Sundstrom, Eric
1986 Work Places. Cambridge:
Cambridge University Press.
Sundstrom, Eric, R. Kring Herbert,
and David Brown
1982 "Privacy and communication
in an open-plan office: A case
study." Environment and Be-
havior, 14: 379-392.
Sutton, Robert I., and Anat Rafaeli
1987 "Characteristics of work sta-
tions as potential occupational
stressors." Academy of Man-
agement Journal, 30:
260-276.
Szilagyi, Andrew D., and Winford
E. Holland
1980 "Changes in social density:
Relationships with functional
interaction and perceptions of
job characteristics, role stress,
and work satisfaction." Journal
of Applied Psychology, 65:
28-33.
Thompson, James D.
1967 Organizations in Action. New
York: McGraw-Hill.
Weber, Max
1947 The Theory of Social and Eco-
nomic Organization. New York:
Free Press.
Wells, B. W. P.
1965 "The psycho-social influence
of building environments: So-
ciometric findings in large and
small office spaces." Building
Science, 1: 153-165.
Wicker, Allan W.
1979 "Ecological psychology: Some
recent and prospective devel-
opments." American Psychol-
ogist, 34: 755-765.
Williamson, Oliver E.
1975 Markets and Hierarchies:
Analysis and Antitrust Implica-
tions. New York: Free Press.
Wofford, Jerry C., Edwin A.
Gerloff, and Robert C. Cummins
1977 Organizational Communication:
The Keystone to Managerial
Effectiveness. New York:
McGraw-Hill.
399/ASQ, September 1987