Behavior Research Methods, Instruments, & Computers
2000, 32 (3), 432-435
PROXIMITY: An update and expansion
ofthe PROXTIME computer program
University ofNorth Carolina, Charlotte,North Carolina
BENJAMIN B. CHASE
Concord, North Carolina
Anewprogram, calledPROXIMITY, was createdas atoolfor proximity calculation and to update and
expand upon Kirste and Monge's(1983)PROXTIME program. The purpose of PROXIMITY is to calcu-
late the fluctuating proximity between individuals within organizations. PROXIMITY provides output
forthree types of relationships: (l) an overall organizationalproximity, (2)pairwise proximity between
individuals, and (3) individual proximity to multiple others. PROXIMITY also updates some of PROX-
TIME'sfeatures such as the computerplatform, the type ofdatathe program can handle, and the form
of output available.
Interpersonal proximity in organizations is an impor-
tant factor in understanding a variety of outcomes in
fields such as environmental planning, communication,
social psychology,and organizational behavior.In environ-
mental planning, researchers are interested in how prox-
imity affects crowding (e.g., Baum & Paulus, 1987; Grif-
fitt & Veitch, 1971;Moos, 1976),privacy(e.g., Sundstrom,
1987), and stress (e.g., Evans & Cohen, 1987), as well as
human comfort, territoriality, and spatial planning (e.g.,
Altman, 1975; Sundstrom, 1986). Communication re-
searchers are interested in how proximity affects general
communication patterns (e.g., Smith & Kearny, 1994),
preferences for amount and recipient of social interac-
tions (e.g., Zahn, 1991),and type ofcommunication (work-
related versus casual/informal) (e.g., Becker & Steele,
1995; Steele, 1986). Researchers in social psychology
are interested in how proximity affects group idea gen-
eration (e.g., Valacich, George, Nunamaker, & Vogel,
1994), and interpersonal attraction (e.g., Arkin & Burger,
1980). Organizational behavior researchers are interested
in how proximity affects attitudes about co-workers (e.g.,
Moos, 1976),organizational andjob satisfaction (e.g., Old-
ham & Fried, 1987), and task performance (e.g., Becker
& Steele, 1995; Smith & Kearny, 1994).
Thoughproximity can be operationalizedin many ways
(e.g., linear distance, functional distance, and psycho-
logical distance), one ofthe most useful operationalizations
This research was partially funded by a grant to the first author from
the University of North Carolina at Charlotte Junior Faculty Summer
Fellowship Program. We thank Matt Kurbat, Sarah Breedin, and Paula
Goolkasian for their assistance in completing this publication. A copy
of the program and additional user-oriented documentation can be ob-
tained at no cost, except for postage and inclusion of a disk, from B.N.
Pollack, 12 15thAvenue,San Francisco, CA 94118 (e-mail: bonnienp@
ofinterpersonal proximity in organizations was developed
by Monge and Kirste (1980), which they described as
"fluctuating distance." This operationalization of prox-
imity improves upon earlier measures because it captures
the dynamic nature ofpeople's movement over time, and
recognizes that individuals are concurrently proximate to
many others in an organizational setting. In 1983, Kirste
and Monge introduceda computer program, called PROX-
TIME, to calculate fluctuating proximity between indi-
viduals within organizations. Fluctuating proximity is
"the probability ofpeople being in the same 'communi-
cation location' during the same interval oftime" (Monge
& Kirste, 1980, p. 112).
Kirste and Monge's (1983) PROXTIME program was a
good initial step toward calculating fluctuating proxim-
ity,and some research has made use ofit (e.g., Monge &
Kirste, 1980, using an earlier version called PROXVAL;
Monge, Rothman, Eisenberg, Miller, & Kirste, 1985).
Wewanted to create a proximity calculation program for
the current computing environment that also expanded
upon PROXTIME's capabilities. Thus, we created a new
program, called PROXIMITY. Below, we discuss some
ofthe features ofPROXTIME. After each feature, we ex-
plain the modifications that we have made and incorpo-
rated into our new PROXIMITY program. They include
(1) the computer platform used, (2) the unit of time for
the output, (3) the addition ofa third proximity calculation
to the original two, (4) the form ofthe input data, (5) the
way in which the work week is defined, and (6) the choice
ofcalculations for the three different types ofproximity.
The first thing that we modified in creating this new
program was the platform (and thus the computer-user
interface as well). Probably the greatest limitation for
current computer users is that PROXTIME was written in
the FORTRAN computer language and must be used with
a mainframe computing system that supports a FOR-
Copyright 2000 Psychonomic Society, Inc. 432
Sections of an Example Spreadsheet for
PROXIMITY Input Data
proximity to multiple others values, the user can allow
the length ofthe work weekto be determined by start and
end days and times, or bythe length oftime when at least
one person was present at the site. For the pairwise prox-
imityvalues, theusercanallowthe lengthofthe workweek
to be determined by start and end days and times, or by
the amount ofoverlapping hours between the twopeople
in the dyad (i.e., the hours in which they were both pre-
sent at the site). These choices allow the researcher to
tailor the proximity measures to the research question
Generaldescription. PROXIMITY calculates (1)orga-
nizational proximity, which is the degree to which peo-
ple within an organization share the same physicalloca-
tions during the same periods oftime for the entire work
week; (2) pairwise proximity, which is an individual's
proximity to each other individual in the organization
(dyad-level information) for the entire work week; and
(3) proximity to multiple others, which is an individual's
proximity to all other individuals in the organization for
the entire work week. PROXIMITY can handle an unlim-
ited number ofindividuals at 2-min or greater time inter-
vals over a 7-day, and up to a 24-h/day work week.
Structure of input data for PROXIMITY runs.
PROXIMITY takes a spreadsheet or tab-delimited lines
as input. An example spreadsheet is given in Table 1.
This spreadsheet needs to be five contiguous columns
wide by some number of contiguous rows long. No head-
ing should be included. The columns should contain the
data in the following order: person's name, day, room
name, start time, and end time. Individuals and room lo-
cations can be written as names or numbers. The day
field is an integer in the range 0 to 6 (i.e., Sunday to Sat-
urday). The start and end time is in an HHMM 24-h/mil-
itary format (i.e., HH varies from 0 to 23 and MM varies
from 00 to 59). Start and end times cannot overlap; in
other words, a person cannot be recorded as occupying
two different rooms at the same time. So room occupa-
TRAN IV compiler. PROXIMITY is a user-friendly
desktop program (i.e., no understanding ofcomputer lan-
guages is needed for one to use it) that operates on Mac-
intosh computers. The program has also been ported to
a UNIX system (Solaris 2.5.1) and can be used on this
platform as well.
The second modification we made was to change the
program's output from dozens of individual results in
time units of 15 min each to single summary results av-
eraged together from part ofa day (minutes or hours) to
a full work week. Although a time unit as fine grained as
15min can be interesting in some types ofresearch (e.g.,
Monge et aI., 1985), many organizational researchers are
interested in results across much larger units of time,
such as full days or entire work weeks (e.g., Kirste &
Monge, 1983;Monge & Kirste, 1980;Zahn, 1991).PROX-
1MITy allows the user to determine what block oftime
is most useful and provides output for various proximity
measures, ranging from those as fined grained as 1 min
to a summary for the entire work week.
The third change that we made was to add the compu-
tation ofa third type ofproximity.PROXTIME calculates
pairwise or dyad-level proximity (an individual's prox-
imity with each other person in the organization) and
proximity to multiple others (an individual's proximity
to all others in the organization). Another useful type of
proximity is what Kirste and Monge (1983, p. 89) and
Monge et al. (1985, p. 1129)described as "organizational
proximity"(thedegree towhichpeoplewithin an organiza-
tion share the same physical locations during the same pe-
riods oftime). PROXTIME does not compute its value, so
in addition to pairwise proximity and proximity to mul-
tiple others, we programmed PROXIMITY to calculate
this overall organizational proximity. Mathematical defi-
The fourth modification concerns the format ofthe in-
put data. In PROXTIME, specifications for how the data
have to be entered are fairly rigid (see Kirste & Monge,
1983, for details). PROXIMITY has the following new
features not present in PROXTIME: Data are now taken
from aspreadsheet; individuals and locations can be given
full names; the length of time intervals can be deter-
mined by the user; all 7 days ofthe week are acceptable
for input; start and end times for the work days can be de-
termined by the user; the length ofthe work week is au-
tomatically determined by the given start and end days
Fifth, in the new program we give the user more con-
trol over how the work week is defined. In PROXTIME,
the work week could only be 50 h, only Monday through
Friday, only 8 a.m. to 6 p.m. PROXIMITY, on the other
hand, allows the user to pick any choice ofstart and end
daysandtimes.This enablestheresearcher to selecta work
week that is meaningful in the situation under study.
Sixth,inthe newprogram wegivethe user more choices
for calculating each ofthe three different types ofprox-
imity. For the overall organizational proximity and the
434POLLACK AND CHASE
Group Proximity is 0.0337.
Valueused for group proximity denominator was 132.
The value printed below, next to each person, is that person's proxim-
ity to all others. This is also called "proximity to multiple others." The
denominator used for these is 132.
Sections of Example PROXIMITY Output
for the Three Types of Proximity
The dyads shown below are for certain pairs of people PI and P2. Be-
cause the value should be the same for the pair PI ,P2 and the pair
P2,PI, only one ofthese two pairs is shown.
Dyad Proximity Value
Proximity to Multiple Others
imum of 2 min). For each sample interval, the program
reads the data once to see which room a particular person
occupies. These are called samples, which are then used
in the proximity calculations.
The first kind of proximity value is the proximity of
the entire group ofN workers (i.e., the whole organiza-
tion). At a given moment in time, the locations ofall the
workers are noted. For every location where there is
more than one person, the number ofpersons is squared
and then divided by the square ofN, and all of these val-
ues are summed. The work week average is obtained by
summing its value over all the sample times and then di-
viding by the number ofsamples. In Table 2, the organi-
zational proximity, labeled "Group Proximity," equals
The second kind ofproximity value is the pairwise or
dyad-level proximity. At a given moment in time, for a
particular pair of workers, if they are in the same loca-
tion, their pairwise proximity is 1;otherwise, it is O. This
value is calculated for all pairs. The work week average
is obtained by summing the values for a particular pair
for some number of samples and then dividing by the
number ofsamples. The user can choose the number of
samples to be either for the entire work week, or just
those samples when both workers were somewhere at
work. In Table 2, the dyadic proximity values are found
in the column labeled "Dyad Proximity Value."
The third kind ofproximity value is the proximity to
multiple others. At a given moment in time, for a partic-
ular worker, his/her proximity to multiple others is the
square of the number of people in his/her location, di-
vided by the square of N, unless he/she is either not at
work or working alone, in which case his/her proximity
to multiple others at that time is O. The values for a par-
ticular person are averaged over all the samples in the
work week to obtain an average proximity to multiple
tion data should be recorded, for example, as a person
being in Room I from 9:00 to 9:29 a.m., and then Room 2
from 9:30 to 9:59 a.m., instead ofRoom 1 from 9:00 to
9:30 a.m. and Room 2 from 9:30 to 10:00 a.m. Each row
contains information for, within a particular day, what
room a person occupies between which times.
Error checking. Extensive error checking is auto-
matically conducted throughout the program, with the
user having the option ofviewing it. At the beginning of
the program run, the user has the option of specifying
choices of "normal output," "verbose output," or "very
verbose output." Each choice respectively gives the user
more information. For all three output choices, the pro-
gram provides basic information about the data so that the
user can examine it and verify its accuracy.When PROX-
IMITY detects an error, it will print out an appropriate
error message to guide the user. Ifthe user enters an in-
valid choice in response to a question, the program will
prompt the user to try again.
Limitations. Unlike PROXTIME, PROXIMITY has
no limitations on the number of individual identifiers
(i.e., number of people) or location/time combinations.
It does have some other limitations, though. Like PROX-
TIME, it can only process data for a maximum of1work
week (7 days), and the time must be specified in military
time. The unit of time is minutes, and the smallest time
intervals that the program will accept are 2 min long.
People and location names are limited in length to 100
characters. Unlike PROXTIME, PROXIMITY will allow
data for 24-h days and workshifts that go around the
clock. In thiscase, the end time must bespecifiedas 23:59
ofone day, and the next start time must be specified as
00:00 of the next day, with data from each of the days
placed on separate lines on the Excel spreadsheet.
Output. The program's output is captured in the user's
input file (i.e., all input and output information is con-
tained in the same scrolling text window), so the user has
complete information about a particular run in a single
file. First, the program displays some descriptive statis-
tics about the input data, so that the user may verify its
accuracy. Then the three proximity values are calculated
and displayed. At the end ofthe run, the user may print
the entire file, save it as a text file, and/or cut and paste
it into a wordprocessing program. As asaved file, the user
can access it within a word processing program.
Calculations. The program calculates three different
kinds ofproximity values. Each ofthese kinds is calcu-
lated at specific times during the work week and then av-
eraged over these times. The second and third type de-
scribed beloware the same as Kirste and Monge's (1983),
except that the results are for the entire work week. The
first type described below is our added proximity, for the
entire organization. An example ofsample output is given
in Table 2.
Before any proximity calculations are made, the pro-
gram reads the data in samples. The user specifies the
lengthofthesampleintervals, whichcanbeasfine grained
as the intervals atwhich the data were collected (to a min-
others. InTable2, the proximity to multiple others values
are found in the column labeled "Proximity to Multiple
Hardware and transportability. The program was
compiled on a 68040 Macintosh and will operate on
newer models of Macintosh. It was tested on a Macin-
tosh Performa 636 running Operating System 7.5.3 and
on a Power Macintosh 6100 running Operating Sys-
tem 7.5.1.The source code is written in ANSI C. The pro-
gram was developed under CodeWarrier, and for the pro-
gram to be modified, CodeWarrier is needed. To simply
run the program, the user does not need the authoring
program; the program is a self-contained, stand-alone ap-
plication. The program can be used on UNIX systems.
Program availability. A copy ofthe program and ad-
ditional user-oriented documentation can be obtained at
no cost, except for that ofpostage and the inclusion ofa
disk, from the first author at 12 5th Avenue, San Fran-
cisco, CA 94118.
ALTMAN, I. (1975). The environment and social behavior. Monterey,
ARKIN, R.M., & BURGER, J. M. (1980). Effectsof unit relation tendencies
on interpersonal attraction. Social Psychology Quarterly, 43, 380-391.
BAUM, A., & PAULUS, P. B. (1987). Crowding. In D. Stokols & I. Alt-
man (Eds.), Handbook ofenvironmental psychology (pp. 533-570).
BECKER, E, & STEELE, E (1995). Workplaceby design: Mapping the
high-performance workscape. San Francisco: Jossey-Bass.
EVANS, G. W.,& COHEN, S. (1987). Environmental stress. In D.Stokols
& I. Altman (Eds.), Handbook ofenvironmentalpsychology(pp. 571-
610). New York: Wiley.
GRIFFlTT, W., & VEITCH, R.(1971). Hot and crowded: Influences of pop-
ulation density and temperature on interpersonal affective behavior.
Journal ofPersonality & Social Psychology, 17,92-98.
KIRSTE, K. K., & MONGE, P R. (1983). Computing dynamic organiza-
tional proximity: The PROXTIME computer program. Behavior Re-
search Methods & Instrumentation, 15, 89-90.
MONGE, P.R., & KIRSTE, K. K. (1980). Measuring proximity in human
organization. Social Psychology Quarterly, 43, 110-115.
MONGE, P. R., ROTHMAN, L. W., EISENBERG, E. M., MILLER, K. I., &
KIRSTE, K. K. (1985). The dynamics of organizational proximity.
Management Science, 31,1129-1141.
Moos, R. H. (1976). The human context: Environmental determinants
ofbehavior. New York:Wiley.
OLDHAM, G. R., & FRIED, Y.(1987). Employee reactions to workspace
characteristics. Journal ofApplied Psychology, 72, 75-80.
SMITH, P, & KEARNY, L. (1994). Creating workplaces wherepeople can
think. San Francisco: Jossey-Bass,
STEELE, E I. (1986). Making and managing high-quality workplaces:
An organizational ecology. New York: Columbia University, Teach-
ers College Press.
SUNDSTROM, E. (1986). Workplaces: The psychology ofthe physical
environment in offices and factories. Cambridge: Cambridge Uni-
SUNDSTROM, E. (1987). Work environments: Offices and factories. In
D. Stokols & I. Altman (Eds.), Handbook ofenvironmental psychol-
ogy (pp. 733-782). New York:Wiley.
VALACICH, J. S., GEORGE, J. E, NUNAMAKER, J. E, & VOGEL, D. R.
(1994). Physical proximity effects on computer-mediated group idea
generation. Small Group Research, 25, 83-104.
ZAHN, G. L. (1991). Face-to-face communication in an office setting:
The effects ofposition, proximity, and exposure. Communication Re-
search, 18, 737-754.
(Manuscript received November IS, 1999;
revision accepted for publication March 3, 2000.)