Professional Documents
Culture Documents
2, 1972
INTRODUCTION
The concept of privacy, although of theoretical importance to social
psychology and practical importance to environmental design, has received little
empirical investigation. The social psychologist is primarily interested in privacy
as it relates to the processes of withdrawal from interaction and to the exercise
of control over involvement with others. The development of a concept of
privacy would help the environmental designer, on the other hand, to design
physical spaces that could be more easily used to provide privacy.
Although a variety of situations and meanings is covered by the single term
privacy, the central theme seems to be the ability to control the degree to which
people and institutions impinge upon one's life (Westin, 1967; Proshansky et al.,
1970) and the ability to adjust the level of privacy to changing needs. Choice
seems to be essential to the attainment of privacy: it is not enough simply to be
alone, for example; one must be alone when one chooses to be.
94
Marshall
95
others have to oneself (Marshall, 1970; Westin, 1967). In the context of artificial
phenomena, the form this behavior takes, whether oriented away from others or
presenting barriers to the behaviors of others oriented toward oneself, would be
a function of the environmental setting as well as of an internal disposition to
prefer one means of control over another, or to prefer a particular level of
privacy. Thus, one might expect the choice of means of achieving privacy to be
largely a function of the opportunities and constraints of the setting.
Past Experience
Privacy orientations were related to past experiences of density and
crowding as variables potentially important in the acquisition of privacy
preferences. Discussions of privacy in the popular literature generally assume a
drive reduction theory of motivation and predict that deprivation of opportunities to obtain privacy because of crowding will be reflected in increased
preference for privacy. In contrast, it was predicted from adaptation level theory
(Helson, 1964) that preferences represent an adaptation toward present and past
levels of input. Thus, a decrease in available privacy should lead to a
corresponding decrease in preferred levels of privacy. Specifically, it was
hypothesized that preferences for privacy-oriented behavior would reflect
adaptation to the levels of density and crowding within the home, within the
neighborhood, and within the surrounding metropolitan area.
Environmental Fit
Subjects
A sample of second-semester sophomore students (n=149, 73 male and 76
female) from a commuter junior college in the outer suburbs of San
Francisco-Oakland and a sample of their parents (n=101,44 male and 57 female)
were selected. The junior college chosen draws students from upper middle class
suburbs and from lower middle class suburbs closely connected with local
industries. There were two major criteria used to select these samples, namely,
that the subjects be more representative of the general population than a
96
Marshall
university sample and that age groups other than college students be included.
The environmental variance was reduced by limiting the sample to persons living
in a suburban family setting.
The socioeconomic status (SES) of the samples was computed using
Hollingshead's (1965) Two-Factor Index. Scale scores on occupation and
education were summed to obtain a total score, in turn divided into five social
classes. The Index distinguishes seven levels of education: professional, 4-year
college graduate, 1-3 years college, high school graduate, 10-11,7-9, and under 7
years of schooling. Holiingshead's seven-level occupational status scale differentiates among kinds of professionals and the size and economic strength of
businesses. To be classified as the highest SES, a subject would have to be in the
highest occupational level (higher executives of large concerns, proprietors,
major professionals) and be at least a college graduate; to be classified as class 5,
a subject would have to be in the lowest occupational level (unskilled
employees) and not have completed high school. The distributions among SES
were approximately the same for the adult and student samples despite the 40%
return rate for the former: 15% were in class 1 (highest), 31% in class 2, 35% in
class 3, 17% in class 4, and 2% in class 5.
Materials
97
The principal components analysis of the PPS produced six major factors
appearing for both the adult and student samples. In order to obtain a
conservative estimate of real differences between adult and student factor
patterns, the loadings obtained from the original Varimax rotations were rotated
using a procedure developed by Bloxom (1968) to find maximum similarity
between factor pattern matrices for subpopulations. Six subscales were
constructed based on items that had loadings above 0.40 on each factor;
contents are summarized below (for a complete listing of the items with their
factor loadings, see Marshall, 1970).
Intimacy
The items loading on the Intimacy factor were largely concerned with
privacy for units larger than one, e.g., with intimate friends or family. The
differences in the factor analyses of the adult and student samples suggested that
they differed in the nature of the desired intimate relationship; while the student
would tend to disclose a great deal about himself in order to gain intimacy, the
adult was concerned with retaining a degree of privacy even within the intimate
relationship. Sample items were
People should respect another's right to be individual and different.
(Agree)
It is important to be able to confide in someone and know that your
confidence will be kept secret. (Agree)
I occasionally enjoy getting away from the rest of the world with an
intimate friend. (Agree)
Not Neighboring
The Not Neighboring items dealt with disliking the tendency of friends or
neighbors to drop in without warning and with a preference for noninvolvement
with neighbors. The factor contained two not necessarily related means of
gaining privacy: through expressed or implied norms about when others might
visit and through choice of persons other than neighbors as close friends,
98
Marshall
Item
Environmental privacy composites
Distance neighbors
Visual privacy
Noise/odor
Perceived crowding
Perceived privacy
Privacy/neighbors
Job density
Town size
Mobility
Open space
Own room
No. siblings
Number of siblings
Sibling intimacy
Sibling disagreement
Past perceived
crowding
99
distance presumably allowing more control over the relationship. Sample items
were
I want my friends to feel that they can drop in at my house any time they
like. (Disagree)
I would prefer a neighborhood where neighbors had a tendency to drop in
all the time to one in which it was difficult to get to know them. (Disagree)
I really enjoy being able to loan things to friends. (Disagree)
Seclusion
Items loading on the Seclusion factor dealt with having a home secluded
from the sight and sound of neighbors and traffic. It included items indicating a
tolerance for being alone and a willingness or preference to be unacquainted
with neighbors. Sample items were
I would like to live in a secluded house out of sight of any other houses.
(Agree)
A house should be so far away from a neighbor that only by yelling at the
top of one's lungs can one be heard. (Agree)
Solitude
The Solitude factor contained items reflecting a desire to be alone at
times, without differentiating between being alone but with others nearby (as in
one's room) and being far from others. The possibility of being alone
"mentally," with others present but not intruding on one's thoughts, was
included. Sample items were
I sometimes want to get away from everyone for a while, even my close
friends. (Agree)
It is important to me to be able to be alone when I want to be. (Agree)
Anonymity
The central theme of the A n o n y m i t y factor was the anonymity of urban
living. Items dealt with being able to a t t a i n privacy in a large city because
"everyone wouldn't know everything about y o u " as opposed to the interest in
and involvement with others in a small town. The student factor, unlike the
adult, also reflected a liking for the noise and traffic of the city; noise did not
seem to be a drawback to city life for students. Sample items were
I would like to live in a large city because neighbors and acquaintances
there would probably be less concerned about my private life. (Agree)
I would not like to live in a small town because there is too much gossip
about your private life. (Agree)
100
Marshall
Reserve
Items loading on the Reserve factor dealt with a preference not to disclose
much about oneself to others, particularly if the others were not known very
well. Several types of items involving self-disclosure were written, covering the
preferred depth of disclosure (how much one was willing to disclose to even the
closest friends) as well as the breadth of disclosure (how much one preferred to
disclose to friends, acquaintances, and strangers). Many of the items in the latter
category, particularly those indicating a preference to be "open and honest with
others" or to have few really intimate friends, did not load on this factor. The
difference in emphasis of the adult and student factors was similar to that on the
Intimacy factor. The adults emphasized not revealing much about themselves to
"friends" until they had known them a long time, while students emphasized
not revealing themselves to "persons I don't know very well." Sample items were
I dislike talking about personal matters to a friend in a crowded place
where other people can overhear us. (Agree)
I usually don't tell people I don't know very well personal things about
myself. (Agree)
I don't like to talk about personal things with friends until I have known
them a long time. (Agree)
Intercorrelations of the six subscales showed two major groupings (Table
II). The first centered on' strong relationships among Anonymity and Not
Neighboring, predictable from the item overlap (two items appeared on both
subscales), and Reserve. Logically; this grouping contains the orientations
involving reserve as a means of attaining privacy, if one interprets controlling
involvement with neighbors (Not Neighboring and Anonymity) as a means of
controlling self-disclosure. The second grouping consisted of a strong relationship between the Solitude subscale and both Seclusion and Intimacy. This
subgrouping centered on orientation toward placing physical barriers, including
distance, between oneself and others; the inclusion of Intimacy indicated that
the unit for which privacy is desired may include one or more intimate others.
The relationship between Solitude and Intimacy suggests that preference for
solitude in this sample did not imply complete withdrawal from others or
reclusiveness. Rather, the same persons who preferred times of solitude preferred
to be with intimates at other times.
Orientations and Environment
The relationships between privacy preferences and opportunities to gain
privacy in both the past and present home environment were examined, first by
predicting present privacy preferences from density or crowding in the past
environment (assuming that past crowding indicates lack of certain kinds of
101
Table II. Reliabilities of and Intercorrelations Among Subscales, Correlations with PPS,
Adult and Student Samples
Scales
1. Privacy
preference
2. Anonymity
Adult
Student
3. Not neighboring
Adult
Student
4. Reserve
Adult
Student
5. Solitude
Adult
Student
6. Seclusion
Adult
Student
0.51 b
0.65 b
0.52 b
0.63 b
0.56 b
0.47 b
0.61 b
0.54 b
0.35 b
0.29 b
0.17
0.04
0.10
0.05
0.14
0.12
0.72
0.71
0.27 b
0.24 b
0.34 b
0.17 a
0.36 b
0,18a
0.18
0.06
0.78
0.76
0.11
0.26 a
0.08
0.20 a
0.28 b
0.23 b
0.57
0.63
0.49 b
0.45 b
0.40 b
0.39 b
0.75
0.67
0.35 b
0.24 b
0.70
0.65
7. Intimacy
Adult
Student
Coefficient
0.56
0.56
ap ( 0 . 0 5 .
bp <0.01.
privacy) and second by examing the closeness o f fit between privacy preferences
and environmental opportunities for obtaining privacy. The correlations
included in the analysis are contained in Table III; to facilitate hypothesis
generation, rather low rs have been presented if their significance reached the
0.05 level.
-0.30 c
0.22 b
-0.19 b
-0.20 b
PPS
-0.20 b
N-N
-0.25 b
-0.28 e
0.24 b
Subscales
Sol
-0.24 b
0.24 b
Sec
0.30 e
-0.20 b
0.30 c
0.35 e
-0.25 b
-0.23 b
-0.26 c
Anon
-.20 b
.24 b
.24 b
.23 b
.23 b
-0.23 b
Res
apPS, total score; N-N, not neighboring; Sec, seclusion; Sol, solitude; Anon, anonymity; Res, reserve; I, intimacy.
bp < 0.05.
Cp( 0 . 0 1 .
Past environment
Time in rural area
Time in metropolitan area
Composite: town size
Single-family dwelling, childhood
Distance from area where spent childhood
Present environment
Items
0.22 b
0.29 c
0.22 b
-0.21 b
0.28 c
-0.24 b
0.34 c
0.27 c
-0.27 c
Table III. Significant Past and Present Environmental Correlates of PPS Total Score and Privacy Subscales, Adult Samplea
~r
103
factors. Predictions were made from two variable sets: actual and perceived
density levels (a) of the current home and (b) of the childhood home.
Focusing on residual stimuli from the past environment, it was
hypothesized that persons growing up in a high-density home, without a room of
their own and a relatively large number of siblings, would have adapted to the
attendant lack of privacy and learned to prefer low levels of privacy. They were
particularly expected to score lower on the Solitude, Reserve, and Intimacy
orientations. None of the expected relationships of the Own Room and Number
of Siblings variables (see Table I for items in the composites) with privacy
orientations were significant, although the largest correlations that did appear
were in the predicted direction. That is, possession of a room of one's own in
childhood showed some relationship to preference for solitude (r = 0.16)and
having a few siblings to high Intimacy scores (r = 0.14).
Similar hypotheses were made relating privacy preferences to the past
environment on a larger scale, with the amount of open space in the
neighborhood and percentage of life spent in towns of various sizes as variables.
AL theory would predict that city dwellers would prefer more anonymity than
small town residents as a result of adapting to the relative availability of
anonymity in the city. In line with the AL predictions, persons spending more of
their lives in cities were significantly more likely to prefer anonymity and
privacy for intimate relationships, and they tended to prefer noninvolvement
with their neighbors.
The only other significant privacy preference correlates of childhood
environmental density were with measures of perceived crowding rather than
measures of actual crowding; that is, persons who perceived their childhood
home as crowded preferred Anonymity and Intimacy. Perceived crowding during
childhood was, however, significantly related to several of the measures of actual
crowding during childhood, including having a room of one's own, living in a
single-family dwelling, number of siblings, and amount of open space. Perceived
crowding thus seemed to act as an intervening variable in the relationship
between actual crowding and privacy preferences, in that perceived crowding
was related both to density measures and to preference for privacy, but density
and privacy preferences were not directly related.
Finally, the relationship between density or crowding in the present
physical environment and privacy preferences was compared with adaptation
level predictions. Density of the present environment would be classified as a
background stimulus within AL theory. In addition to the variables measuring
available privacy within the home and neighborhood listed in Table III,
information was obtained about the degree of overlook of neighbors' entrances,
the size of the lot, and the distance from the nearest neighbor.
Working from the content of the six privacy orientations, it was possible
to determine environmental density variables that would affect the likelihood of
104
Marshall
achieving each of the six kinds of privacy. AL theory would predict that these
preferences represent an adaptation to the amount of relevant privacy available
in the environment; these predictions were tested against the environmental
density correlates of the orientations.
It was predicted that PPS score would be negatively correlated to density
in general; as shown in Table III, the two significant correlations with density
variables were in the hypothesized direction.
Scores on the Solitude, Intimacy, and Reserve subscales were expected to
be related to density levels within both the home and the neighborhood. The
prediction from neighborhood density assumed that density would affect,
respectively, (a) the ability to be alone outside the home, (b) the ability of the
intimate family group to achieve privacy from neighbors, and (c) the ability to
control interaction with and self-disclosure to neighbors. The predictions from
within the home density followed from similar considerations involving the
privacy of individual family members within the home. Of the ten significant
correlations between these three subscales and environmental density variables,
eight were in the direction predicted (Table III).
Orientations toward Seclusion, Anonymity, and Not Neighboring were
expected to be related to density in the neighborhood but not the home setting.
Seclusion in the sense that it appears on this subscale is not likely to be available
in a high-density neighborhood. The relationship of neighborhood density to
Anonymity was expected ~to be positive, since higher density allows anonymity
to occur. By the same token, the relationship with Not Neighboring sould be
curvilinear, with high density making it difficult to achieve privacy by avoiding
involvement with neighbors until density reaches the point that anonymous
relationships appear. Of the three significant density correlates of Seclusion and
Not Neighboring, one related to the home rather than neighborhood setting and
one was not in the direction predicted. The finding that persons who could see
few other houses from their windows preferred seclusion was in accord with the
prediction. All of the significant density correlates of Anonymity dealt with
density within the home rather than the neighborhood.
105
items measuring the amount of privacy available in the present environment (see
Table I) by means o f multivariate analysis o f variance. This analysis involves
computing a multivariate F as a test for significant mean differences between the
three levels of privacy preference for the nine dependent variables simultaneously, and computing stepdown F ratios, independent estimates o f the
significance of effects of individual variables.
PPS total score was found to be negatively related to perceived privacy
within the home (stepdown F = 3.02, p < 0.05) and to the number o f persons
living in the home (stepdown F = 5.21, p < 0.01), although the multivariate F
was not significant. The mean scores indicated that responses of persons with
high preference for privacy fell on the average between "about right" and "not
quite enough" perceived privacy, while persons with tow preference for privacy
tended to respond that they had about the right amount o f privacy. The fit
between preference and environment, then, was better for the group showing
low preference for privacy even though persons with high preference for privacy
tended to have more actual privacy (they lived with few other people and had
fewer rooms that were visually open to each other).
Canonical correlation was used to compare the two sets of variables (Table
IV), privacy orientations and environmental privacy composites (Table I). This
technique expresses the relationship between the variable sets as a single
correlation coefficient and as standardized regression weights indicating the
amount that each variable contributes to the overall correlation. Two linear
Intimacy
Anonymity
Not neighboring
Solitude
Seclusion
Reserve
0.710
-0.558
0.444
0.405
-0.198
0.057
II
Environmental
privacy
II
0.572
-0.543
0.422
-0.402
-0.192
0.185
-0.102
-0.080
-0.062
-0.292
0.369
0.168
-0.569
-0.280
0.078
0.482
-0.258
0.614
aThe canonical rs were computed from a' correlation matrix; the weights refer to
standardized variables. I: R c = 0.61, df = 54, = 108.91,p = 0.000. II: R e = 0.50, df =
40, x 2 = 65.17, p = 0.007.
106
Marshall
107
and source of friends. With regard to the last, it was assumed that if one's friends
came from many sources (neighbors, coworkers, persons one grew up with,
members of the same organizations), they would be less likely to be acquainted
with the person as a whole, but would know him in only one or a few of his
many roles. Although this relationship did not appear, there was a tendency for
persons high on Anonymity not to choose neighbors as friends (r = -0.20, p <
0.05). As expected, high Anonymity scores were related to town size.
Since Reserve involves control over verbal disclosure about oneself, no
particular physical environmental correlates as assessed by this questionnaire
were expected.
An environment compatible with orientation toward Intimacy was
expected to contain opportunities for separation of small groups within the
home and also for separation of the family unit from neighbors. The former
was supported by significant relationships with number of rooms and bedrooms
per person and the latter by a positive relationship with visual privacy for
sunbathing.
On the whole, the fit between preferences and environment was not close,
and seemed to fall below the expectations of those preferring high levels of
privacy, particularly of Solitude and Seclusion.
108
Marshall
neighbors (r--- 0.26, p < 0.01) and having many houses visible from the living
room windows (r--- 0.23, p < 0.05). Persons who felt crowded within the home
had also spent more time in small towns (r = 0.29, p < 0.01) and had lived in
their present neighborhood for a shorter time (r = -0.26, p < 0.01).
Persons who thought they had enough or too much privacy within the
home tended to have a house that insulated noisy activities from quiet ones (r =
0.32, p < 0.01), to have few houses visible from the living room (r = -0.33, p <
0.01), and to be far from their neighbors (r = 0.23, p < 0.05). In addition, they
rated themselves as not easily distractible (r = -0.33, p < 0.01).
Perceived privacy from neighbors was positively related to noise from
traffic or neighbors (r = 0.21, p < 0.05) and negatively related to drawing the
drapes during the day (r = 0.28, p < 0.01). Among the significant correlates of
perceived crowding in childhood were sharing a room (r = 0.40, p < 0.01),
having many siblings (r = 0.25, p < 0.05), and having little open space around
the house (r=-0.25, p < 0.05). Since the answers to all of these items were based
on recall, the correlations serve as a rough estimate of the real relationship.
DISCUSSION
9Means of Attaining Privacy
109
In evaluating the fit between environmental privacy and privacy preferences, the relative homogeneity in environment of the sample should be noted.
Although the sample contained a range of social classes, the sample was
relatively homogeneous in the amount of privacy available in their homes and
neighborhoods. Since homogeneity of the sample lowers correlations, the low
but significant correlations reported should perhaps be given more weight than
would otherwise be true.
Although there was some evidence of a fit between amount of privacy
available in the physical environment and specific privacy preferences, perhaps
the most notable finding was that persons with high privacy preferences often
felt they did not have enough privacy even though they had more than did
persons with low privacy preferences. This was most apparent in the canonical
correlations.
The findings relating environmental variables to perceived privacy and
crowding have implications for environmental designers concerned with increasing opportunities for privacy and reducing perceived crowding through design.
For this relatively low-density suburban sample, noise of, distance from, and
visibility of neighbors were the major physical variables related to low perceived
privacy within the neighborhood. The relationship with drawing the drapes
110
Marshall
REFERENCES
Bloxom, B. (1968). Factorial rotation to simple structure and maximum similarity.
Psyehometrika 33: 237-247.
Helson, H. (1964). Adaptation Level Theory. Harper & Row, New York.
Holfingshead, A. (1965). Two Factor Index of SocialPosition. Yale Station, New Haven.
Kuper, B. (1967). Privacy and Private Housing. Building Design Parternship Graphic Unit,
London,
Levy, L. (1967). The quality of urban fife: An anaylsis from the perspective of mental
health. In Urban America: Goals and Problems. U.S. Government Printing Office,
Washington, D.C., pp. 100-112.
Lewin, K. (1936). Principles of Topological Psychology. McGraw-Hill, New York.
Marshall, N. J. (1970).. Orientations toward privacy: Environmental and personality
components. Unpublished doctoral dissertation in psychology. University of California, Berkeley, 145 pp.
Milgram, S. (1970). The experience of living in cities. Science 167: 1461-1468.
Proshansky, H. M., Ittelson, W. H., and Rivlin, L. G. (1970). Freedom of choice and
behavior in a physical setting. In Proshansky, H. M., Ittleson, W. H., and Rivlin, L. G.
(eds.), Environmental Psychology: Man and His Physical Setting. Holt, Rinehart and
Winston, New York, pp. 173-183.
Simmel, G. (1950). The stranger. In Wolff, K. (ed.), The Sociology of Georg Simmel. Free
Press, Glencoe.
Simon, H. A. (1969). The Sciences of the Artificial. MIT Press, Cambridge, Mass.
Westin, A. F. (1967). Privacy and Freedom. Atheneum, New York.
Wirth, L. (1938). Urbanism as a way of life. American Journal of Sociology 44: 1-24.