Journal of Personality and Social Psychology
1975, Vol. 31, No. 4, 654-660
Antipollution Measures and Psychological Reactance Theory:
A Field Experiment
Michael B. Mazis
College of Business Administration, University of Florida
A field experiment was conducted to determine if hypotheses derived from
psychological reactance theory could explain response to the implementation
of an antiphosphate ordinance. Deprived housewives expressed more positive
attitudes toward the eliminated alternative than did control subjects thereby
supporting reactance theory predictions. Within the experimental group, subjects
were divided into two groups based on their degree of choice deprivation.
As predicted, subjects forced to switch from their preferred detergent
brand expressed less favorable attitudes about the effectiveness of no-phosphate
versus phosphate detergent than subjects who could maintain brand continuity.
Reduced attractiveness of forced alternative rather than enhancement of
forbidden alternative was the principal mode of response resulting from psychological
reactance.
Psychological reactance theory asserts that
when a person believes himself free to engage
in a given behavior and his freedom is eliminated
or threatened with elimination, the individual
experiences psychological reactance, a
motivational state directed toward reestablishment
of the threatened or eliminated freedom
(Brehm, 1966). One major determinant of
reactance, external pressure applied through
social influence or persuasive communications,
has been studied extensively. Laboratory experiments
have typically manipulated social
influence by having confederates restrict subjects'
freedom of choice through freedom
threatening statements (Brehm & Sensenig,
1966; Worchel & Brehm, 1971).
Recent research exploring the influence of
persuasive communications has focused on
factors attenuating reactance effects. Conformity
pressures (Grabitz-Gniech, 1971;
Pallack & Heller, 1971), the involvement of
peers in eliminating alternatives (Worchel &
Brehm, 1971), and individual difference
variables, including manipulated felt compeThis
study was supported by a grant from the
College of Business Administration, University of
Florida.
The author thanks Robert B. Settle for aiding
conceptualization of the study and Dennis C. Leslie
for assisting in data analysis.
Requests for reprints should be addressed to
Michael B. Mazis, College of Business Administration,
University of Florida, Gainesville, Florida 32611.
tence (Wicklund & Brehm, 1968), feelings of
inadequacy (Grabitz-Gniech, 1971), and locus
of control (Biondo & MacDonald, 1971) have
been found to influence reactivity.
While the loss of options may be a major
determinant of psychological reactance also,
it has not been fully explored in reactance
research. Grabitz-Gniech (1971) and Brehm,
Stires, Sensenig, and Shaban (1966) examined
the effect of reducing the choice of records
and paintings, respectively, among college
students in laboratory experiments. While
statistically significant reactance effects were
found, both experiments failed to present an
ego-involving situation in which subjects lost
an important freedom. The possibility of demand
characteristics producing a reactive response
must be considered. Perhaps a more
appropriate laboratory for experimentation is
the marketplace where consumers' freedom
of choice has been frequently restricted
through governmental action such as requiring
safety devices on automobiles; recalling
soups, cranberries, tuna, drugs, and automobiles;
and restricting the use of energy. Unfortunately,
the only field study (Weiner &
Brehm, 1966) examining reactance effects
was debilitated by the failure to apply unobtrusive
pressure to subjects, thereby negating
a major advantage of field experimentation.
The elimination of free choice by authority
sources has been explored in reactance research,
but the only authority source studied
654
ANTIPOLLUTION MEASURES AND REACTANCE THEORY 655
has been the psychological experimenter
(Hammock & Brehm, 1966). For research
involving authority sources to have generalizability,
particularly for social issues,- a wider
sampling of authority sources is needed. For
example, Davis and Eichhorn (1963) have
studied the issue of compliance and noncompliance
with physicians' regimens. The current
experiment is directed toward determining
if responses to the imposition of a "socially
beneficial" governmental action (antiphosphate
ordinance) might be predicted
from psychological reactance theory.
Obviously, not all individuals react against
all restrictions of free choice. It remains for
research to specify which subjects and under
which restrictions psychological reactance is
manifested when options are eliminated or
curtailed by authority sources.
The principal consequence of reactivity
needs to be specified more precisely also. The
present study explores two major manifestations
of reactivity, enhanced attractiveness of
restricted alternatives and reduced desirability
of alternatives forced upon subjects. Most
reactance studies have failed to separate these
two effects.
The present investigation concernsresponse
to an event affecting many households
throughout the United States—imposition of
an antiphosphate law. On Jaunary 1, 1972,
Dade County (Miami), Florida, began prohibiting
the sale, possession, or use of laundry
detergents and other cleaning products
containing phosphates. Since only a small
number of popular brands were available in
no-phosphate formulations, shoppers found
their choice of laundry detergent drastically
diminished. These dramatic changes in the
number of choice alternatives provided an
opportunity to examine psychological reactance
theory predictions in a field setting.
According to reactance theory, Miami
households would be motivationally aroused
due to the phosphate restrictions and feel an
increased desire to have the forbidden detergent
as contrasted with households in a
control city (Tampa) whose freedom of choice
was unrestricted. The reactance aroused in
Miami subjects should result in higher effectiveness
attributed to phosphate detergents as
compared with Tampa subjects and more
negative attitudes toward governmental regulation
of environmental matters which should
influence the product evaluations of Miami
consumers.
Since Miami households could be classified
into three major subgroups, additional reactance
theory predictions may be made. The
first group consisted of housewives who were
able to continue purchasing products labeled
by their favorite brand names, but which
were now being sold without phosphates;
these households are referred to as "nonswitchers."
Since one leading detergent manufacturer
quickly began distributing no-phosphate
reformulationsof all its existing laundry
detergent brands, users could continue topurchase
their regular brands after the antiphosphate
law went into effect.
Several other detergent manufacturers did
not begin distribution of no-phosphate reformulations
for several months and therefore
users were forced to switch brands after
the phosphate restrictions went into effect.
These consumers, who were forced to switch
from their regular brands, are referred to as
"switchers."
Based on their dissimilar motivational
states, it is anticipated that (a) switchers
would express more negative attitudes about
the effectiveness of no-phosphate products
and/or more positive attitudes about the
effectiveness of phosphate products than
would nonswitchers and (b) switchers would
have a more negative attitude toward the
antiphosphate law than would nonswitchers.
Switchers should feel a greater restriction
of their freedom of choice than should nonswitchers
as a result of being unable to purchase
a laundry detergent sold under their
favorite brand label. It must be remembered
that both switchers and nonswitchers were
using a totally new product; however, in one
case consumers were using a product with a
"new" label or brand name, while in the other
they were able to use a reformulated product
being sold under their favorite brand name.
The third group of Miami households are
those who defied the antiphosphate ordinance
by smuggling detergent into Miami from surrounding
counties or who accummulated large
amounts of phosphate detergent before the
January enforcement date. Since these smug-
656 MICHAEL B. MAZIS
TABLE i
MEANS AND STANDARD DEVIATIONS FOR EFFECTIVENESS
RATINGS CF PHOSPHATE DETERGENTS
Characteristic
Whiteness
Freshness
Cleans in cold water
Brightness
Stain removal
Pours easily
Gentleness
Miami
(n = 76)
M.
8.68
8.77
8.52
8.31
8.00
9.45
8.81
SD
1.87
2.08
2.15
2.09
2A3
2.05
1.80
Tampa
(« = 45)
M
8.27
7.87
7.47
7.84
6.96
9.07
8.71
SD
1.56
1.51
1.83
1.80
2.05
1.77
1.51
Note. Based on an 11-point scale with 11 labeled ''absolutely
perfect" and 1 labeled "poor."
glers and hoarders may have had very favorable
attitudes toward phosphate detergents
before the enactment of the no-phosphate
restrictions, no predictions can be made concerning
their attitudes about the effectiveness
of phosphate brands based on reactance theory.
Their predisposition toward phosphate
brands may have been instrumental in their
being categorized as "violators." On the other
hand, switchers and nonswitchers were placed
into those categories as a result of detergent
manufacturers' decisions about whether to
produce no-phosphate reformulations of detergent
brands.1
METHOD
An instrument was designed to determine attitudes
about laundry detergents and related products,
opinions about laws regulating the use of phosphates,
and demographic characteristics. Seventy-six
interviews were completed by four female interviewers
in Miami and 45 completed questionnaires were
returned by the two female interviewers in Tampa.
Nine questionnaires were discarded as a result of
incomplete responses. Interviews were conducted
from a period from 7 to 9 weeks after the antiphosphate
statute became effective.
Since only limited funding was available to defer
the cost of interviewing, sample selection in both
Tampa and Miami emphasized subject homogeneity
to restrict the impact of extraneous variables. To
enhance the similarity between Tampa and Miami
samples, all respondents interviewed were: (a) English-speaking
Caucasian women; (b) with family
incomes of $7,500 to $15,000 per year; (c) who had
at least one child under 16 years of age living at
home; (d) who resided in single-family dwellings;
and (e) who used primarily phosphate detergent
brands during the precceding 6 months. Comparisons
between Tampa and Miami samples revealed no significant
income, age, or educational differences.
Two middle-income census tracts were chosen in
both cities and blocks were randomly selected within
tracts. Two sampling points were randomly designated
within each block and interviewing commenced
from each point in a clockwise direction
until two respondents were found who met the criteria
enumerated above.
RESULTS
Miami versus Tampa
At the beginning of each interview, subjects
were asked to rate the effectiveness of
the phosphate laundry detergent they had
used most during the previous 6 months.
Seven brand characteristicswere evaluated on
an 11-point scale labeled "absolutely perfect"
and "poor" at the end points. Salient
characteristics were determined through a
free association proceduresimilar to that used
by Fishbein (1967).
Reactance theory predictions are supported
by the data in Tables 1 and 2. On all seven
characteristics, Miami subjects gave higher
mean effectiveness ratings to phosphate
brands than did Tampa subjects.
The two way Groups X Trials analysis of
variance (Veldman, 1967, pp. 247-257) in
Table 2 provides statistical support for the
proposition that Miami housewives were in a
reactive motivational state. Overall, Miami
subjects rated phosphate detergents as being
more efficacious than did Tampa subjects
(F = S.42, p< .05).
The significant interaction between cities
and characteristics (F = 2.27, p < .05) shows
that the reactivity of Miami housewives was
not manifested uniformly across all characterTABLE
2
ANALYSIS OF VARIANCE FOR EFFECTIVENESS RATINGS
OF PHOSPHATE DETERGENTS
1
No subjects classified as nonswitchers switched to
another manufacturer's brand after enactment of the
antiphosphate ordinance.
Source
Groups: Miami vs. Tampa
Error (Groups)
Trials : Characteristics
Groups X Trials
Error (Trials)
MS
77.74
14.34
35.06
3.97
1.75
F
5.42*
20.07**
2.27*
* p < .05.
**p < .001.
ANTIPOLLUTION MEASURES AND REACTANCE THEORY 657
TABLE 3
ATTITUDE TOWARD GOVERNMENT AND
ANTIPHOSPHATE LAWS
,. . . .
VctriiiiDlC
The government
should play an important
role in protecting
our water
from pollution.
Legal restrictions
should be imposed
against the sale of
detergents containing
phosphates.
Miami
M
2.39
2.76
SD
.90
.93
Tampa
M
1.64
2.20
SD
1.20
1.17
£
3.62*
2.72*
Note. Responses on Likert-type scale with "1" indicating
strong agreement and "5" indicating strong disagreement.
*t <-01.
istics. Also, there was a significant main effect
for characteristics indicating that subjects did
discriminate among various detergent attributes
by not rating'detergent brands as being
equally effective on all characteristics (F =
20.07, p< .001).
Miami subjects expressed a less optimistic
view about the success of governmental action
in solving water pollution problems and
toward the usefulness of phosphate content
regulation than did Tampa subjects (see
Table3).
Tampa residents stated a stronger degree of
agreement with the statement, "The government
should play an important role in protecting
our water from pollution," than did
Miami subjects (t — 3.62, p < .01). In addition,
Tampa housewives had greater expectations
about the usefulness of laws restricting
the sale of detergents containing phosphates
than did Miami respondents (t = 2.72, p <
.01), thereby sustaining the view that Miami
housewives were experiencing psychological
reactance.
Switchers versus Nonswitchers
Brand ratings. According to psychological
reactance theory, the amount of freedom
eliminated directly influences psychological
reactance. Since switchers were experiencing
more psychological choice deprivation, they
should provide lower mean effectiveness ratings
to no-phosphate brands than should non-
switchers.
The data in Table 4 shows that of the 76
Miami households responding (9 of which had
incomplete data), the 44 switchers provided
lower average effectiveness ratings across all
seven detergent attributes than did the 23
nonswitchers. The Groups X Trials analysis
of variance in Table 5 reveals significantly
lower attractiveness estimates for no-phosphate
products by switchers than by nonswitchers
(F = 4.48, p < .OS).
The greater degree of psychological reactance
being experienced by switchers may
take the form of increased attractiveness of
the restricted alternative (phosphate detergent),
as well as decreased attractiveness of
'the alternative forced upon subjects (nophosphate
detergent). Similar ratings of phosphate
detergent effectiveness between Tampa
and Miami subjects are shown in the middle
sections of Tables 4 and S (F = .00, p <
TABLE 4
MEANS AND STANDARD DEVIATIONS FOR EFFECTIVENESS RATINGS OF DETERGENTS
Characteristic
Whiteness
Freshness
Cleans in cold water
Brightness
Stain removal
Pours easily
Gentleness
No-phosphate detergents
Switchers
M
7.07
7.11
6.36
6.30
6.14
8.45
7.23
SD
1.60
1.70
2.12
2.00
2.53
2.01
1.92
Nonswitchers
M
7.87
7.85
7.61
7.43
7.41
9.17
8.04
SD
1.72
1.85
2.37
2.13
1.96
2.37
2.08
Phosphate detergents
Switchers
M
8.59
8.75
8.72
8.41
7.79
9.30
8.93
SD
1.63
1.61
1.93
1.78
2.00
1.41
1.70
Nonswitchers
M
8.61
8.78
8.43
8.17
8.13
9.78
8.74
SD
1.51
1.41
1.42
1.61
1.95
1.89
1.21
Difference scores
Switchers
M
1.52
1.63
2.36
2.27
1.66
.84
1.70
SD
.79
1.17
1.43
1.51
1.27
1.05
1.23
Nonswitchers
M
.74
.91
.83
.74
.70
.61
.70
SD
1.70
1.88
2.43
2.26
2.01
1.62
2.01
Note. Based on an 11-point scale labeled "absolutely perfect" and "poor" at end points.
658 MICHAEL B. MAZIS
TABLE 5
ANALYSIS or VARIANCE FOR EFFECTIVENESS RATINGS OP DETERGENTS
0
Groups : Switchers vs.
nonswitchers
Error
Trials : Characteristics
Groups X Trials
Error
No-phosphate detergents
if
1
65
6
6
390
MS
99.06
22.09
35.32
.94
1.40
F
4.48*
25.29***
.67
Phosphatedetergents
df
1
65
6
6
390
MS
.05
12.88
15.33
1.32
1.10
F
.00
13.92***
1.20
Difference scores
df
1
65
6
6
390
MS
99.27
13.21
8.31
3.22
1.60
F
7.51**
5.20***
2.02
* p < .05.
**/> <.01.
***# <.001.
.99). These nonsignificant findings are explored
in the final section of this article.
In a within-subjects experimental design,
the "total reactance effect" (Hammock &
Brehm, 1966) can be examined by analyzing
difference scores generated by subtracting
each subject's rating of no-phosphate detergent
from her rating of phosphate detergent
for each characteristic. In this way both hypotheses
derived from reactance theory (enhancement
of forbidden alternative and derogation
of forced alternative) can be tested
simultaneously. The last sections of Tables 4
and 5 indicate that switchers felt there was a
substantial difference in efficacy between phosphate
and no-phosphate formulations whereas
nonswitchers felt that no-phosphate brands
were only slightly less effective than phosphate
variations (F = 7.S1, p < .01).
Attitudes toward detergent use. In addition
to rating phosphate and no-phosphate detergents
on seven characteristics, subjects evaluated
the potency and cost of alternative
formulations. As a result of being in a reactive
motivational state, subjects who were
forced to switch from their favorite brands
would be expected to express more negative
attitudes about the amount of detergent and
extra ingredients required for each washload
and about the cost of washing clothes when
using no-phosphate detergent than would non-
switchers.
Table 6 indicates that while nearly a third
of. switchers felt they had been using more
no-phosphate detergent, x2
(2) = 9.4S, p <
.01, and additional ingredients (e.g., bleach
or fabric softener), x
2
(2) = 4.15, P < .05,
with each washload than they did with phosphate
detergent, less than 10% of nonswitchers
felt this was the case. There is a similar
disparity about perceptions of the relative
cost of using phosphate and no-phosphate
products. Over 71% of switchers believed
that no-phosphate brands were more expensive
to use than products containing phosphates,
but only 30% of nonswitchers expressed negTABLE
6
ATTITUDE TOWARD DETERGENT USE
Group
Switchers
Nonswitchers
Amount of
detergent used"
More
14
(32%)
(4%)
Less
4
(10%)
5
(28%)
Same
26
(58%)
17
(68%)
Amount of extra
ingredients usedb
More
14
(32%)
2
(8%)
Same
or Les8d
30
(68%)
21
(92%)
Cost of
no-phosphate detergent"
More
31
(71%)
7
(30%)
Same
or Lessd
13
(29%)
16
(70%)
•X'(2) = 9.45, p < .01.
b XJ?1) = 4.15, p < .05. , . .
« xHl) = 9.58. P < -001.d
Categories combined to provide sufficient expected value per cell.
ANTIPOLLUTION MEASURES AND REACTANCE THEORY 659
ative feelings about the cost of using their
reformulated no-phosphate brands, x2
(1)=
9.58, p < .001.
DISCUSSION
While student subjects in laboratory experiments
have often behaved as predicted
by psychological reactance theory, the current
study provides strong support for the theory's
predictions among housewives in a field setting.
As predicted, Miami subjects rated
phosphate detergents as being more effective
ithan did Tampa subjects. Hypotheses derived
from psychological reactance theory concerning
Miami versus Tampa subjects' attitudes
toward government's role in protecting the
public from water pollution and in imposing
legal restrictions against the sale of products
containing phosphates were sustained as well.
Since adaptation effects (Helson, 1964)
may have contributed to the enhanced attractiveness
of phosphate detergent as a result
of expanded use of the less effective nophosphate
brands, the most compelling evidence
supporting a reactance theory interpretation
is found in the attitudinal differences
of switchers and nonswitchers. Analysis of
effectiveness ratings revealed that there were
no significant mean differences in the evaluation
of alternative brands of no-phosphate
detergent. As a result, the differential attitudes
expressed by switchers and nonswitchers
can be attributed to psychological reactance
rather than to physical differences in
the brands used.
This research does make several contributions
to the psychological reactance literature.
First, it demonstrates that psychological reactance
is not solely a laboratory phenomena.
Housewives did feel more psychological choice
deprivation when they were forced to switch
from their favorite phosphate brands to unfamiliar
brands of no-phosphate laundry de-
tergent.
Second, greater insight into the consequences
of reactance were obtained. Negative
attitudes about the effectiveness of the alternative
forced upon housewives rather -than
enhancement of the eliminated alternative
was found.
Since most subjects expressed satisfaction
with their laundry detergents, psychological
reactance would be less likely to take the form
of increased desirability of the eliminated alternative
as compared with decreased attractiveness
of the forced alternative due to ceiling
effects. The mean rating for control group
(Tampa) subjects was 8.02 on an 11-point
scale. While the circumstances surrounding
this experiment may be unique, it is possible
that reactance often takes the form of a reduction
in the attractiveness of the forced alternative
since people are satisfied with the
products they use as a result of dissonance
reduction. More valuable than such speculation
would be further research on this topic.
Third, public policy implications of psychological
reactance research have been made
apparent. From a societal perspective, the
most appropriate laboratory for psychological
reactance research might be the consumer
market. While governmental agencies and legislators
issue administrative decisions and enact
ordinances which restrict freedom of
choice, consumer assessment after restriction
is rarely undertaken. Too much restriction
may result in extensive noncompliance. However,
additional research is needed to assess
the degree of consumer reactivity as a result
of: (a) the issue, (b) the proportion of
choice alternatives remaining and number of
substitutes available, (c) the authority sources
(e.g., federal vs. local and state government;
legislative vs. executive decisions) used to
eliminate freedom of choice, and (d) the permanence
of a reactive motivational state.
REFERENCES
Biondo, J., & MacDonald, A, P. Internal-external
locus of control and response to influence attempts.
Journal of Personality, 1971,39, 407-419.
Brehm, J. W. (Ed.), A Theory of Psychological
Reactance. New York: Academic Press, 1966.
Brehm, J. W., & Sensenig, J. Social influence as a
function of attempted and implied usurpation of
choice. Journal of Personality and Social Psychology,
1966, 4, 703-707.
Brehm, J. W., Stires, L. K., Sensenig, J., & Shaban,
J. The attractiveness of an eliminated choice
alternative. Journal of Experimental Social Psychology,
1966, 2, 301-313.
Davis, M., & Eichhorn, R, Compliance with medical
regimens: A panel study. Journal of Health and
Human Behavior, 1963, 4, 240-249.
Fishbein, M. A behavior theory approach to the
relations between beliefs about an object and the
attitude toward the object. In M. Fishbein (Ed.),
660 MICHAEL B. MAZIS
Readings in attitude theory and measurement.
New York: Wiley, 1967.
Grabitz-Gniech, G. Some restrictive conditions for
the occurrence fo psychological reactance. Journal
oj Personality and Social Psychology, 1971, 19,
188-196.
Hammock, T., & Brehm, J. W. The attractiveness of
choice alternatives when freedom to choose is
eliminated by a social agent. Journal of Personality,
1966, 34, S46-SS4.
Kelson, H. Adaptation-level theory. New York:
Harper & Row, 1964.
Pallack, M. S., & Heller, J. F. Interactive effects of
commitment to future interaction and threat to
attitudinal freedom. Journal of Personality and
Social Psychology, 1970, 14, 39-4S.
Veldman, D. J. Fortran programming for the behavioral
sciences. New York: Holt, Rinehart &
Winston, 1967.
Weiner, J., & Brehm, J. Buying behavior as a function
of verbal and monetary inducements. In J.
W. Brehm (Ed.), A theory of psychological reactance.
New York: Academic Press, 1966.
Wicklund, R. A., & Brehm, J. W. Attitude change
as a function of felt competence and threat to
attitudinal freedom. Journal oj Experimental Social
Psychology, 1968, 4, 64-75.
Worchel, S., & Brehm, J. W. Direct and implied
social restoration of freedom. Journal of Personality
and Social Psychology, 1971, IS, 294-304.
(Received September 20, 1973)