Art Review 4
leedee23
EBSCO-FullText-04_09_2026.pdf
Professional Psychology: Research and Practice 1993, Vol. 24. No. 2,164-170
Copyright 1993 by the American Fsvchological Association, inc 0735-7028/93/S3.00
The Impact of Paradoxical Interventions on Perceptions of the Therapist and Ratings of Treatment Acceptability
Gordon R. Betts and Rory Remer
Ethical questions about the manipulative, deceptive nature of paradoxical interventions were ad- dressed. Perceptions of the therapist and of the acceptability of paradoxical vs. nonparadoxical directives were examined using simulation methodology. Participants, 97 undergraduates, took part in a semester-long family simulation of excessive family arguing associated with a rebellious adolescent daughter. Before experimental manipulation, the "families" participated in 4 role-play- ing exercises and attended a family therapy session. They then received a letter from their therapist containing 1 of the 2 types of directives and, subsequently, responded to the dependent measures. Results indicated (a) paradoxical directives did not negatively influence perceptions of therapist attractiveness, expertness, or trustworthiness; (b) they were judged less acceptable than nonpara- doxical directives, although neither was deemed unacceptable. Implications for clinical practice and future research are discussed.
Paradoxical interventions in psychotherapy have attracted widespread attention in recent years. The label paradoxical in- tervention refers to a wide variety of therapeutic techniques and directives, all of which share the following curious trademark: They are designed to eliminate the client's symptom by encour- aging it, either directly or indirectly.
The current popularity of therapeutic paradox can be traced primarily to the writings of several prominent systems-oriented family therapists (e.g., Haley, 1963,1976; Papp, 1980,1983; Sel- vini Palazzoli, Cecchin, Praia, & Boscolo, 1978; Watzlawick, Weakland, & Fisch, 1974; Weeks & L'Abate, 1982). However, family therapy practitioners do not hold a monopoly on these unorthodox methods. Paradoxical interventions have been practiced by therapists across a wide range of theoretical orien- tations since the early days of psychotherapy, although they have not always been labeled as such (Seltzer, 1986; Weeks & lAbate, 1982). The list of early practitioners of therapeutic tech- niques that can be classified as paradoxical includes such prom- inent psychotherapy figures as Viktor Frankl, Alfred Adler, Frederick "Fritz" Perls, John Rosen, and Knight Dunlap, to name just a few (Seltzer, 1986).
GORDON R. BETTS received his PhD from the University of Kentucky in 1992. He is currently a Counseling Psychologist for the Bluegrass Regional Mental Health/Mental Retardation Board and is in private practice in Lexington, Kentucky. RORY REMER received his PhD from the University of Colorado, Boulder in 1972. He is currently a Professor of Counseling Psychology in the Department of Educational and Counseling Psychology at the University of Kentucky. WE THANK Steve Finger, Wally Reynolds, John Elliot, and Lou Hicks for assisting in data collection. Our appreciation also goes to John F. Crosby, and O'Neal Weeks, Professors in the Department of Family Studies at the University of Kentucky, for their support of the family simulation research project. CORRESPONDENCE CONCERNING THIS ARTICLE should be addressed to Gordon R. Betts, 4501 Brookglen Place, Lexington, Kentucky 40515- 6102.
In the contemporary psychotherapy literature, theoretical discussions of paradoxical strategies and anecdotal clinical re- ports attesting to their efficacy have become increasingly preva- lent since the 1960s. Only in the past decade, however, have researchers launched a serious effort to study the effects of par- adoxical interventions in a controlled way. Most of the research cqnducted thus far has examined the effectiveness of various paradoxical techniques. Indeed, the accumulated treatment outcome research is quite promising. Recent meta-analytic and traditional reviews of treatment outcome studies indicate that paradoxical techniques are effective in treating a variety of hu- man emotional and relationship problems (DeBord, 1989; Dowd & Milne, 1986; Hill, 1987; Katz, 1984;Schotte, Ascher, & Cools, 1989; Shoham-Salomon & Rosenthal, 1987; Strong, 1984).
However, despite their apparent efficacy and their popular- ity across diverse schools of therapy, the use of paradoxical interventions in clinical practice is muddled in controversy due to widespread concerns that such unorthodox methods are overly manipulative, deceptive, and possibly unethical (Brown & Slee, 1986; Cavell, Frentz, & Kelley, 1986; Doherty, 1989; Hunsley, 1988; Johnson, 1986; Mahoney, 1986; Ridley & Tan, 1986; Schmidt, 1986; Whan, 1983). Central to the critics' charges against therapeutic paradox is the fact that therapists who use paradoxical techniques rarely reveal their "true" inten- tion when clients are asked to continue (or even exaggerate) their symptomatic behavior. Johnson (1986) stated that para- doxical therapists appear to "put on an act" and "deliberately deceive" clients (p. 301). Even when paradoxical directives are made to sound more palatable by including a rationale for their use, the resulting position of the therapist can still appear quite pessimistic and counterintuitive. Detractors of paradox con- tend that this symptom-endorsing stance, coupled with ele- ments of deception, is likely to undermine the therapist's abil- ity to convey sufficient levels of genuineness and empathy, thereby reducing trust in the therapeutic relationship. Some have gone so far as to suggest that the repeated use of such
164
Th is
d oc
um en
t i s c
op yr
ig ht
ed b
y th
e A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. Th
is a
rti cl
e is
in te
nd ed
so le
ly fo
r t he
p er
so na
l u se
o f t
he in
di vi
du al
u se
r a nd
is n
ot to
b e
di ss
em in
at ed
b ro
ad ly
. A
ll rig
ht s,
in cl
ud in
g fo
r t ex
t a nd
d at
a m
in in
g, A
I t ra
in in
g, a
nd si
m ila
r t ec
hn ol
og ie
s, ar
e re
se rv
ed .
PARADOXICAL INTERVENTIONS 165
deceptive methods in clinical practice potentially could under- mine the basic trust that supports the entire profession (e.g., Cavell et al., 1986; Doherty, 1989).
Proponents of paradoxical interventions have developed ar- guments to counter the critics' charges (e.g., Brown & Slee, 1986; Haley, 1987; Hunsley, 1988; Weeks & lAbate, 1982); however, relatively few empirical studies have been designed explicitly to investigate the ethical concerns associated with paradoxical methods. A perusal of the existing literature indicates that re- searchers have adopted two general approaches in addressing this issue. One approach has been to examine the impact that paradoxical interventions have on therapeutic relationship measures, the assumption being that if paradoxical strategies are overly deceptive and manipulative then client perceptions of the therapeutic relationship will be weakened along certain dimensions, such as therapist attractiveness, expertness, and trustworthiness. The results of these studies suggest that para- doxical interventions (compared to nonparadoxical interven- tions) can be administered without damaging perceptions of therapist attractiveness, expertness, and trustworthiness (Cono- ley & Beard, 1984; Conoley & Garber, 1985; Newton & Dowd, 1990; Perrin & Dowd, 1986) or ratings of therapist empathy and unconditional regard (Feldman, Strong, & Danser, 1982; Lopez & Wambach, 1982). However, contrary to the above findings, McMillan and Johnson (1990) recently reported that therapists who used paradoxical interventions received lower ratings on attractiveness, expertness, and trustworthiness than therapists who used cognitive-behavioral interventions. In addition, other intervening variables such as the type of rationale provided with the paradoxical intervention (Hills, Gruszkos, & Strong, 1985), previous therapist statements and interventions (Feld- man et al., 1982), and core therapeutic conditions exhibited by the therapist (Conoley & Beard, 1984) can exert significant indi- rect effects on these same client ratings.
A second approach to investigating the potential ethical problems associated with paradoxical treatments has been to evaluate the consumer acceptability of such methods directly. As defined by Kazdin (1981), treatment acceptability "refers to judgments of lay persons, clients, and others of whether proce- dures proposed for treatment are appropriate, fair and reason- able for the problem or client" (p. 493). Preliminary evidence exists indicating that paradoxical interventions are judged gen- erally unacceptable by the lay public and less acceptable than other treatment alternatives (Cavell et al, 1986; Mittl & Robin, 1987).
Although the general izability problems of analogue research are well known, studies investigating treatment acceptability, including those by Cavell et al. (1986) and Mittl and Robin (1987), have traditionally used analogue methods (Cross Cal- vert & Johnston, 1990). Indeed, some researchers recommend a strict analogue approach (i.e, using typewritten case descrip- tions and independent raters) over obtaining ratings from actual therapy clients in investigating treatment acceptability, because factors that are independent of the treatment procedures them- selves but that may influence acceptability ratings (e.g, thera- pist and therapeutic relationship factors) are said to be mini- mized when the former methodology is followed (e.g, Cavell et al, 1986; Kazdin, 1980). Ironically, it is for this very reason that Kolko and Milan (1986) have sharply criticized Cavell et al.'s
(1986) study, arguing that the contextual factors framing any treatment procedure, which inevitably are deemphasized in most analogue studies, exert a powerful impact on perceptions of the treatment's acceptability and therefore should be in- cluded in the experimental design. The simulation methodol- ogy used in the present study is unique in that it retains some of the advantages of analogue methodology (i.e, increased stan- dardization and experimenter control) while also incorporating certain contextual factors, thereby more closely emulating an actual therapeutic situation than does a strictly analogue de- sign.
In summary, the existing research, although still somewhat sparse, tends to suggest that paradoxical interventions can be administered without damaging clients' perceptions of thera- pists who use them. However, evidence also exists indicating that paradoxical interventions are viewed by the lay public as generally unacceptable and less acceptable than other, more conventional interventions. Intuitively, these two findings ap- pear inconsistent, although it is possible that ratings of thera- pists and the interventions they use are more independent than is generally presumed. Perrin and Dowd (1986), for instance, reported that even though paradoxical directives were viewed as more manipulative, tricky, and confusing than nonparadoxi- cal directives in their study, this evidently did not negatively influence clients' ratings of therapists on a variety of measures.
The purpose of this study was to supplement the current research evidence pertaining to the ethics of intervening para- doxically in therapy. Specifically, clients' reactions to a para- doxical directive versus a nonparadoxical directive were com- pared in terms of their perceptions of the therapist and the acceptability of the intervention itself. Although the current study must be classified as analogue research because the par- ticipants were not actual clients involved in therapy, simulation and role-playing procedures were used to enhance realism, the- oretically increasing the generalizability of the results to actual therapy settings. A family therapy format was chosen over an individual therapy format to accommodate the simulation methodology used and because paradoxical interventions are commonly used in a family therapy context.
On the basis of previous research, no statistically significant differences were expected between the paradoxical and non- paradoxical directives on participants' ratings of therapist at- tractiveness, expertness, and trustworthiness. The research on paradox and treatment acceptability is more limited. More- over, the two studies that have been published used strict ana- logue methods compared to simulation methods in this study. Therefore, similar ratings of treatment acceptability were also expected. A second independent variable was the family role adopted by participants in their simulated family (i.e, father, mother, or daughter). Although there appears to be no convinc- ing theoretical basis for making directional hypotheses with respect to this variable, differences were anticipated because of possible disparate attitudes among family members stemming from the initial family role structure provided to participants.
Method Participants
The initial sample comprised 102 (68 female and 34 male) college students enrolled in an undergraduate family studies class at the Uni-
Th is
d oc
um en
t i s c
op yr
ig ht
ed b
y th
e A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. Th
is a
rti cl
e is
in te
nd ed
so le
ly fo
r t he
p er
so na
l u se
o f t
he in
di vi
du al
u se
r a nd
is n
ot to
b e
di ss
em in
at ed
b ro
ad ly
. A
ll rig
ht s,
in cl
ud in
g fo
r t ex
t a nd
d at
a m
in in
g, A
I t ra
in in
g, a
nd si
m ila
r t ec
hn ol
og ie
s, ar
e re
se rv
ed .
166 GORDON R. BETTS AND RORY REMER
versity of Kentucky who elected to participate in a semester-long fam- ily simulation project in lieu of completing an in-depth autobiographi- cal term paper. Although no detailed demographic information was gathered, easily the majority were in the 18-22 age bracket. Partici- pants were assigned to groups of 3, consisting of 1 male student and 2 female students, using stratified random assignment procedures. These groups of 3 made up 34 simulated families, each composed of a father, a mother, and an adolescent daughter. Five participants were excluded from the analysis because of various attrition factors, result- ing in a final sample size of 97 (66 female and 31 male) participants.
Therapists
Therapist pool. Seventeen (4 male and 13 female) therapists were recruited from a pool of current and previous students enrolled in a graduate-level marriage and family therapy class at the University of Kentucky to conduct the simulated family therapy sessions. All of the therapists were enrolled in graduate programs in either counseling psychology (9), school psychology (1), or family studies (7). Levels of actual, nonacademic clinical experience varied: Nine therapists had less than I year, 1 had 1 -2 years, 2 had 2-5 years, and 5 had over 5 years' experience. All of the therapists were White and ranged in age from 25 to 49 years, with an average age of 37. Assignment of families to thera- pists was made randomly—two families per therapist with one family in each treatment condition—to control for therapist differences. Al- though the present research design did not require therapist delivery of the treatment interventions, each therapist had received training in family therapy, including at least an overview of paradoxical interven- tions, as part of their marriage and family therapy class instruction prior to participating in the study.
Therapist instructions. To increase consistency among the initial therapy sessions, therapists received written instructions beforehand. Specifically, they were told to focus on gathering information, getting to know family members, establishing rapport, and assessing the fam- ily's presenting problem. They also were told not to intervene with advice, assignments, or recommendations to the family. The sessions were held in counseling offices on campus and lasted between 40 min and 1 hr. Despite these general guidelines, the importance of spontane- ity and relaxation in carrying out the role play was emphasized. Re- garding students' potential breaking out of role during the simulated session, therapists were encouraged to stay in role in handling such situations if possible. For example, if a student says "This is really stupid," the therapist might respond by saying "It sounds like being here in therapy is uncomfortable for you." Finally, as an added control for potential confounding variables, therapists were blind to which treatment condition each of their simulated families were assigned.
Instrumentation
Counselor Rating Form—Short Version. The Counselor Rating Form—Short Version (CRF-S; Corrigan & Schmidt, 1983) was used to measure participants' perceptions of therapist attractiveness, expert- ness, and trustworthiness. The CRF-S consists of 12 adjectives (4 items per dimension) that are scored on a 7-point Likert scale anchored by the words not very (1) and very (7). The CRF-S is a shortened and revised version of Barak and LaCrosse's (1975) Counselor Rating Form, which has been reported as being the most frequently used measure of client perceptions of the counselor (Heppner & Claiborn, 1989). The CRF-S has reported interitem reliabilities, ranging from .82 to .94, for its three subscales: counselor Attractiveness, Expertness, and Trustworthiness (Corrigan & Schmidt, 1983).
Treatment Evaluation Inventory—Short Form. A slightly modified version of the Treatment Evaluation Inventory—Short Form (TEI-SF; Kelley, Heffer, Gresham, & Elliot, 1989) was used to measure partici-
pants' perceptions of treatment acceptability. The TEI-SF is a short- ened version of the original Treatment Evaluation Inventory (TEI; Kazdin, 1980) and consists of nine items scored on a 5-point Likert scale ranging from strongly disagree(1) to strongly agree (5). Compared to the original TEI, Kelley et al. (1989) reported that the TEI-SF is more readable, quicker to complete, better liked by respondents, and still able to differentiate among alternative treatments, thereby lending support to its construct validity. Coefficient alpha estimates of internal consistency are comparable to the TEI (Kelley et al., 1989). Because both the TEI and the TEI-SF were developed originally to measure differential treatment acceptability among children's behavioral treat- ments, the wording of three items on the TEI-SF were modified to match the family therapy paradigm of this study. Specifically, the word child was replaced by family member. One item was omitted because it was not applicable for adults in treatment. Five items remained un- changed. The Cronbach alpha reliability for the modified, 8-item TEI- SF used in this study was .81.
Other measures. Two ancillary questionnaires—one pertaining to the decision-making processes manifested by the simulated families and one pertaining to the perceived realism of the simulated interac- tions—were administered at selected intervals during the project but were not analyzed. The purpose of gathering this additional data was to assist in evaluating the usefulness of simulation methodology for future research.
Procedure
Students in an undergraduate family studies course were ap- proached to participate in a simulation of family interactions. First, the simulation and requirements for participation were described in writing as part of the class syllabus. Further explanation was provided verbally during the fourth class period, at which time volunteers for the study were elicited. As part of the verbal introduction to the proj- ect, students were given basic information about role playing as well as estimates of the time commitment and work load involved in the simu- lation. They were told that the general focus of the simulation would be on role playing certain family interactions and relationships within a family that had a particular problem (i.e., excessive family arguing associated with a rebellious teenage daughter). Although the specifics of each phase of the simulation were not provided initially (to preserve spontaneity), the experiences were described as nonthreatening and even enjoyable.
Overview of phases. The simulation consisted of four semistruc- tured role-playing exercises and one simulated family therapy session prior to the experimental manipulation. Packets of materials for each phase—including instructions and the appropriate questionnaires— were distributed in class on completion and submission of the mate- rials from the previous phase. No extended check was made to ensure that the groups actually carried out the simulated family interactions other than having them submit their completed questionnaires and brief written descriptions (as well as receipts, when appropriate) of their activities. In addition, the therapists recorded participants" atten- dance of the simulated therapy sessions.
In Phase 1, participants were provided with initial instructions and a brief description of the family—father, mother, and daughter—to be simulated. After deciding on a specific family role to adopt, they were told to expand on the information already given by creating a more detailed "family history" (e.g., deciding where the wife or husband worked, what kinds of things they liked to do, etc.). Next, participants practiced role playing by interacting in their new family roles, culmin- ating in a family decision about where to eat dinner together in the next phase. Phase 2 consisted of going out to dinner together and discussing a family problem (i.e., daughter skipping school). Phase 3 involved planning a family vacation. In Phase 4, the family discussed the idea of
Th is
d oc
um en
t i s c
op yr
ig ht
ed b
y th
e A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. Th
is a
rti cl
e is
in te
nd ed
so le
ly fo
r t he
p er
so na
l u se
o f t
he in
di vi
du al
u se
r a nd
is n
ot to
b e
di ss
em in
at ed
b ro
ad ly
. A
ll rig
ht s,
in cl
ud in
g fo
r t ex
t a nd
d at
a m
in in
g, A
I t ra
in in
g, a
nd si
m ila
r t ec
hn ol
og ie
s, ar
e re
se rv
ed .
PARADOXICAL INTERVENTIONS 167
attending family therapy and related issues. In Phase 5, the family contacted their assigned therapist and attended a simulated therapy session. Phase 6 was the experimental manipulation phase (described below) in which the main independent variable was presented to partic- ipants individually, followed by administration of the experimental questionnaire containing the dependent measures.
Experimental manipulation. The main independent variable was the type of therapy intervention received, resulting in two experimen- tal conditions: the paradoxical directive (PD) condition and the non- paradoxical directive (NPD) condition. Participants in both groups were exposed to the intervention in the form of a typewritten, signed letter from their therapist that they received after the initial simulated therapy session. Half of the participants received each directive, ran- domly assigned within each therapist's pair of families.
The PD letter and NPD letter contained identical opening and clos- ing statements. The body of the PD letter contained a positive reframe of family arguing (i.e., fighting can be a way of expressing care and concern for one another and serves to maintain familial cohesion and communication) and a homework assignment recommending that the family purposefully argue with each other every day at a specified time during the upcoming week. The combined use of reframing and symp- tom-prescribing techniques is recommended by many practitioners of therapeutic paradox (e.g., Papp, 1983; Selvini Palazzoli et al., 1978; Weeks & L'Abate, 1982). The body of the NPD letter contained a similarly scheduled family homework assignment designed to mini- mize arguing and enhance communication. Participants in both con- ditions were instructed to read their letters carefully and then answer the enclosed questionnaire (containing the dependent measures) inde- pendently, without discussing the questions or their responses with their partners in the project.
Analysis
A 2 (Treatment) X 3 (Family Role) fixed effects, multivariate analysis of variance (M ANOVA) was used to examine the overall significance of interaction and main effects for the dependent measures. Univariate analysis of variance procedures were then performed to identify indi- vidual scale contributions to the significant multivariate Fs. An a < .05 level was used to determine statistical significance. Tukey's multiple comparison procedure was planned if necessary to aid in the interpre- tation of statistically significant findings for family role. Potential con- founding influences due to therapist differences and Therapist X Treatment interaction effects were controlled by random assignment of treatments within therapist pairs and standardized presentation of the two treatments through written letters and therefore were not ana- lyzed.
Results
The mean item ratings and standard deviations for each de- pendent variable, broken down by treatment and family role, are presented in Table 1. Results of the two-way M ANOVA revealed that the main effect for treatment was statistically sig- nificant, Wilks' lambda = .88, F(4, 87) = 3.06, p = .02. The main effect for family role was not statistically significant, Wilks' lambda = .89, F(8,174) = 1.35, p = .22. In addition, the Treatment X Family Role interaction effect was not statistically significant, Wilks' lambda = .92, F(8, 174) = 0.88, p = .54. Because the overall effects for family role and for the Treat- ment X Family Role interaction were not statistically signifi- cant, only the main effects for treatment (PD vs. NPD) were subsequently examined in the univariate analyses.
Results of the univariate analyses revealed no statistically
significant treatment effects on the three CRF-S subscales, therapist Attractiveness (F = 0.39, p = .54), Expertness (F = 2.48, p = . 12), and Trustworthiness (F = 0.16, p = .69). How- ever, the effect of treatment on the TEI—SF was statistically significant (F = 5.43, p = .02), suggesting that participants judged the paradoxical directive as less acceptable than the nonparadoxical directive. Results of the univariate analyses for treatment on each of the dependent variables are summarized in Table 2.
Discussion
The results suggest that paradoxical interventions used in a family therapy context do not necessarily diminish clients' per- ceptions of therapist attractiveness, expertness, and trust- worthiness. These findings are consistent with previous re- search examining the relation between paradoxical interven- tions and perceptions of the therapist (Conoley & Beard, 1984; Conoley & Garber, 1985; Feldman et al., 1982; Lopez & Wam- bach, 1982; Newton & Dowd, 1990; Perrin & Dowd, 1986). Despite widespread concern that paradoxical interventions, by their very nature, are apt to undermine the therapeutic rela- tionship, only one published study could be found that yielded empirical evidence supporting such a conclusion (McMillan & Johnson, 1990). Because it appears that a host of mediating variables play an important role in shaping clients' perceptions of therapists (Conoley & Beard, 1984; Feldman et al., 1982; Hills et al., 1985), perhaps subtle differences in the therapist's style or delivery or in the type of paradoxical intervention used in McMillan and Johnson's (1990) study can account for their conflicting findings.
On the other hand, the expectation that both the paradoxical directive and the nonparadoxical directive would receive simi- lar ratings of treatment acceptability was not supported by the data. Specifically, the paradoxical directive was judged less ac- ceptable than the nonparadoxical directive. This finding is consistent with previous research examining the consumer ac- ceptability of paradoxical interventions (Cavell et al., 1986; Mittl & Robin, 1987). In terms of practical significance, how- ever, the current results may be in contrast with those of the earlier studies. In general, researchers investigating treatment acceptability have defined acceptable interventions as those achieving mean scores greater than the midpoint on the accept- ability measure used (Cross Calvert & Johnston, 1990). Kelley et al. (1989) stated that a midpoint rating of 3 on TEI-SF items suggests "moderate" acceptability for a particular intervention. In the present study, the mean item rating on the TEI-SF for the paradoxical directive was above the midpoint (3.18), despite being significantly lower than the mean item rating for the nonparadoxical directive (3.47). Therefore, based on conven- tional formulas for determining a treatment's acceptability, the paradoxical directive was judged acceptable, although less ac- ceptable than the nonparadoxical directive. In the two previous studies by Cavell et al. (1986) and Mittl and Robin (1987), not only did paradoxical interventions receive lower-acceptability ratings than nonparadoxical interventions, but they also were judged as generally unacceptable.
Kolko and Milan (1986) offered several possible explanations for the low acceptability ratings ascribed to paradoxical inter-
Th is
d oc
um en
t i s c
op yr
ig ht
ed b
y th
e A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. Th
is a
rti cl
e is
in te
nd ed
so le
ly fo
r t he
p er
so na
l u se
o f t
he in
di vi
du al
u se
r a nd
is n
ot to
b e
di ss
em in
at ed
b ro
ad ly
. A
ll rig
ht s,
in cl
ud in
g fo
r t ex
t a nd
d at
a m
in in
g, A
I t ra
in in
g, a
nd si
m ila
r t ec
hn ol
og ie
s, ar
e re
se rv
ed .
168 GORDON R. BETTS AND RORY REMER
Table 1 Main Effects, Mean Item Ratings, and Standard Deviations for Dependent Measures by Treatment and Family Role
Dependent measure
CRF-S Attractiveness
M SD
Expertness M SD
Trustworthiness M SD
TEI-SF M SD
PD (« = 49")
6.07 1.12
5.78 0.90
5.99 0.79
3.18 0.66
NPD (n = 48)
5.94 0.91
5.47 0.99
5.93 0.88
3.47 0.51
Father
5.78 1.07
5.38 1.03
5.71 0.77
3.33 0.57
Mother (n = 33}
6.07 1.05
5.68 1.00
6.05 0.89
3.46 0.58
Daughter (n = 33")
6.14 0.93
5.80 0.80
6.11 0.79
3.18 0.64
Note. CRF-S = Counselor Rating Form—Short Version; TEI-SF = Treatment Evaluation Inventory- Short Form; PD = paradoxical directive; NPD = nonparadoxical directive. a n = 48 for the TEI-SF due to missing data. " n = 32 for the TEI-SF due to missing data.
ventions in the Cavell et al. (1986) study, some of which may be applicable to the Mittl and Robin (1987) study as well. Specifi- cally, they charged that Cavell et al.'s (1986) analogue methods placed insufficient emphasis on the context and severity of the actual situation on which the case vignettes were based and provided inadequate detail in describing how the paradoxical interventions were introduced. Questions about the integrity of the analogue methodology used by Mittl and Robin (1987) must also be raised in light of their own finding that one of the case vignettes containing the paradoxical intervention was judged by mental health professionals as being only "somewhat representative" of paradoxical techniques in general.
Unfortunately, direct comparisons between the acceptability ratings for paradoxical interventions found in this study and the two earlier studies are hindered by the use of different measures of treatment acceptability. However, the practical dif-
ferences in acceptability ratings (i.e., "acceptable" vs. "unaccept- able") might be explained by methodological differences (i.e., simulation vs. strict analogue methods). Perhaps physical expo- sure to the therapist and other contextual factors associated with therapy influence treatment acceptability ratings, as Kolko and Milan (1986) have suggested. If this is true, the po- tential benefit of simulation methods (i.e., closer approximation of actual therapy settings) is evident.
Although the combined findings of this study may seem somewhat contradictory, the fact that the mean acceptability rating for the paradoxical directive was still within the accept- able range might explain why the paradoxical directive evi- dently did not damage perceptions of the therapist. In other words, treatments may have to be judged as unacceptable be- fore they begin to have a negative influence on perceptions of therapists. In addition, actual contact and interaction with a
Table 2 Results of One Way Univariate Analyses of Variance (ANOVAs) for Treatment
Dependent variable Source df SS F
Therapist Attractiveness
Therapist Expertness
Therapist Trustworthiness
Treatment Acceptability
Treatment Error
Total
Treatment Error
Total
Treatment Error
Total
Treatment Error
Total
1 95 96
1 95 96
1 95 96
1 94 95
0.40 99.03 99.44
2.20 84.31 86.51
0.11 65.68 65.79
1.90 32.87 34.77
0.40 0.39 1.04
2.20 2.48 0.89
O . l i 0.16 0.69
1.90 5.43 0.35
.54
.12
.69
.02
Note. Only 96 observations were used in the ANOVA test on Treatment Acceptability due to one observa- tion with a missing value.
Th is
d oc
um en
t i s c
op yr
ig ht
ed b
y th
e A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. Th
is a
rti cl
e is
in te
nd ed
so le
ly fo
r t he
p er
so na
l u se
o f t
he in
di vi
du al
u se
r a nd
is n
ot to
b e
di ss
em in
at ed
b ro
ad ly
. A
ll rig
ht s,
in cl
ud in
g fo
r t ex
t a nd
d at
a m
in in
g, A
I t ra
in in
g, a
nd si
m ila
r t ec
hn ol
og ie
s, ar
e re
se rv
ed .
PARADOXICAL INTERVENTIONS 169
therapist, which is afforded when simulation methods are used, may be the primary basis for establishing and solidifying per- ceptions of the therapist, thereby diluting the influence of dif- ferential treatment acceptability ratings.
The role of treatment acceptability scores in ethical decision making is a complex issue. For instance, Cross Calvert and Johnston (1990) noted that effective treatments are not necessar- ily high on acceptability and vice versa. In addition, the lower ratings of treatment acceptability for paradoxical interventions evident in this study and previous research could be an artifact of the counterintuitive nature of paradoxical methods rather than a true indication that such methods are somehow unethi- cal. Cross Calvert and Johnston (1990) suggested that measures of treatment acceptability may include not only judgments about the perceived appropriateness and fairness of a treatment for a particular problem but also judgments about whether the treatment is sensible and concurs with popular notions of what constitutes treatment. Paradoxical interventions, by definition, are unanticipated by clients and are generally perceived as ab- surd or irrational because the symptom is prescribed, encour- aged, or endorsed in some fashion. In actual practice, clinicians have often mentioned the surprise, shock, fear, bewilderment, or confusion displayed by clients immediately after paradoxical interventions are administered (e.g., Weeks & JJVbate, 1982). Therefore, can paradoxical interventions be expected to achieve treatment acceptability ratings as high as nonparadoxi- cal interventions? They probably cannot. Perhaps treatment ac- ceptability ratings in the moderate range of acceptability are sufficient to warrant continued use of paradoxical interven- tions, especially in light of accumulating evidence that percep- tions of the therapist are not necessarily damaged when such interventions are used.
The present study was limited in several ways. Perhaps the most significant limitation involved the study's analogue de- sign, although efforts were made to improve the generalizabil- ity of the results to natural therapy settings by using simulation and role playing. Still, actual family therapy clients may have responded quite differently than did these participants, indi- cating a need for further research with such populations before any firm conclusions can be drawn. Another limitation of the study involved the selection of the research sample. Not only were the participants a fairly homogeneous group with respect to age, race, and marital status, but they also were enrolled in the same undergraduate family studies course. Although class lectures did not include information about specific family ther- apy techniques or paradoxical interventions, some of the mate- rial covered may have influenced participants' attitudes toward family arguing and, indirectly, their reactions to the paradoxi- cal directive used in this study. This potential confound under- scores the need for additional research with other populations.
Future research efforts might also address the fact that only one type of paradoxical intervention for a specific family prob- lem was examined in this study. Moreover, the intervention was presented in an unconventional manner (i.e., in a typewritten letter after only one session). Comparison of different types of paradoxical interventions with a variety of nonparadoxical in- terventions under more realistic conditions is needed.
The construct of treatment acceptability and its measure- ment also needs to be explored and validated further. Treat-
ment acceptability is generally assumed to be related to client attitudes and behaviors, treatment effectiveness, and satisfac- tion with treatment; however, such associations have received minimal empirical investigation (Cross Calvert & Johnston, 1990). More research in this area may help to determine the practical significance of acceptability scores (e.g., the point at which less acceptable ratings begin to have a significant influ- ence on other therapy process variables). The development and validation of treatment acceptability measures for adult popula- tions is also needed. Existing inventories were developed pri- marily to evaluate the acceptability of child behavioral treat- ments (e.g., Kazdin, 1980; Kelley et al, 1989; Witt & Martens, 1983).
Finally, even though no statistically significant differences were found for family role in this study, the use of simulation and role-playing methods in conducting therapy process re- search is worthy of further empirical investigation. Studies are needed that compare simulation methods with strict analogue methods to determine the specific benefits and liabilities of the two approaches. Such research may also help to resolve the controversy about whether treatment acceptability should be investigated using strict analogue methods, as has been recom- mended. Although these methods do an excellent job of isolat- ing treatments from other factors associated with therapy, ex- tensions are needed that more closely approximate natural ther- apy settings (e.g., simulation).
Although much debate regarding the ethical issues involved in using paradoxical interventions exists in the literature, the empirical examination of these issues remains relatively unex- plored. Further research is needed to increase our understand- ing of the impact of paradoxical interventions on the therapeu- tic process and of the underlying change mechanisms asso- ciated with these unorthodox methods. As additional information from controlled studies is produced, more knowl- edgeable and effective use of paradoxical interventions in clini- cal practice can be anticipated.
References
Barak, A., & LaCrosse, M. B. (1975). Multidimensional perception of counselor behavior. Journal of Counseling Psychology, 22, 471-476.
Brown, J. E., & Slee, P. T. (1986). Paradoxical strategies: The ethics of intervention. Professional Psychology: Research and Practice, 17, 487-491.
Cavell, T. A., Frentz, C. E., & Kelley, M. L. (1986). Acceptability of paradoxical interventions: Some nonparadoxical findings. Profes- sional Psychology: Research and Practice, 17, 519-523.
Conoley, C. W, & Beard, M. (1984). The effects of paradoxical inter- vention on therapeutic relationship measures. Psychotherapy, 21, 273-277.
Conoley, C. W, & Garber, R. A. (1985). Effects of reframing and self- control directives on loneliness, depression, and controllability. Journal of Counseling Psychology, 32, 139-142.
Corrigan, J. D., & Schmidt, L. D. (1983). Development and validation of revisions in the Counselor Rating Form. Journal of Counseling Psychology, 30, 64-75.
Cross Calvert, S., & Johnston, C. (1990). Acceptability of treatments for child behavior problems: Issues and implications for future research. Journal of Clinical Child Psychology, 19, 61-74.
DeBord, J. B. (1989). Paradoxical interventions: A review of the recent literature. Journal of Counseling and Development, 67, 394-398.
Th is
d oc
um en
t i s c
op yr
ig ht
ed b
y th
e A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. Th
is a
rti cl
e is
in te
nd ed
so le
ly fo
r t he
p er
so na
l u se
o f t
he in
di vi
du al
u se
r a nd
is n
ot to
b e
di ss
em in
at ed
b ro
ad ly
. A
ll rig
ht s,
in cl
ud in
g fo
r t ex
t a nd
d at
a m
in in
g, A
I t ra
in in
g, a
nd si
m ila
r t ec
hn ol
og ie
s, ar
e re
se rv
ed .
170 GORDON R. BETTS AND RORY REMER
Doherty, W J. (1989). Unmasking family therapy. The Family Therapy Networker, 13. 35-41.
Dowd, E. T., & Milne, C. R. (1986). Paradoxical interventions in coun- seling psychology. The Counseling Psychologist, 14, 237-282.
Feldman, D. A., Strong, S. R., & Danser, D. B. (1982). A comparison of paradoxical and nonparadoxical interpretations and directives. Journal of Counseling Psychology, 29, 572-579.
Haley, J. (1963). Strategies of psychotherapy. New York: Grime & Strat- ton.
Haley, J. (1976). Problem-solving therapy. San Francisco: Josey-Bass. Haley, J. (1987). Problem-solving therapy (2nd ed.). San Francisco: Jo-
sey-Bass. Heppner, P. P., & Claiborn, C. D. (1989). Social influence research in
counseling: A review and critique. Journal of Counseling Psychology, 36, 365-387.
Hill, K. A. (1987). Meta-analysis of paradoxical interventions. Psycho- therapy. 24, 266-270.
Hills, H. I., Gruszkos, J. R., & Strong, S. R. (1985). Attribution and the double bind in paradoxical interventions. Psvchotherapy, 22, 779- 785.
Hunsley, J. (1988). Conceptions and misconceptions about the context of paradoxical therapy. Professional Psychology: Research and Prac- tice, 19, 553-559.
Johnson, M. (1986). Paradoxical interventions: From repugnance to cautious curiosity. The Counseling Psychologist, 14, 297-302.
Katz, J. (1984). Symptom prescription: A review of the clinical out- come literature. Clinical Psychology Review, 4, 703-717.
Kazdin, A. E. (1980). Acceptability of alternative treatments for de- viant child behavior. Journal of Applied Behavior Analysis, 13, 259- 273.
Kazdin, A. E. (1981). Acceptability of child treatment techniques: The influence of treatment efficacy and adverse side effects. Behavior Therapy, 12, 493-506.
Kelley, M. L., Heffer, R. W, Gresham, F. M., & Elliot, S. N. (1989). Development of a modified Treatment Evaluation Inventory. Jour- nal of Psychopathology and Behavioral Assessment, 11, 235-247.
Kolko, D. J., & Milan, M. A. (1986). Acceptability of paradoxical inter- ventions: Some paradoxes of psychotherapy research. Professional Psychology: Research and Practice, 17, 524-527.
Lopez, F. G., & Wambach, C. A. (1982). Effects of paradoxical and self-control directives in counseling. Journal of Counseling Psychol- ogy, 29, 115-124.
Mahoney, M. J. (1986). Paradoxical intention, symptom prescription, and principles of therapeutic change. The Counseling Psychologist, 14, 283-290.
McMillan, D, & Johnson, M. E. (1990). Paradoxical versus cognitive-
behavioral interventions: Effects on perceptions of counselor char- acteristics. Journal of Mental Health Counseling, 12, 67-75.
Mittl, V F, & Robin, A. (1987). Acceptability for alternative interven- tions for parent-adolescent conflict. Behavioral Assessment, 9,417- 428.
Newton, G. R., & Dowd, E. T. (1990). Effect of client sense of humor and paradoxical interventions on test anxiety. Journal of Counseling and Development, 68, 668-672.
Papp, P. (1980). The Greek chorus and other techniques of paradoxical therapy. Family Process, 19, 45-57.
Papp, P. (1983). The process of change. New York: Guilford Press. Perrin, D. K., & Dowd, E. T. (1986). Effect of paradoxical and nonpara-
doxical self-disclosure on counselor social influence. Journal of Counseling Psychology, 33, 207-210.
Ridley, C. R., & Tan, S. Y. (1986). Unintentional paradoxes and poten- tial pitfalls of paradoxical psychotherapy. The Counseling Psycholo- gist, 14, 303-308.
Schmidt, L. D. (1986). Some questions about paradoxical interven- tions. The Counseling Psychologist, 14, 309-3! 1.
Schotte, D. E., Ascher, L. M.,&Cools, J.(1989).Theuseofparadoxical intention in behavior therapy. In L. M. Ascher (Ed.), Therapeutic paradox (pp. 17-31). New York: Guilford Press.
Seltzer, L. (1986). Paradoxical strategies in psychotherapy: A compre- hensive overview and guidebook. New York: Wiley.
Selvini Palazzoli, M., Cecchin, G.. Praia, G., & Boscolo, L. (1978). Paradox and counterparadox. Northvale, NJ: Aronson.
Shoham-Salomon, Y, & Rosenthal, R. (1987). Paradoxical interven- tions: A meta-analysis. Journal of Consulting and Clinical Psychol- ogy, 55, 22-28.
Strong, S. R. (1984). Experimental studies in explicitly paradoxical interventions: Results and implications. Journal of Behavior Therapy and Experimental Psychiatry, 15, 189-194.
Watzlawick, P., Weakland, J. H., & Fisch, R. (1974). Change: Principles of problem formation and resolution. New "York: Norton.
Weeks, G. R., & L'Abate, L. (1982). Paradoxical psychotherapy: Theory and practice with individuals, couples, and families. New York: Brun- ner/Mazel.
Whan, M. (1983). Tricks of the trade: Questionable theory and prac- tice in family therapy. British Journal of Social Work, 13, 321-337.
Witt, J. C, & Martens, B. K. (1983). Assessing the acceptability of behavioral interventions used in the classroom. Psychology in the Schools, 20, 510-517.
Received June 8,1992 Revision received August 28,1992
Accepted September 8,1992 •
Th is
d oc
um en
t i s c
op yr
ig ht
ed b
y th
e A m
er ic
an P
sy ch
ol og
ic al
A ss
oc ia
tio n
or o
ne o
f i ts
a lli
ed p
ub lis
he rs
. Th
is a
rti cl
e is
in te
nd ed
so le
ly fo
r t he
p er
so na
l u se
o f t
he in
di vi
du al
u se
r a nd
is n
ot to
b e
di ss
em in
at ed
b ro
ad ly
. A
ll rig
ht s,
in cl
ud in
g fo
r t ex
t a nd
d at
a m
in in
g, A
I t ra
in in
g, a
nd si
m ila
r t ec
hn ol
og ie
s, ar
e re
se rv
ed .
