Why, Why, Why?

: Reason-Giving and Rumination

as Predictors of Response to Activation- and
Insight-Oriented Treatment Rationales
Ä

Michael E. Addis and Kelly M. Carpenter

Clark University

This study examines the relationships among the reasons a person offers
for depression, the tendency to ruminate in response to depression, and
reactions to activation-oriented (AO) or insight-oriented (IO) treatment ratio-
nales. Adults from the community (N 5 51) completed self-report mea-
sures of reason-giving and rumination and rated the credibility of, and
personal reactions to, AO and IO rationales presented in written and video-
tape formats. Participants who gave more reasons for depression also
tended to ruminate more in response to depressed mood. Reason-giving
and rumination predicted lower credibility ratings and more negative per-
sonal reactions to the AO rationale. Although no relationship was found
between these variables and response to the IO rationale, specific reasons
were associated with different reactions to the two rationales. We discuss
the roles of reason-giving and rumination in predicting responses to psy-
chotherapies for depression. © 1999 John Wiley & Sons, Inc. J Clin
Psychol 55: 881–894, 1999.

Psychotherapists typically assume that clients who agree with the rationale underlying an
intervention will have better outcomes than those who disagree. This clinical observation
has been empirically supported in areas as diverse as depression (Addis & Jacobson,
1996; Fennel & Teasdale, 1987; Ilardi & Craighead, 1994), pain management (Spence &
Sharpe, 1993), and medical treatment (Becker et al., 1977; Centers for Disease Control,
1990; DeGood, 1983). What is less clear is exactly why different clients agree or disagree

This study was supported by a Faculty Development Award from Clark University awarded to Michael E.
Addis.
The authors would like to thank Mary Davis, William Doyle, Sandra Fulton, Christina Hatgis, and Aaron
Krasnow for their help in conducting this study.
Correspondence regarding this article should be addressed to Michael E. Addis, Ph.D., Department of Psychol-
ogy, Clark University, 950 Main St., Worcester, MA 01610.

JOURNAL OF CLINICAL PSYCHOLOGY, Vol. 55(7), 881–894 (1999)

© 1999 John Wiley & Sons, Inc. CCC 0021-9762/99/070881-14
882 Journal of Clinical Psychology, July 1999

with different treatment rationales. Studies addressing this individual difference question
have clinical implications for treatment matching (e.g., Beutler et al., 1991) and can help
to elucidate the effective ingredients in different psychotherapies (Shoham-Salomon &
Hannah, 1991).
Because a treatment rationale includes a model of the etiology of a problem and a
model of change, numerous studies have examined the role of clients’ preexisting etio-
logical beliefs and attributions in determining receptiveness to alternative rationales. For
example, Tracey (1988) found that congruence between therapist and client attribution of
blame for a problem predicted more positive psychotherapy outcomes. Barber and Stol-
tenberg (1994) found that students with an internal emotional locus of control preferred a
description of a cognitive therapy workshop, whereas those with an external locus pre-
ferred a behavioral one. Other studies have focused on similarities in etiological beliefs
between therapist and client (Atkinson, Worthington, Dana, & Good, 1991; Worthington
& Atkinson, 1996), attributional biases among therapists (Baston, Jones, & Cochran,
1979; Murdock & Fremont, 1989; Plous & Zimbardo, 1986), and cross-cultural differ-
ences in causal attributions (e.g., Narikiyo & Kameoka, 1992).

Reason-Giving Versus Causal Attributions and Etiological Beliefs

The body of literature on causal attributions and etiological beliefs has identified the
importance of a match between individual clients’ and therapists’ explanatory frame-
works. However, none of these studies have examined explanation, or reason-giving, as a
socially situated action rather than a reflection of underlying beliefs (Antaki, 1994; Hayes,
1989; Skinner, 1974). Because psychotherapy can also be understood as a socially sanc-
tioned process of interpersonal influence, it is worth considering how reason-giving may
influence responses to different treatment models.
The goal of the present study is to replicate and extend previous research demon-
strating a link between reason-giving and response to different treatments for depression.
In a previous study, Addis and Jacobson (1996) found that the total number of reasons a
client endorses (regardless of the content of specific reasons) predicts response to a behav-
ior change treatment for depression. Clients endorsing more reasons on the Reasons for
Depression Questionnaire (RFD; Addis, Truax, & Jacobson, 1995) had higher depression
scores following treatment, even when controlling for pretreatment levels of depression.
The goal of the present study was to examine specifically the relationship between reason-
giving and response to behavior change versus insight-oriented treatment rationales. We
first review the theoretical and empirical support linking reason-giving to different treat-
ment rationales and then describe the specific questions addressed in this study.

Reason-Giving and Response to Different Treatment Rationales

Why would reason-giving be related to response to different treatment rationales? The
answer lies in examining more broadly how giving reasons functions in changing our
own and others’ behavior. First, giving reasons often serves the function of excusing our
behavior, resisting inducements to change, or both. When a mother asks a child why he or
she spilled the milk, the child offers an explanation that (hopefully from the child’s point
of view) reduces the likelihood of any further negative consequences delivered by the
mother (Anderson, Hawkins, & Scotti, 1997; Hayes, 1989; Skinner, 1974). When a crit-
ical husband asks his wife, “Why are you depressed?” her stated reasons may justify her
behavior and function to decrease the likelihood of future critical comments from her
husband (see Antaki, 1994, for numerous examples of similar processes occurring in
 Reason-Giving and Treatment Rationales 883

natural dialogue). In psychotherapy, reason-giving may serve to avoid a therapist’s induce-

ments to change a client’s behavior (Addis & Jacobson, 1996). To the degree that a client
enters treatment offering multiple reasons for problems, he or she may be less likely to
collaborate with a therapist on changing specific behaviors. Because different treatments
vary in their degree of emphasis on behavioral change, reason-giving may be an impor-
tant individual difference variable in determining responses to specific rationales.
Second, reason-giving may be a private attempt, a public attempt, or both to solve a
problem by discovering its causes. We all have long learning histories in which stating
reasons for problems (to ourselves or to others) helps to identify what needs to be changed.
If our car won’t start, we can’t solve the problem until we know why it isn’t starting.
However, the process may be less productive when it comes to human experiences such
as depression and anxiety.1 Drawing on a history of reason-giving as problem solving can
lead to rumination, or continuous representation 2 of problems as opposed to solutions.
Lyubomirsky and Nolen-Hoeksema (1993) found that experimentally induced rumina-
tion led participants to report less likelihood of engaging in pleasant activities. Partici-
pants felt it would interfere with their efforts to understand themselves (see also Hayes,
1989, and Hayes & Wilson, 1993, for a thorough discussion of the role reason-giving may
play in maintaining behavioral problems). Thus, an additional purpose of this study was
to examine the associations between (a) reason-giving and the tendency to ruminate and
(b) rumination and response to a treatment rationale emphasizing the importance of under-
standing the causes of depression.
To summarize, reason-giving may predict response to different psychotherapies in
ways not addressed by previous research on causal attributions and beliefs. Because it
can serve to justify or exonerate behavior, reason-giving should predict a negative response
to treatment rationales emphasizing a client’s role in changing behavior. Second, if reason-
giving reflects a tendency to try to solve problems by discovering causes, it should pre-
dict negative reactions to change-oriented treatment rationales and positive reactions to
those focusing on insight as a change mechanism. In the present study we sought to
replicate and extend these hypotheses by examining the relationships among reason-
giving, rumination, and response to activation-oriented (AO) versus insight-oriented (IO)
treatment rationales.
We used an analogue design to evaluate reactions to treatment rationales presented in
video and written format. Although the study did not involve actual psychotherapy clients
and treatments, there is consistent evidence that acceptance of particular treatment ratio-
nales is associated with positive treatment outcomes (Addis & Jacobson, 1996; Fennel &
Teasdale, 1987; Spence & Sharpe, 1993). Thus, variables affecting reactions to rationales
are worth investigating in their own right (e.g., Rokke, Carter, Rehm, & Veltum, 1990).
We were interested in exploring two types of reactions to each rationale. The first was the
perceived credibility of a rationale as presented in written form. The second was an
individual’s personal reaction to the same rationale presented in a video format. Although
we did not have a firm basis from which to predict differences between the two types of

1
There are obviously times when it can be helpful to determine the causes of emotional and behavioral prob-
lems. However, it is not always necessary. Regardless of why one is depressed, engaging in pleasant activities,
exercising, combating pessimistic thinking, or taking medication all can be helpful. Nonetheless, most of us
continue to assume that it is necessary to delve deeply into the causes of distress before treatment can effective.
Presumably, this tendency is due to a strong history of reason-giving being functional in solving problems in the
physical world, and our culture’s tendency to reinforce offering private events (e.g., thoughts and feelings) as
causes of behavior. Hayes (1988) and Hayes and Wilson (1993) have discussed this issue in more detail.
2
We are referring here not to symbolic representation as a cognitive process but to re-presentation: the act of
continuing to think and talk (privately or publicly) about distressing events and their causes.
884 Journal of Clinical Psychology, July 1999

ratings, we considered it important to test for possible differences rather than assuming a
priori that the reactions would be similar. Written descriptions emphasize the abstract
elements of a rationale absent of an interpersonal context. Video presentations show the
same treatment in action. Moreover, the degree to which a treatment is perceived as
credible in the abstract may or may not be related to a client’s perception that the same
treatment would be personally helpful. For example, a person might believe that medi-
cation is an effective treatment for depression in general (credibility), and also doubt that
it will work for himself or herself (personal reaction).

Method

Participants

Fifty-one adults (30 women, 21 men) were recruited from the community to participate in
this study. They ranged in age from 20 to 65 (M 5 46, SD 5 11.5 ). Ninety-four percent
of the sample were Caucasian. Thirty-three percent of the sample were married, 49%
were single or divorced, and 18% had been divorced and remarried. Sixty-nine percent of
the participants had seen a psychotherapist in their lifetime, and 31% had seen one in the
past year. Participants were recruited through advertisements in the classified section of
several local papers. The advertisements stated that a faculty member of a nearby uni-
versity was conducting a study of personality and coping strategies and that individuals
would be paid $20 for completing a set of questionnaires and an interview.

Procedure

A graduate student research assistant or the first author introduced each participant to the
study. Participants then completed the questionnaires listed later in the article. Following
completion of the questionnaires, each participant watched two 10-minute videotaped
therapy vignettes (described later). After each tape was viewed, participants rated their
reactions to the treatment rationales presented in the tape.

Materials

Written Treatment Descriptions. We created written descriptions of AO and IO

approaches to treating depression. The descriptions were based on those used in previous
research assessing the credibility of alternative treatments for depression (Rokke et al.,
1990; Rokke & Scogin, 1995). Participants were asked to read the descriptions of each
treatment and answer seven questions (described later). The AO treatment description
stated that people feel the way they do because of the types of activities or behaviors they
engage in. Treatment for depression was described as focusing on increasing pleasurable
activities and solving specific problems. The role of homework assignments and a con-
crete in-session focus was also discussed. The IO description stated that people become
depressed because of thoughts and feelings outside of their awareness and that these
feelings and conflicts are typically caused by childhood experiences. The treatment was
described as not focusing on changing specific behaviors, but on helping a person to gain
insight into his or her unconscious thoughts and feelings.3

3
Copies of the treatment descriptions are available from the authors.
 Reason-Giving and Treatment Rationales 885

Video Treatment Vignettes. Vignettes of a therapist presenting an AO and an IO

treatment rationale were created for participants to watch and respond to. The first author
served as the therapist in each vignette. An actor or actress played the client in both
treatments. Thus, there were four videotaped vignettes with either a man or woman client
and an AO or IO rationale. Women participants saw the actress in both the AO and the IO
treatment. Men participants saw the actor in both treatments. We chose to match the sex
of the participant and the client to maximize the likelihood that the former would identify
with the latter. The actor and actress were asked to play the role indicated in a written
paragraph (presented subsequently). They were instructed to appear depressed and to
describe their problems briefly, but to let the therapist have the floor when presenting the
rationale or making interventions. Finally, they were told not to endorse either treatment
in any way (e.g., head nodding, or saying, “That sounds really good.”).
The order of presentation of the treatments was randomized. Each tape showed a
split-screen image with head and torso profiles of the therapist on the left and the client
on the right. Each tape lasted approximately 10 minutes (M 5 10.4 minutes, SD 5 49
seconds). Several steps were taken to insure that that the vignettes differed only with
respect to the treatment rationale and the type of interventions utilized by the therapist.
The particular issues discussed were matched in each tape. Prior to viewing the first tape,
participants were given a brief written paragraph and told that it would give them some
background information on the client they were about to see. Women participants were
given the following paragraph:

This client is a married woman in her mid-thirties who has sought psychotherapy for feelings of depres-
sion. She feels deeply sad most days, has difficulty sleeping, difficulty remembering things, and is tired
most of the day. The client has recently moved to Massachusetts from the Midwest where she was born,
and where her family and friends still live. She recently started a new job that she finds challenging and
enjoyable, although at times the stress is overwhelming. Her primary difficulty at work is that she often
feels incompetent and is very self-critical of her performance. Her self-critical tendencies continue despite
consistent positive feedback from her employer.

Men participants were given an identical paragraph describing a male client. These par-
ticular issues were chosen because they represent common and credible life events asso-
ciated with depression.
Participants were asked to imagine that they were experiencing the type of therapy
described in each tape and to think about how helpful it would be for them if they were
feeling depressed. On each tape, the therapist spent approximately the first 2 minutes
describing an AO or an IO treatment rationale after which the image faded out.4 The
therapist then spent approximately 4 minutes focusing on symptoms of depression related
to the recent move and missing friends and family, followed by a fade-out. The final 4
minutes were spent discussing the client’s feelings of incompetence at work and per-
ceived criticism from his or her employer. The latter two segments were intended to show
how the different treatment rationales would be implemented in concrete interventions.
In the AO treatment, the therapist approached the issue of moving to a new city and
missing family by suggesting that the client increase pleasurable activities, exercise, and
phone extended family more frequently. The difficulties at work were approached by
focusing on concrete ways to decrease stress and manage time more effectively. In the IO

4
The rationales were virtually identical to those presented in the written treatment descriptions. The AO ratio-
nale focused on the importance of increasing pleasurable activities, solving problems, and getting more active
as an effective way to combat depression. The IO rationale emphasized the importance of understanding the
unconscious causes of depression, the specific feelings involved, and possible past events that may be contrib-
uting to current feelings.
886 Journal of Clinical Psychology, July 1999

treatment, the therapist interpreted the client’s feelings of abandonment by and anger at
the family as causes of the depression. With regard to the conflict at work, the therapist
suggested that perhaps the client perceived his or her boss similarly to someone in the
past who had been critical, perhaps the client’s father. The therapist then suggested that
awareness of this pattern could be helpful in combating feelings of depression.
The mean amount of time talking for the client was 3.4 minutes in the AO and 4.3
minutes in the IO treatment. The mean amount of time talking for the therapist was 6.3
minutes in the AO treatment and 4.7 minutes in the IO treatment. The fact that therapists
spoke more in the AO treatment and clients spoke more in the IO treatment is consistent
with each theoretical approach (i.e., behavior change therapies tend to be more therapist
directed, insight-oriented therapies tend to be more client directed).

Measures
Reasons for Depression Questionnaire. The Reasons for Depression Questionnaire
(RFD; Addis et al., 1995) is a self-report measure of explanations for the causes of
depression. Respondents rate the degree to which different reasons are related to their
depression on a 5-point Likert scale. Factor analytically derived subscales are created by
summing the ratings of individual items (mean number of items per scale 5 5.5, range 5
3 to 10) and dividing by the number of items per scale. The subscales include character-
ological, existential, interpersonal conflict, intimacy, achievement, relationship, physi-
cal, and childhood reasons for depression. Psychometric data suggest that specific RFD
scales are internally consistent and moderately correlated with current depression, with
the association being stronger in a nonclinically selected sample than a depressed sample.
Specific subscales are also related to functioning in corresponding areas of life (Addis
et al., 1995). The total RFD score (summing across subscales) has been shown to predict
response to different psychotherapies for depression (Addis & Jacobson, 1996). Because
the focus of the current study was on reason-giving in general, the majority of analyses
involved total RFD scores. The total score demonstrated high internal consistency (Cron-
bach’s alpha 5 .96) in the current sample.

Beck Depression Inventory. The Beck Depression Inventory (BDI; Beck, Ward, Men-
delson, Mock, & Erbaugh, 1961) is a widely used self-report measure of current symp-
toms of depression. Scores range from 0 to 63, with scores of less than 10 thought to
reflect an absence of clinical depression. Psychometric studies of the BDI indicate accept-
able levels of reliability and concurrent validity (Beck, Steer, & Garbin, 1988).

Response Styles Questionnaire. The Response Styles Questionnaire (RSQ; Nolen-

Hoeksema & Morrow, 1991) is a 33-item self-report measure of the tendency to ruminate
or become distracted in response to depressed mood. Twenty-two items focus on the
tendency to ruminate and 11 items focus on distraction as a coping strategy. This measure
has been used in a number of studies investigating variations in response to depressed
mood (e.g., Nolen-Hoeksema & Morrow, 1991; Nolen-Hoeksema, Parker, & Larson,
1994). Psychometric analyses indicate high internal consistency (Cronbach’s alpha 5
.90) and acceptable convergent and predictive validity (Nolen-Hoeksema & Morrow,
1991).

Credibility Scale. After reading each written treatment description, participants rated
the following questions on a 7-point scale from 1 (not at all ) to 7 (extremely): How
logical does this therapy seem to you? How scientific does this therapy seem to you?
 Reason-Giving and Treatment Rationales 887

How complete does this therapy seem to you? In other words, do you think this therapy
covers all the types of people who become depressed? To what extent would this therapy
help an individual in other areas of his or her life? How likely would you be to go into this
therapy if you were depressed? How effective do you think this therapy would be for
most people? If a close friend or relative were depressed, would you recommend this
therapy to them? These seven items, adapted from Borkovec and Nau (1972), have dem-
onstrated good internal consistency and have been used to assess young adults’ percep-
tions of the credibility of different treatment for depression (Rokke et al., 1990).

Personal Reactions to the Rationales (PRR). After watching each treatment vignette,
participants answered the following five questions on a 7-point scale from 1 (not at all )
to 7 (extremely): If you were depressed and went to see a therapist, how helpful do you
think this therapy would be for you? To what extent do you think that this therapy would
help you to understand the causes of your depression? To what extent do you think that
this therapy would help you learn effective ways to cope with feeling depressed? If you
were to seek therapy for depression, how likely would you be to choose this type of
therapy? If you were to try this type of therapy, how effective would it be in treating your
depression? In contrast to the credibility items for the written descriptions, these items
were chosen for their more personal focus (i.e., do you think this treatment will be effec-
tive for you?)

Results
Individual item ratings on the PRR were highly intercorrelated within the AO (mean r 5
.75, a 5 .94) and IO (mean r 5 .78, a 5 .95) treatments. Due to the high degree of
intercorrelation among the items we chose to sum the individual item ratings for each
video to create single PRR scores for the AO and IO rationales. These scores had a
possible range of 5 to 35 (mean AO 5 24.1, SD 5 7.0, mean IO 5 22.0, SD 5 7.1) with
higher scores indicating more positive reactions.
Total ratings for the credibility items following the written descriptions of the AO
(mean r 5 .49, a 5 .87) and IO (mean r 5 .65, a 5 .93) rationales were computed in the
same fashion. These scores had a possible range of 7 to 49 (mean AO 5 29.0, SD 5 7.5,
mean IO 5 31.6, SD 5 8.0) with higher scores indicating higher credibility ratings. Cred-
ibility and PRR ratings of the same rationale were moderately to highly correlated
(r(AO) 5 .53, p , .001, r(IO) 5 .62, p , .001). The correlations between credibility and
PRR ratings of contrasting rationales were less than .20 and nonsignificant.

Order Effects and Comparisons of the Rationales

We conducted a paired t-test to test for possible order of presentation effects on PRR
ratings of each video vignette. The results indicated significantly higher average scores
for the second vignette (t(50) 5 2.82, p , .01). Comparisons of differences between AO
and IO ratings depending on order of treatment presentation revealed that PRR scores
were higher for the AO vignette when it was presented second (t(49) 5 2.1, p 5 .03). PRR
ratings for the IO vignette did not differ depending on presentation order. The order of
presentation of the vignettes was dummy-coded and treated as a covariate in the remain-
ing analyses involving PRR ratings.
Table 1 shows the means and standard deviations for all variables. The mean BDI
score of 10.4 (SD 5 8.7, range 5 0– 42) suggests that the sample as a whole was not
888 Journal of Clinical Psychology, July 1999

Table 1
Means, Standard Deviations, and Zero-Order Correlations for All Variables

RFD RSQ-R BDI PRR-AO PRR-IO Cred-AO Cred-IO

M (SD) 18.7 (6.7) 43.1 (10.5) 10.4 (8.7) 24.1 (7.0) 22.0 (7.1) 29.0 (7.5) 31.6 (8.1)
RFD — — — — — — —
RSQ-R .65** — — — — — —
BDI .76** .62** — — — — —
PRR-AO 2.29* 2.31* 2.07 — — — —
PRR-IO .13 .07 .12 2.18 — — —
Cred-AO 2.14 2.11 .12 .53** .04 — —
Cred-IO .05 .15 .08 2.08 .62** .03 —

Note. RFD 5 Reasons for Depression total score; RSQ-R 5 Response Styles Questionnaire 2 Rumination Subscale; BDI 5
Beck Depression Inventory; AO 5 Activation Oriented; IO 5 Insight Oriented; PRR 5 personal reaction ratings of video
vignettes; Cred 5 credibility ratings of written treatment descriptions.
*p , .05. **p , .01.

clinically depressed but endorsed more distress than would be characteristic of a strictly
normal population. Scores of 10–17 are thought to reflect dysphoria in nonclinical pop-
ulations (Kendall, Hollon, Beck, Hammen, & Ingram, 1987).5 Forty-three percent of the
sample had BDI scores greater than 9.
A repeated measures analysis of covariance (order of treatment presentation serving
as a covariate) indicated that PRR ratings on the AO vignette (M 5 24.1, SD 5 7.0), were
slightly higher than the IO vignette (M 5 21.2, SD 5 7.1, F(1,49) 5 4.3, p 5 .04). A
within-subjects t-test revealed no difference between the credibility ratings of the written
AO and IO descriptions (t(50) 5 1.76, ns).

Relationships Between Demographic Variables and Treatment Ratings

We calculated zero-order correlations between demographic variables (age, gender, mar-

ital status, current experience in therapy, past experience in therapy, and the number of
separate times in therapy) and PRR and credibility ratings of the AO and IO rationales.
Because we had no a priori hypotheses regarding the associations between these vari-
ables, we adjusted the alpha level to .01 to control for Type I errors. We found no signif-
icant correlations between any of the demographic variables and personal reactions or
credibility ratings of the rationales.

Reason-Giving, Rumination, and Response to Treatment Rationales

Table 1 also shows the zero-order correlations among all measures. We had predicted that
RFD would be positively correlated with the rumination subscale of the RSQ. As Table 1
indicates, our prediction was supported (r(49) 5 .65, p , .01). Table 1 also indicates that
scores on the BDI were highly correlated with the RFD (r(49) 5 .76, p , .01) and the
RSQ (r(49) 5 .62, p , .01), although BDI scores were uncorrelated with any of the
ratings of the two treatment rationales. Although it is common for reason-giving and

5
There were no significant differences between men and women on BDI scores, RSQ-R scores, or credibility
and PRR ratings of the AO and IO rationales.
 Reason-Giving and Treatment Rationales 889

rumination to be associated with depressed mood in a nonclinically selected sample (Addis

et al., 1995), we wanted to see whether or not the relationship between these two vari-
ables was completely accounted for by current depressed mood. Partialling out the BDI
resulted in a smaller, but significant, correlation (r(49) 5 .36, p , .01) between the RFD
and the RSQ. Thus, the relationship between the RFD and the RSQ was not completely
accounted for by current depressed mood.
We performed a series of hierarchical multiple regression analyses to determine
whether RFD and RSQ-R scores predicted PRR and credibility ratings for the two ratio-
nales. PRR or credibility ratings served as the criterion in each analysis. The order of
treatment presentation was entered first in all analyses involving PRR ratings as a crite-
rion. BDI scores were also entered first in all analyses to determine the unique contribu-
tion of reason-giving and rumination to predicting responses to the rationales. As Table 2
indicates, both RFD and RSQ-R scores were significant predictors of responses to the AO
rationale, accounting for between 6% and 13% of the variance in PRR and credibility
ratings. Participants scoring higher on these measures tended to have more negative
reactions to the AO rationale and to rate it as less credible. We found no significant
relationships between the RFD and RSQ-R scores and response to the IO rationales.
We conducted an additional series of hierarchical regressions to test for mediational
effects between reason-giving or rumination and response to the rationales. Table 2 shows
that RFD scores continued to be a significant predictor of lower PRR and credibility
ratings for the AO rationale, even when controlling for RSQ-R scores. Scores on the
RSQ-R did not mediate the relationship between the RFD and response to the AO ratio-
nales. The relationship between RSQ-R scores and PRR ratings of the AO rationales
became nonsignificant when RFD scores were entered first into the equation. However,
the difference in partial correlations when controlling (r(48) 5 2.22) versus not con-
trolling for RFD scores (r(48) 5 2.30) was nonsignificant (Z(difference) 5 .43, ns).
Table 2 shows that similar results were obtained for credibility ratings. Although the

Table 2
Partial Correlations, Beta Weights, and Percentage of Variance Accounted for
by Regressions of Treatment Rationale Ratings on RFD and RSQ-R Scores

Activation Oriented Insight Oriented

Predictor a Partial r b R 2b Partial r b R2

PRR Ratings
RFD 2.31* 2.48 .09 2.01 .02 .00
RSQ-R 2.30* 2.37 .08 2.06 2.07 .00
RFD (controlling for RSQ-R) 2.24* 2.37 .05 2.03 .05 .00
RSQ-R (controlling for RFD) 2.22 2.27 .04 2.07 2.09 .01
Credibility Ratings
RFD 2.35** 2.54 .13 2.02 2.03 .00
RSQ-R 2.24* 2.31 .06 .13 .16 .01
RFD (controlling for RSQ-R) 2.29* 2.46 .08 2.07 2.11 .00
RSQ-R (controlling for RFD) 2.13 2.17 .02 .14 .20 .01

Note. RFD 5 Reasons for Depression Total Score; RSQ-R 5 Response Styles Questionnaire 2 Rumination Subscale.
a
All regression equations with PRR ratings as the criteria have the order of treatment presentation and BDI scores entered prior
to the variables in the table. Equations with credibility ratings as criteria have BDI scores entered first.
b 2
R represents the proportion of variance for each variable beyond that accounted for by others in the equation.
*p , .05. **p , .01.
890 Journal of Clinical Psychology, July 1999

relationship between RSQ-R scores and credibility ratings became nonsignificant when
controlling for RFD scores, the difference in partial correlations was nonsignificant
(Z(difference) 5 .57, ns). Thus, we found no mediational effects between the RFD, the
RSQ-R, and responses to the treatment rationales.
Table 3 shows the partial correlations between specific RFD subscales and reactions
to the two rationales controlling for BDI scores. The results show that although reason-
giving in general was not associated with response to the IO rationale (Table 2), specific
subscales were associated with credibility and PRR ratings of this rationale. Childhood
(r(49) 5 .27, p , .05) and characterological (r(49) 5 .31, p , .05) reasons were asso-
ciated with more positive PRR ratings of the IO rationale. These same subscales pre-
dicted more negative reactions to the AO rationale. Using Fisher’s r to z transformation,
the differences between the correlations in the AO and IO conditions were all significant
( p , .01) with the exception of the relationship between characterological reasons and
credibility ratings ( p 5 .08, two-tailed).

Discussion
The purpose of this study was to replicate and extend previous research demonstrating a
link between reason-giving and response to different treatment rationales. Consistent
with previous research (Addis & Jacobson, 1996) we found that people who give more
reasons and ruminate in response to depressed mood respond more negatively to a treat-
ment rationale that emphasizes a client’s role in changing his or her behavior as a means
of combating depression. Before exploring the results further, we would like to address
some limitations of this study.
A possible source of bias is potential unmeasured differences between the activation-
and insight-oriented video tapes. Although credibility and personal reaction ratings were
similar in the two conditions, it is possible that variation along other dimensions influ-
enced participants’ ratings.
The study was not conducted with a clinically depressed population, and 94% of the
sample were Caucasian. It is possible that the findings will not generalize to the range of
depressed persons seeking psychotherapy. Similarly, the analogue nature of the treatment

Table 3
Partial Correlations Between RFD Subscales and Credibility and PRR Ratings of the Rationales

Activation Oriented Insight Oriented

RFD Subscale Credibility PRR Credibility PRR

Characterological 2.08 2.25* .16 .27*

Conflict 2.07 2.14 2.02 .00
Intimacy 2.17 2.28* 2.10 2.11
Achievement 2.10 2.26* .10 .06
Childhood 2.21 2.32* .30* .31*
Relationship 2.25* 2.16 2.16 2.20
Physical .07 .08 2.09 2.08
Existential 2.09 2.04 .12 .07

Note. All correlations have BDI scores partialled out. RFD 5 Reasons for Depression Questionnaire; Conflict 5 Interpersonal
Conflict.
*p , .05.
 Reason-Giving and Treatment Rationales 891

rationales (video and written presentation) differs from the more typical context in which
a treatment rationale is presented (i.e., within a therapy session). At the same time, there
are characteristics of the participants and the methodology that help to increase the gen-
eralizability of the findings. First, given that BDI scores of greater than 9 are thought to
reflect dysphoria in nonclinical populations, the mean score of 10.4 for our sample sug-
gests a greater severity of depressed mood than would be found in a strictly normal
population. Twenty-seven percent of the sample had BDI scores in the dysphoric range
and 16% scored above the typical cutoff for major depression (Kendall et al., 1987).
Thirty-one percent of the sample had seen a psychotherapist in the past year (69% life-
time). These figures are higher than those found in large-scale studies of normal popula-
tions not selected for any type of psychopathology or treatment history (cf. Howard et al.,
1996). With regard to the treatment rationales, it is increasingly common for written
psychoeducational materials to be used in conjunction with psychotherapy (e.g., Burns,
1980; Craske, Meadows, & Barlow, 1994). Moreover, as psychotherapy treatments become
increasingly structured, therapists necessarily become more explicit in their presentation
of the rationale behind treatment. Thus, although the present study is analogue in nature,
it is not entirely dissimilar to a real-world clinical context.
The most salient finding from this study is that individuals who endorse more rea-
sons for depression have a less positive reaction to an AO treatment rationale. The rela-
tionship holds for personal reactions to a video vignette and credibility ratings in response
to a written treatment rationale. Together with a previous study demonstrating similar
findings (Addis & Jacobson, 1996), these results suggest that clients who offer more
reasons for depression will be more likely to resist an activation-oriented rationale for
treatment. Given that acceptance of a rationale is associated with positive treatment out-
comes (e.g., Fennel & Teasdale, 1987; Ilardi & Craighead, 1994; Spence & Sharpe, 1993),
high scores on the RFD may be considered a negative indicator for an activation oriented
treatment.
Precisely why such an association exists is less clear. One possibility is that reason-
giving is often functioning to justify behavior; in this case depressed behavior. If an
individual feels the need to justify his or her behavior, she or he should be more likely to
resist inducements to take personal responsibility for change. It also may be that people
who perceive the causes of their depression to be complex, and therefore give multiple
reasons, will resist the apparently simple approach of changing specific behaviors. This
interpretation is consistent with the finding from the current study that the tendency to
ruminate is associated with a more negative reaction to a behavior change rationale (see
also Lyubomirsky & Nolen-Hoeksema, 1993). However, our results suggest that rumina-
tion is not the pathway through which reason-giving is associated with resistance to a
behavioral rationale. Although, as one would expect, reason-giving and rumination are
correlated, they appear to have separate effects on receptiveness to a behavior change–
oriented treatment rationale. Future studies might explore in an open-ended format peo-
ple’s reasons for rejecting or accepting various treatment options.
Contrary to what we had predicted, we found no evidence of an association between
either reason-giving or rumination and responses to a treatment rationale emphasizing the
role of insight in treating depression. This was true regardless of whether participants
were rating their personal reactions to a video vignette, or the credibility of a written
description. The tendency to give multiple reasons for a problem may not necessarily
make one more receptive to an insight-oriented rationale. However, characterological
and childhood reasons specifically predict positive personal reactions to an insight-
oriented rationale. In contrast, both characterological and childhood reasons are associ-
ated with a more negative personal reaction to an activation-oriented rationale. It appears
892 Journal of Clinical Psychology, July 1999

that individuals who attribute depression to past experiences in childhood or to stable

aspects of the self (e.g., “I’m depressed because this is the way I’ve always been”) respond
more favorably to the idea of discovering the causes of depression, than to changing one’s
current behavior. It is tempting to speculate about treatment matching here. However,
these results should obviously be interpreted with caution until they are replicated with a
clinical sample.
The relationships between reasons for depression and responses to the treatment
rationales were fairly consistent for both credibility ratings and personal reactions. This
suggests that the relationship between reason-giving and reactions to activation- and
insight-oriented rationales is relatively robust across different ways of presenting the
treatments. Clients who offer more reasons for depression are likely to view behavior
change treatments as both less credible generally, and as less likely to be personally
helpful.
A final purpose of this study was to explore the relationship between reason-giving
and the tendency to ruminate in response to depressed mood. Our results suggest that
individuals who ruminate also endorse more reasons for being depressed. This supports
the validity of the rumination construct because thinking about the reasons one is depressed
is thought to be a central part of rumination (e.g., Nolen-Hoeksema et al., 1994).
We would like to offer one final observation and caveat regarding this line of research.
This study contributes to a growing body of empirical and clinical literature suggesting
that convergence between a client’s and a therapist’s conceptual models of etiology and
treatment facilitates positive treatment responses. From a clinical standpoint, one might
therefore assume that a therapist should always be working to match a client’s conceptual
framework for his or her problems. However, some individuals may ultimately benefit
from precisely the treatment whose rationale they resist. For example, clients who rumi-
nate or give excessive reasons for depression may be exactly the individuals who need to
become active and change their behavior. Future research should focus on identifying the
conditions under which a client’s initial negative reaction to a treatment rationale is pre-
scriptive or proscriptive for a particular treatment. It is also unlikely that personal reac-
tions to a treatment rationale, or assessments of a treatment’s credibility, are static processes.
A client’s attitudes toward a treatment approach may unfold over the course of treatment
and may not follow a linear progression of agreement or disagreement with a particular
rationale. Both clinicians and psychotherapy researchers interested in reactions to treat-
ment rationales should make multiple assessments over time.

References
Addis, M.E., & Jacobson, N.S. (1996). Reasons for depression and the process and outcome of
cognitive-behavioral psychotherapies. Journal of Consulting and Clinical Psychology, 64, 1417–
1424.
Addis, M.E., Truax, P., & Jacobson, N.S. (1995). Why do people think they are depressed?: The
reasons for depression questionnaire. Psychotherapy, 32, 476– 483.
Anderson, C.M., Hawkins, R.P., & Scotti, J.R. (1997). Private events in behavior analysis: Con-
ceptual basis and clinical relevance. Behavior Therapy, 28, 157–179.
Antaki, C. (1994). Explaining and arguing: The social organization of accounts. Thousand Oaks,
CA: Sage.
Atkinson, D.R., Worthington, R.L., Dana, D.M., & Good, G.E. (1991). Etiology beliefs, prefer-
ences for counseling orientations, and counseling effectiveness. Journal of Counseling Psy-
chology, 38, 258–264.
 Reason-Giving and Treatment Rationales 893

Barber, L.C., & Stoltenberg, C.D. (1994). Preference for counseling approach as a function of
emotional locus of control and personal relevance. Journal of Social and Clinical Psychology,
3, 240–251.
Baston, C.D., Jones, C.H., & Cochran, P.J. (1979). Attributional bias in counselors’ diagnoses: The
effect of resources. Journal of Applied Social Psychology, 9, 377–393.
Beck, A.T., Steer, R.A., & Garbin, M. (1988). Psychometric properties of the Beck Depression
Inventory: Twenty-five years of evaluation. Clinical Psychology Review, 8, 77–100.
Beck, A.T., Ward, C., Mendelson, M., Mock, J., & Erbaugh, J. (1961). An inventory for measuring
depression. Archives of General Psychiatry, 4, 561–571.
Becker, M.H., Haefner, D.P., Kasl, S.V., Kirscht, J.P., Maiman, L.A., & Rosenstock, I.M. (1977).
Selected psychosocial models and correlates of individual health-related behaviors. Medical
Care, 15, 27.
Beutler, L.E., Engle, D.E., Mohr, D., Daldrup, R.J., Bergan, J., Meredith, K., & Merry, W. (1991).
Predictors of differential response to cognitive, experiential, and self-directed psychothera-
peutic procedures. Journal of Consulting and Clinical Psychology, 59, 333–340.
Borkovec, T.D., & Nau, S.D. (1972). Credibility of analogue therapy rationales. Journal of Behav-
ior Therapy and Experimental Psychiatry, 3, 257–260.
Burns, D.D. (1980). Feeling good: The new mood therapy. New York: William Morrow.
Centers for Disease Control. (1990). Health beliefs, compliance-hypertension. Journal of the Amer-
ican Medical Association, 264, 2864.
Craske, M.G., Meadows, E., & Barlow, D.H. (1994). Therapist’s guide for the mastery of your
anxiety and panic II & agoraphobia supplement. New York: Graywind.
DeGood, D.E. (1983). Reducing medical patients’ reluctance to participate in psychological ther-
apies: The initial session. Professional Psychology: Research and Practice, 5, 570–579.
Fennel, M.J.V., & Teasdale, J.D. (1987). Cognitive therapy for depression: Individual differences
and the process of change. Cognitive Therapy and Research, 11, 253–271.
Hayes, S.C. (1989). Rule governed behavior: Cognition, contingencies, and instructional control.
New York: Plenum.
Hayes, S.C., & Wilson, K.G. (1993). Some applied implications of a contemporary behavior-
analytic account of verbal events. The Behavior Analyst, 16, 283–301.
Howard, K.I., Cornille, T.A., Lyons, J.S., Vessey, J.T., Lueger, R.J., & Saunders, S.M. (1996).
Patterns of mental health service utilization. Archives of General Psychiatry, 53, 697–703.
Ilardi, S.S., & Craighead, W.E. (1994). The role of nonspecific factors in cognitive-behavior ther-
apy for depression. Clinical Psychology Science and Practice, 1, 138–156.
Kendall, P.C., Hollon, S.D., Beck, A.T., Hammen, C.L., & Ingram, R.E. (1987). Issues and recom-
mendations regarding use of the Beck Depression Inventory. Cognitive Therapy and Research,
11, 289–299.
Lyubomirsky, S., & Nolen-Hoeksema, S. (1993). Self-perpetuating properties of dysphoric rumi-
nation. Journal of Personality and Social Psychology, 65, 339–349.
Murdock, N.L., & Fremont, S.K. (1989). Attributional influences in counselor decision making.
Journal of Counseling Psychology, 36, 417– 422.
Narikiyo, T.A., & Kameoka, V.A. (1992). Attributions of mental illness and judgments about help
seeking among Japanese-American and White American students. Journal of Counseling Psy-
chology, 39, 363–369.
Nolen-Hoeksema, S., & Morrow, J. (1991). A prospective study of depression and distress follow-
ing a natural disaster: The 1989 Loma Prieta earthquake. Journal of Personality and Social
Psychology, 61, 115–121.
Nolen-Hoeksema, S., Parker, L.E., & Larson, J. (1994). Ruminative coping with depressed mood
following loss. Journal of Personality and Social Psychology, 67, 92–104.
894 Journal of Clinical Psychology, July 1999

Plous, S., & Zimbardo, P.G. (1986). Attributional biases among clinicians: A comparison of psy-
choanalysts and behavior therapists. Journal of Consulting and Clinical Psychology, 54, 568–
570.
Rokke, P.D., Carter, A.S., Rehm, L.P., & Veltum, L.G. (1990). Comparative credibility of current
treatments for depression. Psychotherapy, 27, 235–242.
Rokke, P.D., & Scogin, F. (1995). Depression treatment preferences in younger and older adults.
Journal of Clinical Geropsychology, 1, 243–257.
Shoham-Salomon, V., & Hannah, M.T. (1991). Client-treatment interaction in the study of differ-
ential change processes. Journal of Consulting and Clinical Psychology, 59, 217–225.
Skinner, B.F. (1974). About behaviorism. New York: Vintage.
Spence, S.H., & Sharpe, L. (1993). Problems of dropout in the self-help treatment of chronic,
occupational pain of the upper limbs. Behavioural and Cognitive Psychotherapy, 21, 311–328.
Tracey, T. (1988). Relationship of responsibility attribution congruence to psychotherapy outcome.
Journal of Social and Clinical Psychology, 7, 131–146.
Worthington, R.L., & Atkinson, D.R. (1996). Effects of perceived etiology attribution similarity on
client ratings of counselor credibility. Journal of Counseling Psychology, 43, 423– 429.