Behav. Res. Ther. Vol. 33, No. 7, pp.

779-784, 1995
Copyright © 1995 ElsevierScienceLtd
Pe~mon Printed in Great Britain. All fights reserved
0005-7967(95)00016-X
0005-7967/95 $9.50 + 0.00

PERCEIVED RESPONSIBILITY: STRUCTURE AND

SIGNIFICANCE

S. RACHMAN, l DANA S. THORDARSON, I ROZ SHAFRAN 2 and

SHEILA R. WOODY 3
~Department of Psychology, University of British Columbia, 2136 West Mall, Vancouver, B.C., Canada
V6T IZA, 2Department of Psychology, Institute of Psychiatry, De Crespigny Park, London SE5 8AF,
England and 3Department of Psychology, Yale University, P.O. Box 208205, Yale Station, New Haven,
CT 06520-7447, U.S.A.

(Received 23 December 1994)

Summary---Given the postulated significance of inflated responsibility in obsessive compulsive disorder

(OCD), there is a need for clarification of the concept itself and a means for measuring such responsibility.
Two psychometric studies were conducted in order to develop a reliable self-report scale. In the first study
291 students completed the specially constructed Responsibility Appraisal Questionnaire (RAQ). Four
factors emerged: responsibility for harm, responsibility in social contexts, a positive outlook towards
responsibility, and thought-action fusion (TAF). In the second study, 234 students completed a revised
RAQ. Four comparable factors emerged, and the TAF subscale correlated significantly with measures of
obscssionality, guilt, and depression. The correlations between TAF and obsessionality and guilt remained
significant even after BDI scores were controlled. It is concluded that the broad concept of inflated
responsibility needs to be qualified; the connection between inflated responsibility and OCD appears to
be situation-specific and idiosyncratic. There is more inflated responsibility than there is OCD. The
measured concept of inflated responsibility is multifactorial (harm, social, positive, and TAF), not unitary.
The TAF factor appears to be particularly significant in OCD.

INTRODUCTION
There is good reason to assume that an inflated sense of responsibility plays a significant role in
particular types of obsessive compulsive disorder (OCD), especially in compulsive checking. As part
of this thrust towards a cognitive-behavioural analysis of OCD, Salkovskis and Kirk (1989)
postulated that "obsessional thoughts almost always involve the fear of responsibility for harm"
(p. 152), and some extrapolations and implications were set out by Rachman (1993).
Evidence to support the general idea of a connection between inflated responsibility and OCD
comes from clinical findings (Salkovskis, 1985; Tallis, 1994; Van Oppen & Arntz, 1994),
psychometric studies (Salkovskis, 1989; Tallis, 1994; Van Oppen & Arntz, 1994), and experimental
analyses (Lopatka & Rachman, 1995; Rachman & Hodgson, 1980). In a recently completed
experiment, Ss with significant OCD problems reported steep decreases in the compulsive urge to
check after an experimental manipulation produced a temporary reduction in their sense of
responsibility (Lopatka & Rachman, 1995). The checking urges increased following an experimen-
tally induced inflation of responsibility, but the changes failed to reach a significant level--perhaps
because of ceiling effects. Compulsive cleaning did not show a similar relationship to inflated
responsibility.
These results, combined with the findings of the Rachman and Hodgson (1980) experiments,
psychometric results, and Salkovskis's analysis, pointed to the need for a strong and dependable
measure of responsibility. This we set out to produce.
Aim. The main aim of this study was to produce a reliable scale for measuring responsibility.
In order to avoid a confound between responsibility and OCD, we attempted to construct a new
scale that was free of OCD-related items. The need for a 'pure' scale was essential for testing the
relations between responsibility and OCD.
We expected to find a significant relationship between inflated responsibility and OCD, and a
particularly close relationship between inflated responsibility and compulsive checking. We were
779
780 S. Rachman et al.

also curious about the relationships between inflated responsibility and depression, and inflated
responsibility and anxiety.

STUDY 1: DEVELOPMENT OF THE RESPONSIBILITY APPRAISAL

Q U E S T I O N N A I R E (RAQ)
The purpose of this study was to develop a questionnaire to assess responsibility in normal Ss.
A pool of items was written and administered to undergraduate students; their responses were used
to select the best items for a final scale.

Method
Subjects
The Ss were 291 psychology students at the University of British Columbia who completed the
questionnaire package for course credit. Their average age was 19.5 yr; 72% of the sample was
female.

Materials
We devised a pool of 36 items to assess responsibility in five content areas: responsibility for
property damage [e.g. "If I smelled smoke in a store, it is (not) up to me to inform the manager"];
responsibility for physical harm coming to other people [e.g. "When driving, it is (not) up to me
to make sure that my passengers are wearing safety belts"]; responsibility in social contexts (e.g.
"I should not turn down an invitation from a friend"); a positive outlook on responsibility (e.g.
"I would welcome the opportunity to be put in charge of the safety of buildings"); and
thought-action fusion (TAF), which is the belief that thoughts can influence events or are almost
equivalent to actions [e.g. "For me, having a mean thought is (not) as bad as doing something
mean" and/or "My mean thoughts wishing a person harm can increase the chance that something
harmful will happen to him/her"] (see Rachman, 1993). Care was taken to avoid items with
obsessional content. Approximately half the items were reverse-coded. Subjects endorsed each item
on a visual analogue scale from 0 to 100.

Results
A principal components analysis with varimax rotation performed on the 36 items indicated that
the best simple structure was obtained using a 2 or 4 factor solution. The 4 factor solution was
selected because it nearly corresponded to our original 5 domains of responsibility. The 4 factor
solution, which accounted for 30.9% of the variance, approximated the domains of responsibility
for harm (incorporating both physical harm and property damage; 11.4% of variance), social
responsibility (7.4% of variance), positive responsibility (5.5% of variance), and thought-action
fusion (TAF; 6.6% of variance).
From the original pool of 36 items, 18 were retained to form the final scale with subscales based
on the four factors. Three criteria were used in the selection of items: (1) loading only on their
hypothesized factor (harm, social, positive, or TAF), (2) high item-total and item-subscale
correlations, and (3) mean response not extremely high (to maximize the ability of the scale to
discriminate among Ss at high levels of responsibility). To produce an even distribution of items
among the 4 subscales, considerable flexibility in the application of these criteria was required. The
coefficient ~ts were: 0.56 for Harm (5 items), 0.47 for Social (5 items), 0.58 for Positive (4 items),
and 0.51 for TAF (4 items). Subscales were uncorrelated, with the exception of Harm and TAF
(r = 0.18, P < 0.01), and Social and Positive (r = 0.17, P < 0.01).
The revised RAQ was administered to a new sample tO verify its factor structure and investigate
its relationship with other variables such as obsessionality and depression (see Study 2).

STUDY 2: CORRELATES OF RESPONSIBILITY

The association between responsibility, as measured by subscales of the revised RAQ, and
obsessionality, depression, and guilt was investigated.
 Perceived r e s p o n s i b i l i t y 781

Method
Subjects
The subjects were 234 psychology students at the University of British Columbia who completed
the questionnaire package for course credit. Their average age was 20 yr; 68% were female.

Materials
They completed the following questionnaires:
RAQ. The 18 item revised RAQ, with instructions to rate agreement with the statements on a
6 point scale, ranging from 1 (Absolutely Disagree) to 6 (Absolutely Agree).
M O C L The Maudsley Obsessional Compulsive Inventory (Hodgson & Rachman, 1977), a 30
item true/false inventory with subscales of cleaning, checking, doubting/conscientiousness, and
slowness.
BDL The Beck Depression Inventory (Beck, 1967), a 21 item inventory to assess severity of
depression.
GL The Guilt Inventory (Kugler & Jones, 1992), a 45 item questionnaire comprising subscales
of trait guilt, state guilt, and guilt standards. Subjects endorse items on a scale of 1 (Very untrue
of me or strongly disagree) to 5 (Very true of me or strongly agree).
IBRO. The Inventory of Beliefs Related to Obsessionality (Freeston, Ladouceur, Thibodeau &
Gagnon, 1993), a 20 item inventory in which Ss endorse beliefs on a scale of 1 (I believe strongly
that this statement is false) to 6 (I believe strongly that this statement is true).

Results
A principal components analysis with Varimax rotation was performed on the 18 items. A
4-factor solution accounted for 46.2% of the variance, and corresponded to our 4 domains of
responsibility from Study 1. The factors for TAF and Positive Responsibility were the most
coherent, whereas 2 Social and 1 Harm item unexpectedly loaded on factors other than their own.
The means, standard deviations, and internal consistencies of the RAQ subscales are given in
Table 1. MOCI scores ranged from 0 to 20 (M = 7.60, SD = 4.64); BDI scores ranged from 0 to
43 (M = 9.26, SD = 7.08). The means of the MOCI and BDI were, as expected, in the non-clinical
range.
Pearson correlation coefficients were calculated to investigate the relationships between RAQ
subscales and measures of obsessionality, depression, and guilt (see correlation matrix in Table 2).
Of the RAQ subscales, TAF was most strongly correlated with measures of obsessionality,
particularly the IBRO. TAF was also associated with depression and guilt measures. Partial
correlation coefficients, controlling for BDI scores, were also conducted. All significant correlations
between the RAQ subscales, including the TAF subscale, and measures of obsessionality remained
significant even when BDI scores were partialled out. After controlling for the effects of depression,
the correlations between RAQ subscales and guilt were no longer significant.
It is possible that high MOCI scorers, i.e. Ss who score within the clinical range, are qualitatively
different from other Ss. Therefore analyses were conducted comparing Ss scoring above a clinical
cutoff of 15 on the MOCI with the rest of the sample. The TAF scores of high MOCI scorers were
significantly greater than those of the rest of the sample, but there were no differences between the
groups on the other RAQ subscales (Table 3). The results encourage the notion that TAF is
implicated in OCD, but that responsibility in general may not be elevated in OCD.

Table 1. Descriptive statistics for the RAQ subscales: Study 2

Subscale No. items Mean SD Cronbach's ~t
Harm 5 19.98 3.40 0.43
Social 5 19.68 2.97 0.40
Positive 4 15.72 3.31 0.66
TAF 4 8.57 3.32 0.76
782 S. R a c h m a n e t al.

Table 2. Correlations among RAQ subscales and measures of obsessionality, depression

and guilt
Harm Social Positive TAF
MOCI-total 0.05 0.17* - 0.10 0.45* *
MOCI-cleaning 0.06 0.06 -0.04 0.35**
MOCl-checking 0.04 0.09 -0.12 0.41"*
MOCI-doubting 0.12 0.23** 0.02 0.26**
MOCI-slowness 0.04 0.09 -0.02 0.10
IBRO 0.07 0.20* -0.12 0.50**
BDI 0.09 0.12 - 0.17" 0.38**
Guilt trait 0.01 0.16 -0.20* 0.37**
Guilt state -0.02 0,12 -0.13 0.34**
Guilt standards 0.21" 0,13 0.12 -0.01
*P < 0.01.
**P < 0.001.

DISCUSSION
The results of the two studies show the need to qualify the original, broad view of inflated
responsibility. The postulated connection between inflated responsibility and obses-
sional-compulsive problems appears to be more situation-specific and idiosyncratic than was
originally assumed. The idea that people affected by obsessional problems are broadly over-respon-
sible is difficult to support. Furthermore, even some patients with severe OCD problems and
definite evidence of inflated responsibility nevertheless seek and welcome specific types of
responsibility (e.g. a severely affected teacher who nevertheless enjoyed the responsibility of looking
after a class of children).
Inflated responsibility is connected to compulsive behaviour, especially checking behaviour, but
the connections probably are situation-specific and idiosyncratic, rather than broad and commonly
held.
There are seven arguments in favour of a situation-specific explanation. Firstly, there is far more
inflated responsibility than there is OCD. Hence, inflated responsibility by itself is not sufficient
to account for OCD--possibly because inflated responsibility is too broadly conceived. Pre-
sumably, it is only certain manifestations of inflated responsibility that contribute to OCD (e.g.
Rheaume, Ladouceur, Freeston & Letarte, 1995).
Secondly, the psychometric studies failed to reveal a unitary factor of responsibility. Only one
of the factors on the RAQ, thought-action fusion (TAF), correlated strongly with aspects of OCD,
particularly with the measure of beliefs related to obsessionality (IBRO). This relationship is to
be expected, given that the IBRO has a number of items concerning beliefs about thoughts (e.g.
"Guilt is an appropriate response to unacceptable thoughts"). The TAF subscale was also the only
subscale to correlate strongly with guilt (trait and state). It is easy to see how believing in TAF
is liable to result in feelings of guilt. If a person believes that thinking something is almost as bad
as doing it, an unacceptable intrusive thought will cause guilt. Interestingly, the TAF subscale was
not related to guilt standards as assessed by the Guilt Inventory. This suggests that TAF is not
merely an excess of normal morality, but rather a variant of it. In support of the emerging
importance of TAF, high scorers on the MOCI returned significantly larger scores on TAF (and
guilt, depression) than did the low-MOCI scorers.
In the Lopatka and Rachman experiment on induced responsibility, effort and time had to be

Table 3. High MOCI scorers (n = 23) vs rest of sample (n = 210)

High scorers Rest of sample
Scale M (SD) M (SD) t
RAQ-Harm 20.7 (2.8) 19.9 (3.5) 1.0
RAQ-Social 20.0 (2.5) 19.6 (3.0) 0.6
RAQ-Positive 15.0 (2.5) 15.8 (3.4) -l.I
RAQ-TAF 11.7 (3.1) 8.2 (3.2) 5.1'*
IBRO 74.0 (8.1) 64.5 (10.3) 4.2**
BDI 15.2 (9.5) 8.6 (6.5) 4.3**
Guilt trait 64.3 (9.7) 56.6 (12,4) 2.9*
Guilt state 34.7 (6.8) 29.3 (7.6) 3.3*
Guilt standards 48.3 (4,5) 44.8 (5,2) 3.1"
*P < 0.005.
**P < 0.001.
 Perceived responsibility 783

devoted to identifying the exact actions that were responsible for shifts in responsibility; it was
not possible to construct a set of standardised instructions that would apply across the board.
The Ss had specific, idiosyncratic ideas of responsibility and of the misfortunes that might
occur if they failed to exercise their self-determined responsibility. Similarly, in the early
experiments by Rachman and Hodgson (see 1980) on compulsive checking, the variability (and
at that time, the lack of predictability) of the patient's sense of responsibility was a hindrance.
Specific and idiosyncratic expressions of responsibility were common; broad, universally experi-
enced expressions of responsibility were unusual.
Even patients who report high levels of inflated responsibility are easily able to delineate
those tasks or events which evoke no responsibility at all. Further, as noted earlier, patients
who are tormented by a highly irrational sense of responsibility can describe specific situations
in which they welcome responsibility.
Contrary to expectation, the 23 Ss who scored within the clinical range on the MOCI ( > 14)
did not have higher scores than the rest of the Ss on the Harm, Social, or Positive subscales.
The only subscale in which the high MOCI scorers differed from the rest of the sample was
the TAF subscale.
The results do not support the idea that people with OCD may have generally inflated
responsibility. And Ss with MOCI scores within the clinical range were as likely as other Ss
to welcome some forms of responsibility. The presence of OCD tendencies does not necessarily
eliminate or even suppress the acceptance of responsibility. All of this suggests that the inflated
responsibility which plays a part in OCD must be situation-specific; it is not general and does
not preclude acceptance of selected forms of responsibility.
It is of course possible that our RAQ scale is not sufficiently sensitive for the purpose;
only further investigations will enable one to answer that query. Also, Rheaume et al. (in press)
found that pivotal responsibility may be the key factor. The association with OCD may
come into play only when the affected person feels solely or mainly responsible for the
anticipated harmful event (i.e. feels pivotally responsible). The inclusion of items designed to
tap feelings of pivotal responsibility may increase the sensitivity of the RAQ, and then reveal
a degree of association between OCD and inflated responsibility that extends beyond TAF.
It is, however, necessary to balance the need for an assessment of responsibility which is person
and situation specific with an assessment that is not confounded by including obsessional
situations.
The concept of psychological responsibility appears to be multi-factorial. So far, only one
of four psychometric factors, TAF, seems to relate to OCD. From the assessment point of
view, the use of a general measure of responsibility may make little difference, except perhaps
as a baseline. The TAF factor score shows some promise and warrants psychometric devel-
opment and experimental analysis.

REFERENCES
Beck, A. T. (1967). Depression: clinical, experimental, and theoretical aspects. New York: Harper & Row.
Freeston, M. H., Ladouceur, R., Thibodeau, N. & Gagnon, F. (1993). Beliefs about obsessional thoughts. Journal of
Psychopathology and Behavioral Assessment, 15, 1-21.
Hodgson, R. J. & Rachman, S. (1977). Obsessional-compulsive complaints. Behaviour Research and Therapy, 15,
389-395.
Kugler, K. & Jones, W. H. (1992). On conceptualizing and assessing guilt. Journal of Personality and Social Psychology,
62, 318-327.
Lopatka, C. & Rachman, S. (1995). Perceived responsibility and compulsive checking: An experimental analysis. Behaviour
Research and Therapy, 33, 673-684.
Rachman, S. (1993). Obsessions, responsibility and guilt. Behaviour Research and Therapy, 31, 149-154.
Rachman, S. & Hodgson, R. (1980). Obsessions and compulsions. Englewood Cliffs, NJ: Prentice Hall.
Rheaume, J., Ladouceur, R., Freeston, M. H. & Letarte, H. (1995). Inflated responsibility in obsessive-compulsive disorder:
Validation of an operational definition. Behaviour Research and Therapy, 33, 159-169.
Salkovskis, P. M. (1985), Obsessional-compulsive problems: A cognitive-behavioural analysis. Behaviour Research and
Therapy, 25, 571-583.
Salkovskis, P. M. (1989). Cognitive-behavioural factors and the persistence of intrusive thoughts in obsessional problems.
Behaviour Research and Therapy, 27, 677-682.
Salkovskis, P. M. & Kirk, J. (1989). Obsessional disorders. In Hawton, K., Salkovskis, P. M., Kirk, J. & Clark,
784 S. Rachman et al.

D. M. (Eds), Cognitive-behavioural treatment of psychiatric disorders: A practical guide. Oxford: Oxford University
Press.
Tallis, F. (1994). Obsessions, responsibility and guilt: Two case reports suggesting a common and specific aetiology.
Behaviour Research and Therapy, 32, 143-145.
Van Oppen, P. & Arntz, A. (1994). Cognitive therapy for obsessive compulsive disorder. Behaviour Research and Therapy,
32, 79-87.