THE CONCEPT OF AN IDEAL THERAPEUTIC

RELATIONSHIP1
FRED E. FIEDLER2
THE UNIVERSITY OF CHICAGO AND VA MENTAL HYGIENE CLINIC

O
NE of the major differences between tance than therapeutic skill, we will expect
the various systems of psychotherapy therapists of different schools to differ more
is the divergence of views concerning from those of other schools than from ther-
the most desirable therapeutic relationship apists within their own school. If, on the other
that the therapist should create. There has hand, experience and skill are of greater im-
been little, if any, disagreement among psy- portance; that is, if the concept of what a
chotherapists as to the paramount importance therapeutic relationship ideally should be is de-
of a good therapeutic relationship to eventual rived from personal experience and skill, and
cure [1, pp. 32-47; 2; 5, p. 3; 7; 8, pp. 85- only one such type of relationship has been
114]. Rogers has been most insistent that the found maximally effective, we will expect the
relationship is important; so have Thorne [10] more expert to agree among themselves irre-
and Horney [6, pp. 276-305]. Yet, the dif- spective of school, rather than with the less ex-
ferences of opinion as to what constitutes a pert within their own school.
good therapeutic relationship are quite marked. We can assume here that therapists will at-
Rogers favors a relationship which permits the tempt to create the relationship which they
patient most scope, and in which the therapist consider ideal. We can further assume that the
intervenes as little as possible, while Horney skilled therapist is better versed in his school's
and Thorne advocate greater intervention and theory of therapeutic relationships than is the
activity on the part of the therapist. less expert. Consequently, the expert of any
If, however, the therapeutic relationship it- one school can be expected to disagree more
self is of such paramount importance then it with experts of other schools if the schools
seems unlikely that so many different types of differ in theory, since experts are generally
relationships can be equally effective in helping more representative of their school's practice
patients to adjust, especially since all schools and theory than novices of a school.
of therapy have reported successes as well as Stating this hypothesis in more operational
failures among their therapy cases. terms, if there are any real differences in
One question we wish to answer here is, schools as to the relationship which they at-
"Are these differences in theory semantic or do tempt to achieve we will expect a factor anal-
they represent actual divergencies in the goal ysis to yield as many factors as there are points
therapists set for themselves?" If the theory of view. If, on the other hand, these differences
and training of therapists are of greater impor- are present in theory only, and only one type
J
of relationship is actually considered maximally
Reviewed by the Veterans Administration and effective, we will expect to find only one gen-
published with the approval of the Chief Medical
Director. The statements and conclusions published eral factor among therapists of various schools.
by the author are the result of his own study and Some therapists have expressed the convic-
do not necessarily reflect the opinion or policy of tion that the therapeutic relationship is some-
the Veterans Administration.
2 thing unique and does not occur outside of the
The results of this paper were discussed in part
at the Midwestern Psychological Association Meet- therapeutic situation [i.e. 5 and 7]. We may
ing, Chicago, April, 1949. The writer wishes to assume therefore that one must have had ex-
acknowledge the suggestions he received from Mr. perience with therapeutic situations in order to
Ralph Heine and Dr. Morris I. Stein which were
invaluable in clarifying some of the aspects of this be able to describe them. An hypothesis would
investigation. then be that lay persons, who have had no con-
239
240 FRED E. FIEDLER

TABLE 1
PERSONAL DATA OF THERAPISTS AND UNTRAINED RATERS COOPERATING IN STUDY
Academic Years Years Personal Theoretical
Rater Training Experience Therapy Orientation
PA-1« AM 5 1 Psychoanalytic
(Washington School)
PA-2* MD 10 2'/2 Psychoanalytic
(Chicago)
PA-3t MD 10 7 Psychoanalytic
PA-4* MA, SSA 5 3>/2 Psychoanalytic
social service
PA-S AM 2 l'/2 Psychoanalytic
PA-6 MD 3 Psychoanalytic
ND-7* PhD 6 — Nondirective
l'/2
ND-8* PhD 5 _ Nondirective
ND-9 AM 1 Nondirective
ND-10 AM 3 '/4 Nondirective
IP-11* MD 26 Adlerian
Ec-12* AM 1 Eclectic
Ec-13 AM 5 Eclectic
Un-14 PhB 3 II Untrained
Un-15 BS, RN — Untrained
Un-16 BS, EE — Untrained
* Considered to be very good therapist.—
fThis rater thought there "was a catch" to the ratings.
^Experience of this rater has been primarily in social service work.
|| Psychoanalysis.

tact with psychotherapeutic situations, would started his analysis. One psychologist had two
be unable to describe the therapeutic relation- years of intensive and five years of general
ship with any degree of success as compared to therapeutic experience, and 150 hours of per-
experienced therapists. sonal analysis. The other psychologist had little
experience in doing therapy, but had received
METHODS AND RESULTS 120 hours of psychoanalysis.
The method of investigation used here was Of the two nondirective therapists, one had
to have therapists describe the relationship had over 70 hours of personal nondirective
which they consider ideal. This was done by therapy and had practiced for five years. The
means of a series of Q-technique3 [9] ratings other had obtained 12 hours of nondirective
by therapists of different schools and with dif- therapy and had three years of experience.
ferent degrees of training and reputed skill. Neither of the eclectic therapists had received
Two similar investigations were undertaken. personal therapy. One had five, the other one
The First Study. The first utilized the year of experience as a therapist.
ratings of six psychologists and two psychia- The measuring instrument was a series of
trists. Four of these therapists subscribed to a 119 qualitative statements descriptive of pa-
psychoanalytic theory, two were nondirectively tient-therapist relationships. These traits were
oriented, and two considered themselves eclec- randomly selected from a large number of
tics [See Table 1]. Among the four psycho- statements about relationships which were
analytically oriented subjects, one had com- drawn from books, journal articles, statements
pleted his medical and most of his psychoana- by therapists in case records, and in confer-
lytic training, and had practiced for ten years. ences. These 119 statements were mounted on
The other medically trained therapist had been cards, shuffled, and given to each subject with
practicing for three years and had not yet instructions to sort them into seven categories,
8
with 2, 11, 28, 37, 11, and 2 statements in each
Q-technique is a method developed by W. Ste- category, with the statements most character-
phenson which permits correlation between persons
on the basis of traits. It is akin to inverse factor istic of their idea of an ideal therapeutic rela-
analysis. tionship at one extreme, and the statements
 CONCEPT OF IDEAL THERAPEUTIC RELATIONSHIP 241

least characteristic of the ideal therapeutic rela- Inspection of individual items revealed no
tionship at the other extreme of this normal major differences between schools. Below are
distribution. Each of the statements, in accord- the statements which were characterized as
ance with the regular design of Stephenson's most characteristic of an ideal therapeutic re-
Q-technique, received a score from one to lationship.
seven, depending on the category in which the An empathic relationship
card had been placed by the therapist. The Therapist and patient relate well
rating by each therapist could then be inter- Therapist sticks closely to the patient's problems
correlated with that by every other therapist, The patient feels free to say what he likes
and the resulting correlation matrix factored by An atmosphere of mutual trust and confidence
exists
Holzinger's bi-factor method. The table of in- Rapport is excellent
tercorrelations and factor loadings of each The patient assumes an active role
therapist are presented in Table 2. The therapist leaves patient free to make his own
choices
TABLE 2 The therapist accepts all feelings which the pa-
INTERCORRELATIONS* AND FACTOR LOADINGS OBTAINED tient expresses as completely normal and under-
IN THE FIRST STUDY standable
A tolerant atmosphere exists
An understanding therapist
Patient feels most of the time that he is really
understood
FA-1 — 66 72 61 77 59 70 64 Therapist is really able to understand patient
PA-2 — 70 57 72 63 64 60 The therapist really tries to understand the pa-
PA-5 — 61 78 67 69 68 tient's feelings
PA-6 — 68 67 64 60
ND-7
—
64 72 67 The following statements were rated as
ND-10 — 56 48
least characteristic of an ideal therapeutic re-
Ec-12 64
Ec-13 — lationship :
—
Factor A punitive therapist
Saturations :f 83 80 87 75 92 72 82 76
Therapist makes the patient feel rejected
"Correlations were computed by the method of squared The therapist seems to have no respect for the
differences. All correlations are significantly different patient
from zero on the 1 per cent level or above.
fMrosid. = .002 ; SDresId. <= .027. An impersonal, cold relationship
The therapist often puts the patient "in his place"
The results indicate that all therapists cor- The therapist curries favor with the patient
relate positively with each other. Correlations The therapist tries to impress the patient with his
range from .48 to .78 with a median of .64. skill or knowledge
The therapist treats the patient like a child
Reliability on resorting was .80.
The analyzed, psychoanalytically oriented The Second Study. These results posed sev-
therapists PA-1, PA-2, correlated with the eral questions: Are the results presented above
nondirective therapist, who is considered an a reflection of a stereotyped notion among ther-
expert, (ND-7) .77, and .72, while they cor- apists of what a good therapeutic relationship
related with psychoanalytically oriented inex- is? Are the results a function of the individu-
pert therapist, PA-6, .61, and .57. The dif- als who cooperated in this study, and the par-
ferences between these correlations are signif- ticular statements which were used here?
icant at the 5 per cent level or above. The Would an Adlerian expert also correlate high-
nondirective expert, who had correlated with ly with therapists of nondirective orientation?
the analyzed therapists .77 and .72, correlated Finally, if, as some have suggested, the thera-
only .64 with the less well trained nondirec- peutic relationship is something not found out-
tive therapist, ND-10, a difference also sig- side of therapy, will a lay person who had
nificant on the 5 per cent level or above. never been in therapy be unable to describe a
The bi-factor analysis yielded only one gen- therapeutic relationship ?
eral factor. Factor loadings for therapists who A similar study was made to answer these
had received personal therapy are somewhat questions. Seven different subjects cooperated
higher than loadings of others. (Table 1). A different array of 75 statements
242 FRED E. FIEDLER

about therapeutic relationships was selected on 1. Only one general factor was found;
the following basis: the same trends in correlations were apparent
a. A number of therapists had to agree con- (See Table 3).
cerning the aspect of the relationship and the 2. The Adlerian expert, IP-11, correlated
intensity of the statements, in order to make cer- highly with other well trained therapists.
tain that statements had the same meaning to
various therapists. 3. The lay person who had been analyzed,
b. A number of therapists had to agree that Un-14, correlated .84 with one nondirective
all statements were descriptive of therapeutic re- therapist, and .79 with a psychoanalytically
lationships. oriented therapist. We are therefore justified
The relationship was envisiaged as consisting in concluding that this concept of a good
of three dimensions: therapeutic relationship is not peculiar to ther-
a. The therapist's ability to communicate with, apists, especially since one person, who had
and understand the patient. never done, nor been in psychotherapy (Un-
b. The emotional distance which the therapist 15) correlated .76 with a nondirective, .81
takes toward the patient (emotionally withdraw- with an analytically oriented therapist, while
ing, neutral, or close to patient). she correlated only .56 with her husband,
c. The status of the therapist in relation to the
patient (superior to the patient, subordinate to ND-9, who is a nondirective novice.
the patient, or equal with the patient). Inasmuch as only one general factor was ob-
tained, the ratings could be pooled and a com-
To each of these dimensions were allotted bined rating of the ideal could be obtained.
25 statements, and each dimension in turn was This results from the pooling of the ratings of
subdivided into groups of 5 statements each, the four most expert therapists representing
representing five steps on each dimension. the psychoanalytic, the nondirective, and the
Statements were retained only when no dis- Adlerian viewpoint. (PA-1, ND-7, ND-8,
agreement could be found among therapists as IP-11) Since their factor loadings are prac-
to the dimension into which they should be tically as high as the retest reliability of this
placed, and the step within the dimension. second set of statements (.92), better esti-
The instructions to the ten persons rating mates of the Ideal could not be obtained by
the statements were the same as in the previous pooling the ratings of more therapists.
study, except that statements were to be placed This combined concept of the Ideal Thera-
into categories with 1, 7, 18, 23, 18, 7, and 1 peutic Relationship is presented below :*
statements in each category.
The second study confirmed the results of *The numbering system of these statements gives
the first study: the original classification into dimensions and steps
within the dimensions r
TABLE 3 Communication:
INTERCORRELATIONS* AND FACTOR LOADINGS OBTAINED 1-5 No communication is possible
11-15 Communication is poor
IN THE SECOND STUDY 21-25 Some communication exists
31-35 Communication and understanding is good
• > CO 41-45 Communication and understanding is ex-
t> fcl cellent
Emotional distance
PA-1 — 62 66 81 81 66 82 79 73 71 51-55 Therapist draws away from or rejects pa-
PA-3 — 43 50 61 44 62 60 57 58 tient
PA-4 — 70
_ 69 64 63 67 60 61 . 61-65 Therapist is somewhat cool toward patient
ND-7 82 69 76 78 73 72 71-75 Therapist is emotionally neutral
ND-8 — 78 81 84 76 73 81-85 Therapist tends to draw emotionally close
ND-9 63 72 65 56 to patient
—
IP-11
— 72 81 69 91-95 Therapist tends to be too close, is sticky
Un-14
— 79 65
Status
Un-15 .— 69
Un-16
101-105 Therapist feels very inferior and insecure
— 111-115 Therapist tends to look up and defer to
Factor patient
Saturations! 90 66 73 88 93 75 88 89 85 79 121-125 Therapist maintains peer relationship with
*AI1 correlations are significantly different from
patient
fcero on the 1 per cent level or above. 131-135 Therapist tends to look down on patient
tMro.id. = .003 ; SDresid. = .033. 141-145 Therapist feels very superior to patient
 CONCEPT OF IDEAL THERAPEUTIC RELATIONSHIP 243

Category 7 — (Most characteristic of an Ideal 62. The therapist at times draws emotionally
Therapeutic Relationship:) away from the patient
42. The therapist is able to participate completely 63. The therapist occasionally makes the patient
in the patient's communication angry
Category 6 — (Very characteristic) 64. The therapist feels somewhat tense and on
41. The therapist's comments are always right in edge
line with what the patient is trying to convey 65. The therapist seems to be a little afraid of
32. The therapist is well able to understand the the patient
patient's feelings 74. The therapist accepts all of the patient's
33. The therapist really tries to understand the statements in a noncommittal manner
patient's feelings 83. The therapist is pleased with the patient
34. The therapist always follows the patient's 84. The therapist is trying to establish an emo-
line of thought tionally close relationship with the patient
45. The therapist's tone of voice conveys the com- 85. The therapist is sympathetic with the patient
plete ability to share the patient's feelings 93. The therapist greatly encourages and re-
121. The therapist sees the patient as a co-worker assures the patient
on a common problem 94. The therapist expresses great liking for the
123. The therapist treats the patient as an equal patient
95. The therapist is deeply moved by the patient
Category 5 — (Somewhat characteristic) 102. The therapist tries to sell himself
21. The therapist reacts with some understanding 111. The therapist seems hesitant about asking
of the patient's feelings questions
22. The therapist is able to keep up with the 112. The therapist readily accedes to the patient's
patient's communication much of the time wishes
24. The therapist's reactions are neither particu- 114. The therapist assumes an apologetic tone of
larly favorable or unfavorable in permitting voice when commenting
free communication by the oatient 115. The therapist tries to please the patient
25. The therapist usually maintains rapport with 131. The therapist acts toward the patient in a
the patient somewhat protective manner
31. The therapist is usually able to get what the 134. The therapist treats the patient like his pupil
patient is trying to communicate 135. The therapist directs and guides the patient
35. The therapist usually catches the patient's
feelings Category 3 — (Somewhat inapplicable)
43. The therapist is never in any doubt about 3. The therapist somehow seems to miss the pa-
what the patient means tient's meaning time and again
44. The therapist's remarks fit in just right with 5. The therapist reacts in terms of his own
the patient's mood and content problems
71. The therapist is interested but emotionally 13. The therapist is unable to understand the pa-
uninvolved tient on any but a purely intellectual level
72. The therapist's feelings do not seem to be 14. The therapist finds it difficult to think along
swayed by the patient's remarks the patient's lines
73. The therapist maintains a friendly, neutral 53. The therapist is rejecting to the patient
attitude throughout 61. The therapist is somewhat cool toward the
75. The therapist shows little positive or negative patient
emotion in his reaction to the patient 92. The therapist showers the patient with af-
SI. The therapist seems to like the patient fection and sympathy
82. The therapist is pleasant to the patient 101. The therapist treats the patient like an
113. The theranist lets the patient determine the honored guest
course of the session 103. The therapist treats the patient with much
122. The therapist gives and takes in the situations deference
124. The therapist acts neither superior nor sub- 104. The therapist curries favor with the patient
missive to the patient 105. The therapist always apologizes when mak-
125. The therapist treats the patient like a friend ing a remark
132. The therapist tends to look down on the pa-
Catetriry 4 — (Middle category) tient
11. The therapist often flounders around before 141. The therapist talks down to the patient as if
petting the patient's meaning he were a child
12. The therapist often misses the point the pa- 143. The therapist is very condescending to the
tient is trying to get across patient
15. The therapist's comments tend to divert the 144. The therapist puts the patient in his place
patient's trend of thought 145. The therapist gives the impression of feeling
23. The therapist's understanding of the patient's very much above the patient in social and
feelings is neither particularly good or bad intellectual status
 244 FRED E. FIEDLER

Category 2 — (Uncharacteristic) who are considered less expert. (The dif-

1. The therapist cannot maintain rapport with ferences between these correlations are also
the patient significant at the 5 per cent level or above.)
4. The therapist's own needs completely inter- The findings suggest that therapists general-
fere with his understanding of the patient
51. The therapist feels disgusted by the patient ly agree on the most effective type of thera-
52. The therapist is hostile toward the patient peutic relationship, and that theoretical dif-
54. The therapist is punitive ferences are the result of poor communication
55. The therapist is very unpleasant to the pa- between therapists of different schools.
tient It may be objected that the statements here
142. The therapist acts in a very superior manner
toward the patient used would naturally yield high correlations
between persons since statements fall naturally
Category 1 — (Least characteristic of an Ideal
into.place. The fact yet remains that experts
Therapeutic Relationship)
2. The therapist shows no comprehension of the do have somewhat higher factor saturations,
feelings the patient is trying to communicate that only one general factor was found, and
that significant differences were found between
DISCUSSION OF THE RESULTS experts and nonexperts of the same school, as
In the two related studies which have here against correlations between experts of differ-
been presented, we have attempted to answer ent schools. Finally, two different sets of state-
two questions: ments were utilized, and a total of 16 different
1. Do theory and technique influence thera- persons rated the statements. About ten other
pists' beliefs about the nature of the ideal persons rated the ideal therapeutic relation-
therapeutic relationship ? ship on the basis of these statements. Although
2. Is the therapeutic relationship a unique these ratings were not incorporated into the
phenomenon which exists only within the factor design, the ratings thus obtained were
therapeutic situation? thoroughly consistent with those here pre-
sented.
Both studies have indicated that theoretical
allegiance to one system of psychotherapy does It may be objected that the statements do
not change the therapist's goal with respect not really describe therapeutic relationships.
to the relationship which he strives to create Yet, a subsequent investigation which is now
with his patient. The differences in factor under way [4] definitely indicates that thera-
loadings and correlations are not striking, and peutic relationships can be assessed by means
suggest that it is expertness which determines of these statements by outside observers on the
the type of relationship which is set as a goal by basis of wire recordings of therapeutic inter-
therapists. One would have expected at least views.
some of the wide theoretical differences which The second question which this investigation
exist, for instance, between nondirective and posed concerned the uniqueness of the thera-
Adlerian therapeutic theories and techniques peutic relationship. Since the naive raters are
to be reflected in the assessments obtained practically indistinguishable from therapists on
from therapists of these schools. Yet, as in the the basis of their ratings, it is difficult to see
first study, the two nondirective therapists, how such high agreement could be obtained if
ND-7 and ND-8 correlate significantly higher the therapeutic relationship actually were a
(at the 5 per cent level or above) with the unique phenomenon. These data rather support
Adlerian therapist IP-11 than with the non- the hypothesis that a good therapeutic rela-
expert of the nondirective school, ND-9. Sim- tionship is very much like any good interper-
ilarly, they correlate more highly with the sonal relationship. The latter is, no doubt, dif-
better trained psychoanalytically oriented ther- ficult to find in our society, yet one toward
apist in this study than with the nonexpert of which our lay subjects can and do aspire.
their own school. On the other hand, the psy- The general factors which emerged from
choanalytically trained therapist PA-1 corre- the correlation matrices do not seem to be as-
lates more highly with the Adlerian and the sociated with the therapist's theory, methods,
nondirective experts than with the psychoan- or academic training. The only conclusion
alytically oriented therapists PA-3 and PA-4, which we can draw from this is that all thera-
 CONCEPT OF IDEAL THERAPEUTIC RELATIONSHIP 245
pists and naive subjects had essentially the same 2. The ability to describe this concept is
ideal in mind. probably a function of expertness rather than
These studies leave us with several problems theoretical allegiance.
which remain to be solved: 3. Nontherapists can describe the ideal ther-
1. Do therapists of the various schools ac- apeutic relationship in the same manner and
tually create an essentially similar relationship about as well as therapists. The therapeutic
with their patients? relationship may therefore be but a variation
2. Are the experts better able to approxi- of good interpersonal relationships in general.
mate this generally held ideal than the less ex- 4. A composite rating of the Ideal Thera-
pert are capable of doing? peutic Relationship can be obtained by means
3. Is a good interpersonal relationship really of pooled ratings. This composite rating has
like a good therapeutic relationship? here been presented.
An investigation by the writer is now under Received September 20, 1949.
way to answer the first two questions.
REFERENCES
SUMMARY AND CONCLUSIONS 1. ADLER, A. Individual psychology, translated
by P. Radin. London: Kegan Paul, 1929.
Two investigations of Q-technique design 2. ESTES, S. G. The therapeutic relationship In
were undertaken to ascertain whether ther- the dynamics of cure. J, consult. Psycho!.,
apists with divergent theoretical views and 1948, 12, 76-81.
therapeutic techniques also differed in their 3. FIEDLER, F. £. Factors operative in shaping
concept of an ideal therapeutic relationship. It the concept of the ideal therapeutic relation-
was further attempted to answer the question ship. Paper read at the Midwest Psychol.
Ass., Chicago, April, 1949.
whether the therapeutic relationship is unique 4. FIEDLER, F. E. A comparative investigation of
to psychotherapy, or whether it resembles in early therapeutic relationships created by ex-
general a good interpersonal relationship. perts and non-experts of the psychoanalytic,
A number of therapists of different schools non-directive, and Adlerian schools. Unpub-
and with different degrees of training were lished Ph.D. dissertation, Univ. of Chicago,
asked to sort arrays of statements descriptive 1949.
5. FREUD, S. Vorletungen xur Einfuehrung in die
of therapeutic relationships into categories Psychoanalyse. Vienna: Internationaler psy-
ranging from the most characteristic to the choan. Verlag, 1922.
least characteristic. The ratings were then in- 6. HORNBY, KAREN, Are you considering psycho-
tercorrelated and the correlation matrices fac- analysis? New York: Norton, 1945.
tor analyzed. In both studies only one general 7. MURRAY, J. M., AND BROSIN, H. W. General
principles of psychotherapy. In H. L. Witmer
factor was found. The better trained therapists (Ed.) Teaching psychotherapeutic medicine.
of different schools agreed more highly with New York: Commonwealth Fund, 1947. Pp.
each other than they agreed with less well 273-288.
trained therapists within their own school. 8. ROGERS, C. R. Counseling and psychotherapy.
Naive subjects were well able to describe the Boston: Houghton Mifflin, 1942.
ideal therapeutic relationship. 9. STEPHENSON, W, Methodological considera-
tions of Jungian typology. /. ment. Set., 1939,
The following conclusions could be drawn. 85, 185-205.
1. Therapists of different schools do not dif- 10. THORNE, F. C. Principles of directive coun-
fer in describing their concept of an ideal seling and therapy. Amer. Psychologist, 1948,
therapeutic relationship. 3, 160-165.