The study reported here tried to answer the question to what extent judgments about personality derived from dream series alone agree with those derived from the same dream series when free associations to the dreams are also evaluated. Before this work is described in detail, some underlying assumptions will be discussed.
The traditional and most frequently used technique to interpret dreams since the publication of Freud's book on dreams in 1900 is to collect free associations to the dream as a whole, its parts, and the feelings in the dream. From these data the dream analysant and analyst in a cooperative enterprise try to decipher the motivational sources of the dream.
While perhaps not experimentally "proven" this technique has shown its usefulness in clinical practice so repeatedly that it has become part of the working equipment of most psychotherapists, though they may differ in some details of usage.
Though this technique is used widely in psychotherapy it does have certain shortcomings, especially when the objective is not psychotherapy but limited to the understanding of personality dynamics. In the latter instance, the collecting of free associations to dreams is often not economical in terms of time and effort on the part of subject and investigator since less time consuming techniques of personality diagnosis, (Rorschach, for example) are available. Another disadvantage with free association is that people vary a great deal in their aptitude for free associations. Therapists and investigators too, differ in their aptitude to elicit free associations.
Thus it would be worthwhile to investigate how much can be learned about the personality of people solely from their dreams. But over and above these practical considerations the question whether dreams alone can reveal the personality of the dreamer is a significant theoretical problem.
There has been in recent years a more open interest in manifest dreams as indicated by published material. Erikson's  approach to the manifest dream consists of a careful examination of certain formal aspects such as partial and emotional content of the manifest dream. It must be noted that the subject of the study here is the individual dream.
A somewhat different approach was made by Hall  who studied dream series as a whole. His technique was to collect a number of dreams from subjects and then examine them as a unit, starting with a "spotlight dream" i.e., a dream with important obvious significance, and fitting the rest together in as internally consistent a picture as possible. This method is not unlike that used in putting together a jigsaw puzzle where one usually starts with a prominent piece and then fits the rest around it. The criterion is how well the pieces fit together and whether they make a meaningful picture. In this method the degree of internal consistency is the criterion of validity of the interpretation unless other data such as tests are available. The assumption underlying the acceptance of internal consistency as criterion for validity is that there is only one optimal fit which represents the correct understanding of the meaning of the dream series. Hall used in his work a fairly orthodox Freudian theoretical framework.
The Problem and the Subjects
The question immediately arises to what degree does Hall's method of dream interpretation produce results comparable to those achieved by the use of dreams and free associations? Specifically, to what extent in qualitative and quantitative terms does the picture of a dreamer's personality, as derived from a series of dreams alone, agree with that derived from the same dreams if free associations to these dreams are also evaluated? This was the problem of the study. It was carried out in partial fulfillment of the requirements for the Ph.D. degree in psychology by the writer at Western Reserve University from June, 1949, through September, 1950, under the supervision of Dr. Calvin S. Hall. It should be noted that the validity of the personality picture derived from dreams and free associations is not an issue of this study, but for the purposes here is taken for granted.
The subjects were 24 college students, 12 male and 12 female, who volunteered for the study. The age range was 19 through 27 for the males, 19 through 21 for the females. The mean age of the total sample was 20.7 years, being 21.5 years for the boys and 19.9 for the girls. Thirty-seven and one-half per cent of the subjects were psychology majors, 42.7% of the males and 33.3% of the females.
Design of the Study
The study consisted of the following parts:
This research plan will now be discussed in detail.
The Rating Scale
The writer had to develop his own rating scale because existing scales were not adequate for the purpose. The rating scale should reflect the psychoanalytic personality theory since this is a theoretical framework within which this writer thinks. The rating scale should also be adapted as much as possible to the way a practical clinical psychologist operates in his daily work. Specifically, the clinician does not think in terms of isolated traits, but in terms of the total personality. For this reason the categories should be as global as the thinking of the psychoanalytically oriented clinician who describes personality in terms of orality, oedipus complexes, etc. Lastly, since this writer feels that reliable global judgments are possible, provided raters are taught the technique in personal training sessions, rather than by a scoring manual, the units of the rating scale should reflect a fairly high level of abstraction. The following rating scale while far from perfect seemed more suitable for the purposes of this work. (See Table I: Rating Scale)
To demonstrate the point that global judgment of this nature can be made reliably from dream series alone as well as from dream series with free associations co-workers Lane and Johnson-Dyer were instructed in the technique of rating dream series alone and dream series with free associations respectively on this rating scale, and their ratings compared statistically with those of the writer. The results of these reliability studies are reported in the next section. Lane, who has since then received her Ph.D. in psychology, was at the time a graduate student in psychology at Western Reserve University. Johnson-Dyer was at the time an undergraduate with a major in psychology at Western Reserve University.
The rating of dreams with and without free associations was done in the following manner:
It took, on the average, four to five hours to rate a dream series.
It is estimated that it took, on the average, eight to ten hours to rate an entire dream series and associations.
The Study Proper
As written down here, the process seems mechanical. Nothing could be farther from the truth. If one were to conduct the session in a mechanical fashion, the response, for example, to the stimulus word, "chair," would be associations like "table," "desk," "couch," etc. However, this investigator was not after associations per se, but after meaningful associations, i.e., associations that fit into an internally coherent personality picture. To this end the subject was at all times urged to integrate the material produced. Much of the time of each session consisted more in a dream inquiry than in the production of associations. When there was a block in the dream inquiry, the emphasis was shifted to free association. In some instances the attempt to integrate interfered with the production of free association. In these instances the subject was urged to forget about the dream temporarily and was later guided back to a more integrative set.
Underlying all the work with the subject was the attitude, which was carefully implanted and fostered, that he, and only he "knew," in some recess of his mind, the meaning of the dream and that he was the one to discover it. He was told that a dream is comparable to a stage production, of which he is the director, and that he was the one to answer all questions of the why and how of the production, granting that this might involve finding it out by roundabout ways. If the subject did offer an interpretation, the attitude of the investigator was one of skepticism, which was expressed as follows: "Is that what it means?" "How do you know?" "Let's see why." "Can you prove it?"
The responses of the subjects to their dreams were recorded as faithfully as possible by the investigator in longhand. The recording turned out to be by far the most thorny part of the whole study.
Many subjects used the sessions for discussion of their personal problems, though this had not been suggested, at least not consciously. The attitude on the part of the investigator which made subjects feel free to discuss their problems is considered a necessary condition to elicit the quality of material which the investigator wanted to and did obtain from the subject. If the investigator had indicated more or less subtly, "We have only time for dreams," or, "Let's discuss this later," it would have meant a great loss in the intensity of rapport and a concomitant reduction of the depth and meaningfulness of the material produced.
After the collection of the free associations to the ten dreams, the investigator rerated the dream series on the personality scale in the light of the free associations.
In this section, Study I will refer to the first reliability study carried out by co-worker Lane by the rating of twenty-four comparable but different dream series prior to the study, and comparing her ratings with those of the investigator on the same data. Study II will refer to the reliability study carried out by co-worker Johnson-Dyer, by rating of ten dream series (six male and four females) and pertaining associations which are part of this study, and comparison of her ratings with those of the investigator based on the same material. Study III will refer to the comparison of the two sets of ratings of the investigator on the twenty-four subjects (twelve males and twelve females) based on the dreams alone with those based on dreams with free associations.
The concept of reliability, as it applies to this thesis, is discussed in detail in Lane's  paper. We shall, therefore, be brief in our treatment.
Each of the co-workers rated each case on all twenty-five ratings, i.e., they were forced to make a rating regardless of whether they felt sufficient evidence was available for a rating. The investigator rated only in those instances where he felt a rating was justified by evidence. Thus, each co-worker had on each item of the scale three choices, "a," "b," and "c," while the investigator had four choices, i.e., the same as the co-workers plus "no rating."
A complete agreement between two ratings like "a"-"a," "b"-"b," or "c"-"c," was counted one point (1.00). A one step disagreement like "a"-"b," "b"-"a," "b"-"c," or "c"-"b," was counted half a point (0.50). Total disagreement like "a"-"c," or "c"-"a," was counted zero point (0.00).
Table I gives the rating scale and numbers assigned to each item on the scale.
Tables with the Quantitative Results
Table II gives the complete results of Studies I, II, III. In this table the column headings and abbreviations have the following meanings:
Discussion of the Results
In Table II we have given the levels of confidence with which the chance hypothesis can be rejected. The underlying assumption is that all ratings (a,b,c) have an equal probability of occurring, i.e., the rater has no preference for a particular rating. This assumption might be challenged on the grounds that, for example, ratings "a" and "c" do not occur at all on item one, which is incidentally the only item on which this happened. Of course, it is perfectly true that this item does not differentiate, and cannot for that reason be used for that purpose. It would have been very unwise to force the statistical curve on the data, which would have not been difficult to do. As a matter of fact, during the study the raters never investigated how often they had use t e various ratings (a,b,c), but made the judgment which seemed to fit the case best. Actually the universal rating of "b" in this case is very meaningful, suggesting that our sample as a whole had unresolved Oedipus complexes. In other words, the investigator holds that the absence of ratings "a" and "c" in this case is not due to a reluctance on the part of the rater to. use these categories, but to the distribution of the characteristics in question in our sample.
As far as changes in the sense of addition are concerned, Study III offered opportunities for six hundred ratings (twenty-five ratings on twenty-four subjects). After reading of the dreams alone the investigator made five hundred fifty-four or 92.33 per cent of the possible ratings. After additional study of the associations he made five hundred eighty-five ratings, that is, 97.50 per cent of the possible ratings, or an addition of 5.17 per cent to the ratings made on the study of the dreams alone.
This shows that after study of the free associations the judges were not able to rate many more variables than they had been rating after study of the dreams alone. If the objective of collecting free associations to dreams had been to permit more variables to be evaluated the small number of additional variables which could then be rated would not be considered a worthwhile return for the time and effort invested.
The ratings do not suggest that there were systematic differences in the percentage agreements on male subjects versus female subjects either in Study II or III.
Even if all the items were rated reliably, knowledge of all the ratings on a subject would not reveal enough about the individual to give the reader an adequate picture of the subjects personality. More specifically, the rating scale is not adequate to give a rounded picture on two counts:
While there was good agreement between the ratings on dream series alone and those on dream series with free associations, it must not be assumed that the collection of free associations is not useful for the understanding of the personality. An infinitely richer picture of the personality, so to speak in flesh and blood, was derived after the study of the free associations. From the dreams alone, a shadowy blueprint of the personality with many ambiguities developed before the dream interpreter's eyes, while the addition of the free associations gives one the feeling of knowing the person, seeing him clearly delineated in the matrix of his environment.
From the work with these dreams and associations many possibilities for future research were suggested....
Summary and Conclusions
The personalities of twenty-four college students were rated on a scale of twenty-five items from dream series of ten dreams each, with and without free associations.
It was demonstrated by the study proper (Study III) and the reliability studies (Studies I and II), (the latter carried out by two raters other than the investigator) that: (a) many ratings can be judged reliably from dream series alone; (b) many ratings can be judged reliably from dreams and free associations; (c) and many of the ratings derived from dreams series alone agreed with those derived from dreams with free associations; (d) dreams alone suggest the blueprint of the personality while free associations add the brick and mortar, so to speak.
Go back to the Dream Library index.