CHAPTER III

PRESENTATION AND ANALYSIS OF DATA

Description of Subjects

     There was an equal number of male (18) and female (19) subjects.  None of the subjects had had previous psychotherapy, dianetic or otherwise.  The average age of the subjects was thirty-four years, with 70% falling between twenty-two and forty-seven years.
     Half of the group had either started high school or graduated from high school.  The other half had completed at least two years of college, and two of the eighteen had Master of Arts degrees.  These were distributed approximately equally in each of the three groups.
     From the average scores of the tests of intellectual functioning used in this study, it was noted that the average subject was placed in the eighty-fourth percentile of the general population.  The mean score for personality conflicts placed the average subject in the sixty-first percentile of the general population, which indicates a significant degree of personality disturbance according to the data published by Rotter
.

Intellectual Functioning

     The raw score of intellectual functioning on each of the two tests was obtained for each subject.  These raw scores were then transferred into standard scores by dividing the difference from the published normative mean by the standard deviation for the respective test.  The negative values were then coded out by the addition of a constant which was a number greater by one than the lowest value.  This same procedure was repeated for the results of the second testing session which were obtained after the therapeutic interval.
     For each subject, the difference between the average score on the first test and the average score on the second test was calculated.  The sign was kept so that a positive value stood for an increase in score.  These scores were thereafter coded so that negative values were obviated.
     The final result was a coded difference score of intellectual functioning for each subject.  These were tabulated in accordance with the experimental design and are presented in Table I.
     The division of a variable into sub-categories makes the assumption in sampling that although the means of the categories are significantly different (by construction), the sampling of these categories should be from a common population.  Since, in this study, no controls, other than the variables themselves were used, randomness is a tenable a priori assumption.  However, there is always a possibility that a statistical artifact has occurred in non-randomness of the scores of the subjects who were randomly selected.  This must always be tested first before the variance of the scores is analyzed, since it is a necessary condition of the analysis.  The test of homogeneity was performed for each of the major variables and the results are given in Table II.  From Table II it is noted that the probability levels for the first three variables (therapy, education, and age) indicate a sufficient degree of tenability for the hypothesis of homogeneity between the subcategories.
     However, the scores of random sequence are significantly different among the sub-categories of first, second, and third choice for each of the twelve sets of conditions.  It should be remembered that the subjects were assigned randomly to their positions in the sets of conditions.  Thus, a lack of randomness in the scores of these subjects is most likely a statistical artifact rather than a reflection of the systematic influence of some extrinsic factor.  In this case, following common procedure,1 an attempt was made to reduce the artifact in the direction of homogeneity by transformations of the data.
     The first of these transformations— indicated in Table II by random sequence (2)— utilized the square roots of the data.  This was chosen because it is recommended2 when the means and the variances tend to be proportional.  This condition obtained in these data.  However, the probability associated with this transformation still did not permit the hypothesis of homogeneity.
     The second transformation utilized logarithmic values.  This is also recommended when the means and the variances tend to be proportional.3  Again, the hypothesis of homogeneity was found to be untenable.
     Following the rationale of Peters and Van Voorhis4 for log logs which approximate the inverse sine, the third transformation utilized the inverse sines of the data.  But, once more, the hypothesis of homogeneity was found to untenable.
     Since the means were not only proportional to the variances but were also similar to asymtotic relationships, the reciprocal values were used for the next transformation.5  Still, the hypothesis was found untenable.
     Because none of these transformations had performed the necessary stabilization of variance, reduction of skewness and normalizing of the scores,6 the data were converted to "z" values under the normal curve.  This was tested— indicated in Table II by random sequence (6)— and the hypothesis of of homogeneity was found to be tenable.
     Using these tested data, the various combinations of variance were calculated.  The results are presented in Table III.
     From Table III it is noted that the only significant difference resides between the sub-categories of the age variable.  This demonstrates that, for this study, intellectual functioning varies with a change in age.  The average score for the younger group is 87.7 while the average score for the older group is 78.1.  The average age of the younger group in twenty-seven while the average age for the older group is forty-one.  Thus we find that persons of forty-one do not score on tests of intellectual functioning as well as persons of twenty-seven years of age.  Of course, this finding cannot be justifiably extended to the general population because this sampling is not representative of the general population— or, at least, this has not been demonstrated.  Since the scores on these tests of intellectual functioning are not corrected for age differences within the adult population, it is not totally surprising that there is a decrement after the age of thirty-five.  David Wechsler reports7 a decrement for the general population on his test of intellectual functioning between the ages of twenty-seven and forty-one which is 9 points.  The data from this study show a decrement of 9.6 points.  It seems that these findings are mutually supporting.
     Referring back to the first problem (Chapter I: Specific Problems) for which this study was designed to provide an answer, the data in Table III support the conclusion that there is no systematic influence exerted by dianetic therapy upon the level of intellectual functioning.  This evidence does not support that the claims for dianetic therapy in this area are true.  In addition, the data hold as unsupportable any claim that dianetics will favorably affect intellectual functioning, but this statement obtains only within the population which is sampled by this study (persons from twenty-to to forty-seven years of age who have had at least some high school education regardless of their sex).

Arithmetical Ability

     The raw score of arithmetical ability on each of the two tests was obtained for each subject.  These raw scores were then transformed into standard scores through the division of the difference from the published normative mean, by the standard deviation for the respective test.  The negative values were then coded out by the addition of a constant— a number greater by one than the lowest value.  This same procedure was repeated for the results of the second testing situation which were obtained after the therapeutic interval.
     For each subject, the difference between the average score on the first test and the average score on the second test was calculated.  The sign was kept so that a positive value stood for an increase in score.  These scores were coded so that negative values were obviated.  The final result was a coded difference score of mathematical ability for each subject.  These scores were tabulated in accordance with the experimental design and are presented in Table IV.
     Following the rationale presented in the discussion of the analysis of the scores of intellectual functioning, the data of Table IV were tested for homogeneity of variance.  The results are given in Table V.
     From Table V it was noted that the probability level associated with each of the major variables is of a degree sufficient to allow as tenable the hypothesis of homogeneity between sub-categories.  This satisfied the prerequisite condition for the analysis of the variance.  Using these tested data, the various combinations of variance were calculated.  The results are presented in Table VI.
     The data of Table VI provide an answer to the second problem (Chapter I: Specific Problems) which this study was designed to to investigate.  The data show no evidence of a systematic influence exerted by dianetic therapy upon the level of mathematical ability.  This evidence denies that the claims for dianetic therapy in this area are true.  In addition, the data hold as unsupportable any contention that dianetic therapy will favorably affect mathematical ability, but this statement holds only within the limits of the population which was sampled by this study (persons from twenty-two to forty-seven years of age who have had at least some high school education regardless of their sex).

Personality Conflicts

     The raw score expressing the degree of personality conflicts was obtained for each subject from Rotter's test for this characteristic.  This information was obtained again from the results of the second testing session which took place after the therapeutic interval.  For each subject, the difference between the first test score and the second test score was calculated.  The sign was kept so that a positive value stood for an increase in score (since the score itself is of an unfavorable characteristic, an increase in score is an unfavorable outcome).  These scores were coded so that negative values were obviated.  The final result was a coded difference score of personalty conflicts for each subject.  These scores were tabulated in accordance with the experimental design and are presented in Table VII.
     Following the rationale presented under the foregoing discussion of the analysis of the scores of intellectual functioning, the data of Table VII were tested for homogeneity of variance.  The results are given in Table VIII.
     From Table VIII it was noted that the probability level associated with each of the major variables is of a degree sufficient to allow as tenable the hypothesis of homogeneity between sub-categories.  This satisfied the prerequisite condition for the analysis of the variance.  Using these tested data, the various combinations of variance were calculated.  The results are presented in Table IX.
     The data of Table IX provide an answer to the third problem (Chapter I: Specific Problems) which this study was designed to to measure.  The data support the finding that there is no systematic influence exerted by dianetic therapy upon the degree of personality conflicts.  This evidence denies that the claims for dianetic therapy in this area are true.  In addition, the data hold as unsupportable any claim that dianetic therapy will favorably affect the degree of personality conflicts, but this statement holds only within the limits of the population which was sampled in this study (persons from twenty-two to forty-seven years of age who have had at least some high school education, regardless of the sex of these persons).


Table of Contents |  Index | Chapter IV


 1. A. Edwards, Experimental Design in Psychological Research, pp. 198-199.
 2. Ibid., pp. 199-200.
 3. Ibid., pp. 202-203.
 4. Statistical Procedures and Their Mathematical Bases, p. 438.
 5. Edwards, op. cit., p. 203.
 6. Ibid., p.199.
 7. Measurement of Adult Intelligence.  Baltimore: The Williams and Wilkins Company, 1944, p. 29.