THEORY AND RESEARCH ON THE NATURE AND ASSESSMENT OF HUMOR
Despite the wide use of humor tests, the state of the art of assessment of humor appreciation has not developed very far. In most of the investigations, the "humor test" used is an ad hoc measure of unknown psychometric properties. Most frequently a couple of jokes and cartoons was employed, which were selected by several "experts" as being representative for certain theoretical humor categories. Neither an empirical test of the homogeneity of these scales was
applied nor an investigation of the validity or the comprehensiveness of the categories used was undertaken. Most frequently, the Freudian categories of harmless, sexual, and aggressive humor were used. Some of these humor measures were employed by one author only, or even worse, by one author in one study only. Consequently, results obtained in different studies were not comparable and thus this approach did not lead to much accumulation of knowledge on the nature of humor.

A Pioneer
Other instruments are based on more careful and solid work. In this tradition, factor analysis was used to derive an exhaustive taxonomy of humor. Next guided by psychometric principles scales were designed and validity studies were undertaken. As an example of this approach the IPAT Humor Test of Personality (Cattell & Tollefson, 1966) deserves to be mentioned. This humor test is based on factor analytic studies of humor by Cattell and colleagues as well as by other researchers (Andrews, 1943; Cattell & Luborsky, 1947; Eysenck, 1942, 1943; Yarnold & Berkeley, 1954). Cattell arrived at a taxonomy of humor consisting of 12 factor analytically derived categories which should not only provide a profile of an individual's humor but also an assessment of dynamic or temperamental dimensions. For example, humor factor 1 ("anxious considerateness vs. debonair sexual and general uninhibitedness") is considered to be synonymous with the second order questionnaire factor of extraversion. This assumption is based on the correlation of this scale with the 16PF markers of extraversion, A (cyclothymia), F (surgency), and H (venturesomeness). In addition to the 12 humor factors, a further scale (factor 13: dullness vs. general intelligence) was added to measure, roughly, general intelligence. These items were initially selected on the basis of their correlation with intelligence measures.
The final version of the IPAT humor test consists of two forms: Form A, employing eight pairs of jokes for each of the 13 factors (104 items) and utilizing a "forced choice" design; and Form B, comprised of 10 items per factor to be rated individually (for a total of 130 items).
Whereas in Form A, subjects indicate which of the pair is funnier, in Form B they mark whether they consider the joke or cartoon as funny or dull (i.e., above or below their average) after being instructed to use each of the two labels about equally frequently. For both forms there is a test booklet containing the jokes and cartoons and a separate answer sheet. Separate norms for male and female high school and college students are added. The test is considered to be suitable for individual and group administration. For the latter there is the instruction not to disturb others by laughing aloud or attempting to share a joke. Split-half coefficients are between .18 and .50, stability coefficients for an interval of two weeks are between .20 and .58 and the correlations between the two forms range between .19 and .43.
Despite the efforts spent for the construction of the test, the IPAT humor test today plays a minor role in contemporary research on both personality and humor. Although the humor test was used in several studies (e.g., Breme, 1976; Carroll, 1989; Mones, 1974; Saper, 1984) its overall impact is low and seems to be obsolete. Considering contemporary knowledge on the assessment of humor, one can assume that the selected "forced choice" answer format is responsible for the problems with the IPAT humor test, such as the apparent overextraction, preoccupation with minor sources of variance, neglect of structural characteristics in humor, lack of reliability, or unproved validity. Cattell and Tollefson applied these answer formats because they assumed that the shape of the profile (i.e., the preference for certain humor
categories) already contains the relevant information whereas the interindividual differences in the level of the profile are irrelevant. However, this format eliminated the strong factors and thus probably hindered the discovery that the first factors appearing in factor analysis of jokes and cartoons are characterized by the items structure and not by their content. Ruch (1981, 1984) as well as Herzog and Larwin (1988) report the existence of one or two factors appearing which deal with a variety of themes but are characterized by similar structure. The common denominator of these jokes and cartoons refers to similar cognitive processes rather than to thematic content. Whereas the format hindered the extraction of major humor factors, it raised the relative importance of minor sources of variance, which apparently are less reliable as reflected in the coefficients reported above. The absence of strong factors impaired the distinction in major and minor sources of variance. The overextraction can be seen in the facts that in subsequent studies between three and four meaningful factors were extracted and that with the possible exception of the sexual humor category no other of Cattell's factors was replicated. The answer format also seems to be responsible for a further specifity of the IPAT humor test, the bipolarity of the humor dimensions. There is a convergence in the results of the studies from Andrews (1943) onwards. Whenever there is no restriction in the number of items one is allowed to find funny, the overwhelming majority of the intercorrelations between pairs of jokes or cartoons are positive and negative coefficients seldomly exceed the value of -.20. As a consequence only unipolar factors appear and tend to be orthogonal or positively correlated but never negatively correlated.
Despite the persuasive arguments favoring humor tests as a means for the objective assessment of personality the IPAT humor test has not found widespread use in personality research. Maybe this is due to the failure of Cattell and Tollefson to present convincing tables of correlation coefficients showing that the relationship between humor and personality is strong and replicable. From the perspective of humor research the explicit link to theory is also lacking.
The work of Cattell and colleagues was stimulated by the Freudian (1905) hypothesis that repressed needs find relief in jokes. Thus, they expected that application of factor analysis will help to find such repressed areas in the realm of humor. Whereas some factors (e.g., Factor 5: urbane pleasantness vs. hostile derogation) relate to sex or aggression and thus have face-validity, others (e.g., Factor 10: cheerful independence vs. mistreatment humor) are less easily
integrated in taxonomies of needs. Thus, there are no apparent links between humor categories measured by the IPAT humor test and general humor theories, especially since influence of Freudian thinking on humor research ceased and alternative theories were developed (see McGhee, 1979).

The Current Standard in the Assessment of Appreciation of Humor
Although the IPAT humor test now seems to be obsolete, its basic ideas have been taken up and pursued by more recent humor tests. The use of factor analysis to derive a taxonomy of humor and the belief that the appreciation of humor reflects aspects of personality continues to inspire research. However, the investigation of individual differences in humor is presently more integrated into general humor research than in the pioneer years. There is a fruitful mutual
exchange between general humor theories and taxonomic studies of humor.
Ruch (1980) proposed that a comprehensive assessment of humor should not only cover a taxonomy of humor stimuli but also an investigation of the dimensionality of the responses to humor. The taxonomy of humor stimuli was achieved by a set of factor analytic studies of differing but overlapping sets of jokes and cartoons. In order to get a robust taxonomy samples differing with regard to sex, age, occupation, health status and other variables were used. Most
importantly, the first construction samples covered Austrian as well as German subjects. After establishing the taxonomy the items were translated into English, French, Hebrew, Russian, and Turkish in order to be able to test its cross-cultural stability. Similarly, the dimensions of appreciation were obtained by correlational and factor analytic studies of several rating scales covering different aspects of the responses to humor.