Chapter 30.    Factor Analysis: Dimensions of Intoxication

ALL THE DATA on marijuana intoxication presented so far have been based on how experienced users describe their experiences; i.e., it is primarily descriptive. One naturally wonders if there are more basic dimensions of the intoxication experience that could account for the many different specific effects, that would reduce many effects to a smaller, more basic number.
Factor analysis is a statistical technique that begins to answer such questions. All the different items of information are correlated with each other; a factor analysis then ascertains whether some sets of items form natural groupings that might represent more basic dimensions.
The data format of the present study is not well suited to factor analysis; the five-point frequency scale is rather limited, is not normally distributed, and does not constitute an interval or ratio scale. Thus the data given below are the weakest of the present study and are presented only for the sake of completeness and the guidance of other investigators.
In order to reduce the number of items to a level the computerized factor analysis program of the University of California at Berkeley's Computer Center could handle, every other item, starting with Q1, was selected. This included two items (Q67 and Q167) that were not answered in the same form as the others and so are not considered in interpreting the factors. Thus the analysis is based on 104 frequency of occurrence items.

RESULTS

The principal components analysis revealed one main factor (Eigenvalues for the first twelve factors were 13.953, 4.842, 3.956, 3.489, 3.191, 2.927, 2.758, 2.673, 2.443, 2.377, 2.255, and 2.171). Items loading .400 or greater on this first factor are presented in Table 30-1. The usual type-style code for overall frequency of occurrence is used. AS principal component rotations are primarily of mathematical rather than psychological interest, no interpretation of this factor will be made.
A Varimax rotation for 12 factors revealed one small-sized factor and eleven others that were not too much smaller. The proportion of the total communality accounted for by each factor was .078, .048, .044, .035, .033, .032, .031, .030, .028, .028, and .025. Each factor is described in Table 30-2.
Factor I seems to consist mainly of feelings of competence, perceptiveness, and intuitive-archetypal approaches to the environment. We might call it "Being High," as it fits many descriptions by users of the virtues of marijuana intoxication.
Factor II reflects enhancement of touch, taste, and smell sensations and imagery. It has been tentatively called "Enhancement of Nondominant Senses," as these senses generally play a minor role compared to vision and hearing in most of our transactions with the world. One could also consider these as close-up or "intimate" senses, as contrasted with the functioning of vision and hearing at much greater distances.
Factor III deals primarily with increased awareness of various internal processes, such as body tensions, dreams, feelings of the location of consciousness, and thoughts. It has tentatively been named "Enhanced Internal Awareness."
Factor IV consists of items describing increased absorption in internal processes and a (consequent) loss of contact with the external world. It has tentatively been called "Internalization of Awareness."
Factor V does not Seem to have a common theme, and Factor VI has only three items loading heavily enough to define it. It seems to represent unpleasant, dysphoric effects. Factor VII seems to represent perceptual instability in the visual system.
Factor VIII does not show any clear pattern, unless it he memory decrement. Factor IX also seems to represent memory decrement, although it is defined by only two items. The remaining three factors that were analyzed for in the Varimax rotation show no particular patterns that can readily be named.