Gerald S. Kenyon

A Conceptual Model for Characterizing Physical Activity

Abstract It is the purpose of this paper to report efforts to construct a model characterizing physical activity as a sociopsychological phenomenon. The work was based upon the assumptions that physical activity can be reduced to more specific components, i.e., a set of all physical activities can be reduced to logical subsets; and that a meaningful basis for such a procedure is the instrumental value physical activity is perceived to have for the individual. A rationale is given for each of the six dimensions of the model. In addition to construct validity, the integrity of the model was assessed by determining the degree of univocality and independence of the postulated subdomains. I

Six Scales for Assessing Attitude toward Physical Activity

Abstract Based upon a multidimensional model for characterizing physical activity, the development of two forms (college men, college women) of an attitude inventory is described. Items held to be representative of the six dimensions of the model were evaluated on the basis of factor and item analyses using data generated by the use of preliminary forms. Hoyt reliabilities ranged from .72 to .89 for the six scales. Scales were able to differentiate between appropriate high and low preference groups for a particular type of activity in all cases except physical activity perceived as “catharsis.” The stability of the instrument is shown by comparing measures of central tendency, variability, and reliability of the basic sample with those drawn from a second, but similar population. Normative data, instructions, and a list of items, are provided.

Effects of Short Periods of Physical Training upon Body Weight of White Mice

Abstract Fifty-one white mice were randomly assigned to three groups. The animals in group 1 were exercised daily on a motor-driven treadmill for eighteen successive days. The mice in group 2 were required to swim daily for eighteen successive days. The members of group 3 received no special exercise. Using covariance procedure, the adjusted mean body weights of the three groups were compared. It was concluded that, despite the ad libitum provision for food, the body weights of the exercised mice were lower than those of the sedentary group. Further, the animals who swam weighed less than those trained on the treadmill.

Multiple Comparisons and the Analysis of Variance: An Empirical Illustration

Abstract To illustrate the multiple comparisons problem in experiments involving three or more groups, 12 samples of 25 observations each were randomly drawn from a population of 1,000 measures. When the significance of differences among the 12 means was tested with a one-way analysis of variance, the null hypothesis was clearly tenable. However, when t tests were used to test the significance between each pair of the same means, the null hypothesis was rejected. This result was not unexpected in that there was no attempt to distinguish between “error rate per experiment” and “error rate per comparison.” Since the probability of committing a Type I error increases as a function of the number of comparisons, it behooves the investigator to account for this phenomenon through the careful formulation of experimental hypotheses and the use of appropriate multiple-comparison procedures.