Carol J. Feltz

AN ASYMPTOTIC TEST FOR THE EQUALITY OF COEFFICIENTS OF VARIATION FROM k POPULATIONS

The coefficient of variation is often used as a guide of the repeatability of measurements in clinical trials and other medical work. When possible, one makes repeated measurements on a set of individuals to calculate the relative variability of the test with the understanding that a reliable clinical test should give similar results when repeated on the same patient. There are times, however, when repeated measurements on the same patient are not possible. Under these circumstances, to combine results from different clinical trials or test sites, it is necessary to compare the coefficients of variation of several clinical trials. Using the work of Miller, we develop a general statistic for testing the hypothesis that the coefficients of variation are the same for k populations, with unequal sample sizes. This statistic is invariant under the choice of the order of the populations, and is asymptotically chi 2. We provide an example using data from Yang and HayGlass. We compare the size and the power of the test to that of Bennett, Doornbos and Dijkstra and a statistic based on Hedges and Olkin.

The Effects of Team-Based Functional Assessment on the Behavior of Students in Classroom Settings

We examined the impact of functional assessment interventions on both the appropriate and challenging behaviors of groups of students within preschool classrooms for children with special needs and for children at risk. We also examined the effectiveness of a training model to teach school-based teams to conduct functional assessment. The results indicated that school-based teams were able to conduct functional assessment during intervention and maintained functional assessment skills during follow-up observations. In addition, the functional assessment procedures resulted in a decrease in challenging behavior and nonengagement and an increase in active engagement and peer interaction for groups of students within classroom settings. The levels of appropriate and challenging behavior observed during intervention and maintenance within at-risk and special education classrooms were similar to those observed in early childhood control classrooms.

Asymptotic inference for coefficients of variation

Asymptotic inference results for the coefficients of variation of normal populations are presented in this article. This includes formulas for test statistics, power, confidence intervals, and simultaneous inference. The results are based on the asymptotic normality of the sample coefficient of variation as derived by Miller (1991). An example which compares the homogeneity of bone test samples produced from two different methods is presented.

Theory & Methods: Generalizations of the δ‐Corrected Kolmogorov‐Smirnov Goodness‐of‐Fit Test

This paper generalizes the δ-corrected Kolmogorov–Smirnov (K–S) test into a family of tests, and investigates the behaviour of some of the members of the family, comparing them with the usual K–S test, as well as the δ-corrected K–S tests suggested by Harter et al. (1984) and Khamis (1990, 1992, 1993). These investigations and power studies suggest which tests are most powerful for which alternative hypotheses.

Customizing Generalizations of the Kolmogorov-Smirnov Goodness-of-fit Test

This paper presents a method of customizing goodness-of-fit tests that transforms the empirical distribution function in such a way as to create tests for certain alternatives. Using the @ , g transform described in Blom(1958), one can create non-parametric tests for an assortment of alternative distributions. As examples, three new ( f , g )-corrected Kolmogorov-Smirnov tests for goodness-of-fit are discussed. One of these tests is powerful for testing whether or not the data come from an alternative that is heavier in the tails. Another test identifies whether or not the data come from an alternative which is heavier in the middle of the distribution. The last test identifies if the data come from an alternative in which the first or third quartile is far from the corresponding quartile of the hypothesized distribution. The behavior of the three new tests is investigated through a power study.