A new procedure for testing equivalence in comparative bioavailability and other clinical trials

The problem of testing for equivalence in clinical trials is restated here in terms of the proper clinical hypotheses and a simple classical frequentist significance test based on the central t distribution is derived. This method is then shown to be more powerful than the methods based on usual (shortest) and symmetric confidence intervals.

We begin by considering a noncentral t statistic and then consider three approximations to it. A simulation is used to compare actual test sizes to the nominal values in crossover and completely randomized designs. A central t approximation was the best. The power calculation is then shown to be based on a central t distribution, and a method is developed for obtaining the sample size required to obtain a specified power. For the approximations, a simulation compares actual powers to those obtained for the t distribution and confirms that the theoretical results are close to the actual powers.     

Examining the impact of family services on homeless children

This study examined the impact on children of services delivered through a collaborative initiative designed to provide supports to working poor families who are homeless or at immediate risk of becoming homeless. Overall, a number of positive findings can serve to inform the development of promising intervention strategies for children in homeless families and highlight areas for further research. Along with improved housing status, there was a statistically significant improvement in the economic self‐sufficiency of the families served, regardless of whether they completed the full programme cycle. In addition, youth self‐report data indicated significant improvements in externalizing and internalizing behaviour. The ability to combine programme data with school data for participating children also made it possible to explore school performance outcomes. Although these results did not indicate a change in participants’ levels of attendance or reading performance, there was a statistically significant improvement in students’ math scores. Moreover, when participating children were compared with a group of students matched on their level of academic performance at baseline who had not received housing or homeless services, no significant differences were detected between groups post‐intervention, indicating that services may help reduce the risk of homeless students’ falling further behind their peers academically.     

A bayesian approach for testing periodicity of dichotomous observations

A Bayesian approach is utilized to test for periodicity in a dichotomous time series. Dichotomous data arise in a variety of circumstances when a variable takes on only two possible values. Conjugate and noninformative priors are considered as well as a hierarchical Bayes approach; the latter is considered the superior Bayes methodology. The situation of stochastic period lengths is also discussed. The generalization to the multinomial model is investigated to allow for the case that a variable takes on more than two possible values. In all cases decisions are made based on a Bayes factor. The proposed procedures are demonstrated on earthquake data in the central Virginia seismic zone     

College students' living arrangements and attitudes toward family roles: Demographic and value correlates

Using data from the college population, Pearson correlations and regression were used to examine three hypotheses. The major hypothesis was that “conformity-oriented” values are associated with college students' traditional living arrangements and family-related attitudes, while “self-directed” values are associated with nontraditional attitudes and behavior. Two other hypotheses focused on interrelationships among demographic variables, values, family-related attitudes, and living arrangements. General support was found for the major value-orientation hypothesis and mixed support for the secondary hypotheses. Comparing these findings with those of several other studies, we propose that there may be a subset of human values associated with general tendencies toward confomity and resistance to change on the one hand, and experimentation and receptivity to change on the other. A typology of such value orientations is put forward for future research in areas of sex roles, social change, sociopolitical conflict, and attitude and/or behavior change.