A study of several useful tests for ordered alternatives in the randomized block design

This study investigates the small sample powers of several tests designed against ordered location alternatives in randomized block experiments. The results are intended to aid the researcher in the selection process. Toward this end the small sample powers of three classes of rank tests — tests based on ‘within-blocks’ rankings (W-tests), ‘among-b locks’ rankings (A-tests), and ‘ranking after alignment’ within blocks (RAA-tests)— are compared and contrasted with the asymptotic properties given by Pirie (1974) as well as with the empirical powers of competing parametric procedures.     

Causes of death contributing to changes in life expectancy in New York City between 1983 and 1992

Recent changes in life expectancy among race and sex groups in New York City were evaluated by analyzing the relative effects of different causes of death in 1983 and 1992, a period in which life expectancy at birth declined by 1.1 years among white males, remained unchanged among black males, and increased 1.2 years among white and black females. Heart disease was found to be the leading cause of death making positive contributions to changes in life expectancy regardless of race or sex, and HIV/AIDS was the leading negative contributor. Overall, deaths from infectious diseases and external causes are becoming more important compared to degenerative conditions in explaining trends in life expectancy in New York City. Past improvements in survival due to reductions in infant deaths are being reversed due to an increase in deaths from preventable causes such as violence and AIDS. Future gains in longevity may require a greater emphasis on policies and programs emphasizing conflict resolution and HIV prevention.     

Designing completely randomized experiments to detect ordered treatment effects: an empirical study

In designing experiments the researcher frequently must decide as to how to allocate fixed resources among k factor levels (Cox (1958)). This study investigates the effects on the power of a test caused by changes in: the sample size (n); the number of factor levels (k); the allocation of fixed total observations (N) among k and n: the shift parameter (ø); the type of parent population sampled; and, the type of ordered location alternative involved. Using Monte Carlo methods the powers of eight test procedures specifically devised to detect ordered treatment effects under completely randomized designs were evaluated along with those of the more general one-way F test. The results are of interest to researchers in all fields of application.     

The Role of Parent and Peer Support in Predicting Adolescent Depression: A Longitudinal Community Study

This study examines whether perceived parent support, peer support, and the interaction between them predict depression symptoms and depression diagnosis 2 years later in a community sample of 389 adolescents. Controlling for Time 1 depression, parent support and anticipated peer support were not independently related to Time 2 depression in either linear or logistic regression analyses. However, there was a significant interaction between the two support variables, suggesting that parent support moderates the relationship between anticipated peer support and depression symptoms and diagnosis. Anticipated peer support is protective among adolescents with high parental support, but may act as a risk factor for adolescents with low parental support. Regarding developmental differences, low anticipated peer support at Time 1 was a stronger predictor of Time 2 depression symptoms among older, compared with younger, adolescents. These findings highlight the importance of parent and peer support in predicting future depression among community adolescents.     

Some useful nonparametric tests for ordered alternatives in randomized block experiments

Testing for ordered alternatives in randomized block designs has been a problem of interest for almost three decades (Jonckheere (1954)). Three classes of rank tests have evolved—tests based on “within-blocks” rankings (W-tests), tests based on “ranking after alignment” within blocks (RAA-tests), and tests based on “among-blocks” rankings (A-Tests). This paper focuses on the latter. A simplified version of the Skillings-Wolfe generalized Purl test (1977) is suggested and two very useful A-tests—a generalized Johnson-Mehrotra “Optimal contrast” procedure and a generalized Tryon-Hettmansperger rank test—are developed. These procedures are compared and contrasted with other recent competitors presented by Skllllngs and Wolfe (1978) and by Salama and Quade (1981).     

Absent Fathers as Providers: Race/Ethnic Differences in Support for Adolescent Mothers

Race/ethnic differences in the prevalence of and characteristics associated with low support provided by fathers of babies born to pregnant adolescents were examined among 719 adolescent mothers who were interviewed postpartum. Low support was reported by 41.9% of African-American, 26.2% of Mexican-American, and 19.4% of Caucasian adolescent mothers. Across race/ethnic subgroups, fathers who provided limited support were characterized by chronic substance use, gang or police involvement, or partner-directed violence. Policies advocating universal father involvement without concomitant programs designed to address their many needs may ultimately be detrimental to the welfare of mother and child.