To meet the needs of a diverse student population while addressing the knowledge and skills outcomes of the curriculum, faculty should develop a repertoire of instructional models. Research-based instructional models from the fields of education and health care have been developed to support various learner outcomes. Emphasis on evidence-based practice in the field of communication sciences and disorders should include increased study of research-based instructional practices in higher education. The present study examined the models of instruction currently implemented by professors in speech-language pathology graduate programs. Problem-based learning and induction were used most frequently, followed by direct instruction. Cooperative learning, memory strategy instruction, and role play were used less frequently. Results suggest that instructors are utilizing a variety of instructional models to meet learning outcomes. Implications for the use of models of instruction, as well as future directions for research, are discussed.

Even after having taken an introductory statistics class, many students continue to struggle with the concepts of sampling distributions, sampling error, and statistical inference. As a first step in identifying ways of more effectively approaching these topics, the poster will examine methods currently used in popular text books and propose an alternative based on using binomial and Poisson distributions. The examination of textbooks will identify variability or lack of variability in the approaches to teaching these topics with particular attention being paid to the use of statistical language. The poster will also illustrate how exercises useful in teaching the three topics of sampling distributions, sampling error, and statistical inference.

Adverse reproductive outcomes (AROs) disproportionately affect black American infants and significantly contribute to the U.S. infant mortality rate. Without accurate understanding of AROs, there remains little hope of ameliorating infant mortality rates or eliminating infant health disparities. However, despite the importance of monitoring infant mortality rates and health disparities, birth record data quality is not assured. Racial disparities in the reporting of birth record data have been documented, and missing birth record data for AROs appears to be disproportionate. Due to the extent of missing birth record data, innovative strategies have been developed to evaluate relationships between maternal socioeconomic status (SES) and community-based ARO rates. Because addresses convey aggregate information about income level, education and occupation, ZIP codes, census tracts and census block-groups have been applied to geocoding efforts. The goals of this study are to: 1) analyze the extent of missing birth record data for New Jersey areas with high rates of an ARO (preterm birth), 2) evaluate associations between the extent of missing birth record data and other AROs, and 3) consider how geocoding strategies could be applied to provide a basis for understanding maternal SES risk factors and ARO resource allocation for at-risk communities.

This poster presented results from a survey of college faculty regarding expectations for students who have completed introductory students. The sample included faculty who are not statisticians or statistical educators. Nevertheless, their expectations for literacy are consistent with those of several statistical educators.

Graduate students in applied areas such as psychology and public health often collect data for their dissertation projects and are typically faced with constraints that result in their samples being viewed as mere convenience samples. In describing their sampling plan and its limitations, students often struggle to evaluate the quality of their sampling procedures. This poster takes the view that all samples of human participants are convenience samples to some degree, if for no other reason than ethical considerations make participation voluntary and financial limits make pure random sampling exorbitantly expensive. According to this view, psychology and other applied fields are disciplines built on convenience samples. In spite of the use of convenience samples, applied statistics and data analysis procedures are useful in making advances in applied research. Because some convenience samples may be better than others, this poster session will examine factors and issues in sample selection. The aim is initiate discussion that will result in a framework that graduate students can use to address how to generalize their results.

The Consortium for the Advancement of Undergraduate Statistics Education (CAUSE) and the International Statistics Project emphasize the need for statistical literacy. Definitions of literacy tend to be general statements that need to be expanded to address specific skills. This poster presented a draft survey for identifying key specific elements of literacy. The intent of the surveys will be to make comparisons of expecta- tions held by faculty and by employers of recent gradu- ates. The academic areas planned to be surveyed in- clude: business, psychology, and public health. While Fulcomer et al. (2005) noted similarities in various statistics courses that cross content lines, the current effort is in identifying similarities, differences, and potentially missing elements in statistics curricula. The session was used to encourage comments on the draft instrument.