We report results from two studies testing the Mediation Model of Research Experiences, which posits that science (or engineering) self-efficacy and identity as a scientist (or engineer) mediate the association between support programs and students' commitment to science, technology, engineering, and mathematics (STEM) careers. Study 1 included 502 matriculated and recently graduated undergraduate STEM students. Structural equation modeling analyses indicated that research experience, instrumental mentoring, and involvement in a community of

No abstract

In recent years, at the same time that performance assessments in science have become more popular, the number of English language learners (ELLs) (i.e., students whose native language is other than English) served by the U.S. educational system has also increased rapidly. While the research base is growing in each of these areas independently, little attention has been paid to their intersection. This case study of the use of a science performance assessment with 96 ELLs in five high school science classes investigated the face, construct, and consequential validity of this intersection. Qualitative and quantitative data analyses showed that both teachers and students had an overall favorable response to the assessment, although students' English comprehension and expression skills were determining factors for certain items. While most responses were reliably scored, ELL spelling and syntax on certain responses were significant sources of error. The degree of specificity of teachers' guidance also significantly affected students' scores. Recommendations from this study include increasing the clarity of an assessment's design, allowing ELLs more time to complete assessments, and scoring by raters who are knowledgeable about typical patterns in written English for this student population. Furthermore, it is recommended that the use of performance assessments with ELLs be exploratory until such time as their validity and reliability with this population can be more adequately established.

Science performance assessments (SPAs) are designed to elicit a wider range of scientific knowledge and abilities than ordinarily measured by more traditional paper‐and‐pencil tests. To engage in SPAs and thus demonstrate abilities such as scientific inquiry, students must interact with various participants and communicate in a variety of ways. Using an analytical framework that draws on functional and interactional views of language, we analyzed video footage of one SPA, The World Down Under, enacted in three fifth‐grade classrooms. The video analysis revealed a wide variety of language demands that students encountered during the SPA, including a range of participant structures (whole class, small group, and individual), communicative modes (interpreting language, presenting language, and engaging interpersonally), and 18 different texts. Moreover, the language demands differed by classroom. A closer look at two video episodes illustrated how this more expansive communicative context afforded two English Learners (ELs) with opportunities to navigate the language demands of the assessment. However, the same language demands could present challenges for ELs with lower levels of English proficiency. The findings contribute to science education by bringing attention to the nature and role of language in SPAs, which has only begun to be described in the literature. © 2012 Wiley Periodicals, Inc. Sci Ed 96:631–651, 2012

This article report on the development and use of an analytical framework intended to map the language demands encountered by English learners as they engage in science performance assessments. Grounded in functional and interactional views of language and language use, the authors—two science education researchers and a language scholar—developed the framework via an inductive, iterative, and systematic review of written assessment materials associated with three fifth grade science performance tasks. The resulting Science Assessment Language Demands (SALD) framework is organized around three dimensions: participant structures, communicative modes, and written texts and genres that students are called upon to read and produce. The authors used textual analysis to conduct an expert review of the written documents associated with the three assessment tasks. The results indicate that the framework can be used to document a wide range of functional and interactional language demands involved in science performance assessments. The demands revealed by the SALD framework highlight both potential challenges facing English learners during science performance assessments as well as opportunities afforded by such assessments for demonstrating their knowledge and skills and further developing language proficiency. A major implication of the study is the potential use of the framework to evaluate the language demands and opportunities of science assessments used in classrooms with English learners. © 2010 Wiley Periodicals, Inc. J Res Sci Teach 47: 909–928, 2010