"Time to Act," a 2009 report of the Carnegie Corporation's Council on Advancing Adolescent Literacy, concludes that U.S. students are ill-prepared for the literacy challenges of 21st century higher education, employment, and citizenship. Word Generation (WG) is a research-based vocabulary program for middle school students designed to teach words through language arts, math, science, and social studies classes. The program consists of weekly units that introduce 5 high-utility target words through brief passages designed to spark active examination and discussion of contemporary issues. WG was designed with the understanding that promoting classroom discussion can result in particular kinds of academic benefits, such as improved word knowledge, reasoning, and expression. This paper presents new data from the first two years of the WG evaluation focusing on: (1) basic psychometric and descriptive information (including variance decomposition), by grade level, for measures of academic language and perspective taking as well as standardized assessments of student vocabulary and reading comprehension (e.g., Word Generation Academic Vocabulary Assessment and Global Integrated Scenario-based Assessments, GISA); relationships among these measures and the standardized assessments within and across time, and by grade level; and (2) results from impact analyses focusing on WG intent-to-treat effects on (a) standardized assessments of student vocabulary and reading comprehension, and (b) new measures of academic language and perspective taking after one and then two years for the first cohort of schools, and after one year for the second cohort of schools. Twenty four schools in total were randomized within the four districts, within pairs of schools matched on several characteristics such as enrollment, socio-demographics, and standardized test scores. The program was implemented and data were collected on all students in grades 4 through 7. Data were collected for three years (2011-2014) with the final wave of data collection completed in June 2014. For this paper the authors employ data collected in the first two years of the study focusing on 4th graders in cohort 1 in year 1, following those children into 5th grade in year 2 (with two data collection points, fall and spring, in each year), and on 4th and 5th graders in cohort 2 in year 2 (again with two data collection points, fall and spring). All students in the participating grades completed the following group-administered assessments: (1) WG Academic Vocabulary; (2) Global, Integrated Student Assessments (GISA); (3) Core Academic Language Skills Instrument (CALS-I); and (4) The Perspective Taking Survey (PTS). Preliminary data from the first year (cohort 1) of the WG evaluation indicate that, as expected, mean levels of the primary constructs of focus increase significantly over the course of one school year, and are significantly higher for older grades than for younger (e.g., 4th graders). In addition, the measures of academic language and perspective taking were correlated with standard assessments (e.g., with GISA, ~0.5-0.6), and the intercorrelations did not vary substantially by grade.