The authors describe successes and challenges in developing a QL-friendly course at the University of Arkansas. This work is part of a three-year NSF project Quantitative Reasoning in the Contemporary World (QRCW) that supported the expansion of the course. The course, MATH 2183, began experimentally in Fall 2004 as a section of finite mathematics known informally as “News Math” for 26 students from arts and humanities disciplines. Over the past six years, the course has evolved and now MATH 2183 is approved to satisfy the College of Arts and Sciences mathematics requirement for the Bachelor of Arts degree. In 2009-2010, it was offered in 16 sections to about 500 students. The course,, which is designed so that students work collaboratively in groups of three to four to discuss and answer questions related to quantitative information found in newspaper and other media articles, has encountered a variety of challenges that exemplify broader questions confronting interactive teaching of mathematics in context. Many students possess deeply held views regarding mathematics and struggle with the departure from traditional, lecture-driven mathematics classes. Available curricular materials that engage undergraduate students to reason in real-world settings are limited. Solving new problems on quizzes and examinations is challenging and uncomfortable for students, but necessary as QL requires “authentic” tasks. The variety of contexts in which QR is needed tests the instructor’s flexibility and knowledge. Many of the challenges have been ameliorated by putting together Case Studies for Quantitative Reasoning: A Casebook of Media Articles which bases sets of questions upon quantitative content derived from media articles. Learning gains measured by pre- and post-course tests are modest (two percentage points for the mean), but still larger than in other control groups. Faculty advisors’ attitudes about the course are overall positive. Persistence beyond the course, as measured by a survey of 300 former students, seems positive; for example, 29 of the 42 respondents state that their confidence with QR had increased since the course. Similar courses are being taught at Central Washington University and Hollins University by Stuart Boersma and Caren Diefenderfer, respectively, who are also co-PIs on the NSF QRCW project.