Despite growing promotion of student-centered teaching, few studies have assessed how widely these practices have diffused throughout computer science (CS) higher education. Studies on teaching practices often treat instructor- and student-centered learning as being dichotomous. Recent research indicates that this reductionist model misinterprets what actually goes on in CS classrooms. For example, a recent study found that most CS faculty categorized their teaching practices as equally student-centered and instructor-centered. We contribute to this scholarship by reporting the results of a survey disseminated to 821 CS educators in 595 universities, colleges, and community colleges in the U.S. Participants were asked to report how frequently they employed several instructor- or student-centered teaching practices and tools during the most recent semester that they taught a specific course. Results indicate that when asked to reflect on specific behaviors within a certain timeframe, faculty reported using instructor-centered teaching practices, especially delivering content via lecture, more than student-centered practices. However, most faculty also reported using at least one student-centered technique, albeit less frequently, throughout the semester. This study found few differences across certain situational and demographic variables (e.g., institution type, tenure status, etc.). Faculty justify their teaching choices with concern for negative evaluations, desire for student learning, large enrollments, heavy teaching loads, and retaining underrepresented students.