Writing is critical for college and career readiness, yet secondary students in America are not good writers (National Center for Education Statistics, 2012). Unfortunately, researchers know relatively little about secondary students' writing skills, and even less about their digital writing. In this study, we explored prior computer use, keyboard activity during writing, and their relations to writing achievement using the 8th grade 2011 National Assessment of Educational Progress writing assessment, the first national computer-based writing assessment for U.S. secondary students. We found that prior computer use predicted students' writing skills directly (0.08) and indirectly (e.g., keypresses, 0.14) via keyboard activity during the test. We found differential effects for certain groups including current English learners and disadvantaged students. We also found a small positive interaction effect of prior use and keypresses on writing. That is, the benefits of prior computer use for school writing and the value of students' additional keypresses on writing achievement were amplified when both were present.