This paper reports on the replication of a study of novice programmers, looking for relationships between ability to 'explain in plain English' the meaning of a code segment and success in writing code later in the semester. This study explores the question in a different learning environment and qualitatively evaluates 'explain in plain English' responses to identify implications for teaching. Statistical results from this study are similar to those of the earlier work. Results highlight students' fragile knowledge, particularly for students excluded from the primary analyses by a set of screening questions, and suggest the need for assessment and instruction of basic concepts later into the term than instructors are likely to expect.