This paper reports a study in a Chinese context of L2 learner variables and their relationship to English achievement. To establish the variables, a questionnaire was administered to 242 students, the entire cohort of second-year English majors from five tertiary institutions in Nanjing and Shanghai. The subjects' Chinese and English language proficiency on entry to tertiary education was established through scores on the matriculation examinations. Their English achievement was established by their scores on the nation-wide standardized English proficiency test taken on completion of their two-year intensive English program, the first part of their four-year BA degree. Sixteen learner variables were established and a hypothetical causal model was constructed. Relationships within the model and the relationships between learner variables and achievement were then examined through Parial Least Squares (PLS) analysis. It was found that six variables had direct effects on English achievement. Three were traits which existed prior to the students' admission to tertiary education, that is, Sex, and L1 and L2 Proficiency as measured in the matriculation examination. The other three were formed by clusters of strategies relating to Vocabulary learning, Tolerating ambiguity (Risk-taking), the only negative direct effect, and Mother tongue avoidance. Management strategies had the strongest indirect effect on English achievement. Finally, the direct effects of belief variables on strategy variables were examined and found to be strong and consistent. Qualitative data collected from selected subjects are used throughout to provide further insight into and illustrations of the differences in strategy use between more and less successful learners.