Online higher education professors may find their teaching approaches conflict with the learning preferences of their globally dispersed students, which can impede academic performance. In this empirical study of 254 international doctorate students (across 23 cultures), a model was developed to assess how learning expectations affected dissertation performance. Five indicators were validated from a survey and split sample: mentoring, rendering, interpreting, constructing, and schemata. Contemporary higher education performance-related factors and instruments were also discussed, namely: attrition, competence, motivation, supervision, discipline difference, learning style, and culture. Limited experimental control consisted of same university context and program. The hypotheses were tested using structural equation modeling; revealing three latent factors (method, supervision, and quality) explained 56% of the variance effect on candidate performance. (Contains 4 tables and 2 figures.)