The present study examines productive and receptive knowledge of PVs among Saudi undergraduates learning English as a Foreign Language (EFL). It uses a mixed-methods approach to elicit two kinds of data: Firstly, 195 Saudi undergraduates in Saudi universities were asked to complete three multiple-choice (MC) tasks designed to assess their productive and receptive skills, and to measure their PV avoidance behavior. The design of the MC tasks was informed by an analysis of a specially-constructed corpus of Saudi EFL textbooks, and taking into account the most frequent PVs in Liu’s (2011) corpus-based study, to maximize the likelihood that students were only presented with familiar PVs – those that they have been introduced to as well as ‘high-frequency PVs’, which many learners at this stage of learning (i.e. undergraduate students) are likely to have encountered, and which they might then either use or avoid. Secondly, data are drawn from a self-built corpus of written compositions, the Saudi Learners of English Corpus (SLEC), comprising over 175,000 words written by 741 Saudi undergraduate EFL students. In both cases, the research attempts to trace the influence of the following variables on learners’ use (or avoidance or underuse) of PVs: students’ proficiency level (beginner vs. lower intermediate); students’ gender; and the semantic type (literal vs. figurative) of the PV in question. In addition, the analysis of the SLEC data investigates the impact of text genre (argumentative vs. narrative vs. descriptive) on learners’ use of PVs. With respect to overall PV frequency, the results of the corpus analysis indicated that there was a relatively low use of PVs in the learners’ English production. Furthermore, the results from the MC tasks support those of earlier studies, particularly at the production level. Participants showed better receptive than productive knowledge with an average percentage of correct answers of 72% for the receptive task and 59% for the productive task. In addition, the results for both MC tasks and corpus analysis further suggest that proficiency level, PV type, and text genre play a significant role in Saudi learners’ use of PVs, while gender is found to be significant only in the corpus analysis. Meanwhile, proficiency level and PV type had a significant effect on the frequency of PV avoidance.