Scientific knowledge is not communicated through the verbal mode only but through the integrated use of tabular, graphical and verbal forms, a fact that has positioned visual understanding as a major element of literacy demands to acquire advanced knowledge. In spite of this, empirical studies on how scientists develop meaning out of visual data are scarce. Grounded in the Tunisian EAP (English for Academic Purposes), the present study aims to explores the relative contributions of background knowledge and language effect to the Tunisian specialists' strategy use; the purposeful actions they use to understand visuals. Based on their answer in a questionnaire probing their visual skills and strategies, eight specialists from the field of management belonging to two areas of specialization and displaying different layers of visual proficiency are recruited in this study. They were invited to verbalize their thoughts about visuals from research articles within and outside their area of specialization written in French and English. At a first stage of analysis, the participants' idea units were classified into six categories. At a second stage, the clustering of idea units for each visual is analysed and divided into three categories. The participants strategies were analyzed in number and quality and were associated with the study variables. The results showed that background knowledge is a more predictive factor of participants' performance than language effect. They also indicated weaknesses in participants' use of strategies for meaning-building. Their strategy use is most often confined to low-level idea units (for example, restatement and main idea). The present study argues for a systematic implementation of visual teaching in EAP classes.