Witten language examination is still the conventional, albeit universal, tool to measure students' performance in language teaching, where the targeted cognitive ability is conceptualized through required learning outcomes. As it is the case with any other subject area, to assess the ability of the EFL written examination to evaluate students' abilities is contingent upon the questions included in the exam. In testing, it is a rule that any exam sheet has to include questions varying in their levels of complexity to accommodate students' individual differences. The current study aims at identifying these levels in EFL exam sheets, based on the criterion keywords found in them. The method used was a content analysis formula to assess how exam questions, students' answers, and EFL course learning outcomes interact. The findings showed that: (a) there was an inadequate matching between expected course learning outcomes and exam questions; and (b) students, when given a chance to choose, tended to choose lower-order questions. The study is significant to EFL instructors and those responsible for EFL testing. The study then concluded with relevant insights, presenting some recommendations with respect to a more appropriate EFL assessment.