Assessment is a complex task that requires defining clear and transparent criteria, so that all actors understand it as a credible act, of educational and social utility. However, the assessment is not being encompassed in an exact science, it is naturally a process of subjective character. In higher education the study of assessment in mathematics disciplines is almost non-existent. SOLO (Structure of Observed Learning Outcome) taxonomy can be applied in assessing the quality of examinations and identification of cognitive complexity levels required in the preparation of assessments. In this paper the adopted design is described for structured creation of questions in a mathematics examination. Exam questions created should flow in a structure that starts with some sub-questions with levels of minor complexity and gradually become more defiant. Students in responding to questions with a variety of levels of progressive cognitive complexity, they increase their confidence as they advance to more complex and challenging questions.