This paper deals with cognitive diagnostic analysis of the reading comprehension part of the Standardized Test of English Proficiency (STEP-RC) developed and administered by the National Center for Assessment in Saudi Arabia. The Least Squares Distance Model (LSDM) was used to obtain probabilities of correct performance on cognitive attributes of STEP-RC items across levels of the underlying ability measured by the test. Among other results, cutting scores on the scale for correct performance on targeted attributes are provided for the purposes of criterion-based diagnostic decisions. Aspects of the validity of the cognitive attributes are also examined in the LSDM framework. The methodology and procedures illustrated in this study can be applied in different areas of assessment and cognitive analysis.