Reading difficulties in school are very challenging for teachers due to many different reader subtypes in one and the same class. Moreover, there are few easy-to-use tools enabling teachers to assess reading ability. According to the Simple View of Reading (Hoover and Gough in "Reading and Writing", 2(2), 127-160, 1990), efficient reading comprehension is the result of an interaction between word reading (through three word representation levels: orthographic, phonological, and semantic) and comprehension (through three processing types: literal, text-connecting, and gap-filling inferences). Difficulties in one of these components, or in both, should lead to difficulties in reading comprehension and bring about different reader subtypes. This study aims, first, to examine the validity of the tool and, second, to explore performance reading patterns of children with reading difficulties. A population of 485 typically developing French children from grade 2 to grade 9 was tested using three computerized tasks that recorded accuracy and speed: lexical quality to examine the three levels of word representation; silent reading and listening comprehension to examine both literal and inferential processing. Results showed the appropriateness of the tool but also identified a number of limits. It was possible with the results to detect 76 children with reading comprehension difficulties and to divide them into 5 clusters essentially according to their word reading performances. The results are discussed in relation to the theoretical frameworks used to build the tool.