To measure reading comprehension, we must define a set of concepts applied here in a theoretical framework derived from studies on the psychology of language. Understanding a text or document is a complex activity whose ultimate aim is to develop a consistent mental representation of the content (Gernsbacher, 1994; Kintsch, 1994 and 1998). Beyond reading the words that compose the document, comprehension draws on high-level processes that analyze the sentence in syntactical-semantic terms and handle the document as a whole. The reader seeks consistency for documents comprising purely verbal information and documents combining verbal and visual-spatial information. We analyzed the underlying dimensions of the “upper” module of the 2004 IVQ Survey from data collected on 7,389 respondents. Because of the small number of module items, we should be highly cautious in interpreting the results. Ultimately, however, only two distinct factors emerge, one visual (maps and charts), the other strictly verbal. The relevance of these two dimensions would appear to indicate a specialization in the processing of information depending on its type—verbal or visual-spatial. Specialization seems to be corroborated by the study of differential gender-related functioning. A classic result is the consistent “bias” in favour of men in the visual-spatial tests (village map and chart) (the bias is defined by the pass rate on each item by gender, for an identical performance level by both genders). We do observe biases in favour of women on some items involving verbal information, but they are few in number.