Many recent studies have aimed to demonstrate that specific types of reading comprehension depend on different underlying cognitive abilities. In these studies, it is often implicitly assumed that reading comprehension is a multidimensional construct. The general aim of this study was to examine the dimensionality of a large pool of reading comprehension items differing according to text and question type. The items were administered to 996 fourth-grade children. We used multitrait, multimethod modeling to test for the existence of specific text and question types. In addition, the correlations of factor scores, reflecting the different measures of reading comprehension, with word reading speed, vocabulary, and working memory were examined. Confirmatory factor analyses revealed that the specific measures of comprehension, differing according to text and question type, hardly reflected systematic variation, after a general factor of reading comprehension was taken into account. Reading comprehension items thus largely reflect a common factor. Factor scores that were supposed to reflect specific comprehension factors were not reliable and were hardly related to word reading speed, vocabulary, and working memory.