We provide a meta-analytic review of all group-comparison studies that used reading-level match design, were conducted in highly consistent European orthographies, included children with dyslexia younger than 13 years of age as participants, and included measures of one or more of the potential causes of dyslexia. We identified 21 studies meeting these criteria that examined one or more of phonological awareness, rapid naming, verbal short-term memory, or auditory temporal processing. A random effects model analysis showed first that the groups were matched imperfectly and they differed significantly in word reading measures not used for matching. Second, there were no significant differences between the individuals with dyslexia and their reading-level-matched controls in rapid naming, phonological memory, and auditory temporal processing. Finally, the analyses for phonological awareness showed a significant effect for comparisons that involved manipulating phonemes but not for tasks that involved manipulating syllables. The results are compatible with phonological deficit theories of dyslexia, but this conclusion is qualified by observed differences in reading skills and sample selection concerns.