A substantial body of research, including narrative and meta-analytic reviews, has established concurrent and longitudinal relationships between sensitivity to speech prosody and reading ability. This study synthesised research findings pertaining to a possible impairment of prosodic skills in dyslexia. Three-level meta-analysis was utilised to synthesise 124 effect sizes from 37 studies comprising data from 1771 participants (mean age = 13.7 years, range: 6-26). A moderate-to-large impairment of prosodic competence relative to chronological-age-controls was identified. The effect emerged consistently across languages and different types of prosodic competence task, individual studies were adequately powered to detect effects of this magnitude, and there was no evidence of publication bias. There was no significant impairment relative to reading-age-controls. There was also no impairment for tasks tapping the implicit processing and mental representation of prosody. However, studies in the literature are not sufficiently powered to detect smaller differences between reading groups. The evidence is consistent with a developmental delay in prosodic competence in dyslexia. One interpretation of this is that prosodic competence is in some way dependent on reading experience. Findings also support theoretical models of dyslexia which emphasise explicit processing of phonology, rather than impaired phonological representations or underlying sensory difficulties. However, fully addressing questions of causality will require additional studies that incorporate reading-age-controls and possess sufficient statistical power to detect small effects.