The ability to reason about relations is relevant for many spatial cognitive processes. This can involve: (i) to represent spatial information mentally, (ii) to manipulate the spatial representation, and (iii) to infer new spatial information. Several cognitive theories make assumptions and predictions about the underlying processes. A detailed and systematic overview and analysis of ireliable effects across studies is missing. This article presents a meta-analysis of 35 studies about spatial relational reasoning. Studies were classified according to different factors including the ambiguity of the spatial description, i.e., if it the description allows for more than one representation, the presentation of information, i.e., if the information has been presented auditorily or in a written form, and the task, i.e., if a conclusion or model of the premises needs to be generated or verified. Implications of the findings for the mental model theory and working memory are discussed.