The aim of this thesis was to investigate whether individuals with autism lack a natural drive towards central coherence as predicted by weak central coherence theory (Frith, 1989). To investigate the nature of this difficulty, two main areas of proposed impairment, context (Experiments 1-9) and global processing (Experiment 10), were separated out. In Experiments 1-3 a method was developed to test the ability to use context information in the visual and verbal domains. Results showed that the performance of individuals with autism was facilitated by the provision of visual and verbal context information. Experiment 4 showed that the same children were able to use semantic category information to aid recall. However in Experiment 5 the same individuals had difficulties with a sentence processing task when using sentence context to disambiguate homographs. Experiments 6-9 examined two possible alternative explanations for the difficulties in Experiment 5. The results of these experiments indicated that the difficulties in Experiment 5 were not due to a failure is using context when targets are embedded in the context (Experiment 6). It remains to be tested whether the difficulties are related to the processing of ambiguous information as Experiments 8-9 failed to develop a method to examine this possibility. Having established that the impairment in using context was highly specific to the use of sentence context to disambiguate homographs, Experiment 10 moved away from examining context to study a different measure of central coherence, global processing by use of a face recognition task. The results of this study have confirmed that there is a deficit in the ability to process global information in autism. However further confirmation of a global impairment will be needed by use of non-facial tasks. In summary, the findings fail to support the claim of a general impairment in autism regarding central coherence.