My suggestion will depend on the type of data in the reports. For the commentary you can apply text sensitive coding for example with the use of Nvivo. Establish a comparative framework i.e. type of sector, program type etc. for each text and then direct the Nivivo program to search for key terms of corruption (bribery a.o) comes up and systematically compare across your groups. I am currently using this procedure in relation to 1.000+ audit reports on various development projects. Let me know if you could use short suggestion.
If data consists of numbers (fx. expenses) I believe you have to establish a kind of yardstick to compare with.
Keep in touch and let me know more. I will certainly appreciate information on the procedure you develop.
Do you mean that you need a method for approaching analysis of previously audited financial statements? And also further procedures for identifying possible signs of corruption within the financial statements?
OR
Are you looking for signs of corruption within the Audit Reports?
OR
Are you wanting to correlate something in the Audit Reports with other information, such as the front-matter of Annual Reports?
That is, is the data you are working with numeric or text?
One of the best way to analyze audit reports is the use of content analysis. I had a good experience in using such methodology for analyzing the content of audit committee minutes for an article where I investigated the role of audit committee in real life and how they follow the requirements stated in code of corporate governance and their own codes. It was very interesting to see how extensive members of the committee consider some of the roles and responsibilities of the audit committee and for others how they forget about some of the important responsibilities.