Abstract:

Collection of sufficient information on a single-source survey is a well documented barrier to multi-media evaluation. There are a number of methods used to attempt to ameliorate this information problem by collecting information on two or more separate surveys and then combining them. The authors describe these as 'survey merging methods'. There are three main methods of survey merging. These can be described as: a) one-way ascription; b) two-way ascription or fusion; and c) calibration. Method c), calibration, has not previously been used (to the authors' knowledge) in this survey merging context, but it can be invaluable and superior to the other approaches. In some cases it is clear which method has to be used but often one can choose between two, or even all three, of the methods. Which method is the superior, when two or all three are available, is rarely a straightforward decision. It requires considerable expertise and experience with all three methods to decide which approach is the best for a particular situation. This paper discusses the factors which influence the decision.