The problem of missing data is quite common in data mining. One of the first problems that a researcher is faced with when analyzing results is that of an incomplete dataset and the presence of errors. This generally happens because whoever collects the data has not correctly interpreted the structure, accidentally commits some errors, does not want to deliberately insert that data because of an error in the encoding tool, or is someone who instead deals with the data entry.
There is no single technique or methodology to approach the problem of how to monitor the effect of missing data; each situation is a case in itself. In general, it is always advisable to test the survey instrument with pilot surveys ...
