The valid per cent is simply the proportion of a sample that is valid. Data can be invalid for a variety of reasons. Some data are simply impossible, such as negative heights or weights. Some data can be shown to be invalid by comparing them with other data. For example, a person might be two years old, and a person might be a widow. But it is hard to conceive of a person who is a two year old widow! Finally, some data can be identified as machine error or data entry error.
Write down the total sample size. For example, you might have 1000 cases.
Write down the number that are invalid. For example, there might be 92 invalid cases for one reason or another.
Subtract the result in step 2 from that in step 1. For instance 1000 - 92 = 908.
Divide the result in step 3 by the result in step 1 and multiply by 100. 908/1000 = .908. .908*100 = 90.8. Therefore 90.8 per cent of our data are valid.