Given the concerns raised over data plausibility and completeness
http://plosmedicine.org/article/info:doi/10.1371/journal.pmed.1001214#article1.body1.sec4.p2

It should be recognized that all data, perhaps especially in international development, have quality issues. A key question is whether that data causes bias. The language here seems to suggest that any mismeasurement should undermine the quality of the results. That is not so. The purpose of statistics is to ferret signals out of noise. Standard errors are computed to estimate how much the noise blurs the conclusions. Where it is proper to be concerned about the data quality per se is if it become so bad that a few outliers can completely distort results. That is a legitimate concern here, but should be articulated more precisely.