That doesn't make sense at all. First of all, you must ask yourself: Normal in what respect, and with regards to what population (what's normal in a Brazilian kindergarten will almost certainly be grossly abnormal in a Russian science institute). Then you go measure that property in a significant chunk of that population, and find it's (almost certainly) a bell curve.
Now, normal people would fall within one standard deviation from the average. But that only constitutes 67% of the population. But people between the first and second standard deviation are only somewhat abnormal, as they do infact fall into a category that encompasses 27% of the population. If you go beyond two standard deviations, the abnormality is significant. Only 5% of people would fall in this category.
That being said, the concept of normality is easily abused if you choose strange properties and populations. Having any given job will for example be abnormal, as most of the population will not have the same job.