What different between parametric and non-parametric test?

When using parametric test, the populations need to follow some assumptions, but non-parametric test does not need them.
1. The data has been obtained from a population which is considered "normal" in the statistical sense. In other words if a sufficiently large sample of data was obtained from the population is it likely that it would be normally distributed. This may be known from similar research, but if not, and provided the sample is reasonably large, it may be possible to estimate from the sample whether this assumption is reasonable.
2. The populations from which the samples are drawn should have equal variances. This can be determined by inspection of the data, looking at the spread or standard deviation of the data. The F - test can also be used to test the hypothesis that the samples have been drawn from populations with the equal variance.
3. The data should be measured, at the very least, on an interval scale. Even this is not always easy to determine, but a useful "rule of thumb" is to compare two scores, (say) a 5 and a 10, and to ask yourself if a score of 10 means that there is exactly twice as much of that particular attribute compared with a score of 5. If the data was temperature, then this is obviously the case, but what if the data represented a measure of patient satisfaction. A score of 10 may indicate more satisfaction than a 5, but not necessarily twice as much; i.e. this may be an ordinal scale rather than interval.