Obviously, one would like a distribution that is symmetric and light in the tails.
If there is a large study with a finely graded scale, it may make sense to look at skewness and kurtosis and see whether there are deviations from normality (e.g., through a kolmogorow-smirnov test).