Answer
There are two differences: (a) Parametric tests
give accurate probabilities of a Type I error if
the populations the samples are drawn from
have the characteristics that parametric tests
assume (e.g., ANOVA assumes the populations
are normally distributed and have equal variances).
Nonparametric tests do not have
assumptions about the populations built in.
(b) The null hypothesis for parametric tests is
that the population means are equal; for nonparametric
tests, the null hypothesis is that the
population distributions are the same.
Work Step by Step
There are two differences: (a) Parametric tests
give accurate probabilities of a Type I error if
the populations the samples are drawn from
have the characteristics that parametric tests
assume (e.g., ANOVA assumes the populations
are normally distributed and have equal variances).
Nonparametric tests do not have
assumptions about the populations built in.
(b) The null hypothesis for parametric tests is
that the population means are equal; for nonparametric
tests, the null hypothesis is that the
population distributions are the same.
