PARAMETRIC AND NONPARAMETRIC TESTS
Parametric and nonparametric tests
In order to make a generalization regarding a population from a sample, statistical tests are used. The two hypothetical testing include parametric and nonparametric tests. The parametric test is referred to as a hypothesis test that generalizes by making statements regarding the mean of the sample population. The parametric test uses a t-test and t-statistic. The t-statistic is based on the underlying assumptions that there is a normal distribution of variables and the mean is assumed to be known. Also, the variables of interest in a population are measured on an interval scale. The measure of central tendency in this hypothesis test is mean. The parametric test is applied only on variables. Parametric tests are best used in nonmoral distributions such as income distribution. The parametric test is appropriate to use if the mean accurately represents the center of distribution and the sample size is huge enough since the parametric test is more powerful (Pataky, Vanrenterghem & Robinson, 2015).
On the other hand, nonparametric tests or distribution-free-test is where the researcher has no information about the population. Also, the nonparametric test is not based on the underlying assumptions. The nonparametric test assumes that the research variables are measured on an ordinal or normal level. The nonparametric test is mainly used in situations where independent variables are non-metric. The measure of central tendency in the nonparametric test is median. Lastly, the nonparametric test applies both attributes and variables. A good example of when to use nonparametric tests is when dealing with a center of a skewed distribution such as income which can be measured using median. The appropriate test to use when