Week 3

Can the formal hypothesis testing approach be used for nonparametric tests?

Hypothesis tests require and estimation of one or more unknown parameters and often make unrealistic assumptions about the normality of the underlying population or require large samples. Nonparametric tests focus on the sign or rank of the data rather than the exact numerical value of the variable, do not specify the shape of the parent population, can often be used in small samples, and can be used for ordinal data. For this reason statisticians are attracted to nonparametric tests. Many of the nonparametric tests are similar to parametric tests. So yes a formal hypothesis can be used for nonparametric tests.

How are parametric and nonparametric statistics different?

Parametric and nonparametric statistics differ in the amount of data required. A nonparametric test is also used if you have nominal or ordinal data, where a parametric test uses ratio or interval data. In addition if there nonparametric tests are used when you have rank data. Parametric tests also make more assumptions then a nonparametric test does and follows the normal t distribution.

How are parametric and nonparametric statistics similar?

Both the parametric and nonparametric use information from the distribution of the population that was taken, along with using assumptions and variance.

Parametric and nonparametric statistics are both considered inferential statistics. We use inferential statistics find out things about a population through the use of a sample population. When using parametric tests, we assume that the distribution is rather normal and need to know certain things like population mean or the variance. When using nonparametric tests, we do not assume normality of distribution in our population in question. According to Doane & Seward (2007), nonparametric tests can be “more powerful than parametric tests when normality cannot be assumed” (p. 699). Nonparametric tests do not require as...