NON PARAMETRIC TESTS

All  practical  data follow normal distribution under  such situations can estimate the  parameters such  as mean  variance etc ,,, and use the  standard  test they are  known as parametric test. The practical data may be non normal and or it may not be possible to estimate the  parameter of the data. The tests  which  are sued for such  situations are called  nonparametric tests. Since  these  tests  are  based  on the  data which  are free  distribution and parameter these  tests are known as nonparametric  tests.

The  nonparametric  tests require less calculation because there is no need to compute  parameters. Also  these  test can  be applied to very  small samples more specifically during pilot   studies in market research.

The  test techniques  makes  use of  or more values obtained from  sample data often called test statistic to arrive at a probability statement about  the hypothesis. But  no such  assumptions are  made in  case of  nonparametric  tests. In a statistical test two kinds of  assertions  are involved viz... an  assertion directly  related to the  purpose of investigation and  other assertions to make  a probability  statement. The  former  is an  assertion to be tested  and is technically called a hypothesis whereas the set  of al other assertions  is  called the model. When we apply a test to ( to test the hypothesis ) without  a model it is known as  distribution free test or the non parametric test. Non  parametric tests do not make an assumption  about  the parameters of the population  and thus do not make an assumption distribution. In other  words under nonparametric or distribution free tests  we do not  assume that a  particular distribution is applicable  or that a certain value is attached to a parameter of the  population.