Hypothesis Testing Basics

[sushant_kulkarni]

I have currently study Hypothesis testing.
Need help regarding following things,

1) There are n number of testing methods are available ? t-test, F-test, ANOVA, 2-Porportion test, Bartellte etc.
What is criteria for selection of test methods?

2) Is Normal Distribution data is essential for all Hypothesis testing or Any specific test required Normally distributed data?

3) Sample Calculation for testing been required any pre-requisite?

4) When One trial test and Two trial test been used?

5) Large Sample size and small sample size variation tackle with which test?

6) Where variance comparison is used for F test where as mean comparison is used by student t test?


Any books available or training source will be welcome.

Regards,
Sushant

 

[Miner]

1) There are n number of testing methods are available ? t-test, F-test, ANOVA, 2-Porportion test, Bartellte etc.
What is criteria for selection of test methods?

There are hundreds of possible hypothesis tests available. This site helps you to decide among the more common ones. Selection is made based on the type of data, 1 vs. 2 samples, paired vs. unpaired, relationships, etc.

2) Is Normal Distribution data is essential for all Hypothesis testing or Any specific test required Normally distributed data?

Most parametric tests have an assumption of normality of the data (or of the residuals). Parameteric means that the test assumes a specific distribution (usually normal). Certain parametric tests are robust against some deviation from normality under certain conditions such as larger sample sizes. A nonparameteric test does not assume normality, but may have other assumptions such as symmetry or similar distributions.

3) Sample Calculation for testing been required any pre-requisite?

This is a good practice. If this is not done there is a discrepancy between statistical significance and practical importance.

4) When One trial test and Two trial test been used?

Do you mean 1-sample and 2-sample? 1-sample is used to determine whether the mean/median is different from a reference value. 2-sample is used to determine whether the means/medians are different from each other.

5) Large Sample size and small sample size variation tackle with which test?

Sample size does not determine the test.

6) Where variance comparison is used for F test where as mean comparison is used by student t test?

The goal (comparing variance vs. comparing means) determines the test.

Any books available or training source will be welcome.

Two good introductory texts are:

• Your Statistical Consultant
• 100 Statistical Tests