The Two Variable Model: Hypothesis Testing

 It is a method used to make decision using experimental data. It is also known as confirmatory data analysis. It is widely used in frequency probability; the decisions are made using these null-hypothesis tests. In the case of null-hypothesis test an answer is assumed to the problem and it is found what the probability of getting the desired answer is. If the probability is more than the desired one then the solution is used else the alternate solution is used.  This process is called as hypothesis testing.

Steps in hypothesis testing are

• Stating the null and alternative hypothesis. These two must be mutually exclusively. Mutually exclusive in the sense occurrence of null will always happen in the absence of occurrence of the alternative hypothesis.
• Identify analysis plan. In this step the way how the experimental data’s are to be analyzed in determined. Various hypothesis testing methods are available based on the sample size. Suitable testing method has to be determined from these methods.  
• Analyses the experimental data using the above identified method.
• Interpret the result. Identify whether the null hypothesis goes on with the condition. If it does not goes well with the condition then reject the null hypothesis and accept alternate hypothesis.

In hypothesis testing there is a possibility for the occurrence of two types of error, they are,

• Type I error occurs when the analyzer who is analyzing the experimental data using hypothesis testing rejects the null hypothesis even thought it should be accepted. It is usually denoted by a term called α.
• Type II error occurs when the analyzer rejects the null hypothesis even though he/she must have accepted it. This type of error are usually denoted by β

Decision rule are the rules followed by statistician while making decision. There are basically two types of decision rules they are,

• Acceptance of the null hypothesis with respect to P rule is nothing but process of finding out the p value and comparing it with the value of s called as test statistics. If the value of P is equal or greater than s then the hypothesis is accepted else it is rejected.
• Acceptance of null hypothesis if it falls within the region of acceptance. In this method the null hypothesis is accepted if the calculated value falls in the region of acceptance, else it is rejected.

Question

• What are two types of error that might take place in the hypothesis testing method?
• What is null and alternative hypothesis?