Beginner’s Guide Hypothesis Testing

Statistics' main purpose is to verify or refute a notion. For example, you may do research and discover that a particular medicine is useful in the treatment of headaches.

No one will believe your findings if you can't replicate the experiment. Because no one was able to repeat the results, the discovery of cold fusion, for example, sank into obscurity.

What exactly is a hypothesis?

An informed assumption about anything in the world around you is referred to as a hypothesis. It should be able to be put to the test, either by experiment or observation. Consider the following scenario:

·       You've discovered a new treatment that you believe will work.

·       You believe there is a better approach to teach.

·       A more equitable method of administering standardized assessments.

It may actually be anything as long as it can be put to the test.

Hypothesis Testing:

Hypothesis Testing is a statistical test that determines if the hypothesis assumed for a sample of data holds true for the full population or not. Simply said, a hypothesis is a statement that is put to the test in order to identify the link between two sets of data.

In hypothesis testing, two competing hypotheses about a population are generated, such as the Null Hypothesis (H0) and the Alternative Hypothesis (H1).

The null hypothesis states that there is no difference between the sample statistic and the population parameter, and it is the one that is tested, but the alternative hypothesis asserts If the null hypothesis is rejected, it is untrue.

To test the assumption, the following Hypothesis Testing Procedure is used:

·       Create a Hypothesis

·       Create a Significance Level that is appropriate for your situation.

·       Choosing an appropriate Test Statistic

·       Identifying the Critical Zone

·       Completing calculations

·       Decision-making

An individual may make the following sorts of errors while testing the hypothesis:

1. Error of Type I:

True Null hypothesis is rejected, i.e. when the hypothesis should be accepted, it is rejected. The likelihood of making a type-I error is expressed by and is referred to as a significance level.

If Pr[type-I error] = Pr [reject H0/H0 is true], then

If (1-α) = Pr[accept H0/H0 is true], then

(1-α) = correlates to the Confidence Interval concept.

2. Type-II Error:

A Type-II Error occurs when a False Null hypothesis is accepted when it should be rejected. The likelihood of making a type-II error is given by β.

If β = Pr[accept H0/H0 is false] = Pr[type-II error]

So (1-β) = Pr[reject Ho/H0 is false, 

Then, (1-β) = statistical test power

As a result, hypothesis testing is an essential tool in statistical inference for determining how much the sample data deviates from the population value. Hypothesis testing are commonly used in business and industry to help make important business choices.

How Hypothesis Testing Is Conducted:

An analyst performs hypothesis testing on a statistical sample with the purpose of proving the null hypothesis's plausibility.

A hypothesis is tested by taking measurements and examining a random sample of the population under investigation. All analysts utilize a random population sample to test two hypotheses: the null hypothesis and the alternative hypothesis.

The null hypothesis is often a hypothesis of population parameter equivalency; for example, the population mean return might be zero. An alternate hypothesis is the polar opposite of a null hypothesis (e.g., the population mean return is not equal to zero).

Hypothesis Testing in Four Steps:

A four-step approach is used to examine all hypotheses:

1.  The analyst must first state the two hypotheses such that only one may be correct.

2.  The following step is to create an analysis strategy that explains how the data will be analyzed.

3.  The third stage is to put the strategy into action and assess the sample data physically.

4.  The last stage is to assess the results and either reject or accept the null hypothesis based on the available evidence.