Member-only story
Statistics
Hypothesis Testing In Statistics with Examples
A concept in statistics for data-driven decisions
This article will be fun for all readers
Hypotheses testing is an idea to be tested in statistics on observed data points. It is all about guessing the things that can be work or not to make meaningful results.
A good hypothesis contains “if” and “then” words. For example, if the temperature is increased then the solid will melt.
Pre-requisite for this article:
- Knowledge of Descriptive Statistics
- Knowledge of Inferential Statistics
Steps involve in decision making are shown below:
- Formulate a hypotheses
- Find the right test
- Execute the test
- Make a decision
When we always do hypotheses we have to know what is our null hypothesis. For example, if we say the petrol price is higher then this is just a statement. What if we say, the petrol price in one state is 2 rupees higher than the other state. So, this statement shows a point to do a hypotheses where we have some data for analyses.
Example:
The mean (or null hypothesis) salary of a manager is around 100K and the alternative hypothesis is the mean salary is not around 100K. So, we have to check the null hypothesis to be true.
If the mean salary is to be near 100K then we will accept the null hypothesis and if not then we will reject the null hypothesis.
The above distribution shows the confidence interval area that is the mean value falls under it and the rejection area denoted by the alpha i.e the significance level area. The alpha value tells the probability to reject the null hypothesis if it is true. The typical value of alpha are 0.1, 0.5, 0.01 i.e. 10%, 5%, 1% and most commonly used value is 5%.
