Best IT Training Institutes in Chennai with Placement | DeepNeuron

Introduction

Hypothesis testing is one of the most important techniques applied in various fields such as statistics, economics, pharmaceutical, mining and manufacturing industries. Suppose we want to know if something took place if certain medicines are effective, if groups differ from each other or if one variable predicts another variable.

     Source: ety.com

All in all, we want to predict if the data collected is statistically significantly different from another. This article is for anyone who wants to know and understand the concept of hypothesis testing, which is a significant component of inferential statistics. The 5 steps taken to conduct the hypothesis testing have been explained in detail.

Alright, let’s begin!

What is Hypothesis Testing?

Hypothesis Testing is an inferential statistical method that is required to use sample data to solve assumptions about a population parameter (a characteristic that describes a population).

Inferential Statistics Definition

I want to define inferential statistics as the main reason we study statistics to make inferences about our population of interest using sample data. Hence, inferential statistics is a branch of statistics that refers to the probability that the sample data can represent the population data from which the sample was obtained. In other words, it’s like saying what can we conclude about the population based on what we know about the sample. For instance, if you wanted to know the average salary of a domestic worker, if covid has less impact on people who take vitamin C if a manufacturer produces the same quantity of 50 litres of milk and many other examples where inferential statistics is applicable.

5 Steps taken to Conduct a Hypothesis Testing

The following is a graphical representation of the steps to conduct hypothesis testing.

Step 1: Select the appropriate Statistics

The test is chosen based on whether you are looking for relationships or differences between groups and whether the assumptions for a particular test have been satisfied. However, this step involves exploratory data analysis, which helps us choose the test to conduct the hypothesis.

° Looking for Differences between groups. Hypothesis tests that can be used are as follows:

✓ The independent sample t-test is a parametric procedure or test used to compare two independent samples where the mean from one group has no effect on the mean from the other group.

✓ In other terms, it tests whether or not the sample means are statistically significantly different from each other. For example, measuring women’s weight from one group will not affect the importance of men from another group.

 

✓ The dependent sample t-test is a parametric procedure or test, also referred to as the matched paired sample t-test, related sample t-test and repeated measures t-test.

✓ It is used to compare means between two related samples; in short, it tests whether the average difference between means is different from zero. For instance, before and after experimental designs, matched pairs and pre-test and post-tests experiments use the dependent sample t-test.

✓ The One-Sample t-test is a parametric procedure or test that allows us to compare a sample mean to a known population mean provided the population standard deviation is unknown and checks if the means are statistically significantly different.

✓ The Mann-Whitney U test is a non-parametric test that corresponds to the independent sample t-test.

✓ It is categorised under the group of ranked sum tests. It is used to measure two independent samples that have been calculated on an ordinal level and follow a ranked order.

√ The Wilcoxon signed-rank test is a non-parametric procedure or test that corresponds to the one-sample t-test and can be used to test if the sample median is statistically significantly different from the hypothesised population median value.

✓ It is a non-parametric procedure or test corresponding to the dependent sample t-test.

✓ It measures whether there is a difference between two related/dependent/matched samples that have been calculated on an ordinal level and follow a ranked order.

✓ The Friedman test is a non-parametric procedure or test that corresponds to the repeated measures ANOVA.

✓ It is used to measure whether or not there is a statistically significant difference between medians of three groups and the same subjects or scores show up in the same groups or classes.

✓ It is a non-parametric procedure or test corresponding to the one-way ANOVA.

✓ It is used to measure whether or not there is a statistically significant difference between medians of three independent groups.