Why the null hypothesis (Ho) should never be ‘accepted’?

People may have some level of difficulty in making inferences using hypothesis testing.

Let me put forth a common mistake that we do while making conclusions.

We know that when the p-value is less than the level of significance (α), we reject the null hypothesis. Commonly we take α as 0.05. So when p< 0.05, we reject the null hypothesis.

But, when the p> 0.05 we can not reject the null hypothesis.

At this point, many of us tend to say that ‘we accept the null hypothesis‘ and this statement is not correct.

Reason:

We make conclusions about the population based on a random chance that could say that Ho is true. But this random chance is at the chosen level of α. There is a possibility that some other randomly chosen sample gives a different conclusion at a different level of significance.


Thus, when p> α, we fail to reject Ho


When p is greater than α we should write the results in neutral language. For example,

Ho: Machine 1 is producing spindles with average length of 30 cm

Ha: Machine 1 is not producing spindles with average length of 30 cm

If p< 0.05, At the significance level of 0.05, we reject the null hypothesis that machine 1 is producing spindles with an average length of 30 cm. This also means that we accept (and not prove) the alternate hypothesis,  a nice explanation can be found here for this case.

When p> 0.05, at the significance level of 0.05, we can not reject the null hypothesis that machine 1 is producing spindles with an average length of 30 cm.


This simply means that at a significance level of 0.05, we can not conclude whether the machine is producing an average length of 30 cm or not.


So, we should always keep in mind that when we are not able to reject the null hypothesis, we should not conclude in any other way but in a neutral manner.

Do share in comments, your experience and other mistakes that you may have once committed/ witnessed.



Learn all about applied statistics in our upcoming training on Lean Six Sigma (green belt). Enroll * for the training now and get a 20% discount when you pay.

Our free basic training available-

Lean six sigma yellow belt training
Quality Management
Statistics for Managers
* T& C apply

 

Advertisements

Sampling Basics

There are certain vital techniques which must be understood well if we want to understand the subject of statistics and measurements. Sampling is one such very important topic, which we’ll be covering in a couple of articles.

Some terminologies first…

To understand the concept, we should also understand a few frequent terms.

  • Element– an object on which a measurement is taken
  • Population– a collection of elements about which we wish to make an inference
  • Sampling units non-overlapping collections of elements from the population that cover the entire population
  • Sampling frame– a list of sampling units
  • Sample- a collection of sampling units drawn from a sampling frame
  • Parameter: numerical characteristic of a population
  • Statistic: numerical characteristic of a sample

What is Sampling?

The activity in which elements, from a population, are collected so as to represent the population. In this video, a very good introduction to sampling has been provided.

Why Sampling?

Most of the times, the population is too large to measure all of its elements, thus sampling is done. A sample reflects the characteristics of the population from which it is drawn. For example, a machine produces 1000 spindles a day. It may be difficult to measure all of them, so we take samples and measure them.

It is very crucial that samples are selected carefully. Incorrect sampling may lead to incorrect inferences about the population.

Sampling has many advantages over exhaustive sampling, which covers the whole population.

  • Sampling can save money
  • Sampling can save time
  • In case of destructive inspection, it is not prudent to do exhaustive sampling

In our next article, we’ll learn about the common sampling techniques.


Learn all about sampling techniques along with statistical techniques in our upcoming training on Lean Six Sigma (green belt). Enroll for the training now and get a 20% discount when you pay.

Our free basic training available-

Lean six sigma yellow belt training
Quality Management