Error in hypothesis testing

During hypothesis testing we assume something, we then test our assumption (or Null Hypothesis) using sample data. We either reject the null hypothesis when find it to be false or we are not able reject (we accept) when it is true .

As we test these hypothesis on a sample and draw our conclusion, there is always a risk associated. It might happen, that our conclusion doesn't correspond well to overall population characteristics and hence there is possibility of an error. We might reject the null hypothesis while it is true or accept it while it is false; these are called Type I and Type II errors respectively.

 Let's try to understand these errors : Type I and Type II, with an interesting business example.



Let's understand the errors with an example :

Suppose, a statistician test the null hypothesis on a sample and decides whether to pass the stock for sales or not. Businessman, on the other hand, knows his business in and out and in this example let's consider him to be population.

and hence

Although, I have tried to make it quite simple to understand, yet you might find it quite not digestible. I totally understand, as I also belong to your league. Better we try to understand with an example.










This is a wonderful and humorous way of remembering the same.

Well, it is not my idea. I  have taken it from one  of the blogs of  Paul Ellis

In above example the null hypothesis is "Not Pregnant"

H0 : Someone is not pregnant.
H1 : He or she is pregnant.

In the first picture, The null hypothesis of not being pregnant is actually true, yet being rejected by Doctor. So he is committing a Type I error.
In second one, the null hypothesis of not being pregnant is actually false (woman is pregnant), But doctor is accepting it. Doctor is committing Type II error. 

Hope you are clear about the Type I and Type II error now.


Enjoy reading our other articles and stay tuned with ...

Kindly do provide your feedback in the 'Comments' Section and share as much as possible.