# Application / Use of Hypothesis Testing in any Six Sigma Black Belt Project

#### AMIT BALLAL

Trusted Information Resource
Hi all !!

Can anyone share the use of hypothesis testing in any six sigma black belt project with some simple examples?
I am little confused about how I can use hypothesis testing in a project, this will help me to get some clarity.

Any help will be appreciated.

Thanks,
Amit

#### Miner

##### Forum Moderator
Staff member
There are an almost unlimited number of ways to use hypothesis testing. A simple example is making a modification to a fixture (or a new design for a fixture). Then comparing the variation of the modified/new fixture to that of the old fixture using a test for equal variances. If significant, the test would validate that the improvement is real.

C

#### Curtis317

If you have MiniTab you can get into the help files and learn how to use it. It is also a great to review the materials if you have done any of the statistical tools in a while.

#### julsbear

##### Involved In Discussions
You can use hypothesis testing to confirm that your change has indeed resulted in an improvement over the existing system. You can test if the mean has improved or if the variance has improved. If you can show improvement in both, you may have hit the jackpot.

#### Bev D

##### Heretical Statistician
Staff member
Super Moderator
it may be more helpful for you if you could explain what your project is or what type of project you will be assigned.

#### AMIT BALLAL

Trusted Information Resource
it may be more helpful for you if you could explain what your project is or what type of project you will be assigned.
Thanks to all for help!!!

I have taken the Project of Improving Training Effectiveness, problem considered is trainings given are not effective / trainings given are not effectively utilized.

Can you suggest how to use hypothesis for this project / training process ?

Thanks,
Amit

#### Miner

##### Forum Moderator
Staff member
You could run Before/After tests for comprehension and/or observe for desired behavior changes. These could be tested for significance using the paired t-test.

#### Bev D

##### Heretical Statistician
Staff member
Super Moderator
Miner has a good example. I have used the paired t test frequently when assessing whether or not a particular training approach worked - at least long enough to take the test.

what makes the paired t test appropriate is that each individual will come to the training at different levels of knowledge and/or skill. so their 'improvement' difference will be different (think about a group of people who decide to lose weight).

The longer term effectiveness of any training in changing behaviors and ultimately results is probably better assessed with a control chart than a hypothesis test.

As with any improvement project, the statistical test of significance that you use are of far less importance than how you determine the cause(s) of poor performance, how you incorporate good science into determining the solutions and how you structure the experiments to determine if the solutions will work...

#### AMIT BALLAL

Trusted Information Resource
Thank you all for quick response & help provided.

I'll utilize the help provided & will share my experience with you.

Thanks,
Amit