Why do we talk about hypothesis tests? Some companies are ‘third party’ friendly organization; which means they are very dependent on vendors, consultants or freelancers to provide the solution for them.
Most of the time, the third party was not there to validate the results from the solution given. Sometimes it ends up with the company hires another vendor to validate and another vendor might propose the new solution.
As the internal resource of the company, you might responsible to validate the results from the solution.
But how the earth should I validate the results?
Have you heard about Hypothesis tests?
Hypothesis tesst is very useful in Lean Six Sigma projects to compare the means of one variable for two groups of cases. But in this situation, it can be used as validation tools.
A hypothesis tests examines two opposing hypotheses about a population:
-
The null hypothesis (Hₒ) and the alternative hypothesis. The null hypothesis is the statement being tested. Usually the null hypothesis is a statement of “no effect” or “no difference”.
-
The alternative hypothesis (Hₐ) is the statement you want to be able to conclude is true.
Basically, there are 2 categories of hypothesis test; test to compare mean and test to compare variation. But today, I want to focus more on how to utilize hypothesis test for comparing mean.
A corporation has several web servers; there have been consistent complaints about the efficiency of these servers – None perform significantly better or worse than the others
The manager responsible for the operation of these servers decided to spend RM20,000 modifying one of them (Server A) to improve efficiency – Before spending more money, time, and resources n modifying the rest of the servers the manager wants to know if she has “significantly” improved the efficiency of Server A (modified) against Server B (original spec)
Practical Question:
Will the modifications on Server A improved its efficiency when compared to the current process, represented by Server B?
Hₒ = Performance efficiency Server A ≤ Server B
Hₐ = Performance efficiency Server A > Server B
Based on our 2 sample T test (comparing 2 population means), it shown p value is 0.099 (p>0.05).
Hence, based on our hypothesis test we can conclude that no evidence to conclude performance of Server A is more efficient compared to Server B. Hence, it might a good decision to spend money to modified another server.
You get the idea, right?
Take every objection as a question. Always measure and validate the result.