Hypothesis testing : Case Study
In this interesting situation, we have 2 queries; one probably was more important than other. Now, according to statistical analysis, if we apply Hypothesis test we can look at how it would be scored on evaluation parameters.
Usually Hypothesis testing starts with an assumption which is called Null hypothesis.
Note: Please click on the image to enlarge it...
We set out on to see if we can prove it wrong and hence vote for Alternative Hypothesis.
Nully Hypothesis : It is equal to nothing/ no defect or in our case : For the player both the queries are equally important.
Alternative Hypothesis here is: That for the player the query 1 was more important than the query 2.
Now there are 2 errors that can happen:
Type I error: For the player both the queries are equally important. But, we decide that Query 1 is more important than Query 2.
Type II error: For the player Query 1 was more important that Query 2. But, we decide that both the queries are equally important.(Hence prefer to score the contact as: Avg)
Looking at it graphically:
Type I Error is not accepted in Governance. We do not want an innocent to get punished.
But for business which look to make results better by the day, they would not mind taking up something that is not serious to the customer as to be serious, rather than missing out on something that is a concern to the customer.
Type II Error is acceptable in Governance as they don’t mind letting a criminal go free but never want to punish an innocent. For businesses Type II error can be suicidal.
Conclusion : Hence companies would not mind living with Type I errors as they do not cause any damage to the business(any damage that can be caused to the employee if any is taken care of by the laws of averages) Mostly companies try ignoring outlier instances unless they is 100% evidence.
So, in the contact that we have below are we willing to live with a:
Type I error (I assume the first concern was more important to player: Hence I penalize the associate) or a
Type II error(I assume that for the player both the queries are equally important, when it may not be so and hence cause damage to business, as we are giving a decent score and hence are lowering the incentive to avoid such errors in future).
Please think about it…
We set out on to see if we can prove it wrong and hence vote for Alternative Hypothesis.
Nully Hypothesis : It is equal to nothing/ no defect or in our case : For the player both the queries are equally important.
Alternative Hypothesis here is: That for the player the query 1 was more important than the query 2.
Now there are 2 errors that can happen:
Type I error: For the player both the queries are equally important. But, we decide that Query 1 is more important than Query 2.
Type II error: For the player Query 1 was more important that Query 2. But, we decide that both the queries are equally important.(Hence prefer to score the contact as: Avg)
Looking at it graphically:
Type I Error is not accepted in Governance. We do not want an innocent to get punished.
But for business which look to make results better by the day, they would not mind taking up something that is not serious to the customer as to be serious, rather than missing out on something that is a concern to the customer.
Type II Error is acceptable in Governance as they don’t mind letting a criminal go free but never want to punish an innocent. For businesses Type II error can be suicidal.
Conclusion : Hence companies would not mind living with Type I errors as they do not cause any damage to the business(any damage that can be caused to the employee if any is taken care of by the laws of averages) Mostly companies try ignoring outlier instances unless they is 100% evidence.
So, in the contact that we have below are we willing to live with a:
Type I error (I assume the first concern was more important to player: Hence I penalize the associate) or a
Type II error(I assume that for the player both the queries are equally important, when it may not be so and hence cause damage to business, as we are giving a decent score and hence are lowering the incentive to avoid such errors in future).
Please think about it…
posted by LearnersParadise @ 1:10 PM
0 comments
