iSixSigma

Hypothesis Testing

Six Sigma – iSixSigma Forums Old Forums General Hypothesis Testing

Viewing 15 posts - 1 through 15 (of 15 total)
  • Author
    Posts
  • #40684

    Nitesh Sondhi
    Participant

    Hi
    Please tell me how to do Hypothesis Testing ?
    I am having problems with this .
    Please help
    Thanx

    0
    #126814

    AB
    Participant

    Nitesh
    Have you received any formal six sigma training? Do you have access to a Black Belt/ Master Black Belt. If you are doing self study, you will need to read a lot.
    Frankly, looking at this question and the other one you asked about analysis phase makes me infer that you need training and not just a clarification. If I were you, I would refer to the training material or look at a case study or go and ask the BB/MBB in my company.
    AB

    0
    #126824

    Mikel
    Member

    Let’s take and example from Ab’s post. Ab has a hypothesis that you have not been trained.
    Ho: Nitesh’s Knowledge = No training (training hours = 0)
    Ha: Nitesh’s Knowledge > No training (training hours > 0)
    First of all this will be known as a one sided hypothesis leading to a one sided test (it is hard to have less than no training).
    We go collect data and we find that Nitesh is currently in on line GB training provided by Darth’s emloyer. Nitiesh has logged on and recorded 20 hours sitting in front of a computer with this robust training package. We find that 15 of Nitesh’s associates have also logged 20 hours. We find that 20 hours should have everyone through 1 sample t, 2 sample t, paired t, chi squared, and F tests.
    Based on the data, that there is collectively 320 hours of GB training, we have to conclude that 320 does not 0 hours of training and reject the null.
    Which leads us to the next hypothesis. Does the training suxx or did Nitesh just fall asleep? Darth, you want to handle that analysis?

    0
    #126827

    Nitesh
    Participant

    Thanx Stan and AB but you are not getting what I have asked
    I had just a little Confusion over Hypothesis testing t-test and nothing more than that .
    Anyways
    thanx a lot for your reply
     

    0
    #126828

    Mikel
    Member

    What specifically do you want to know about t test?

    0
    #126845

    Ken Feldman
    Participant

    Stan, thanks for the plug.  In reality, we did a pretest and post-test with Nitesh.  Our Null Hypothesis is that there is no difference in Nitesh’s scores before and after his attending the online course.  In other words, Knowledge Before = Knowledge After.  Our Alternative Hypothesis is that his Knowledge Before does not equal Knowledge After.  I am interested in the two tail test since I want to see if there was a change in either direction, either more or less knowledge as a result of the online course. We opted for a .05 alpha error, in other words a 95% confidence level.  We chose to use a paired t-test since the exam questions were consistent before and after so we considered them to be paired.  In other words, we will be testing whether the difference between before and after is equal to zero.  We ran the test and came up with a p-value of .345 thus concluding that we don’t have enough evidence to reject the null hypothesis.  In doing an analysis as to why this might have happened, we examined the connection time with Nitesh’s computer and found that while he was online as he was supposed to be, there was little if any data exchange indicating that your initial assumption that he fell asleep was indeed valid.  Hope this helps.

    0
    #126846

    Mikel
    Member

    Very good analysis. May I suggest that your training include a link to an online dating service on every page. Not only will it help keep Nitesh’s attention, it will augment the unreasonably low cost of your employer’s offerings.

    0
    #126850

    AB
    Participant

    Darth
    Lovely analysis.Would you also kindly throw some light on the alpha error and beta error? Nitesh is entitled to knowing the probability of our conclusion being incorrect.
    Thanks,
    AB

    0
    #126853

    thevillageidiot
    Member

    Stan caved….big softy.

    0
    #126854

    AB
    Participant

    TVI…please don’t provoke him. The forum has been working hard on him.
    Stan – nice work. Keep it up.

    0
    #126856

    Joe BB
    Participant

    What’s wrong with Stan? he is being decent with people now. I am glad that he is giving heed to my advice. We can get along now.
    You know that Nitesh was asking a very simple question that made me question his intellectual abilities.

    0
    #126857

    Ken Feldman
    Participant

    We could declare that Nitesh didn’t learn anything when in fact he did.  We could, on the other hand, declare that Nitesh did learn something when, in fact, he didn’t.  Both are mistakes in my decisionmaking.  The first error is called my beta error.  The second error we can call alpha error.  I already decided that my alpha error will be .05.  If I wanted to fix my beta error at, let’s say .2, we could determine how big a sample we would need to be confident that both error levels are considered.

    0
    #126858

    HF Chris
    Participant

    Darth,
    Has anyone questioned his heterogeneity or homogeneity? He may be an outlier?
    Chris

     

    0
    #126863

    sharma
    Participant

    dear…Nitesh…….i m really feel sorry for ..u….
    n  u all…so called……james bond
    tht fellow ask …abt Hypothesis testing ..n u all….find the platform. to show urs  expertise…n in between forgot y…u r posting the reply…..if u all have capability then…..solve her query…with  minimum sentences . can any one do it
     

    0
    #126870

    Nitesh
    Participant

    Thanx Anu
    Thank you for ur reply that was great
    The confusion that I was having got cleared now
    Thanx
    Nitesh

    0
Viewing 15 posts - 1 through 15 (of 15 total)

The forum ‘General’ is closed to new topics and replies.