iSixSigma

Hypothesis testing HELP required

Six Sigma – iSixSigma › Forums › Old Forums › General › Hypothesis testing HELP required

Viewing 6 posts - 1 through 6 (of 6 total)
  • Author
    Posts
  • September 15, 2008 at 3:54 pm #50942

    Steve Harbron
    ★ 10 Years ★
    Member

    I have a sample of 70 complaints received from customers who appear not to understand the product they have purchased.  Having done Cause and Effect Diagrams the bottom line is we believe this is either down to the way the dealer sells the product, the literature, or the customer.
    I want to test whether the dealer is of significance in the sample but allow for the fact that all dealers sell differing quantities and will, therefore, naturally show differing volumes of complaints against them.
    Any advice on how I go about this please?

    0
    September 15, 2008 at 5:27 pm #175770

    Van Loon
    ★ 10 Years ★
    Participant

    Mayby a chi-square on the table :
    Dealer Sold Conplaints
    1   

    0
    September 15, 2008 at 5:30 pm #175771

    Van Loon
    ★ 10 Years ★
    Participant

    Sorry, dont know what happende with the 3 posts……
    Dealer   Sold   Complaints
    1           xx        yy
    2           zz        pp
    3           kk       ll
    A chi-square test should show if there are diferences between
    the dealers.

    0
    September 15, 2008 at 6:26 pm #175775

    Darth
    ★ 10 Years ★
    Participant

    I would be worried about advice from a poster that takes four times to try and give you an answer. On the surface it does appear to be a Chi Square issue. With only 70 data points you may get a nasty message from Minitab that you have less than 5 expected in each cell. You can also use some simple tests of proportions.

    0
    September 16, 2008 at 12:07 am #175793

    Vallee
    ★ 10 Years ★
    Participant

    Unless I have read into this too deep, it appears that you are assuming that the products and complaints can be lumped together in one analysis… the proportion of products sold per dealer is just part of the issue. Start with the customer and get more information about the complaint examining: their expectancy of the product, their source of product information prior to purchase, the dealer purchased from (don’t assume product info came from the same dealer purchased from), the dealer response to customer complaint, and your follow up with the dealer.
    After this is done, how you plot the data can be discussed. If you have already completed this, elaborate more.
    HF Chris Vallee

    0
    September 16, 2008 at 6:16 am #175795

    Chandran
    ★ 10 Years ★
    Participant

    Hi, I would like to add few more points to review:
    1. Whether the complains are from one product or various products. (It could be variants on the main product). Try to visualise pareto for any clues
    2. What type of customer complaints? Can you categorise these customer complaints? If yes, again apply pareto for any clues?
    3. Mode of Purchase made by Customers? It could visiting showroom, or through internet or through telecalling or through reference or through exhibition sale etc. This could also lead on which type of sale the complaints are more?
    4. All these Customers what they actually needed; but what was sold to them? Trying to findout customer needs and wants could help you to findout the root cause that caused in misunderstanding / communication happened during sales processes.

    0
  • Author
    Posts
Viewing 6 posts - 1 through 6 (of 6 total)

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