iSixSigma

Guidance Required for Selecting a Database for Telecalling

Six Sigma – iSixSigma Forums Operations Call Centers Guidance Required for Selecting a Database for Telecalling

Viewing 6 posts - 1 through 6 (of 6 total)
  • Author
    Posts
  • #54206

    Shiva
    Member

    Dear All,

    I am currently working with an NGO. We have a telemarketing division for propagating the cause and raising donations. We do purchase database for the calling process. The problem we face is when we get the sample before buying the database, it yields good results. But once the database is purchased, same results are not obtained.

    Is there a way to statistically check if the sample represents the population. I was thinking of using Hypothesis testing for the same. Since this is discrete data, again the confusion of selecting the appropriate test for the same.

    Require guidance or suggestions for sorting out this.

    Thanks,
    Shiva

    0
    #194233

    Prabhu V
    Participant

    Hi Shiva,

    Based on your base inputs, it seems that Chi-square test for goodness of fit can be useful for your situation.

    Please try for it.

    0
    #194234

    Shiva
    Guest

    Thanks Prabhu for your suggestion. But Sir, please correct me if I am wrong, Chi-Square goodness of fit can be used to calculate the differences between the expected and the observed ones right?

    Here the case is that say for example in my sample of 1000 phone numbers, my connectivity is 180. How do I conclude that I get the same connectivity ratio if i decide to purchase the database population of 100000 phone numbers.

    How can we make sure that the sample represents the population statistically so that I can make a decision to go ahead and purchase it based on the sample.

    Hope you understand the scenario.

    Thanks in advance.

    Shiva

    0
    #194259

    Prabhu V
    Participant

    Shiva,

    On the above post, the details regarding the sample selection is found missing.

    Can you provide information on how are you deriving the samples from the vendor?

    Is your sample size statistically significant?

    0
    #194265

    Robert Butler
    Participant

    What exactly do you mean by “…when we get the sample before buying the database, it yields good results. But once the database is purchased, same results are not obtained.”

    Do you mean you have a quantified, statistically significant difference as identified by a test of two proportions or do you just mean that you have seen some differences in proportions?

    If it is the latter then you should first test for significance. If it is the former then there are other questions:

    1. Do you always use the same supplier?

    2. If it is the same supplier do you have the option of trying someone else?

    3. Regardless of supplier is the difference always in the same direction?

    4. If you are using multiple suppliers and if the differences are always in the same direction then what can you do with respect to checking for similarities/differences in the provided data base – i.e. are they always from the same geographic area, socio-economic strata, etc.

    5. What guarantees do you have that you are not just getting resamples of the same population time after time?

    6. What guarantees do you have that your group will handle a small sample in the same fashion as they would handle a full data base? If usage methods change as the size of your donor data base changes the differences may have nothing to do with the supplied data and everything to do with the way you deal with it.

    If you can provide an answer to these questions perhaps I or someone else may be able to offer additional thoughts.

    0
    #197980

    Sir,
    i required data for religious product …. but whenever i purchase data from the vendor , i find the same bad experience that, there is difference in btw sample and the after purchasing the data …. both response are not same …..

    please help me out from the situation

    0
Viewing 6 posts - 1 through 6 (of 6 total)

You must be logged in to reply to this topic.