iSixSigma

Simple or Random Sampling?

Six Sigma – iSixSigma Forums Old Forums General Simple or Random Sampling?

Viewing 7 posts - 1 through 7 (of 7 total)
  • Author
    Posts
  • #38832

    ANDREA
    Participant

    Hello
    This is all new to me, and I can’t seem to get my head around it.  Can someone help, please?
    I have a population of 26 million.  Assuming a 99% confidence level and a 4% error margin, the number of people I need to survey to be rep of the total pop works out at 1037.  Is this correct??  Is this (1037) my sample size or just the number of responses I must have to make the survey representative?
    The population of 26 million covers 4 regions but I know the population for each region.  I want to identify a sample size for each region and combine them to get a single sample size which is representative of the demographic grouping?  Is it correct to use a simple random selection of the population of each area or does this fall within stratified random sampling?
    Thanks.

    0
    #116807

    NS
    Participant

    Andrea – 1037 would be the required responses that you need to receive.
    This site shows the calculation in more detail.
    http://www.custominsight.com/articles/random-sample-calculator.asp
     

    0
    #116812

    Darth
    Participant

    I calculated 1041.  That will be the final number of people providing data.  If you are doing a mail survey then you will likely need to send out a whole bunch more depending on your response rate.  Be aware of dealing with some non response bias as well.  Will the characteristics of those responding be different than those that don’t?
    Why do you want to stratify?  Do you suspect some difference between the regions?  If money is no object, do the regions independently and then do some Chi Square analysis to see if they are different.  If no difference then you can say something about the combined population. If different then you can say something about each region.  I am assuming you are using some kind of discrete data given the sample size calculations you used.
    This is really too big and complicated a subject to deal with in an online Forum.
     

    0
    #116813

    Mikel
    Member

    I got 1039.432.

    0
    #116814

    Darth
    Participant

    Idiot!!!! You can’t have a .432 person!!!!  Gee, that felt good.  Now I know you feel when you post sometimes. Hope all is well with the Mrs.  U coming down my way for Spring Break or anything?

    0
    #116816

    ANDREA
    Participant

    Thanks, All.
    Yes, Darth, I suspect some differences in the regions, but I would prefer not to stratify since I have time and cost constraints.  I really need to know whether I can safely use a simple random sampling procedure for each region separately although effectively I am conducting a single survey research, i.e. the results will be published in one report.
    I can’t seem to find any text that answers my question even vaguely.  Obviously if my design/methodology is wrong, it will call into question the accuracy and validity of my research findings. Can’t afford for my work to go belly up at this stage.
    Thanks for the link, NS.  I used it earlier today.

    0
    #116817

    Tim F
    Member

    Andrea,One caution as you do your study that no one else has mentioned. Ideally, you send out 1037 surveys at random and get back 1037 responses – this will acheive the 99% confidence 4% margin of error you want. If you don’t get 100% response, then ANY results will be questionable. Consider a restaraunt that sets out 100,000 customer satisfaction surveys and gets back 1037. 50% of those customers say they were very dissatisfied. Does that mean that you are 99% confident that 50 +/- 4% are very dissatisfied? Not at all! Dissatisfied customers are much more likely to fill out the form. Similarly, if the people who respond to your survey are self-selected (i.e. they can choose to respond or not) and if a significant fraction don’t respond, then you have to be VERY careful about believing what the sample is telling you. Tim F

    0
Viewing 7 posts - 1 through 7 (of 7 total)

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