iSixSigma

2 Sample t (sample size)

Six Sigma – iSixSigma Forums Old Forums Healthcare 2 Sample t (sample size)

Viewing 3 posts - 1 through 3 (of 3 total)
  • Author
    Posts
  • #25806

    MatthewG
    Participant

    I’m trying to calculate a required sample size, but have a feeling I don’t have enough information.
    Situation: Two Randomized Groups: 1 no intervention, 1 with intervention.  If there is a 10% or greater increase in size, the intervention has made a difference.
    I believe that is a 2-Sample t Test (1 tailed). I assume Power = .8, but since there hasn’t been sampling, I don’t know SD.
    If someone has a moment, can you let me know what I’m missing… it just isn’t computing (literally).
    Cheers,
    M
     

    0
    #62504

    Darshan
    Participant
    #62505

    StuW
    Member

    It is somewhat difficult from your description to tell exactly what test is being conducted.  The calculator provided in the reply given is useful if you are measuring a percentage, but I’m not sure that is what you are measuring.   Some questions that you need to answer to process this correctly:
    – Are the units “paired” or independent samples?   In other words, if the measurements are made on “like” units, then you can make the comparisons within each unit and have a single group test on the differences instead of two independent samples.   This impacts the sample size calculations.
    – You mention a 10% increase in size, but what is the base measurement?  If it is a proportion, then use a sample size calculator based on 1 or 2-sample proportions.    If it is a measurement, then the test should be done using the estimate of a 10% change relative to the sigma expected.    For example, suppose the base measurement is 100, so you are looking for a 10 unit increase.   However, you need the estimate of variation in these measurements to apply a sample size calculator.  The ratio of the 10 unit change to the process sigma is what you would need to enter in most SS calculators.  You might need to do a few measurements to estimate this variation, otherwise your sample size estimation may be way off.
    Hope this helps. 

    0
Viewing 3 posts - 1 through 3 (of 3 total)

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