Results 1 to 4 of 4
  1. #1
    4 Star Lounger
    Join Date
    May 2002
    Posts
    410
    Thanks
    1
    Thanked 0 Times in 0 Posts
    example: I would like to take the median of 3 year's worth of data to compare to another three year's worth of data. The data vary seasonally.

    Data set 1: The 1st year has 287 daily data points, the 2nd 77 daily data points, and the 3rd year 325 daily data points.
    Data set 2: The 1st year has 347 daily data points, the 2nd 365 daily data points, and the 3rd year 366 daily data points. The data also vary seasonally.

    I'm good at obvious: in layman's terms, the data will be unequally weighted in the first data set


    (1) is this a situation where it is stupid to use median?
    (2) assuming it's not, my thought is to look at the monthly medians and compare then. This may or may not help, depending on the distribution of the data throughout the year for the data sets with fewer data points.
    (3) something else?

    Suggestions appreciated.
    Thanks.

  2. #2
    Plutonium Lounger
    Join Date
    Mar 2002
    Posts
    84,353
    Thanks
    0
    Thanked 29 Times in 29 Posts
    I'd start by plotting the data in a scatter chart to see what patterns emerge.

  3. #3
    Platinum Lounger
    Join Date
    Feb 2001
    Location
    Weert, Limburg, Netherlands
    Posts
    4,812
    Thanks
    0
    Thanked 0 Times in 0 Posts
    Hi,

    Whether or not you need the median function or an "ordinary" average depends on the nature of the data, I'm no expert there.
    To detect whether or not te difference between two averages is statistically significant you need the so-called t-test.
    I found a good description here:

    http://www.stattutorials.com/EXCEL/EXCEL_TTEST2.html

    and a tutorial here:

    http://www.nursing.ucla.edu/researchoffice.../Stat_Excel.pdf

    And finally, this resource has all the technical information you'd need:

    t-test for independent samples (click the basic statistics button on the right)
    Jan Karel Pieterse
    Microsoft Excel MVP, WMVP
    www.jkp-ads.com
    Professional Office Developers Association

  4. #4
    4 Star Lounger
    Join Date
    May 2002
    Posts
    410
    Thanks
    1
    Thanked 0 Times in 0 Posts
    [quote name='pieterse' post='800078' date='27-Oct-2009 11:16']Hi,

    Whether or not you need the median function or an "ordinary" average depends on the nature of the data, I'm no expert there.
    To detect whether or not te difference between two averages is statistically significant you need the so-called t-test.
    I found a good description here:

    http://www.stattutorials.com/EXCEL/EXCEL_TTEST2.html

    and a tutorial here:

    http://www.nursing.ucla.edu/researchoffice.../Stat_Excel.pdf

    And finally, this resource has all the technical information you'd need:

    t-test for independent samples (click the basic statistics button on the right)[/quote]

    Thanks to both of you.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •