+ Reply to Thread
Results 1 to 14 of 14

Converting irregular scores into % then grouping them

  1. #1
    Registered User
    Join Date
    06-16-2007
    Posts
    9

    Converting irregular scores into % then grouping them

    I've been trying to solve this myself for sometime and failed so I'm asking for some help please.

    I have two measures which output an absolute score and relative score at irregular time points for the length of trial. Some trials take longer than others so I would like to align them by expressing the absolute and relatives scores as a percentage of the trial.
    Relatively easy you'd think...
    Measure A starts at time point (00:00:54 and ends at 00:02:05) whilst measure B starts at 00:00:03 and ends at 00:01:16)
    Still not that hard...
    Because of the irregular time points sometimes a row consitutes a 2 sec gap but sometimes it is 5 sec gap.
    Getting tougher...
    Once I have these figures expressed as a % of the trial I would then like to average them and group them into 5% of trial 'bins'
    ...And now I'm stuck!
    If anyone could help me with this problem I've attached a sheet which I think might also help with what I'm looking for.
    Thanks
    Crash
    Attached Files Attached Files

  2. #2
    Forum Guru DonkeyOte's Avatar
    Join Date
    10-22-2008
    Location
    Northumberland, UK
    MS-Off Ver
    O365
    Posts
    21,531

    Re: Converting irregular scores into % then grouping them

    could you outline at least a handful of expected results for the same data ?

    For ex. when you say you want to express as % of the trial - are you saying you want to weight the score in accordance with duration of snapshot relative to total duration ?

    For ex. Measure A takes 71 seconds to complete and aggregate absolute score is 75.772 meaning per second the average score would be 1.07
    Are you then looking to weight the % based on duration of reading in some way ? (ie compare the score to the duration * average per second ?)
    Or, are you simply taking the absolute score and expressing that as a % of the total irrespective of duration ?

    I think a few expected results will better help others follow the necessary logic.
    Last edited by DonkeyOte; 07-26-2010 at 12:18 PM.

  3. #3
    Registered User
    Join Date
    06-16-2007
    Posts
    9

    Re: Converting irregular scores into % then grouping them

    Thanks for taking up the challenge donkeyOte.

    What I want to do is calculate the average absolute and relative values of measures A and B express that as percentages of the trial.

    I can see Step one is convert the times in column A into percentages of the trial ie the 00:54 of cell B3 is equivalent to the start or zero point of the trial and B4 is equivalent to 1.8% of the trial up to the value in B41 which is equivalent to 100 % of the trial.
    Step two would be to do the same with Measure B. Step three would then be averaging the absolute and relative scores of measures A and B, and Step four would be batching the averages into 5% bins to be reported in column K. Ultimately I would end up with a score for each 5% of the trial in K3,K4,K5 and in L3,L4,L5 etc etc.

    I'm struggling because the time scores in columns A and F don't start at the same time point or go up in the same increments. Although there are 39 rows of data each, Measure A is 71 seconds long, whilst Measure B is 73 seconds long, hence why I think they need converting to percentages of trial first. I then run into another brick wall when trying to find a convenient way of batching the averages into 5% bins for columns K and L.

    Hope that helps with the puzzle.
    Sorry for the delay in answering

    Crash

  4. #4
    Forum Guru DonkeyOte's Avatar
    Join Date
    10-22-2008
    Location
    Northumberland, UK
    MS-Off Ver
    O365
    Posts
    21,531

    Re: Converting irregular scores into % then grouping them

    Not to sound like a broken record but with this types of calculations it's generally advisable to post a handful of manually calculated values - this makes it much easier to the XL nerds amongst us to validate what we're doing.

    You state B4 is 1.8% of trial - I presume you meant A4 but wouldn't the % be nearer 2.8% ? ie 2 seconds of a 71 second trial ?

    For ex.:

    D2: =A$41-A$3
    D3: =(A3-A$3)/D$2
    copied down

    (D2:D41 copied to I2:I41)

    would give you your percentages - I think - is that along the lines of where you were intending to start from ?

    To reiterate though - expected values help to clear up any discrepancies between narrative and sample data should they exist.

  5. #5
    Registered User
    Join Date
    06-16-2007
    Posts
    9

    Re: Converting irregular scores into % then grouping them

    Thanks, that formula is a great start!

    I've attached a sheet with my long hand way of doing it thus far. Hopefully it demonstrates the expected values I'm trying to get at.
    Attached Files Attached Files

  6. #6
    Forum Guru DonkeyOte's Avatar
    Join Date
    10-22-2008
    Location
    Northumberland, UK
    MS-Off Ver
    O365
    Posts
    21,531

    Re: Converting irregular scores into % then grouping them

    OK - now I've seen your expected results here is something I prepared earlier which would appear to be along the right track.

    edit: below formula not in attachment....

    to have the last record(s) in the 95% bucket it would be simplest to revise D2/I2 such that

    D2: =(A$41-A$3)+"0:0.001"
    copied to I2

    this way the % are just under 100%
    Attached Files Attached Files
    Last edited by DonkeyOte; 08-10-2010 at 12:02 PM.

  7. #7
    Registered User
    Join Date
    06-16-2007
    Posts
    9

    Re: Converting irregular scores into % then grouping them

    That looks like the business!

    I'll test it out with some bigger data sets tomorrow, but I think you've cracked it.
    Thank you

  8. #8
    Registered User
    Join Date
    06-16-2007
    Posts
    9

    Re: Converting irregular scores into % then grouping them

    I've had a play and I'm getting there.

    I need to incorporate 9 measures (measure a-i) of 4 variables (ab, rel, re, he) in the average. I only gave 2 of each in the eg because it was an eg, and I figured that that might be something I could sort myself! Turns out I was wrong.

    When I try and edit the average formula to include these additional variables the { } around the formula disappears and therefore beyond row 3 the averages stop averaging. When I manually put {} back in there's a space that appears at the start of the formula that I can't seem to delete and the formula turns to text. Is there a way of stopping that happening?

    Thanks

  9. #9
    Forum Guru DonkeyOte's Avatar
    Join Date
    10-22-2008
    Location
    Northumberland, UK
    MS-Off Ver
    O365
    Posts
    21,531

    Re: Converting irregular scores into % then grouping them

    You will I think need to post a sample that better reflects the set up you're trying to work with - I confess I can't visualise the above given earlier sample.

  10. #10
    Registered User
    Join Date
    06-16-2007
    Posts
    9

    Re: Converting irregular scores into % then grouping them

    Found a few more measures so its now 10 not 8 measures that I need to average.
    The original solution you provided is in the first worksheet, the solution required is on a second worksheet.
    I've tried to copy across the solution you gave earlier, but as you can see I can't seem to make it work or include the extra measures required in the average. expected scores are given in the comments box.

    Happy to hear what I'm doing wrong
    Thanks
    Crash

  11. #11
    Forum Guru DonkeyOte's Avatar
    Join Date
    10-22-2008
    Location
    Northumberland, UK
    MS-Off Ver
    O365
    Posts
    21,531

    Re: Converting irregular scores into % then grouping them

    Remember this formula is an Array and you must commit Arrays using CTRL + SHIFT + ENTER (enter alone will not suffice).
    you can tell if the Array has "set" as the formula will be encased within { }

    In theory you can replicate the approach you have presently adopted and replicate the IF for each measure...however it will be a very long formula...

    Given the consistency of the headers in your table (for each measure) I suspect you will find that you can use:

    Please Login or Register  to view this content.
    Note that as before the 100% bucket is defunct given the last record falls into the 95% bucket (by design)

  12. #12
    Forum Expert shg's Avatar
    Join Date
    06-20-2007
    Location
    The Great State of Texas
    MS-Off Ver
    2003, 2010
    Posts
    40,678

    Re: Converting irregular scores into % then grouping them

    Very nice formulation, DO.
    Entia non sunt multiplicanda sine necessitate

  13. #13
    Forum Guru DonkeyOte's Avatar
    Join Date
    10-22-2008
    Location
    Northumberland, UK
    MS-Off Ver
    O365
    Posts
    21,531

    Re: Converting irregular scores into % then grouping them

    Thanks shg - I will wait for the OP to confirm it works before being too pleased with myself [until such time as someone improves upon it!]

  14. #14
    Registered User
    Join Date
    06-16-2007
    Posts
    9

    Re: Converting irregular scores into % then grouping them

    I've no idea how you did that but I am truely impressed!
    Works a treat. Thanks.

+ Reply to Thread

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Bookmarks

Posting Permissions

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

Search Engine Friendly URLs by vBSEO 3.6.0 RC 1