6 {{ upvoteCount | shortNum }}
6 {{ upvoteCount | shortNum }}

Question About Statistical Mean

by Thomas
5 replies
  • OFF TOPIC
    • |
Hello Warriors:

Do any of you brainy people know if a mean score can be legitimately converted to a percentage figure by diving it by the number of terms?

Does that even make sense?
  • Profile picture of the author HeySal
    You can do that assuming that your sampling is actually representative of larger sampling numbers. If you pole 5 people and get a stat assuming that it is representative of 5 million people,
    you might find your numbers really unrepresentational and someone might just notice it.
    Signature

    Sal
    When the Roads and Paths end, learn to guide yourself through the wilderness
    Beyond the Path

    {{ DiscussionBoard.errors[56394].message }}
  • Profile picture of the author Thomas
    Mmmmmm... as usual, I phrased my question arse-ways.

    I was actually thinking of something else.

    Suppose Group A has mean rating of 9, on a 12-point scale and Group B has a mean rating of 5.

    How would you express that difference in percentage terms? (i.e. Group A had an average rating X% higher than Group B).

    Is it as straight-forward as it looks?

    Tommy.
    {{ DiscussionBoard.errors[56551].message }}
    • Profile picture of the author HeySal
      Is it as straight-forward as it looks?
      Yes it is -- as long as you calculate the percentages correctly and don't do an Al Gore with them.
      Signature

      Sal
      When the Roads and Paths end, learn to guide yourself through the wilderness
      Beyond the Path

      {{ DiscussionBoard.errors[56590].message }}
      • Profile picture of the author myob
        Using your example specifically, you would find the mean as a percentage of the scale for each group, and take the difference.

        ie Suppose Group A has mean rating of 9, on a 12-point scale and Group B has a mean rating of 5.

        Percentage mean of Group A is 9/12, and percentage mean of Group B is 5/12, assuming that both groups are on the same scale. The difference is 1/3 so Group A has a 33% higher mean average than Group B.
        {{ DiscussionBoard.errors[57513].message }}
    • Profile picture of the author colinredk
      Thomas, it seems that the stats you're using are non-parametric. It means that you put a significance or interpretation to each number on the scale. In which case, the mean you are trying to extract would lead you to wrong conclusions. An Al Gore thing. You would be better off trying to analyze it using some non-parametric statistical test, which doesn't really need the mean (per se) in the first place.
      Signature

      {{ DiscussionBoard.errors[63024].message }}

Trending Topics