Skip to main content

View Diary: Some statistics on pie (163 comments)

Comment Preferences

  •  Psuedo-science... (4.00)
    your comment "...REAL science..." seems to imply you have the "hard science" (i.e., chemistry, physics, etc) bias against "soft science" (i.e., sociology, economics, etc).  This would imply that you prefer objective analysis over subjective analysis, yet you cite "real variables" such as experience, benefits, and personal choices that...to me...are subjective and subject to bias.

    Your bias shows in the example you chose to share.  Set and example and show us what you consider REAL science using REAL controls and REAL statistical analysis of the results.

    How would you control the experiment without psuedo-replication in your experimental design or analysis?

    Do what you can, with what you have, where you are. - T. Roosevelt

    by ranger31 on Wed Jun 08, 2005 at 03:12:37 PM PDT

    [ Parent ]

    •  little harsh (none)
      This response seems a little harsh.  I think the main point is that the more detailed the statisitcal study, the more that can been gleaned from it.

      You are correct in that things like "experience" can be subjective.  But the object of the researcher would be to find the least subjective variable to describe "experience".  Is it years?  Is it college rank and years, etc. etc.  

      Leaving out significant variables is one of the primary sins of statisitics (so is including nonsense variables...).  The more general an analysis is, the less conclusions can be drawn.  The more detailed the analysis, the more careful the choice of variables. In hard or soft science, it is the same problem.

      •  Point taken... (none)
        and I agree, to some extant, but my read was that the example used to illustrate the point was guilty of the same sin he/she was calling out in his/her comment.  The numbers they both cite may be "real", but without disclosure of the variables used in experimental/survey design or data analysis there is no way to discern whether a significant statistical relationship exists as used in their comment.  You know, application of the old lies, damn lies, and statistics quote.

        Do what you can, with what you have, where you are. - T. Roosevelt

        by ranger31 on Wed Jun 08, 2005 at 07:26:16 PM PDT

        [ Parent ]

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site