1. Forum
  2. Usenet
  3. COMP.SOFT-SYS.STAT.SPSS

  • I am getting an error message when I do a reliability analysis

    From [email protected]@21:1/5 to All on Sun Mar 15 00:37:12 2020
    I am trying to apply reliability analysis but I get reliability in negative

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Rich Ulrich@21:1/5 to All on Tue Mar 17 12:50:09 2020
    On Sun, 15 Mar 2020 00:37:12 -0700 (PDT), [email protected]
    wrote:

    I am trying to apply reliability analysis but I get reliability in negative

    Rescore the particular items which are opposite in direction
    from how you want to interpret your overall score.

    If an item is scored from 1-5, you can reverse the scoring by
    COMPUTE rvar= 6- var . And so on.

    In order to keep your file accurate with its original variables,
    you should NOT simple reverse-score an item in place, unless
    you make sure NOT to save the file with the reversed score.

    When there are only a couple of vars that need reversing, I've
    been satisfied with creating new versions of only those couple,
    so my file might end up with (say) "var1 to var10, rvar4, rvar8".

    Reliability is computed from formulas that have several important
    assumptions, and add and subtract "variance components".
    The important failure-of-assumption is that your correlation
    matrix apparently does have sizable negative r's.

    LOOK at the correlation matrix that can come with the analysis,
    to make sure that you are looking at a consistent set.

    --
    Rich Ulrich

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From [email protected]@21:1/5 to All on Fri Apr 3 18:08:49 2020
    Me too

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Rich Ulrich@21:1/5 to All on Mon Apr 6 18:33:59 2020
    On Fri, 3 Apr 2020 18:08:49 -0700 (PDT), [email protected]
    wrote:

    Me too

    I see approximately 33 follow-ups to this thread from its
    first posting in 2009. See -

    https://groups.google.com/forum/#!topic/comp.soft-sys.stat.spss/IVp-wCTSxR8

    I found this by Googling on
    < "error message when I do a reliability analysis" >


    --
    Rich Ulrich

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Amy Hester@21:1/5 to Rich Ulrich on Fri Jan 7 14:33:44 2022
    On Monday, April 6, 2020 at 5:34:06 PM UTC-5, Rich Ulrich wrote:
    On Fri, 3 Apr 2020 18:08:49 -0700 (PDT), [email protected]
    wrote:

    Me too

    I see approximately 33 follow-ups to this thread from its
    first posting in 2009. See -

    https://groups.google.com/forum/#!topic/comp.soft-sys.stat.spss/IVp-wCTSxR8

    I found this by Googling on
    < "error message when I do a reliability analysis" >


    --
    Rich Ulrich
    I followed all the guidance here, but my problem was different. When I exported my dataset from the survey software used to collect instrument scoring responses, it coded the responses accurately, but the dataset was inversely coded where the columns
    should have been the rows and the rows should have been the columns. This may be what others referred to as reflection or reflecting above. I was testing a gold standard tool, had the same issues, corrected my columns to rows and vice versa and it worked
    perfectly.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)