SPSS Statistics

 View Only
  • 1.  Question about interpreting AMOS Output

    Posted Thu March 09, 2023 08:13 AM

    A recent SEM output  for Bias-corrected Bootstrap Estimates (β), SEs and 95% CIs for Independent Variables after conducting a Bollen-Stine bootstrap provided me with a "..." under the "Lower" value listed for one of my variables. Can someone please tell me what "..." means in this context? Thank you! 



    ------------------------------
    Matthew Marshall
    ------------------------------



  • 2.  RE: Question about interpreting AMOS Output

    Posted Thu March 09, 2023 12:28 PM

    Hi @Matthew Marshall 
    I think we need more context here.  I'm not clear on what you mean by the "Lower" value.

    Here is the  IBM SPSS Amos 29.0 User Guide .  Examples 19, 20, and 21 deal with Bootstrapping.

    Searching the online Help within the product I see:

    Is this perhaps your issue?

    I hope this helps!



    ------------------------------
    David Dwyer
    SPSS Technical Support
    IBM Software
    ------------------------------



  • 3.  RE: Question about interpreting AMOS Output

    Posted Thu March 09, 2023 03:35 PM

    Hello- this is Matthew. I accidentally locked myself out of my account. 

    Thank you for your reply. I did repeat the analysis to obtain bootstrap confidence intervals. 

    My issue is that when I loaded the output for the bootstrap confidence intervals (I will paste an example of a similar looking output below but without the missing value), one of the values listed under the "Lower" column is missing and just has a "..." in it's place. I'm wondering how to interpret the "...". 



    ------------------------------
    Katie Cassidy
    ------------------------------



  • 4.  RE: Question about interpreting AMOS Output

    Posted Thu March 09, 2023 04:46 PM
    Edited by David Dwyer Thu March 09, 2023 05:54 PM

    Hi,
    Are you asking about the causality arrow pointing from your exogenous 'author' variable to your endogenous 'contacttot' variable?

    ---> is just an arrow.

    Are you saying one of the values in the actual "Lower" bound column of the table has '...' in it instead of a value?  Is this value clickable in the Text Output Viewer interface?  When such a value happens ('***')in the "P" column, it means that the value is too small to display (e.g. p<0.001).  Perhaps the '...' here is similar, meaning the value is too small to display or perhaps the value could not be computed.  Do you have other errors or warnings? Are you running with Maximum Likelihood estimation for missing data?   

    Have you "allowed" non-positive definite matrices or checked the box to try and fit unidentified models 

    I just ran Example 19.  In it I can see the '...' value under the "P" column.  Because this happens for the regression estimates that had been fixed at 1, I think it is safe to conclude the value could not be computed.



    ------------------------------
    David Dwyer
    SPSS Technical Support
    IBM Software
    ------------------------------