SPSS Statistics

SPSS Statistics

Your hub for statistical analysis, data management, and data documentation. Connect, learn, and share with your peers! 

 View Only

  • 1.  variables (values, types)

    Posted Sun March 05, 2023 12:23 PM

    Would you please respond to a few questions regarding SPSS basics? I greatly appreciate your responses as I am in a crunch to complete my analyses for my dissertation deadline coming within a few weeks. I did not realize there were so many details I had not paid attention to previously. BY THE WAY, perhaps you would also recommend an SPSS professional (in Long Beach, CA area), who would be willing to review my analyses within a short time frame (a week).

    1. My study looks into 2 conditions (condition 1 and condition 2) should I assign conditions 1 and 2 as "split" or "input" types of variables?

    2. I have missing data; apparently, I have previously learned to assign a random number as a missing value within the variable value, ex. 999 Missing value; after assigning this in the variable view, I inserted 999 for missing values in the data view.

    3. Apparently, from another source, I read, I should have left "." (in the Data view), and the SPSS program would recognize. "." as the missing value. Which approach is correct? Are they both correct?

    I appreciate your response, Ewa



    ------------------------------
    Ewa Burchard
    ------------------------------


  • 2.  RE: variables (values, types)

    Posted Sun March 05, 2023 03:34 PM
    If you use SPLIT FILES, with 1 or more variables named as the split variables, then most procedures will provide separate results for each split.  If you are testing for differences with t tests, regression, etc, that is not what you want.  You would use the variable as the group variable or as a regressor.

    With missing data, you would assign a number like 999 or a string value such as "" as user missing and declare that in the DE Variable View (or by syntax).  That will keep procedures from treating that value as valid.  Most will exclude such cases by default, but you can in some cases use the other values in the case.  A blank such as "" is NOT automatically assumed to be missing.
    Your missing values might already be SYSMIS, which shows as a period in the DE.  In that case those cases will automatically be excluded.  However string variables do not have a SYSMIS value.

    You  also have the option of imputing the missing values via either single or multiple imputation.  Multiple imputation is generally the best approach if you want to use imputation, but it is not supported in all procedures.

    --



    Original Message


Related Content