SPSS Statistics

 View Only
  • 1.  Logistic Regression

    Posted Wed May 24, 2023 11:01 AM

    Hello! I am new to research and I want to see if this makes sense. I have a dichotomous dependent variable (i.e., vaccination status) and a categorical independent/predictor variable (vaccine hesitancy). The categories for vaccine hesitancy are yes (i.e., vaccine hesitant) and no (i.e., vaccine hesitant). In completing a automatic recode for my independent variables, they become numeric. From there, is it okay to run a regression? If so, does it matter if I run a linear or logistic regression? 

    Similarly, I am doing the same thing for multiple predictors/independent variables (i.e., social need; housing, employment, childcare, economic stability, etc.) with vaccination status (dependent/dichotomous). I manipulated the independent variables to denote need (so "yes" for any need endorsed and "no" for lack of need). So, if someone endorses a need for employment then "yes" was recoded to a numeric value, for example. Am I now able to run a linear or logistic regression? 



    ------------------------------
    ashley craft
    ------------------------------


  • 2.  RE: Logistic Regression

    Posted Wed May 24, 2023 12:59 PM

    Hi @ashley craft 

    Yes.

    Here are the pages in our documentation covering Logistic Regression



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



  • 3.  RE: Logistic Regression

    Posted Wed May 24, 2023 01:38 PM

    Thank you. I took a look and it leaves me with another question. Do I need to run something else before the logistic regression such as a Chi Square? If so, why is that? 



    ------------------------------
    ashley craft
    ------------------------------



  • 4.  RE: Logistic Regression

    IBM Champion
    Posted Wed May 24, 2023 07:12 PM
    I would like to suggest that you go to Help > Case Studies and find the case study for logistic regression.  That will walk you through the process and explain how to interpret the results.

    --





  • 5.  RE: Logistic Regression

    Posted Sat May 27, 2023 12:00 PM
    Maximum iterations reachedHello. I was unable to find this on the case studies. Are you aware of what this means? Is my analysis invalid based on this? 


    ------------------------------
    ashley craft
    ------------------------------



  • 6.  RE: Logistic Regression

    Posted Sat May 27, 2023 11:54 AM

    I did not see this in the "logistic regression" link above. Are you familiar with what this may mean? Is my analysis invalid? 



    ------------------------------
    ashley craft
    ------------------------------



  • 7.  RE: Logistic Regression

    IBM Champion
    Posted Wed May 24, 2023 01:10 PM
    With a dichotomous dependent variable, you would use logistic regression rather than regular regression, which would not be suitable.

    With only one dichotomous independent variables as the regressors, you could run a logistic regression, but you could just as well run a 2x2 crosstab.

    With multiple independent variables, logistic regression would be needed to account for all the variables simultaneously.

    --