SPSS Statistics

 View Only
  • 1.  Comparing Variables Measured on Different Scales for Multiple Regression Analysis in SPSS

    Posted 15 days ago

    Hi everyone, If anybody could give some advice you'd be my life saver. I measured three variables in an online questionnaire. One with the Likert scale, one with a modified Likert scale with 6 levels and one, the NPI-15 as a forced-choice format with dichotomous answers to each item. I wanted to analyze the relationship between the variables using a multiple regression analysis in SPSS. Now for the NPI, you normally simply calculate the grand sum for evaluation, since every item can give 1 or 0 points, depending on the choice, the max. would be 15. Is it even possible to compare two variables that have been measured using a Likert scale (the variable gets calculated using the MEAN function in SPSS) to one with a forced-choice-format (the variable gets calculated using the SUM function in SPSS)? Or could I just calculated the mean for the NPI-15 anyway despite of the format used in collecting the data? Or will this mess up my data? This is my first time working with SPSS, please be kind if my question is super dumb.



    ------------------------------
    belulu Stud
    ------------------------------


  • 2.  RE: Comparing Variables Measured on Different Scales for Multiple Regression Analysis in SPSS

    IBM Champion
    Posted 15 days ago
    I don't see why you would need multiple regression here.  I'd suggest looking at nonparametric correlation coefficients - Kendall's tau-b or Spearman. (Analyze > Correlate > Bivariation)

    You might want to go further and look at heterogeneous correlations (Analyze > Correlate > Heterogenous Correlation).  For this, you need to be sure that the measurement levels are set correctly via the Data Editor Variable View.

    You would also need to install the SPSSINC HETCOR extension command via Extensions > Extension Hub.  If you have Statistics V29, you can then just run that, but if your version is older, there is some complicated setup that probably isn't worth the trouble here.

    --





  • 3.  RE: Comparing Variables Measured on Different Scales for Multiple Regression Analysis in SPSS

    Posted 15 days ago

    Hi Jon,

    thank you for your reply. I am writing my Bachelor thesis and we received the specification that I need to use multiple regression in my case. Is that possible when yes do I need to change or transform data? Is there a process that I need to follow?



    ------------------------------
    belulu Stud
    ------------------------------



  • 4.  RE: Comparing Variables Measured on Different Scales for Multiple Regression Analysis in SPSS

    IBM Champion
    Posted 15 days ago
    Are there other variables involved?  What equation did they want you to run?--





  • 5.  RE: Comparing Variables Measured on Different Scales for Multiple Regression Analysis in SPSS

    Posted 15 days ago

     I measured three variables in an online questionnaire. One with the Likert scale, one with a modified Likert scale with 6 levels and one, the NPI-15 as a forced-choice format with dichotomous answers to each item. I don't quite understand the second question



    ------------------------------
    belulu Stud
    ------------------------------



  • 6.  RE: Comparing Variables Measured on Different Scales for Multiple Regression Analysis in SPSS

    IBM Champion
    Posted 14 days ago
    If you are going to run a linear regression, what would the dependent variable be?  Are there other variables that you are supposed to include in the equation?

    --





  • 7.  RE: Comparing Variables Measured on Different Scales for Multiple Regression Analysis in SPSS

    Posted 14 days ago

    Thank you for your replies Jon!

    Dependant variable (ambivalent sxsm, AS S) = metric, continuous
    Independent variable 1 (Honesty-Humility, HEXACO) = metric, continuous
    Independent variable 2 (narcissism, NPI-15) = forced-choice format with dichotomous answers to each item (Now for the NPI, you normally simply calculate the grand sum for evaluation, since every item can give 1 or 0 points, depending on the choice, the max. would be 15)



    ------------------------------
    belulu Stud
    ------------------------------



  • 8.  RE: Comparing Variables Measured on Different Scales for Multiple Regression Analysis in SPSS

    IBM Champion
    Posted 14 days ago
    Well, you can just run the equation, but ask for collinearity statistics and residual plots too.--





  • 9.  RE: Comparing Variables Measured on Different Scales for Multiple Regression Analysis in SPSS

    Posted 13 days ago

    Thank you for your replies Jon!

    Dependant variable (ambivalent sexism, ASS) = metric, continuous
    Independent variable 1 (Honesty-Humility, HEXACO) = metric, continuous
    Independent variable 2 (narcissism, NPI-15) = forced-choice format with dichotomous answers to each item (Now for the NPI, you normally simply calculate the grand sum for evaluation, since every item can give 1 or 0 points, depending on the choice, the max. would be 15)



    ------------------------------
    belulu Stud
    ------------------------------



  • 10.  RE: Comparing Variables Measured on Different Scales for Multiple Regression Analysis in SPSS

    Posted 13 days ago

    Hi Jon thank you for your replies!
    Dependant Variable (ambivalent sexism, ASS) = metric, continuous
    -Independent Variable 1 (Honesty-Humility, HEXACO) = metric, continuous
     Independant variable 2 (narcissism, NPI-15) = forced-choice format with dichotomous answers to each item (Now for the NPI, you normally simply calculate the grand sum for evaluation, since every item can give 1 or 0 points, depending on the choice, the max. would be 15)



    ------------------------------
    belulu Stud
    ------------------------------