SPSS Statistics

 View Only
Expand all | Collapse all

Confidence and Prediction Intervals for simple linear regression

  • 1.  Confidence and Prediction Intervals for simple linear regression

    Posted Thu April 22, 2021 07:11 PM

    Hi,

    How can I find the predicted y value for a given x using the regression model calculated by SPSS? I have my slope and y-intercept calculated by SPSS so can now input x value and calculate the predicted y value manually but I was wondering if this is possible to do on SPSS? If so, I would like to see the confidence intervals for the predicted y value (given certain x value) so that I can generalise it to the population.

    Thank you






    #SPSSStatistics
    #Support
    #SupportMigration


  • 2.  RE: Confidence and Prediction Intervals for simple linear regression

    Posted Thu April 22, 2021 07:17 PM

    Export your model as XML (on the Save subdialog) and then look at the Scoring Wizard on Utilities. You can also use a Selection variable in the Regression dialog and generate predictions for the rest of the sample.






    #SPSSStatistics
    #Support
    #SupportMigration


  • 3.  RE: Confidence and Prediction Intervals for simple linear regression

    Posted Fri April 23, 2021 06:23 AM

    Hi Jon,

    Thank you for your answer. This is still not what I am looking for. I want to find a predicted value y for an x value that is currently not in my dataset. I then want to see what is the confidence interval of that predicted value y.

    When I use the Utilities -> Scoring Wizard it only lets me calculate predicted values y for the x value in my dataset, not for any other value that is within that range but is not exactly there.

    Would you be able to help?

    Thank you






    #SPSSStatistics
    #Support
    #SupportMigration


  • 4.  RE: Confidence and Prediction Intervals for simple linear regression

    Posted Fri April 23, 2021 01:01 PM

    That is not correct. After you save the estimated model as xml, you can activate a different dataset and apply the model to it using the Scoring Wizard. The only problem would be if you have a categorical variable (as dummies) but in the new dataset, there is a value for it that was not in the estimation data.






    #SPSSStatistics
    #Support
    #SupportMigration