Advertisement
If you have a new account but are having problems posting or verifying your account, please email us on hello@boards.ie for help. Thanks :)
Hello all! Please ensure that you are posting a new thread or question in the appropriate forum. The Feedback forum is overwhelmed with questions that are having to be moved elsewhere. If you need help to verify your account contact hello@boards.ie

SPSS multiple regression graphs

Options
  • 18-02-2008 2:11pm
    #1
    TheNibbler
    Registered Users Posts: 271 ✭✭


    Wasn't sure whether to put this in the maths thread or here but thought I'd go for here as my question is to do with presentation of results rather than the maths behind them. Also assume that like me, most people studying Psychology have to use SPSS a lot!

    I'm doing some research at the mo and have just carried out a backward stepwise multiple regression using SPSS with three predictor variables. Have put the results into a table but I know the person correcting my paper is a big fan of graphs so was wondering if anyone knows how best to present these type of results in a graph (or graphs)??


Comments

  • Options
    #2
    LilMrsDahamsta
    Registered Users Posts: 451 ✭✭LilMrsDahamsta


    Andy Field wrote the SPSS bible. This is the website he put together for his students and anyone else in need of statistical assistance. Hopefully you will find what you are looking for there.


  • Options
    #3
    2Scoops
    Registered Users Posts: 1,845 ✭✭✭2Scoops


    There's no real way to graph a multiple regression - it's just a straight line. A table is the best way to present the results. But you can still use Powerpoint to demonstrate that you understand how it worked with judicious use of highlighting/arrows or animation.

    Also, you can graph other things like the linear relation between the dependent and each predictor separately and all the residual vs. predictor stuff that SPSS can do from regression menu... if your supervisor is an easily-impressed anal retentive! ;)

    Edit: Lightbulb! Maybe a graph of percentage variance explained by the model, increasing with the addition of each predictor? That would look nice.


  • Options
    #4
    TheNibbler
    Registered Users Posts: 271 ✭✭TheNibbler


    LilMsMelodrama wrote: »
    Andy Field wrote the SPSS bible.

    Actually sitting here with his book beside me!! Will definitely have a look at his website too though,thanks.

    That graph of increasing variance explained sounds good but......(really going to show my ignorance of stats now!)..........do you think it would make a difference that I did a backwards stepwise but I would be presenting the steps in the graph in the opposite way??! I suppose it probably wouldn't but just thought I should check!


  • Options
    #5
    2Scoops
    Registered Users Posts: 1,845 ✭✭✭2Scoops


    TheNibbler wrote: »
    That graph of increasing variance explained sounds good but......(really going to show my ignorance of stats now!)..........do you think it would make a difference that I did a backwards stepwise but I would be presenting the steps in the graph in the opposite way??! I suppose it probably wouldn't but just thought I should check!

    If anything, it's exposing my ignorance of stats because it will make a difference! So with backward, it puts all 3 predictors in and removes them in steps if they do not add a significant amount of predictive power to the model.

    The most important point is this: it really depends on your results and what you want to emphasize. So, you could graph the amount of variance explained by each predictor alone and then compare to the overall model.

    Or, let's say one of your predictors was shown to be quite useless once you had the other two, the graphs could be used to show how small its contribution was.


Advertisement