+ Reply to Thread
Results 1 to 7 of 7

Thread: Comparing Ordinal IV and Ordinal DV

  1. #1
    Points: 415, Level: 8
    Level completed: 30%, Points required for next Level: 35

    Posts
    10
    Thanks
    5
    Thanked 0 Times in 0 Posts

    Comparing Ordinal IV and Ordinal DV




    Hey guys

    another question

    So I have set up this experiment about the effects of Virtual Surgical Planning on various aspects of outcomes of the procedure.

    I need to figure out whether Virtual Surgical Planning has significant effects on "Flap Loss" (medical jargon for the skin graft to cover the wound was rejected or became necrotic and didn't survive the procedure).

    I ALSO need to see if co-morbidities such as smoking, hypertension, diabetes, etc, had significant effects on flap loss as well.

    So how I have my variables in SPSS set up is

    VSP - ordinal
    0 = did not receive VSP
    1 = did receive VSP

    This is my independent variable

    Then I have

    Flap Loss - ordinal
    0 = no flap loss
    1 = partial flap loss
    2 = total flap loss

    Hypertension - ordinal
    0 = no hypertension
    1 = hypertension

    Tobacco - ordianl
    0 = does not use tobacco
    1 = uses tobacco

    etc etc etc

    So if I am trying to analyze my independent variable against one dependent variable at a time, what test would I run? Chi-square is giving me weird results, something due I think to the frequency being less than 5 for a lot of these variables. And I don't think I can do Kruskal-Wallace, if that would even be what I need to do.

    So basically, comparing an ordinal independent variable against 1 ordinal dependent variable.

    Any help would be greatly appreciated.

    I attached my data sheet if you are interested (it's from SPSS but I just threw it into an Excel sheet)
    Attached Files

  2. #2
    Fortran must die
    Points: 58,790, Level: 100
    Level completed: 0%, Points required for next Level: 0
    noetsi's Avatar
    Posts
    6,532
    Thanks
    692
    Thanked 915 Times in 874 Posts

    Re: Comparing Ordinal IV and Ordinal DV

    I would run binary logistic regression since your DV flap loss has two levels. You will run you IV as a dummy variable in this case. I would think your comorbidities such as smoking are independent variables not dependent ones since you appear to be interested in their effect on flap loss. But I may not understand your question.
    "Very few theories have been abandoned because they were found to be invalid on the basis of empirical evidence...." Spanos, 1995

  3. The Following User Says Thank You to noetsi For This Useful Post:

    mjbarteau (08-22-2013)

  4. #3
    Points: 415, Level: 8
    Level completed: 30%, Points required for next Level: 35

    Posts
    10
    Thanks
    5
    Thanked 0 Times in 0 Posts

    Re: Comparing Ordinal IV and Ordinal DV

    Quote Originally Posted by noetsi View Post
    I would run binary logistic regression since your DV flap loss has two levels. You will run you IV as a dummy variable in this case. I would think your comorbidities such as smoking are independent variables not dependent ones since you appear to be interested in their effect on flap loss. But I may not understand your question.
    you know you are totally right

    yes I also forgot to mention the flap loss has multiple levels

    thanks for pointing that out, it's been a long day and this leg of my data crunching has really gotten to me

    thanks!

  5. #4
    Points: 415, Level: 8
    Level completed: 30%, Points required for next Level: 35

    Posts
    10
    Thanks
    5
    Thanked 0 Times in 0 Posts

    Re: Comparing Ordinal IV and Ordinal DV

    Okay so I ran the binary regression against basically all the DV's I have entered so far

    I had to change "Flap Loss" to just 0 or 1. 1 meaning "partial or total flap loss"; when I had multiple levels to this DV it wouldn't let me do the regression.

    ANYWAY

    so this data looks like total crap right? Is that what this table at the bottom is saying? my sig is like astronomical...

    It's just frustrating, my boss wants me to run all these tests but it's pretty apparent just from looking at it that there is nothing in the numbers, yet I have to keep digging

    attached the regression output

    again thank you so much
    Attached Files

  6. #5
    Points: 415, Level: 8
    Level completed: 30%, Points required for next Level: 35

    Posts
    10
    Thanks
    5
    Thanked 0 Times in 0 Posts

    Re: Comparing Ordinal IV and Ordinal DV

    Just a quick bump.... any other test that would be better?

    I'm just getting gigantic numbers for my results

  7. #6
    RoboStataRaptor
    Points: 11,861, Level: 71
    Level completed: 53%, Points required for next Level: 189
    bukharin's Avatar
    Location
    Sydney, Australia
    Posts
    1,312
    Thanks
    12
    Thanked 326 Times in 315 Posts

    Re: Comparing Ordinal IV and Ordinal DV

    You only have 3 events out of a total of 22 cases. Unfortunately this is not anywhere near enough to construct a meaningful regression model (a common rule of thumb is that you need 10 events, or non-events [whichever is lower], for each covariate in the model).

  8. The Following User Says Thank You to bukharin For This Useful Post:

    mjbarteau (08-25-2013)

  9. #7
    Points: 415, Level: 8
    Level completed: 30%, Points required for next Level: 35

    Posts
    10
    Thanks
    5
    Thanked 0 Times in 0 Posts

    Re: Comparing Ordinal IV and Ordinal DV


    Awesome! thanks for the rule of thumb. As you can tell I'm pretty new to this

    Thanks

+ Reply to Thread

           




Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats