Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Ordered Logit Regression Interaction Term

    I am using ordered logit regression with interaction term for my dissertation, however I am new at this and would like some help interpreting the results.
    My results are from a survey.
    My dependent variable is DGoodIdea 9B, which is the change in belief about whether it is a good idea, where 1 is negative change, 2 is no change and 3 is positive change.
    Frame is 0,1 where 0 is they received the loss frame and 1 is they received the gain frame.
    LengthLONG is 0,1 where 0 is working for a short time at the company and 1 is working for a long time at the company.
    Gender is 0,1 where 0 is male and 1 is female.

    My hypothesis: the longer the people have been with their current company, the less susceptible they are to the frame changing their minds.

    I need help on how to interpret the two significant results? Thank you!

    Code:
     ologit DGoodIdea9B i.Frame#i. LengthLONG i.Gender, or
    
    Iteration 0:   log likelihood = -565.18388  
    Iteration 1:   log likelihood = -559.38126  
    Iteration 2:   log likelihood = -559.36882  
    Iteration 3:   log likelihood = -559.36882  
    
    Ordered logistic regression                       Number of obs   =        564
                                                      LR chi2(4)      =      11.63
                                                      Prob > chi2     =     0.0203
    Log likelihood = -559.36882                       Pseudo R2       =     0.0103
    
    ----------------------------------------------------------------------------------
         DGoodIdea9B | Odds Ratio   Std. Err.      z    P>|z|     [95% Conf. Interval]
    -----------------+----------------------------------------------------------------
    Frame#LengthLONG |
                0 1  |   1.345841   .3044854     1.31   0.189     .8638065    2.096868
                1 0  |   2.049425   .5006418     2.94   0.003     1.269689    3.308012
                1 1  |   1.769977   .4059233     2.49   0.013     1.129158    2.774474
                     |
            1.Gender |   1.181873    .192872     1.02   0.306     .8583451    1.627346
    -----------------+----------------------------------------------------------------
               /cut1 |   -1.26181   .1966678                     -1.647271   -.8763479
               /cut2 |   1.115826   .1940876                      .7354211     1.49623
    ----------------------------------------------------------------------------------

  • #2
    Dear Irene, are you sure DGoodIdea is an ordered variable? To me, it looks more nominal variable and I would suggest to use multinomial logit and not ordered one, as there is no accepted ordered between the values. You should better choose the multinomial logit and set as base category the second value (2 "no change") and compare it to negative or positive change.

    Comment

    Working...
    X