Hi,
I am trying to obtain some simple slopes in exponentiated form for the effect of a continuous variable in two treatment groups. Margins with the predict(xb) option gives the log-odds effects and matches up with that calculated from the regression table. However, when I try to obtain OR's using expression(exp(predict(xb))), it doesn't do what I expect - i.e exp(.0235472) = 1.0238 and exp(-.0385977) = 0.9621.

Am I specifying something incorrectly?

Thanks.

Code:
. logit lapse_first_14_days i.treatment_condition##c.numberofpreviousdetoxes if seq == 1 

Iteration 0:   log likelihood = -207.77746  
Iteration 1:   log likelihood = -204.98423  
Iteration 2:   log likelihood = -204.98247  
Iteration 3:   log likelihood = -204.98247  

Logistic regression                             Number of obs     =        300
                                                LR chi2(3)        =       5.59
                                                Prob > chi2       =     0.1334
Log likelihood = -204.98247                     Pseudo R2         =     0.0135

---------------------------------------------------------------------------------------------------------------
                          lapse_first_14_days |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
----------------------------------------------+----------------------------------------------------------------
                          treatment_condition |
                                intervention  |  -.3104472   .2813866    -1.10   0.270    -.8619548    .2410603
                      numberofpreviousdetoxes |   .0235472    .048203     0.49   0.625     -.070929    .1180233
                                              |
treatment_condition#c.numberofpreviousdetoxes |
                                intervention  |  -.0621448   .0616824    -1.01   0.314      -.18304    .0587504
                                              |
                                        _cons |   .2476101   .1981055     1.25   0.211    -.1406696    .6358898
---------------------------------------------------------------------------------------------------------------




. margins treatment_condition, dydx(numberofpreviousdetoxes) predict(xb)

Average marginal effects                        Number of obs     =        300
Model VCE    : OIM

Expression   : Linear prediction (log odds), predict(xb)
dy/dx w.r.t. : numberofpreviousdetoxes

-----------------------------------------------------------------------------------------
                        |            Delta-method
                        |      dy/dx   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
numberofpreviousdetoxes |
    treatment_condition |
               control  |   .0235472    .048203     0.49   0.625     -.070929    .1180233
          intervention  |  -.0385977   .0384862    -1.00   0.316    -.1140292    .0368339
-----------------------------------------------------------------------------------------




. margins treatment_condition, dydx(numberofpreviousdetoxes) expression(exp(predict(xb)))

Average marginal effects                        Number of obs     =        300
Model VCE    : OIM

Expression   : exp(predict(xb))
dy/dx w.r.t. : numberofpreviousdetoxes

-----------------------------------------------------------------------------------------
                        |            Delta-method
                        |      dy/dx   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
numberofpreviousdetoxes |
    treatment_condition |
               control  |   .0323045   .0676398     0.48   0.633    -.1002671    .1648761
          intervention  |   -.033014   .0342515    -0.96   0.335    -.1001457    .0341177
-----------------------------------------------------------------------------------------