4
$\begingroup$

I am totally confused by statistics and I would be glad if you could help me.

I have a difficulties to interpret marginal effects in logit model, if my independent variable is log transformed.

I will illustrate my question on the example from my data below. I run a logistic regression in stata

My dependent variable is dummy indicating whether a game is of X Genre. My independent variable is a continuous and log transformed variable (log heterogeneity)

After I run a logit regression:

logit xGenre logheterogeneity + control variables

I get the following results:

The coefficient of my independet variable is .567

STATA:

-X Genre        Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]

-log heterog|  .5655944  .1741354     3.25   0.001     .2242953    .9068935

In order to be able to interpret the results easier, I should look at the marginal effects.

I have used therefore an mfx command.

My results show following: for my independent variable I get dy/dx = .056

STATA:

-variable       dy/dx    Std. Err.     z    P>|z|  [    95% C.I.   ]      X

-log heter.  .0563382      .01688    3.34   0.001   .023259  .089417   3.51361

(*) dy/dx is for discrete change of dummy variable from 0 to 1

Now I am confused on how to interpret my results. Can I say, If my independent variable increase in 10% (log heterog.), then the probability that my Game will be of Genre X increases by 0.56%.

$\endgroup$
5
$\begingroup$

You know that in a logit:

$$Pr[y = 1 \vert x,z] = p = \frac{\exp (\alpha + \beta \cdot \ln x + \gamma z)}{1+\exp (\alpha + \beta \cdot \ln x + \gamma z )}. $$

After some tedious calculus and simplification, the partial of that with respect to $x$ becomes:

$$ \frac{\partial Pr[y=1 \vert x,z]}{\partial x} = \frac{\beta}{x} \cdot p \cdot (1-p). $$

This is (sort of) equivalent to

$$\frac{\Delta p}{\Delta x}=\frac{\beta}{x} \cdot p \cdot (1-p),$$

which can be re-written as

$$\frac{\Delta p}{100 \cdot \frac{ \Delta x}{x}}= \frac{\beta \cdot p \cdot (1-p)}{100}.$$

This is the definition of semi-elasticity, and can be interpreted as the change in probability for a 1% change in $x$.

Here's an example in Stata.* Note that I am using margins instead of the out-of-date mfx to get the average marginal effect of $x$, $\frac{1}{N}\Sigma_{i=1}^N\frac{\beta \cdot p_i \cdot (1-p_i)}{100}$:

. sysuse auto, clear
(1978 Automobile Data)

. gen ln_price = ln(price)

. logit foreign ln_price mpg weight, nolog

Logistic regression                             Number of obs     =         74
                                                LR chi2(3)        =      57.69
                                                Prob > chi2       =     0.0000
Log likelihood = -16.185932                     Pseudo R2         =     0.6406

------------------------------------------------------------------------------
     foreign |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
    ln_price |   6.851215    2.11763     3.24   0.001     2.700737    11.00169
         mpg |  -.0880842   .1031317    -0.85   0.393    -.2902186    .1140503
      weight |  -.0062268   .0017269    -3.61   0.000    -.0096115   -.0028422
       _cons |  -41.32383   16.24003    -2.54   0.011    -73.15371   -9.493947
------------------------------------------------------------------------------

. margins, expression(_b[ln_price]*predict()*(1-predict())/100)

Predictive margins                              Number of obs     =         74
Model VCE    : OIM

Expression   : _b[ln_price]*predict()*(1-predict())/100

------------------------------------------------------------------------------
             |            Delta-method
             |     Margin   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
       _cons |   .0046371   .0007965     5.82   0.000      .003076    .0061982
------------------------------------------------------------------------------

This means that for a 1% increase in price, the probability that a car is foreign increases by 0.005 on a [0,1] scale. Or a 10% increase in price gives you a 0.05 increase. In this date, about 0.3 of the cars are foreign, so these are economically and statistically significant.


Edit:

A good way to do this in Stata 10 is to install the user-written command margeff:

. margeff, dydx(ln_price) replace

Average partial effects after margeff
      y  = Pr(foreign) 

------------------------------------------------------------------------------
    variable |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
    ln_price |   .4637103   .0796514     5.82   0.000     .3075964    .6198241
         mpg |  -.0059616    .006781    -0.88   0.379    -.0192522     .007329
      weight |  -.0004214   .0000417   -10.11   0.000    -.0005031   -.0003398
------------------------------------------------------------------------------

. lincom _b[ln_price]/100

 ( 1)  .01*ln_price = 0

------------------------------------------------------------------------------
    variable |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .0046371   .0007965     5.82   0.000      .003076    .0061982
------------------------------------------------------------------------------

*This is actually not a great empirical example since the relationship in the data has an inverted-U shape.

$\endgroup$
  • $\begingroup$ Thank You very much for a quick answer and clear example! Very helpful. I will try the margins command instead of mfx. $\endgroup$ – Alina Lobova Apr 22 '15 at 8:32
  • $\begingroup$ Unfortunately I have Stata 10.1, after a quick googeling I found that I cannot use this command in Stata 10. Is there any way to do the same in Stata 10? $\endgroup$ – Alina Lobova Apr 22 '15 at 8:51
  • 1
    $\begingroup$ @AlinaLobova Luckily, it is pretty easy. See the edit above. It would be good practice to include such info in the question going forward. $\endgroup$ – Dimitriy V. Masterov Apr 22 '15 at 15:48
  • $\begingroup$ the commands you have provided for stata10 worked fine! Will definitely include the program version next time. I have another question,considering the economical significance. In your answer you have written In this date, about 0.3 of the cars are foreign, so these are economically and statistically significant. I understand that economical significance, says if my coefficient/result makes sense in practice. Now I am confused, to see that if in my data only 0.3 cars are foreign, even a small increase such as 0.005 is economically significant. $\endgroup$ – Alina Lobova Apr 22 '15 at 18:38
  • $\begingroup$ @AlinaLobova I would define ES as practical significance, as measured by the sign and magnitude of the AMEs, instead of p-values. 0.3 means that about a third of the cars are foreign. If a 10% price increase takes that to 0.35, that is a 17% increase in imports in the sample. $\endgroup$ – Dimitriy V. Masterov Apr 22 '15 at 18:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.