Base level related question

[Deepty Sarder]

What is autometically taken as base level in stata for binary variable while regression (0,1)?

[William Lisowski]

It is not clear to me the exact context of your question.

If you have a binary dependent variable y, then

logistic y x1 x2 ...

models the probability of a non-zero outcome.

If you have a categorical independent variable x1, then

regress y x1 x2 ...

will treat x1 as a continuous variable, while

regress y i.x1 x2 ...

will treat x1 as a categorical variable, and the output of help factor variables tells us that the default is to use the smallest value as the base level.

[Carlo Lazzaro]

Deepty:
as an aside to William's helpful comment, if your (not so clearly posted, though) aim is to understand which level of a given independent categorical variable of your logistic regression Stata takes as reference category, just add the -allbaselevels- option at the end of your code:

sysuse auto.dta
logistic foreign i.rep78 if rep78>=3, allbaselevels

Logistic regression                             Number of obs     =         59
                                                LR chi2(2)        =      21.93
                                                Prob > chi2       =     0.0000
Log likelihood = -27.444671                     Pseudo R2         =     0.2855

------------------------------------------------------------------------------
     foreign | Odds Ratio   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
       rep78 |
          3  |          1  (base)
          4  |          9   6.928203     2.85   0.004     1.990615    40.69094
          5  |       40.5   40.12325     3.74   0.000     5.809942    282.3178
             |
       _cons |   .1111111   .0676201    -3.61   0.000      .033708    .3662539
------------------------------------------------------------------------------
Note: _cons estimates baseline odds.

.

[Deepty Sarder]

[/CODE]
will treat x1 as a categorical variable, and the output of help factor variables tells us that the default is to use the smallest value as the base level.
[/QUOTE]

this is what i needed to know. thanks a lot.....