Virata manual

logit

Explains a yes/no outcome.

What we have

  • the same factor notation as regress
  • coefficients on the log-odds scale

Coming later

  • or, to print odds ratios instead of coefficients
  • predict and margins afterwards

How you write it

logit depvar indepvars [if]

Square brackets mean optional. A vertical bar means choose one.

What it does

When the column you are explaining is 0 or 1 — left or stayed, poor or not poor, passed or failed — a straight line is the wrong shape, because it will happily predict a probability of 1.4. Logistic regression fits an S-shaped curve that cannot leave the range 0 to 1.

The coefficients are on the log-odds scale, which is hard to read directly. What is easy to read is the sign (does this raise or lower the chance?) and, after exponentiating, the odds ratio: a value of 1.5 means the odds are half again as high.

Watch out. Odds ratios are not risk ratios. When the outcome is common, an odds ratio of 2 is a good deal less than twice the risk.

Try it

Every example below was run after sysuse vira.sample02, clear.

. generate poor = mpce < 2500
The same thing in pandas and in R

pandas

import numpy as np
import pandas as pd

def stata_missing(value):
    """Stata counts a missing value as larger than any number."""
    if isinstance(value, pd.Series):
        if pd.api.types.is_numeric_dtype(value):
            return value.astype("float64").fillna(np.inf)
        return value.fillna("")
    if value is None or (isinstance(value, float) and np.isnan(value)):
        return np.inf
    return value

df["poor"] = (stata_missing(df["mpce"]) < 2500)

R

library(dplyr)
stata_missing <- function(x)
    if (is.numeric(x)) ifelse(is.na(x), Inf, x) else ifelse(is.na(x), "", x)
df <- df %>% mutate(poor = (stata_missing(mpce) < 2500))

# This R was run in a real R (4.6.1) and gave the same numbers.
. logit poor age i.sector
Logistic regression                          Number of obs      =        6,000
                                             LR chi2            =       281.73
                                             Prob > chi2        =       0.0000
                                             Pseudo R2          =       0.0466

------------------------------------------------------------------------------
        poor | Coefficient  Std. err.      z   P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
sector=Urban |   -1.266807  .08599801  -14.73   0.000     -1.43536   -1.098254
         age |   -.0074517  .00175454   -4.25   0.000    -.0108905   -.0040128
       _cons |   -.7060428  .08813655   -8.01   0.000    -.8787873   -.5332984
------------------------------------------------------------------------------
The same thing in pandas and in R

pandas

import statsmodels.formula.api as smf

model = smf.logit("poor ~ age + C(sector)", data=df).fit()

R

model <- glm(poor ~ age + factor(sector),
    family = binomial(link = "logit"), data = df)
summary(model)

# This R was run in a real R (4.6.1) and gave the same numbers.

Nearby

regress probit poisson nbreg ologit oprobit