**Version info**: Code for this page was tested in Stata 12.

Multinomial logistic regression is used to model nominal outcome variables, in which the log odds of the outcomes are modeled as a linear combination of the predictor variables.

**Please note:** The purpose of this page is to show how to use various data analysis commands.
It does not cover all aspects of the research process which researchers are expected to do. In
particular, it does not cover data cleaning and checking, verification of assumptions, model
diagnostics and potential follow-up analyses.

## Examples of multinomial logistic regression

Example 1. People’s occupational choices might be influenced by their parents’ occupations and their own education level. We can study the relationship of one’s occupation choice with education level and father’s occupation. The occupational choices will be the outcome variable which consists of categories of occupations.

Example 2. A biologist may be interested in food choices that alligators make. Adult alligators might have different preferences from young ones. The outcome variable here will be the types of food, and the predictor variables might be size of the alligators and other environmental variables.

Example 3. Entering high school students make program choices among general program, vocational program and academic program. Their choice might be modeled using their writing score and their social economic status.

## Description of the data

For our data analysis example, we will expand the third example using the **
hsbdemo** data set. Let’s first read in the data.

use http://www.ats.ucla.edu/stat/data/hsbdemo, clear

The data set contains variables on 200 students. The outcome variable is **prog**, program type. The predictor variables
are social economic status, **ses, ** a three-level categorical variable
and writing score**, write,** a continuous variable. Let’s start with
getting some descriptive statistics of the
variables of interest.

tab prog ses, chi2type of | ses program | low middle high | Total -----------+---------------------------------+---------- general | 16 20 9 | 45 academic | 19 44 42 | 105 vocation | 12 31 7 | 50 -----------+---------------------------------+---------- Total | 47 95 58 | 200 Pearson chi2(4) = 16.6044 Pr = 0.002table prog, con(mean write sd write)------------------------------------ type of | program | mean(write) sd(write) ----------+------------------------- general | 51.33333 9.397776 academic | 56.25714 7.943343 vocation | 46.76 9.318754 ------------------------------------

## Analysis methods you might consider

- Multinomial logistic regression: the focus of this page.
- Multinomial probit regression: similar to multinomial logistic regression but with independent normal error terms.
- Multiple-group discriminant function analysis: A multivariate method for multinomial outcome variables
- Multiple logistic regression analyses, one for each pair of outcomes: One problem with this approach is that each analysis is potentially run on a different sample. The other problem is that without constraining the logistic models, we can end up with the probability of choosing all possible outcome categories greater than 1.
- Collapsing number of categories to two and then doing a logistic regression: This approach suffers from loss of information and changes the original research questions to very different ones.
- Ordinal logistic regression: If the outcome variable is truly ordered and if it also satisfies the assumption of proportional odds, then switching to ordinal logistic regression will make the model more parsimonious.
- Alternative-specific multinomial probit regression: allows different error structures therefore allows to relax the independence of irrelevant alternatives (IIA, see below “Things to Consider”) assumption. This requires that the data structure be choice-specific.
- Nested logit model: also relaxes the IIA assumption, also requires the data structure be choice-specific.

## Multinomial logistic regression

Below we use the **mlogit** command to estimate a multinomial logistic regression
model. The **i.** before **ses** indicates that **ses** is a indicator
variable (i.e.,
categorical variable), and that it should be included in the model. We
have also used the option “**base**” to indicate the category we would want
to use for the baseline comparison group. In the model below, we have chosen to
use the academic program type as the baseline category.

mlogit prog i.ses write, base(2)Iteration 0: log likelihood = -204.09667 Iteration 1: log likelihood = -180.80105 Iteration 2: log likelihood = -179.98724 Iteration 3: log likelihood = -179.98173 Iteration 4: log likelihood = -179.98173 Multinomial logistic regression Number of obs = 200 LR chi2(6) = 48.23 Prob > chi2 = 0.0000 Log likelihood = -179.98173 Pseudo R2 = 0.1182 ------------------------------------------------------------------------------ prog | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- general | ses | 2 | -.533291 .4437321 -1.20 0.229 -1.40299 .336408 3 | -1.162832 .5142195 -2.26 0.024 -2.170684 -.1549804 | write | -.0579284 .0214109 -2.71 0.007 -.0998931 -.0159637 _cons | 2.852186 1.166439 2.45 0.014 .5660075 5.138365 -------------+---------------------------------------------------------------- academic | (base outcome) -------------+---------------------------------------------------------------- vocation | ses | 2 | .2913931 .4763737 0.61 0.541 -.6422822 1.225068 3 | -.9826703 .5955669 -1.65 0.099 -2.14996 .1846195 | write | -.1136026 .0222199 -5.11 0.000 -.1571528 -.0700524 _cons | 5.2182 1.163549 4.48 0.000 2.937686 7.498714 ------------------------------------------------------------------------------

- In the output above, we first see the iteration log, indicating how quickly the model converged. The log likelihood (-179.98173) can be usedin comparisons of nested models, but we won’t show an example of comparing models here
- The likelihood ratio chi-square of48.23 with a p-value < 0.0001 tells us that our model as a whole fits significantly better than an empty model (i.e., a model with no predictors)
- The output above has two parts, labeled with the categories of the
outcome variable
**prog**. They correspond to the two equations below:$$ln\left(\frac{P(prog=general)}{P(prog=academic)}\right) = b_{10} + b_{11}(ses=2) + b_{12}(ses=3) + b_{13}write$$

$$ln\left(\frac{P(prog=vocation)}{P(prog=academic)}\right) = b_{20} + b_{21}(ses=2) + b_{22}(ses=3) + b_{23}write$$

where \(b\)’s are the regression coefficients.

- A one-unit increase in the variable
**write**is associated with a .058 decrease in the relative log odds of being in general program vs. academic program . - A one-unit increase in the variable
**write**is associated with a .1136 decrease in the relative log odds of being in vocation program vs. academic program. - The relative log odds of being in general program vs. in academic program will
decrease by 1.163 if moving from the lowest level of
**ses**(**ses**==1) to the highest level of**ses**(**ses**==3).

- A one-unit increase in the variable

The ratio of the probability of choosing one outcome category over the
probability of choosing the baseline category is often referred to as relative risk
(and it is also sometimes referred to as odds as we have just used to described the
regression parameters above). Relative risk can be obtained by
exponentiating the linear equations above, yielding
regression coefficients that are relative risk ratios for a unit change in the
predictor variable. We can use the **rrr** option for **
mlogit** command to display the regression results in terms of relative risk
ratios.

mlogit, rrrMultinomial logistic regression Number of obs = 200 LR chi2(6) = 48.23 Prob > chi2 = 0.0000 Log likelihood = -179.98173 Pseudo R2 = 0.1182 ------------------------------------------------------------------------------ prog | RRR Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- general | ses | 2 | .586671 .2603248 -1.20 0.229 .2458607 1.39991 3 | .3125996 .1607448 -2.26 0.024 .1140996 .856432 | write | .9437175 .0202059 -2.71 0.007 .9049342 .984163 _cons | 17.32562 20.20928 2.45 0.014 1.761221 170.4369 -------------+---------------------------------------------------------------- academic | (base outcome) -------------+---------------------------------------------------------------- vocation | ses | 2 | 1.338291 .6375264 0.61 0.541 .5260904 3.404399 3 | .3743103 .2229268 -1.65 0.099 .1164888 1.202761 | write | .8926126 .0198338 -5.11 0.000 .8545734 .9323449 _cons | 184.6016 214.793 4.48 0.000 18.87213 1805.719 ------------------------------------------------------------------------------

- The relative risk ratio for a one-unit increase in the variable
**write**is .9437 (exp(-.0579284) from the output of the first**mlogit**command above) for being in general program vs. academic program. - The relative risk ratio switching from
**ses**= 1 to 3 is .3126 for being in general program vs. academic program. In other words, the expected risk of staying in the general program is lower for subjects who are high in**ses**.

We can test for an overall effect of **ses**
using the **test** command. Below we see that the overall effect of **ses** is
statistically significant.

test 2.ses 3.ses( 1) [general]2.ses = 0 ( 2) [academic]2.ses = 0 ( 3) [vocation]2.ses = 0 ( 4) [general]3.ses = 0 ( 5) [academic]3.ses = 0 ( 6) [vocation]3.ses = 0 Constraint 2 dropped Constraint 5 dropped chi2( 4) = 10.82 Prob > chi2 = 0.0287

More specifically, we can also test if the effect of **3.ses** in
predicting general vs. academic equals the effect of **3.ses** in
predicting vocation vs. academic using the **test** command again. The test
shows that the effects are not statistically different from each other.

test [general]3.ses = [vocation]3.ses( 1) [general]3.ses - [vocation]3.ses = 0 chi2( 1) = 0.08 Prob > chi2 = 0.7811

You can also use predicted probabilities to help you understand the model.
You can calculate predicted probabilities using the **margins** command. Below we use the **margins** command to
calculate the predicted probability of choosing each program type at each level
of **ses**, holding all other variables in the model at their means. Since
there are three possible outcomes, we will need to use the margins command three
times, one for each outcome value.

margins ses, atmeans predict(outcome(1))Adjusted predictions Number of obs = 200 Model VCE : OIM Expression : Pr(prog==general), predict(outcome(1)) at : 1.ses = .235 (mean) 2.ses = .475 (mean) 3.ses = .29 (mean) write = 52.775 (mean) ------------------------------------------------------------------------------ | Delta-method | Margin Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- ses | 1 | .3581927 .0726423 4.93 0.000 .2158163 .500569 2 | .2283338 .0451162 5.06 0.000 .1399075 .31676 3 | .1784932 .0540486 3.30 0.001 .0725598 .2844266 ------------------------------------------------------------------------------margins ses, atmeans predict(outcome(2))Adjusted predictions Number of obs = 200 Model VCE : OIM Expression : Pr(prog==academic), predict(outcome(2)) at : 1.ses = .235 (mean) 2.ses = .475 (mean) 3.ses = .29 (mean) write = 52.775 (mean) ------------------------------------------------------------------------------ | Delta-method | Margin Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- ses | 1 | .4396842 .0779925 5.64 0.000 .2868217 .5925466 2 | .4777488 .0552593 8.65 0.000 .3694426 .586055 3 | .7009021 .0663042 10.57 0.000 .5709483 .8308559 ------------------------------------------------------------------------------margins ses, atmeans predict(outcome(3))Adjusted predictions Number of obs = 200 Model VCE : OIM Expression : Pr(prog==vocation), predict(outcome(3)) at : 1.ses = .235 (mean) 2.ses = .475 (mean) 3.ses = .29 (mean) write = 52.775 (mean) ------------------------------------------------------------------------------ | Delta-method | Margin Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- ses | 1 | .2021232 .0599647 3.37 0.001 .0845945 .3196519 2 | .2939174 .0503617 5.84 0.000 .1952103 .3926246 3 | .1206047 .04643 2.60 0.009 .0296037 .2116058 ------------------------------------------------------------------------------

We can use the **marginsplot** command to plot predicted
probabilities by ses for each category of **prog**. Plots created
by **marginsplot** are based on the last **margins** command
run. Furthermore, we can combine the three marginsplots into one
graph to facilitate comparison using the **graph combine **
command. As it is generated, each marginsplot must be given a name,
which will be used by **graph combine**. Additionally, we would
like the y-axes to have the same range, so we use the **ycommon**
option with **graph combine **.

margins ses, atmeans predict(outcome(1)) marginsplot, name(general) margins ses, atmeans predict(outcome(2)) marginsplot, name(academic) margins ses, atmeans predict(outcome(3)) marginsplot, name(vocational) graph combine general academic vocational, ycommon

Another way to understand the model using the predicted probabilities is to
look at the averaged predicted probabilities for different values of the
continuous predictor variable **write**, averaging across levels of **ses**.

margins, at(write = (30(10) 70)) predict(outcome(1)) vsquishPredictive margins Number of obs = 200 Model VCE : OIM Expression : Pr(prog==general), predict(outcome(1)) 1._at : write = 30 2._at : write = 40 3._at : write = 50 4._at : write = 60 5._at : write = 70 ------------------------------------------------------------------------------ | Delta-method | Margin Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- _at | 1 | .2130954 .0774327 2.75 0.006 .0613302 .3648606 2 | .2569932 .0529761 4.85 0.000 .1531619 .3608245 3 | .2543008 .0336297 7.56 0.000 .1883878 .3202138 4 | .2057855 .0371536 5.54 0.000 .1329658 .2786052 5 | .1423089 .0481683 2.95 0.003 .0479007 .2367172 ------------------------------------------------------------------------------margins, at(write = (30(10) 70)) predict(outcome(2)) vsquishPredictive margins Number of obs = 200 Model VCE : OIM Expression : Pr(prog==academic), predict(outcome(2)) 1._at : write = 30 2._at : write = 40 3._at : write = 50 4._at : write = 60 5._at : write = 70 ------------------------------------------------------------------------------ | Delta-method | Margin Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- _at | 1 | .1348408 .0525979 2.56 0.010 .0317507 .2379308 2 | .2808143 .0553213 5.08 0.000 .1723867 .389242 3 | .4773283 .0397591 12.01 0.000 .399402 .5552547 4 | .6680752 .0434689 15.37 0.000 .5828776 .7532727 5 | .8075124 .0545504 14.80 0.000 .7005956 .9144291 ------------------------------------------------------------------------------margins, at(write = (30(10) 70)) predict(outcome(3)) vsquishPredictive margins Number of obs = 200 Model VCE : OIM Expression : Pr(prog==vocation), predict(outcome(3)) 1._at : write = 30 2._at : write = 40 3._at : write = 50 4._at : write = 60 5._at : write = 70 ------------------------------------------------------------------------------ | Delta-method | Margin Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- _at | 1 | .6520638 .0944041 6.91 0.000 .4670353 .8370924 2 | .4621925 .0614388 7.52 0.000 .3417747 .5826102 3 | .2683708 .0342932 7.83 0.000 .2011575 .3355842 4 | .1261393 .03019 4.18 0.000 .0669679 .1853107 5 | .0501787 .0216863 2.31 0.021 .0076744 .092683 ------------------------------------------------------------------------------

Sometimes, a couple of plots can convey a good deal amount of information.
Below, we plot the predicted probabilities against the writing score by the
level of **ses** for different levels of the outcome variable.

predict p1 p2 p3 sort write twoway (line p1 write if ses ==1) (line p1 write if ses==2) (line p1 write if ses ==3), /// legend(order(1 "ses = 1" 2 "ses = 2" 3 "ses = 3") ring(0) position(7) row(1)) twoway (line p2 write if ses ==1) (line p2 write if ses==2) (line p2 write if ses ==3), /// legend(order(1 "ses = 1" 2 "ses = 2" 3 "ses = 3") ring(0) position(7) row(1)) twoway (line p3 write if ses ==1) (line p3 write if ses==2) (line p3 write if ses ==3), /// legend(order(1 "ses = 1" 2 "ses = 2" 3 "ses = 3") ring(0) position(7) row(1))

We may also wish to see measures of how well our model fits. This can be particularly useful when comparing
competing models. The user-written command **fitstat** produces a
variety of fit statistics. You can find more information on **fitstat** and
download the program by using command
**search fitstat** in Stata (see
How can I use the search command to search for programs and get additional help?
for more information about using **search**).

fitstatMeasures of Fit for mlogit of prog fit Log-Lik Intercept Only: -204.097 Log-Lik Full Model: -179.982 D(185): 359.963 LR(6): 48.230 Prob > LR: 0.000 McFadden's R2: 0.118 McFadden's Adj R2: 0.045 ML (Cox-Snell) R2: 0.214 Cragg-Uhler(Nagelkerke) R2: 0.246 Count R2: 0.610 Adj Count R2: 0.179 AIC: 1.950 AIC*n: 389.963 BIC: -620.225 BIC': -16.440 BIC used by Stata: 402.350 AIC used by Stata: 375.963

## Things to consider

- The Independence of Irrelevant Alternatives (IIA) assumption: roughly,
the IIA assumption means that adding or deleting alternative outcome
categories does not affect the odds among the remaining outcomes. Test of
the IIA assumption can be performed
by using the Stata command
**mlogtest, iia**. However, as of April 23, 2010,**mlogtest, iia**does not work with factor variables. There are alternative modeling methods that relax the IIA assumption, such as alternative-specific multinomial probit models or nested logit models. - Diagnostics and model fit: unlike logistic regression where there are
many statistics for performing model diagnostics, it is not as
straightforward to do diagnostics with multinomial logistic regression
models. Model fit statistics can be obtained via the
**fitstat**command. For the purpose of detecting outliers or influential data points, one can run separate**logit**models and use the diagnostics tools on each model. - Pseudo-R-Squared: the R-squared offered in the output is basically the change in terms of log-likelihood from the intercept-only model to the current model. It does not convey the same information as the R-square for linear regression, even though it is still “the higher, the better”.
- Sample size: multinomial regression uses a maximum likelihood estimation method, it requires a large sample size. It also uses multiple equations. This implies that it requires an even larger sample size than ordinal or binary logistic regression.
- Complete or quasi-complete separation: Complete separation implies that
the outcome variable separates a predictor variable completely, leading
to perfect prediction by the predictor variable. Unlike running a
**logit**model, Stata does not offer a warning when this happens. Instead it continues to compute iteratively and requires a manual quit to stop the process. Perfect prediction means that only one value of a predictor variable is associated with only one value of the response variable. But you can tell from the output of the regression coefficients that something is wrong. You can then do a two-way tabulation of the outcome variable with the problematic variable to confirm this and then rerun the model without the problematic variable. - Empty cells or small cells: You should check for empty or small cells by doing a cross-tabulation between categorical predictors and the outcome variable. If a cell has very few cases (a small cell), the model may become unstable or it might not even run at all.
- Perhaps your data may not perfectly meet the assumptions and your standard errors might be off the mark. You might wish to see our page that shows alternative methods for computing standard errors that Stata offers.
- Sometimes observations are clustered into groups (e.g., people within families, students within classrooms). In such cases, you may want to see our page on non-independence within clusters.

## See also

## References

- Long, J. S. and Freese, J. (2006) Regression Models for Categorical and Limited Dependent Variables Using Stata, Second Edition. College Station, Texas: Stata Press.
- Hosmer, D. and Lemeshow, S. (2000) Applied Logistic Regression (Second Edition). New York: John Wiley & Sons, Inc..
- Agresti, A. (1996) An Introduction to Categorical Data Analysis. New York: John Wiley & Sons, Inc.