Title stata.com logit — Logistic regression, reporting ...

1 Logistic regression , reporting coefficientsDescriptionQuick startMenuSyntaxOptionsRemarks and examplesStored resultsMethods and formulasReferencesAlso seeDescriptionlogitfits a logit model for a binary response by maximum likelihood; it models the probabilityof a positive outcome given a set of to nonzero and nonmissing (typicallydepvarequal to one) indicates a positive outcome, whereasdepvarequal to zero indicates a see [R] Logistic ;logisticdisplays estimates as odds ratios. Many users prefer thelogisticcommand tologit. Results are the same regardless of which you use both are the maximum-likelihood estimator. Several auxiliary commands that can be run afterlogit,probit, orlogisticestimation are described in [R] Logistic startLogit model ofyonx1andx2logit y x1 x2 Add indicators for categorical variablealogit y x1 x2 cluster robust standard errors for clustering by levels ofcvarlogit y x1 x2 , vce(cluster cvar)Save separate coefficient estimates for each level _b, by(cvar) saving(myresults): logit y x1 x2 for complex survey design usingsvysetdatasvy.

2 logit y x1 x2 >Binary outcomes> Logistic regression12 logit Logistic regression , reporting coefficientsSyntaxlogitdepvar[indepvars] [if] [in] [weight] [,options]optionsDescriptionModelnoconst antsuppress constant termoffset(varname)includevarnamein model with coefficient constrained to 1asisretain perfect predictor variablesconstraints(constraints)apply specified linear constraintsSE/Robustvce(vcetype)vcetypem ay beoim,robust,clusterclustvar,bootstrap, orjackknifeReportinglevel(#)set confidence level; default islevel(95)orreport odds ratiosnocnsreportdo not display constraintsdisplayoptionscontrol columns and column formats, row spacing, line width,display of omitted variables and base and empty cells, andfactor-variable labelingMaximizationmaximizeoptionscontr ol the maximization process; seldom usednocoefdo not display coefficient table; seldom usedcollinearkeep collinear variablescoeflegenddisplay legend instead of statisticsindepvarsmay contain factor variables; see[U] Factor contain time-series operators; see[U] Time-series ,bootstrap,by,collect,fmm,fp,jackknife,m fp,mi estimate,nestreg,rolling,statsby,stepwis e, andsvyare allowed; see[U] Prefix commands.

3 For more details, see [BAYES]bayes: logitand [FMM]fmm: (bootstrap)andvce(jackknife)are not allowed with themi estimateprefix; see [MI]mi are not allowed with thebootstrapprefix; see [R] (),nocoef, and weights are not allowed with thesvyprefix; see [SVY] ,iweights, andpweights are allowed; see[U] ,collinear, andcoeflegenddo not appear in the dialog [U] 20 estimation and postestimation commandsfor more capabilities of estimation Model noconstant,offset(varname),constraints(c onstraints); see [R] estimation retention of perfect predictor variables and their associated perfectly predicted observationsand may produce instabilities in maximization; see [R] Logistic regression , reporting coefficients 3 SE/Robust vce(vcetype)specifies the type of standard error reported, which includes types that are derivedfrom asymptotic theory (oim), that are robust to some kinds of misspecification (robust), thatallow for intragroup correlation (clusterclustvar), and that use bootstrap or jackknife methods(bootstrap,jackknife); see [R]vceoption.

4 reporting level(#); see [R] estimation the estimated coefficients transformed to odds ratios, that is,ebrather thanb. Standard errorsand confidence intervals are similarly transformed. This option affects how results are displayed,not how they are be specified at estimation or when replaying previouslyestimated ; see [R] estimation :noci,nopvalues,noomitted,vsquish,noempt ycells,baselevels,allbaselevels,nofvlabe l,fvwrap(#),fvwrapon(style),cformat(%fmt ),pformat(%fmt),sformat(%fmt), andnolstretch; see [R] estimation options. Maximization maximizeoptions:difficult,technique(algo rithmspec),iterate(#),[no]log,trace,grad ient,showstep,hessian,showtolerance,tole rance(#),ltolerance(#),nrtolerance(#),no nrtolerance, andfrom(initspecs); see [R]Maximize. These options areseldom following options are available withlogitbut are not shown in the dialog box:nocoefspecifies that the coefficient table not be displayed.

5 This option is sometimes used by programwriters but is of no use ,coeflegend; see [R] estimation and are presented under the following headings:Basic usageModel identificationBasic usagelogitfits maximum likelihood models with dichotomous dependent (left-hand-side) variablescoded as 0/1 (or, more precisely, coded as 0 and not-0).For grouped data or data in binomial form, a probit model can be fit usingglmwith thefamily(binomial)andlink( logit ) 1We have data on the make, weight, and mileage rating of 22 foreign and 52 domestic wish to fit a logit model explaining whether a car is foreign on the basis of its weight and is an overview of our data:4 logit Logistic regression , reporting coefficients. use (1978 automobile data). keep make mpg weight foreign. describeContains data from : 74 1978 automobile dataVariables: 4 13 Apr 2020 17:45(_dta has notes)Variable Storage Display Valuename type format label Variable labelmake str18 %-18s Make and modelmpg int % Mileage (mpg)weight int % Weight (lbs.)

6 Foreign byte % origin Car originSorted by: foreignNote: Dataset has changed since last inspect foreignforeign: Car origin Number of observationsTotal Integers Nonintegers# Negative - - -# Zero 52 52 -# Positive 22 22 -## # Total 74 74 -# # Missing -0 1 74(2 unique values)foreign is labeled and all values are documented in the variableforeigntakes on two unique values, 0 and 1. The value 0 denotes a domestic car,and 1 denotes a foreign model that we wish to fit isPr(foreign= 1) =F( 0+ 1weight+ 2mpg)whereF(z) =ez/(1 +ez)is the cumulative Logistic Logistic regression , reporting coefficients 5To fit this model, we type.

7 logit foreign weight mpgIteration 0: log likelihood = 1: log likelihood = 2: log likelihood = 3: log likelihood = 4: log likelihood = 5: log likelihood = regression Number of obs = 74LR chi2(2) = > chi2 = likelihood = Pseudo R2 = Std. err. z P>|z| [95% conf. interval] .0010116 .0919175 . find that heavier cars are less likely to be foreign and that cars yielding better gas mileage arealso less likely to be foreign, at least holding the weight of the car noteStata interprets a value of 0 as a negative outcome (failure) and treats all other values (exceptmissing) as positive outcomes (successes). Thus if your dependent variable takes on the values 0 and1, then 0 is interpreted as failure and 1 as success.

8 If your dependent variable takes on the values 0,1, and 2, then 0 is still interpreted as failure, but both 1 and 2 are treated as you prefer a more formal mathematical statement, when you typelogity x, Stata fits themodelPr(yj6= 0|xj) =exp(xj )1 +exp(xj )Model identificationThelogitcommand has one more feature, and it is probably the most the model for identification and, if it is underidentified, drops whatever variables and observationsare necessary for estimation to proceed. ( Logistic ,probit, andivprobitdo this as well.)Example 2 Have you ever fit a logit model where one or more of your independent variables perfectly predictedone or the other outcome?For instance, consider the following data:OutcomeyIndependent variablex010100106 logit Logistic regression , reporting coefficientsSay that we wish to predict the outcome on the basis of the independent variable.

9 The outcome isalways zero whenever the independent variable is one. In our data,Pr(y=0|x=1) =1, whichmeans that the logit coefficient onxmust be minus infinity with a corresponding infinite standarderror. At this point, you may suspect that we have a , not all such problems are so easily detected, especially if you have a lot ofindependent variables in your model. If you have ever had such difficulties, you have experienced oneof the more unpleasant aspects of computer optimization. The computer has no idea that it is tryingto solve for an infinite coefficient as it begins its iterative process. All it knows is that at each step,making the coefficient a little bigger, or a little smaller, works wonders. It continues on its merryway until either 1) the whole thing comes crashing to the ground when a numerical overflow erroroccurs or 2) it reaches some predetermined cutoff that stops the process.

10 In the meantime, you havebeen waiting. The estimates that you finally receive, if you receive any at all, may be nothing morethan numerical watches for these sorts of problems, alerts us, fixes them, and properly fits the s return to our automobile data. Among the variables we have in the data is one calledrepair,which takes on three values. A value of 1 indicates that the car has a poor repair record, 2 indicatesan average record, and 3 indicates a better-than-average record. Here is a tabulation of our data:. use , clear(1978 automobile data). tabulate foreign repairRepairCar origin1 2 3 TotalDomestic10 27 946 Foreign0 3 912 Total10 30 1858 All the cars with poor repair records (repair=1) are domestic. If we were to attempt to predictforeignon the basis of the repair records, the predicted probability for therepair=1 categorywould have to be zero.