Stata 15 help for meprobit

[ME] meprobit -- Multilevel mixed-effects probit regression

Syntax

meprobit depvar fe_equation [|| re_equation] [|| re_equation ...] [, options]

where the syntax of fe_equation is

[indepvars] [if] [in] [weight] [, fe_options]

and the syntax of re_equation is one of the following:

for random coefficients and intercepts

levelvar: [varlist] [, re_options]

for random effects among the values of a factor variable

levelvar: R.varname

levelvar is a variable identifying the group structure for the random effects at that level or is _all representing one group comprising all observations.

fe_options Description ------------------------------------------------------------------------- Model noconstant suppress constant term from the fixed-effects equation offset(varname) include varname in model with coefficient constrained to 1 asis retain perfect predictor variables -------------------------------------------------------------------------

re_options Description ------------------------------------------------------------------------- Model covariance(vartype) variance-covariance structure of the random effects noconstant suppress constant term from the random-effects equation fweight(varname) frequency weights at higher levels iweight(varname) importance weights at higher levels pweight(varname) sampling weights at higher levels -------------------------------------------------------------------------

options Description ------------------------------------------------------------------------- Model binomial(varname|#) set binomial trials if data are in binomial form constraints(constraints) apply specified linear constraints collinear keep collinear variables

SE/Robust vce(vcetype) vcetype may be oim, robust, or cluster clustvar

Reporting level(#) set confidence level; default is level(95) nocnsreport do not display constraints notable suppress coefficient table noheader suppress output header nogroup suppress table summarizing groups display_options control columns and column formats, row spacing, line width, display of omitted variables and base and empty cells, and factor-variable labeling

Integration intmethod(intmethod) integration method intpoints(#) set the number of integration (quadrature) points for all levels; default is intpoints(7)

Maximization maximize_options control the maximization process; seldom used

startvalues(svmethod) method for obtaining starting values startgrid[(gridspec)] perform a grid search to improve starting values noestimate do not fit the model; show starting values instead dnumerical use numerical derivative techniques coeflegend display legend instead of statistics -------------------------------------------------------------------------

vartype Description ------------------------------------------------------------------------- independent one unique variance parameter per random effect, all covariances 0; the default unless the R. notation is used exchangeable equal variances for random effects, and one common pairwise covariance identity equal variances for random effects, all covariances 0; the default if the R. notation is used unstructured all variances and covariances to be distinctly estimated fixed(matname) user-selected variances and covariances constrained to specified values; the remaining variances and covariances unrestricted pattern(matname) user-selected variances and covariances constrained to be equal; the remaining variances and covariances unrestricted -------------------------------------------------------------------------

intmethod Description ------------------------------------------------------------------------- mvaghermite mean-variance adaptive Gauss-Hermite quadrature; the default unless a crossed random-effects model is fit mcaghermite mode-curvature adaptive Gauss-Hermite quadrature ghermite nonadaptive Gauss-Hermite quadrature laplace Laplacian approximation; the default for crossed random-effects models -------------------------------------------------------------------------

indepvars may contain factor variables; see fvvarlist. depvar, indepvars, and varlist may contain time-series operators; see tsvarlist. bayes, by, and svy are allowed; see prefix. For more details, see [BAYES] bayes: meprobit. vce() and weights are not allowed with the svy prefix. fweights, iweights, and pweights are allowed; see weight. Only one type of weight may be specified. Weights are not supported under the Laplacian approximation or for crossed models. startvalues(), startgrid, noestimate, dnumerical, and coeflegend do not appear in the dialog box. See [ME] meprobit postestimation for features available after estimation.

Menu

Statistics > Multilevel mixed-effects models > Probit regression

Description

meprobit fits mixed-effects models for binary or binomial responses. The conditional distribution of the response given the random effects is assumed to be Bernoulli, with success probability determined by the standard normal cumulative distribution function.

Options

+-------+ ----+ Model +------------------------------------------------------------

noconstant suppresses the constant (intercept) term and may be specified for the fixed-effects equation and for any of or all the random-effects equations.

offset(varname) specifies that varname be included in the fixed-effects portion of the model with the coefficient constrained to be 1.

asis forces retention of perfect predictor variables and their associated, perfectly predicted observations and may produce instabilities in maximization; see [R] probit.

covariance(vartype) specifies the structure of the covariance matrix for the random effects and may be specified for each random-effects equation. vartype is one of the following: independent, exchangeable, identity, unstructured, fixed(matname), or pattern(matname).

covariance(independent) covariance structure allows for a distinct variance for each random effect within a random-effects equation and assumes that all covariances are 0. The default is covariance(independent) unless a crossed random-effects model is fit, in which case the default is covariance(identity).

covariance(exchangeable) structure specifies one common variance for all random effects and one common pairwise covariance.

covariance(identity) is short for "multiple of the identity"; that is, all variances are equal and all covariances are 0.

covariance(unstructured) allows for all variances and covariances to be distinct. If an equation consists of p random-effects terms, the unstructured covariance matrix will have p(p+1)/2 unique parameters.

covariance(fixed(matname)) and covariance(pattern(matname)) covariance structures provide a convenient way to impose constraints on variances and covariances of random effects. Each specification requires a matname that defines the restrictions placed on variances and covariances. Only elements in the lower triangle of matname are used, and row and column names of matname are ignored. A missing value in matname means that a given element is unrestricted. In a fixed(matname) covariance structure, (co)variance (i,j) is constrained to equal the value specified in the i,jth entry of matname. In a pattern(matname) covariance structure, (co)variances (i,j) and (k,l) are constrained to be equal if matname[i,j] = matname[k,l].

fweight(varname) specifies frequency weights at higher levels in a multilevel model, whereas frequency weights at the first level (the observation level) are specified in the usual manner, for example, [fw=fwtvar1]. varname can be any valid Stata variable name, and you can specify fweight() at levels two and higher of a multilevel model. For example, in the two-level model

. me... fixed_portion [fw = wt1] || school: ... , fweight(wt2) ...

the variable wt1 would hold the first-level (the observation-level) frequency weights, and wt2 would hold the second-level (the school-level) frequency weights.

iweight(varname) specifies importance weights at higher levels in a multilevel model, whereas importance weights at the first level (the observation level) are specified in the usual manner, for example, [iw=iwtvar1]. varname can be any valid Stata variable name, and you can specify iweight() at levels two and higher of a multilevel model. For example, in the two-level model

. me... fixed_portion [iw = wt1] || school: ... , iweight(wt2) ...

the variable wt1 would hold the first-level (the observation-level) importance weights, and wt2 would hold the second-level (the school-level) importance weights.

pweight(varname) specifies sampling weights at higher levels in a multilevel model, whereas sampling weights at the first level (the observation level) are specified in the usual manner, for example, [pw=pwtvar1]. varname can be any valid Stata variable name, and you can specify pweight() at levels two and higher of a multilevel model. For example, in the two-level model

. me... fixed_portion [pw = wt1] || school: ... , pweight(wt2) ...

variable wt1 would hold the first-level (the observation-level) sampling weights, and wt2 would hold the second-level (the school-level) sampling weights.

binomial(varname|#) specifies that the data are in binomial form; that is, depvar records the number of successes from a series of binomial trials. This number of trials is given either as varname, which allows this number to vary over the observations, or as the constant #. If binomial() is not specified (the default), depvar is treated as Bernoulli, with any nonzero, nonmissing values indicating positive responses.

constraints(constraints), collinear; see [R] estimation options.

+-----------+ ----+ SE/Robust +--------------------------------------------------------

vce(vcetype) specifies the type of standard error reported, which includes types that are derived from asymptotic theory (oim), that are robust to some kinds of misspecification (robust), and that allow for intragroup correlation (cluster clustvar); see [R] vce_option. If vce(robust) is specified, robust variances are clustered at the highest level in the multilevel model.

+-----------+ ----+ Reporting +--------------------------------------------------------

level(#), nocnsreport; see [R] estimation options.

notable suppresses the estimation table, either at estimation or upon replay.

noheader suppresses the output header, either at estimation or upon replay.

nogroup suppresses the display of group summary information (number of groups, average group size, minimum, and maximum) from the output header.

display_options: noci, nopvalues, noomitted, vsquish, noemptycells, baselevels, allbaselevels, nofvlabel, fvwrap(#), fvwrapon(style), cformat(%fmt), pformat(%fmt), sformat(%fmt), and nolstretch; see [R] estimation options.

+-------------+ ----+ Integration +------------------------------------------------------

intmethod(intmethod) specifies the integration method to be used for the random-effects model. mvaghermite performs mean-variance adaptive Gauss-Hermite quadrature; mcaghermite performs mode-curvature adaptive Gauss-Hermite quadrature; ghermite performs nonadaptive Gauss-Hermite quadrature; and laplace performs the Laplacian approximation, equivalent to mode-curvature adaptive Gaussian quadrature with one integration point.

The default integration method is mvaghermite unless a crossed random-effects model is fit, in which case the default integration method is laplace. The Laplacian approximation has been known to produce biased parameter estimates; however, the bias tends to be more prominent in the estimates of the variance components rather than in the estimates of the fixed effects.

For crossed random-effects models, estimation with more than one quadrature point may be prohibitively intensive even for a small number of levels. For this reason, the integration method defaults to the Laplacian approximation. You may override this behavior by specifying a different integration method.

intpoints(#) sets the number of integration points for quadrature. The default is intpoints(7), which means that seven quadrature points are used for each level of random effects. This option is not allowed with intmethod(laplace).

The more integration points, the more accurate the approximation to the log likelihood. However, computation time increases as a function of the number of quadrature points raised to a power equaling the dimension of the random-effects specification. In crossed random-effects models and in models with many levels or many random coefficients, this increase can be substantial.

+--------------+ ----+ Maximization +-----------------------------------------------------

maximize_options: difficult, technique(algorithm_spec), iterate(#), [no]log, trace, gradient, showstep, hessian, showtolerance, tolerance(#), ltolerance(#), nrtolerance(#), nonrtolerance, and from(init_specs); see [R] maximize. Those that require special mention for meprobit are listed below.

from() accepts a properly labeled vector of initial values or a list of coefficient names with values. A list of values is not allowed.

The following options are available with meprobit but are not shown in the dialog box:

startvalues(svmethod), startgrid[(gridspec)], noestimate, and dnumerical; see [ME] meglm.

coeflegend; see [R] estimation options.

Remarks on using sampling weights

Sampling weights are treated differently in multilevel models than they are in standard models such as OLS regression. In a multilevel model, observation-level weights are not indicative of overall inclusion. Instead, they indicate inclusion conditional on the corresponding cluster being included at the next highest-level of sampling.

For example, if you include only observation-level weights in a two-level model, sampling with equal probabilities at level two is assumed, and this may or may not be what you intended. If the sampling at level two is weighted, then including only level-one weights can lead to biased results even if weighting at level two has been incorporated into the level-one weight variable. For example, it is a common practice to multiply conditional weights from multiple levels into one overall weight. By contrast, weighted multilevel analysis requires the component weights from each level of sampling.

Even if you specify sampling weights at all model levels, the scale of sampling weights at lower levels can affect your estimated parameters in a multilevel model. That is, not only do the relative sizes of the weights at lower levels matter, the scale of these weights matters also.

In general, exercise caution when using sampling weights; see Survey data in Remarks and examples of [ME] meglm for more information.

Examples

--------------------------------------------------------------------------- Setup . webuse bangladesh

Two-level random-intercept model, analogous to xtprobit . meprobit c_use urban age child* || district:

Two-level random-intercept and random-coefficient model . meprobit c_use urban age child* || district: urban

Two-level random-intercept and random-coefficient model, correlated random effects . meprobit c_use urban age child* || district: urban, cov(unstruct)

--------------------------------------------------------------------------- Setup . webuse towerlondon

Three-level nested model, subject nested within family . meprobit dtlm difficulty i.group || family: || subject:

--------------------------------------------------------------------------- Setup . webuse fifeschool . gen byte attain_gt_6 = attain > 6

Two-way crossed random effects . meprobit attain_gt_6 sex || _all:R.sid || pid: ---------------------------------------------------------------------------

Stored results

meprobit stores the following in e():

Scalars e(N) number of observations e(k) number of parameters e(k_dv) number of dependent variables e(k_eq) number of equations in e(b) e(k_eq_model) number of equations in overall model test e(k_f) number of fixed-effects parameters e(k_r) number of random-effects parameters e(k_rs) number of variances e(k_rc) number of covariances e(df_m) model degrees of freedom e(ll) log likelihood e(N_clust) number of clusters e(chi2) chi-squared e(p) p-value for model test e(ll_c) log likelihood, comparison model e(chi2_c) chi-squared, comparison test e(df_c) degrees of freedom, comparison test e(p_c) p-value for comparison test e(rank) rank of e(V) e(ic) number of iterations e(rc) return code e(converged) 1 if converged, 0 otherwise

Macros e(cmd) meglm e(cmd2) meprobit e(cmdline) command as typed e(depvar) name of dependent variable e(wtype) weight type e(wexp) weight expression (first-level weights) e(fweightk) fweight variable for kth highest level, if specified e(iweightk) iweight variable for kth highest level, if specified e(pweightk) pweight variable for kth highest level, if specified e(covariates) list of covariates e(ivars) grouping variables e(model) probit e(title) title in estimation output e(link) probit e(family) bernoulli or binomial e(clustvar) name of cluster variable e(offset) offset e(binomial) binomial number of trials e(intmethod) integration method e(n_quad) number of integration points e(chi2type) Wald; type of model chi-squared e(vce) vcetype specified in vce() e(vcetype) title used to label Std. Err. e(opt) type of optimization e(which) max or min; whether optimizer is to perform maximization or minimization e(ml_method) type of ml method e(user) name of likelihood-evaluator program e(technique) maximization technique e(datasignature) the checksum e(datasignaturevars) variables used in calculation of checksum e(properties) b V e(estat_cmd) program used to implement estat e(predict) program used to implement predict e(marginsnotok) predictions disallowed by margins e(marginswtype) weight type for margins e(marginswexp) weight expression for margins e(asbalanced) factor variables fvset as asbalanced e(asobserved) factor variables fvset as asobserved

Matrices e(b) coefficient vector e(Cns) constraints matrix e(ilog) iteration log (up to 20 iterations) e(gradient) gradient vector e(N_g) group counts e(g_min) group-size minimums e(g_avg) group-size averages e(g_max) group-size maximums e(V) variance-covariance matrix of the estimators e(V_modelbased) model-based variance

Functions e(sample) marks estimation sample


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index