Stata 15 help for estimation options

[R] estimation options -- Estimation options

Syntax

estimation_cmd ... [, options]

options Description ------------------------------------------------------------------------- Model noconstant suppress constant term offset(varname_o) include varname_o in model with coefficient constrained to 1 exposure(varname_o) include ln(varname_e) in model with coefficient constrained to 1 constraints(constraints) apply specified linear constraints collinear keep collinear variables

Reporting level(#) set confidence level; default is level(95) lrmodel perform likelihood-ratio model test instead of the default Wald test nocnsreport do not display constraints noci suppress confidence intervals nopvalues suppress p-values and their test statistics noomitted do not display omitted collinear variables vsquish suppress blank space separating factor variables or time-series variables noemptycells do not display empty interaction cells of factor variables baselevels report base levels for factor variables and interactions allbaselevels display all base levels for factor variables and interactions nofvlabel display factor-variable level values rather than value labels fvwrap(#) allow # lines when wrapping long value labels fvwrapon(style) apply style for wrapping long value labels; style may be word or width cformat(%fmt) format for coefficients, standard errors, and confidence limits pformat(%fmt) format for p-values sformat(%fmt) format for test statistics nolstretch do not automatically widen coefficient table for long variable names

Integration intmethod(intmethod) integration method for random-effects models intpoints(#) use # integration (quadrature) points

coeflegend display legend instead of statistics -------------------------------------------------------------------------

Description

This entry describes the options common to many estimation commands. Not all the options documented here work with all estimation commands. See the documentation for the particular estimation command; if an option is listed there, it is applicable.

Options

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

noconstant suppresses the constant term (intercept) in the model.

offset(varname_o) specifies that varname_o be included in the model with the coefficient constrained to be 1.

exposure(varname_e) specifies a variable that reflects the amount of exposure over which the depvar events were observed for each observation; ln(varname_e) with coefficient constrained to be 1 is entered into the log-link function.

constraints(numlist|matname) specifies the linear constraints to be applied during estimation. The default is to perform unconstrained estimation. See [R] reg3 for the use of constraints in multiple-equation contexts.

constraints(numlist) specifies the constraints by number after they have been defined by using the constraint command; see [R] constraint. Some commands (for example, slogit) allow only constraints(numlist).

constraints(matname) specifies a matrix containing the constraints; see [P] makecns.

constraints(clist) is used by some estimation commands, such as mlogit, where clist has the form #[-#][,#[-#] ... ].

collinear specifies that the estimation command not omit collinear variables. Usually, there is no reason to leave collinear variables in place, and, in fact, doing so usually causes the estimation to fail because of the matrix singularity caused by the collinearity. However, with certain models, the variables may be collinear, yet the model is fully identified because of constraints or other features of the model. In such cases, using the collinear option allows the estimation to take place, leaving the equations with collinear variables intact. This option is seldom used.

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

level(#) specifies the confidence level, as a percentage, for confidence intervals. The default is level(95) or as set by set level.

lrmodel specifies to conduct a likelihood-ratio test of the full maximum-likelihood model versus the restricted model that includes only a constant term in the regression equation instead of conducting the default Wald test that all coefficients are zero. This option can substantially increase estimation time.

lrmodel may not be specified with the vce(robust), vce(cluster clustvar), vce(jackknife), vce(bootstrap), or noconstant option.

lrmodel may not be combined with constraints; see constraints(constraints). In some cases, a likelihood-ratio test is valid for models with constraints. To compute a likelihood-ratio test when constraints have been applied during estimation, use lrtest; see [R] lrtest.

lrmodel may not be specified with the jackknife, bootstrap, svy, or mi estimate prefix.

nocnsreport specifies that no constraints be reported. The default is to display user-specified constraints above the coefficient table.

noci suppresses confidence intervals from being reported in the coefficient table.

nopvalues suppresses p-values and their test statistics from being reported in the coefficient table.

noomitted specifies that variables that were omitted because of collinearity not be displayed. The default is to include in the table any variables omitted because of collinearity and to label them as "(omitted)".

vsquish specifies that the blank space separating factor-variable terms or time-series-operated variables from other variables in the model be suppressed.

noemptycells specifies that empty cells for interactions of factor variables not be displayed. The default is to include in the table interaction cells that do not occur in the estimation sample and to label them as "(empty)".

baselevels and allbaselevels control whether the base levels of factor variables and interactions are displayed. The default is to exclude from the table all base categories.

baselevels specifies that base levels be reported for factor variables and for interactions whose bases cannot be inferred from their component factor variables.

allbaselevels specifies that all base levels of factor variables and interactions be reported.

nofvlabel displays factor-variable level values rather than attached value labels. This option overrides the fvlabel setting; see [R] set showbaselevels.

fvwrap(#) specifies how many lines to allow when long value labels must be wrapped. Labels requiring more than # lines are truncated. This option overrides the fvwrap setting; see [R] set showbaselevels.

fvwrapon(style) specifies whether value labels that wrap will break at word boundaries or break based on available space.

fvwrapon(word), the default, specifies that value labels break at word boundaries.

fvwrapon(width) specifies that value labels break based on available space.

This option overrides the fvwrapon setting; see [R] set showbaselevels.

cformat(%fmt) specifies how to format coefficients, standard errors, and confidence limits in the coefficient table. The maximum format width is 9. See [R] set cformat.

pformat(%fmt) specifies how to format p-values in the coefficient table. The maximum format width is 5. See [R] set cformat.

sformat(%fmt) specifies how to format test statistics in the coefficient table. The maximum format width is 8. See [R] set cformat.

nolstretch specifies that the width of the coefficient table not be automatically widened to accommodate longer variable names. The default, lstretch, is to automatically widen the coefficient table up to the width of the Results window. To change the default, use set lstretch off. nolstretch is not shown in the dialog box.

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

intmethod(intmethod) specifies the integration method to be used for the random-effects model. It accepts one of four arguments: mvaghermite, the default for all but a crossed random-effects model, performs mean and variance adaptive Gauss-Hermite quadrature; mcaghermite performs mode and curvature adaptive Gauss-Hermite quadrature; ghermite performs nonadaptive Gauss-Hermite quadrature; and laplace, the default for crossed random-effects models, performs the Laplacian approximation.

intpoints(#) specifies the number of integration points to use for integration by quadrature. The default is intpoints(12); the maximum is intpoints(195). Increasing this value improves accuracy but also increases computation time. Computation time is roughly proportional to its value.

The following option is not shown in the dialog box:

coeflegend specifies that the legend of the coefficients and how to specify them in an expression be displayed rather than displaying the statistics for the coefficients.

Examples

--------------------------------------------------------------------------- Setup . webuse dollhill3 . generate double lnpyears = ln(pyears) . constraint 1 smokes#3.agecat = smokes#4.agecat

Fit a Poisson regression, specifying an exposure of pyears . poisson deaths smokes i.agecat, exposure(pyears)

Same as above . poisson deaths smokes i.agecat, offset(lnpyears)

Replay results, but with 99% confidence intervals . poisson, level(99)

Replay results, suppressing the blank space separating factor-variable terms from other variables . poisson, vsquish

Replay results, showing coefficients, standard errors, and confidence limits to 4 decimal places . poisson, cformat(%8.4f)

Display coefficient legend, showing how to specify coefficients in an expression . poisson, coeflegend

Fit a Poisson regression, constraining the smoking effects on age categories 3 and 4 to be equal . poisson deaths c.smokes#agecat i.agecat, exposure(pyears) constraints(1)

Same as above, but suppress the display of the constraint above the coefficient table . poisson deaths c.smokes#agecat i.agecat, exposure(pyears) constraints(1) nocnsreport

--------------------------------------------------------------------------- Setup . webuse union, clear

Fit random-effects probit model, using 20 quadrature points instead of the default of 12 . xtprobit union age grade i.not_smsa south##c.year, intpoints(20)

Fit random-effects probit model, specifying the ghermite integration method be used . xtprobit union age grade i.not_smsa south##c.year, intmethod(ghermite)

Replay results, showing base levels of not_smsa and south in the table . xtprobit, baselevels

Same as above, but also show the base level for south#c.year . xtprobit, allbaselevels

---------------------------------------------------------------------------


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