.- help for ^rfprobit^ (STB-26: sg41) .- Random-effects probit --------------------- ^rfprobit^ depvar [indepvars] [^if^ exp] [^in^ range] ^,^ [ ^i(^varname^)^ ^q^uadrat^(^#^)^ ^nochi^sq ^nolo^g ^l^evel^(^#^)^ maximize_options ] This command shares the features of all estimation commands; see help @est@. To reset problem-size limits, see help @matsize@. Description ----------- ^rfprobit^ estimates a random-effects probit model for cross-sectional time- series data sets using maximum likelihood estimation. The likelihood (for an independent unit i) is expressed as an integral which is computed using Gaussian-Hermite quadrature. This computational procedure is only accurate when there are a small-to-moderate number of time periods T_i per unit i. It is recommended that ^rfprobit^ only be used when max(T_i) <= 50. Options ------- ^i(^varname^)^ specifies the variable corresponding to an independent unit (e.g., a subject id). This variable represents the i in x_it. Either this option must be specified or i must be set using the ^iis^ command; see help for @xt@. ^quadrat(^#^)^ specifies the number of points to use for Gaussian-Hermite quadrature. Default is 6. Increasing this value slightly improves accuracy, but also increases computation time. Computation time is roughly proportional to its value. ^nochisq^ omits the estimation of the constant-only model. This will reduce computation time at the cost of not being able to calculate the model chi-squared or pseudo R^^2. ^nolog^ suppress the display of the likelihood iterations. ^level(^#^)^ specifies the significance level, in percent, for confidence intervals of the coefficients; see help @level@. maximize_options control the maximization process; see [7] maximize. Use the ^trace^ option to view parameter convergence. The ^ltol(^#^)^ option can be used to loosen the convergence criterion (default is 1e-6) during specification searches. Examples -------- . ^rfprobit y x, i(id)^ . ^iis id^ . ^rfprobit y x^ . ^rfprobit y x, i(id) nochisq^ . ^rfprobit y x, i(id) quad(8) nolog^ . ^rfprobit y x, i(id) trace^ . ^rfprobit^ Method ------ ^rfprobit^ uses the deriv1 method (analytic first derviatives) of Stata's ^ml^ commands. See Butler and Moffitt (1982) for details. Author ------ Bill Sribney Stata Corporation 702 University Drive East College Station, TX 77840 Phone: 409-696-4600 800-782-8272 Fax: 409-696-4601 email: tech_support@@stata.com Reference --------- Butler, J.S. and R. Moffitt. 1982. A computationally efficient quadrature procedure for the one-factor multinomial probit model. Econometrica 50: 761-764. Also see -------- STB: STB-26 sg41 Manual: [5s] xt, [5s] xtreg, [7] maximize On-line: help for @xt@, @xtreg@