Question:

I recently ran a do-file that contained several tobit and xttobit models and later decided to rerun one of the xttobit models. I was surprised to find that I got different results. The puzzle to me is that the iterations that fit the full model are identical to what I estimated the other day (through iteration 6), and then they diverge. The data have not changed, and the correct sample and variables are being used. The only difference is that I am running one regression instead of many. Is there some random component due to the use of quadrature?

Answer:

For those not familiar with xttobit, it and several other commands that estimate random-effects models use Gauss–Hermite quadrature and adaptive quadrature to approximate the high-dimension integrals that are part of the likelihood for these models. The quadrature approximation can be poor for some datasets, and I suspect this is what this user is encountering.

As the user suggests, there is a random component to quadrature in that the within-panel sort will almost certainly be different unless you start Stata fresh and run exactly the same commands before running xttobit. Sort order is not important to the likelihood, but if the likelihood for the dataset cannot be approximated well by quadrature, the order can affect the quadrature computation (more on this later).

Quadrature is one of the most accepted approaches to estimating these models, but there are three cases where it often breaks down: (1) large panel sizes, (2) high within-panel correlation, or (3) variables that are constant or near constant within panel. I don’t know if any of these are true for these data, but any observation that contributes in an extreme way to the likelihood can cause problems. See [XT] quadchk for a good discussion of these issues.

Stata’s quadchk command can help tremendously in assessing whether your data are appropriate for estimation using the quadrature approximation. quadchk works with all of the estimation commands that use quadrature and I definitely recommend that the user try quadchk on the model. I also heartily recommend that people estimating a random-effects model by quadrature check whether the quadrature is stable for their model. If you’re using Stata, use quadchk to do this.

We have tried to point everyone using commands that employ quadrature to quadchk by providing a Technical Note or example in the manual entry. In hindsight, these suggestions could have been stronger.

We at StataCorp could have artificially forced Stata to produce the same answer always from xttobit by performing a sort during quadrature, but we purposely did not do that. That strikes us as ducking the issue. If quadrature is not stable, better not to hide it.

Commands using quadrature have been the source of some debate around StataCorp. None of us are wholly comfortable with estimators that are prone to instability, even if that instability arises only in extreme cases. That is why we feel so strongly about providing quadchk to assess the appropriateness of the estimator for a given dataset. The near consensus here is that these estimators are valuable to those who need them even though they require care from all who use them. They are stable for most datasets. Admittedly, these are leading-edge models, and estimating them requires more understanding of numerical and approximation issues on the part the user than do most other estimation commands.

Stata help for estimation commands

help estimation commands 
-------------------------------------------------------------------------------
Title

 [I] estimation commands -- Quick reference for estimation commands
Description

 For a discussion of properties shared by all estimation commands see 
 estcom.

 This entry provides a quick reference for Stata's estimation commands.
 Since enhancements to Stata are continually being added, type search
 estimation commands for possible additions to this list; see search.

 command description
 -------------------------------------------------------------------------
 anova Analysis of variance and covariance
 arch ARCH family of estimators
 areg Linear regression with a large dummy-variable set
 arima ARIMA, ARMAX, and other dynamic regression models
 asmprobit Alternative-specific multinomial probit regression

 binreg Generalized linear models: Extensions to the
 binomial family
 biprobit Bivariate probit regression
 blogit Logistic regression for grouped data
 bootstrap Bootstrap sampling and estimation
 boxcox Box-Cox regression models
 bprobit Probit regression for grouped data
 bsqreg Quantile regression with bootstrap standard errors

 ca Simple correspondence analysis
 camat Simple correspondence analysis of a matrix
 canon Canonical correlations
 clogit Conditional (fixed-effects) logistic regression
 cloglog Complementary log-log regression
 cnreg Censored-normal regression
 cnsreg Constrained linear regression

 dprobit Probit regression, reporting marginal effects

 eivreg Errors-in-variables regression

 factor Factor analysis
 factormat Factor analysis of a correlation matrix
 fracpoly Fractional polynomial regression
 frontier Stochastic frontier models

 glm Generalized linear models
 glogit Logit and probit for grouped data
 gnbreg Generalized negative binomial model
 gprobit Weighted least-squares probit regression for
 grouped data

 heckman Heckman selection model
 heckprob Probit model with selection
 hetprob Heteroskedastic probit model

 intreg Interval regression
 iqreg Interquantile range regressions
 ivprobit Probit model with endogenous regressors
 ivreg Instrumental variables (two-stage least-squares)
 regression
 ivtobit Tobit model with endogenous regressors

 jackknife Jackknife estimation

 logistic Logistic regression, reporting odds ratios
 logit Logistic regression, reporting coefficients

 manova Multivariate analysis of variance and covariance
 mds Multidimensional scaling for two-way data
 mean Estimate means
 mfp Multivariable fractional polynomial models
 mlogit Multinomial (polytomous) logistic regression
 mprobit Multinomial probit regression
 mvreg Multivariate regression

 nbreg Negative binomial regression
 newey Regression with Newey-West standard errors
 nl Nonlinear least-squares estimation
 nlogit Nested logit regression

 ologit Ordered logistic regression
 oprobit Ordered probit regression

 pca Principal component analysis
 pcamat Principal component analysis of a correlation or
 covariance matrix
 poisson Poisson regression
 prais Prais-Winsten and Cochrane-Orcutt regression
 probit Probit regression
 procrustes Procrustes transformation
 proportion Estimate proportions

 _qreg Internal estimation command for quantile regression
 qreg Quantile (including median) regression

 ratio Estimate ratios
 reg3 Three-stage estimation for systems of simultaneous
 equations
 regress Linear regression
 rocfit Fit ROC models
 rologit Rank-ordered logistic regression
 rreg Robust regression

 scobit Skewed logistic regression
 slogit Stereotype logistic regression
 sqreg Simultaneous-quantile regression
 stcox Fix Cox proportional hazards model
 streg Fit parametric survival models
 sureg Zellner's seemingly unrelated regression
 svar Structural vector autoregression models
 svy: heckman Heckman selection model for survey data
 svy: heckprob Probit regression with selection for survey data
 svy: intreg Censored and interval regression for survey data
 svy: ivreg Instrumental variables regression for survey data
 svy: logistic Logistic regression, reporting odds ratios, for
 survey data
 svy: logit Logistic regression, reporting coefficients, for
 survey data
 svy: mean Estimate means for survey data
 svy: mlogit Multinomial logistic regression for survey data
 svy: nbreg Negative binomial regression for survey data
 svy: ologit Ordered logistic regression for survey data
 svy: oprobit Ordered probit regression for survey data
 svy: poisson Poisson regression for survey data
 svy: probit Probit regression for survey data
 svy: proportion Estimate proportions for survey data
 svy: ratio Estimate ratios for survey data
 svy: regress Linear regression for survey data
 svy: tabulate oneway One-way tables for survey data
 svy: tabulate twoway Two-way tables for survey data
 svy: total Estimate totals for survey data

 tobit Tobit regression
 total Estimate totals
 treatreg Treatment-effects model
 truncreg Truncated regression

 var Vector autoregression models
 var svar Structural vector autoregression models
 varbasic Fit a simple VAR and graph impulse-response
 functions
 vec Vector error-correction models
 vwls Variance-weighted least squares

 xtabond Arellano-Bond linear, dynamic panel-data estimation
 xtcloglog Random-effects and population-averaged cloglog
 models
 xtfrontier Stochastic frontier models for panel data
 xtgee Fit population-averaged panel-data models using GEE
 xtgls Fit panel-data models using GLS
 xthtaylor Hausman-Taylor estimator for error-components
 models
 xtintreg Random-effects interval data regression models
 xtivreg Instrumental variables and two-stage least squares
 for panel-data models
 xtlogit Fixed-effects, random-effects, and
 population-averaged logit models
 xtmixed Multilevel mixed-effects linear regression
 xtnbreg Fixed-effects, random-effects, and
 population-averaged negative binomial models
 xtpcse OLS or Prais-Winsten models with panel-corrected
 standard errors
 xtpoisson Fixed-effects, random-effects, and
 population-averaged Poisson models
 xtprobit Random-effects and population-averaged probit
 models
 xtrc Random-coefficients models
 xtreg Fixed-, between-, and random-effects, and
 population-averaged linear models
 xtregar Fixed- and random-effects linear models with an
 AR(1) disturbance
 xttobit Random-effects tobit model

 zinb Zero-inflated negative binomial regression
 zip Zero-inflated Poisson regression
 ztnb Zero-truncated negative binomial regression
 ztp Zero-truncated Poisson regression

[此贴子已经被作者于2006-5-6 13:08:51编辑过]

In this paper, I will discuss a number of Stata’s capabilities in the area of time series modeling, including data management and graphics. I will focus on a number of user – contributed routines, some of which have found their way into official Stata, with others likely to follow. For brevity, there are some areas I will not cover in this discussion: vector autoregressions and structural VARs, ARCH and GARCH modeling, cointegration tests ( now available in o  cial Stata’s July 2004 update) and panel unit root tests. I concentrate on a number of features and capabilities that may not be so well known, and present some new methodologies for time – series data analysis.

51567.pdf
大小:(183.43 KB)

只需: 1 个论坛币  马上下载

[此贴子已经被作者于2006-5-6 20:13:18编辑过]

Learning Statistical Software Stata

51568.rar
大小:(154.92 KB)

只需: 1 个论坛币  马上下载

本附件包括：

• Learning Statistical Software Stata.chm

[此贴子已经被作者于2006-5-6 20:20:31编辑过]

1. Installing, customizing and updating Stata 3

2. Windows in Stata 5

3. Suggested mode of operation 7

4. Getting help 9

5. Stata file types and names 10

6. Variables and observations 11

6.1. Variable names 11

6.2. Numeric variables 11

6.3. Missing values 12

7. Command syntax 13

8. Getting data into Stata 16

9. Documentation commands 18

10. Modifying data 20

10.1. Calculations 20

10.2. Selections 21

10.3. Renaming and reordering variables 23

10.4. Sorting data 24

10.5. Numbering observations 24

10.6. Combining files 25

10.7. Reshaping data 26

11. Description and analysis 27 11.1. Categorical data 27 11.2. Continuous data 30

12. Regression models 32 12.1. Linear regression 32 12.2. Logistic regression 33

13. Survival and related analyses 34

14. Graphs 38

15. Miscellaneous 54

15.1. Memory considerations 54

15.2. String variables 55

15.3. Dates. Danish CPR numbers 57

15.4. Random samples, simulations 59

15.5. Immediate commands 60

15.6. Sample size and power estimation 61

15.7. Ado- files 62

15.8. Exchange of data with other programs 63

15.9. For old SPSS users 63

16. Do- file examples 65

Appendix 1: Purchasing Stata and manuals 67

Appendix 2: Entering data with EpiData 68

Appendix 3: NoteTab Light: a text editor 70

Alphabetic index 71

51569.pdf
大小:(1.43 MB)

只需: 1 个论坛币  马上下载

Xiaogang Wu.Quantitative Analysis in Social Sciences

51570.pdf
大小:(1.6 MB)

只需: 1 个论坛币  马上下载

[此贴子已经被作者于2006-5-6 20:40:49编辑过]

Top
-------------------------------------------------------------------------------

 Category listings 
 Basics language syntax, expressions and functions, ... 
 Data management inputting, editing, creating new variables, ... 
 Statistics summary statistics, tables, estimation, ... 
 Graphics scatterplots, bar charts, ... 
 Programming and matrices do-files, ado-files, Mata, matrices 
 
 Help file listings 
 Language syntax advice on what to type 
 Manual datasets download datasets from the Reference manuals 
 Copyrights 
 

Stata Help Page: Stata Version 6.0 for Unix

• Using Stata help
• Using Stata search
• Using Stata tutorials
• Using Stata manuals
• Using Stata website
• Using Stata Bulletin

STATA HELP

Stata has extensive help once you are in the program.

help gives an overview of help

help contents gives many pages of commands
[I have saved this as the text-file contents.txt]

help regress for example gives help on the stata command regress for linear regression

STATA SEARCH

This can give more results than Stata help,
including

search statistics for example gives summary of commands for statistical analysis
search simulation for example gives information oin simulation

STATA TUTORIALS

Stata gives several tutorials that demonstrate various modules.

tutorial contents lists available turorials
tutorial survival for example, demonstrates survival commands

Unfortunately unless you use a graphic interface for Stata these tutorials do not show the graphs.

STATA MANUALS

The documentation is extensive.

The starting point is the User's Guide. You should really look at this.
e.g. [U] chapter 17 means User's Guide chapter 17

The Reference Guide is broken into four volumes
e.g. [R] matrix means Reference manual Matrix commands which is in Reference H-O (vol. 2).
The first of the reference manuals has a useful list of contents at the front.

STATA WEB-SITE

The website has a lot of information.
This includes summary of what Stata does.
For answers to frequently asked questions see http://www.stata.com/support/
Within Stata using search will cross-reference material on this website.

STATA BULLETIN

The Stata Bulletin has more recent code that has not yet appeared in Stata.
The website http://www.stata.com/support/stb/faq has an overview.
The programs can be downloaded free of charge.
But to read the accompanying article requires purchase of the Stata bulletin.
Within Stata using search will cross-reference material in the Stata bulletin.

For further information on how to use Stata go to
http://www.econ.ucdavis.edu/faculty/cameron

[分享][下载]陈峰：现代医学统计方法与STATA应用

http://210.72.32.6/cgi-bin/bigate.cgi/b/g/g/http@210.72.32.26/tjshujia/tjjc/t20051212_402295534.htm

[此贴子已经被作者于2006-5-6 21:28:37编辑过]

A little bit of Stata programming goes a long way...

Christopher F Baum1

Abstract

This tutorial will discuss a number of elementary Stata programming constructs and discuss how they may be used to automate and robustify common data manipulation, estimation and graphics tasks. Those used to the syntax of other statistical packages or programming languages must adopt a di erent mindset when working with Stata to take full advantage of its capabilities. Some of Stata’s most useful commands for handling repetitive tasks: forvalues, foreach, egen, local, scalar, estimates and matrix are commonly underutilized by users unacquainted with their power and ease of use. While relatively few users may develop ado- files for circulation to the user community, nearly all will benefit from learning the rudiments of use of the program, syntax and return statements when they are faced with the need to perform repetitive analyses. Worked examples making use of these commands will be presented and discussed in the tutorial.

51598.pdf
大小:(232.86 KB)

只需: 1 个论坛币  马上下载

[此贴子已经被作者于2006-5-7 4:10:38编辑过]

A State Space Approach for Estimating VAR Models

for Panel Data with Latent Dynamic Components

The econometric literature offers various modeling approaches for analyzing micro data in combination with time series of aggregate data. This paper discusses the estimation of a VAR model that allows unobserved heterogeneity across observation unit, as well as unobserved time-specific variables. The time-latent component is assumed to consist of a persistent and a transient term. By using a Helmert-type orthogonal transformation of the variables it is demonstrated that the likelihood function can be expressed on a state space form. The dimension of the state vector is low and independent of the time and cross section dimensions. This fact makes it convenient to employ an ECM algorithm for estimating the parameters of the model. An empirical application provides new insight into the problem of making forecasts for aggregate variables based on information from micro data.

51600.pdf
大小:(453.3 KB)

 马上下载

[此贴子已经被作者于2006-5-7 4:23:13编辑过]

Stata help for hausman

 help hausman dialog: hausman ------------------------------------------------------------------------------- 
 Title 
  [R] hausman -- Hausman specification test 
 
 Syntax 
  hausman name-consistent [name-efficient] [, options] 
  options description ------------------------------------------------------------------------- Main constant include estimated intercepts in comparison; default is to exclude alleqs use all equations to peform test; default is first equation only skipeqs(eqlist) skip specified equations when performing test equations(matchlist) associate/compare the specified (by number) pairs of equations force force performance of test, even though assumptions are not met df(#) use # degrees of freedom sigmamore base both (co)variance matrices on disturbance variance estimate from efficient estimator sigmaless base both (co)variance matrices on disturbance variance estimate from consistent estimator 
  Advanced tconsistent(string) consistent estimator column header tefficient(string) efficient estimator column header ------------------------------------------------------------------------- 
  where name-consistent and name-efficient are names under which estimation results were saved via estimates store. A period (.) may be used to refer to the last estimation results, even if these were not already stored. Not specifying name-efficient is equivalent to specifying the last estimation results as ".". 
 
 Description 
  hausman performs Hausman's specification test. To use hausman, one has to perform the following steps. 
  (1) obtain an estimator that is consistent whether or not the hypothesis is true; (2) store the estimation results under a name-consistent using estimates store; (3) obtain an estimator that is efficient (and consistent) under the hypothesis that you are testing, but inconsistent otherwise; (4) store the estimation results under a name-efficient using estimates  store; (5) use hausman to perform the test 
  hausman name-consistent name-efficient [, options] 
  The order of computing the two estimators may be reversed. You have to be careful though to specify to hausman the models in the order "always consistent" first and "efficient under H0" second. It is possible to skip storing the second model and refer to the last estimation results by a period (.). 
  hausman may be used in any context. The order in which you specify the regressors in each model does not matter, but it is your responsibility to assure that the estimators and models are comparable, and satisfy the theoretical conditions (see (1) and (3) above). 
 
 Options 
  +------+ ----+ Main +------------------------------------------------------------- 
  constant specifies that the estimated intercept(s) be included in the model comparison; by default, they are excluded. The default behavior is appropriate for models in which the constant does not have a common interpretation across the two models. 
  alleqs specifies that all the equations in the models be used to perform the Hausman test; by default, only the first equation is used. 
  skipeqs(eqlist) specifies in eqlist the names of equations to be excluded from the test. Equation numbers are not allowed in this context, as the equation names, along with the variable names, are used to identify common coefficients. 
  equations(matchlist) specifies, by number, the pairs of equations that are to be compared. 
  The matchlist in equations() should follow the syntax 
  #c:#e [,#c:#e[, ...]] 
  where #c(#e) is an equation number of the always-consistent (efficient under H0) estimator. For instance equations(1:1), equations(1:1, 2:2), or equations(1:2). 
  If equations() is not specified, then equations are matched on equation names. 
  equations() handles the situation in which one estimator uses equation names and the other does not. For instance, equations(1:2) means that equation 1 of the always-consistent estimator is to be tested against equation 2 of the efficient estimator. equations(1:1, 2:2) means that equation 1 is to be tested against equation 1 and that equation 2 is to be tested against equation 2. If equations() is specified, options alleqs and skipeqs are ignored. 
  force specifies that the Hausman test be performed, even though the assumptions of the Hausman test seem not to be met, for example, because the estimators were p-weighted. 
  df(#) specifies the degrees of freedom for the Hausman test. The default is the matrix rank of the variance of the difference between the coefficients of the two estimators. 
  sigmamore and sigmaless specify that the two covariance matrices used in the test be based on a common estimate of disturbance variance (sigma2). 
  sigmamore specifies that the covariance matrices be based on the estimated disturbance variance from the efficient estimator. This option provides a proper estimate of the contrast variance for so-called tests of exogeneity and overidentification in instrumental variables regression. 
  sigmaless specifies that the covariance matrices be based on the estimated disturbance variance from the consistent estimator. 
  These options can only be specified when both estimators save e(sigma) or e(rmse), or with command xtreg. e(sigma_e) is saved after command xtreg with options fe or mle. e(rmse) is saved after command xtreg with option re. 
  sigmamore or sigmaless are recommended when comparing fixed-effects and random-effects linear regression because they are much less likely to produce a nonpositive-definite differenced covariance matrix (although the tests are asymptotically equivalent whether or not one of the options is specified). 
  +----------+ ----+ Advanced +--------------------------------------------------------- 
  tconsistent(string) and tefficient(string) are formatting options. They allow you to specify the headers of the columns of coefficients that default to the names of the models. These options will be primarily of interest to programmers. 
 
 Remark: An alternative to hausman 
  The assumption that one of the estimators is efficient (i.e., has minimal asymptotic variance) is a demanding one. It is violated, for instance, if your observations are clustered or pweighted, or if your model is somehow misspecified. Moreover, even if the assumption is satisfied, there may be a "small sample" problem with the Hausman test. Hausman's test is based on estimating the variance var(b-B) of the difference of the estimators by the difference var(b)-var(B) of the variances. Under the assumptions (1) and (3), var(b)-var(B) is a consistent estimator of var(b-B), but it is not necessarily positive definite "in finite samples", i.e., in your application. If this is the case, the Hausman test is undefined. Unfortunately, this is not a rare event. Stata supports a generalized Hausman test that overcomes both of these problems. See suest for details. 
 
 Examples 
  Typing 
  . xtreg lny educ age, fe . est store fixed . xtreg lny educ age sex, re . hausman fixed . 
  presents Hausman's specification test, which tests the appropriateness of the random-effects estimator (xtreg, re). 
  Typing 
  . mlogit travmode age gender income . est store all . mlogit travmode age gender income if travmode != 2 . est store partial . hausman partial all, alleqs constant 
  will perform a Hausman test for independence of irrelevant alternatives (IIA). 
  When one estimator uses equation names and the other does not, specify the equations() option to force the comparison. This is illustrated in the comparison of the OLS estimator and the estimator of the regress part of the heckman model 
  . regress mpg price . est store reg . heckman mpg price, sel(foreign=weight) . hausman reg ., eq(1:1) 
  Comparison of the probit and selection model of the heckman 
  . probit foreign weight . est store probit_for . heckman mpg price, sel(foreign=weight) . hausman probit_for ., eq(1:2) 
 
 Also see 
  Manual: [R] hausman 
  Online: lrtest, suest, test, xtreg, xtregar 
 

[此贴子已经被作者于2006-5-8 2:12:07编辑过]

29.14.1 Linear regression with panel data

29.14.2 Censored linear regression with panel data

29.14.3 Generalized linear model with panel data

29.14.4 Qualitative dependent variable models with panel data

29.14.5 Count dependent variable models with panel data

29.14.6 Random coefficient models with panel data

Paul Allison: Fixed Effects Negative Binomial Regression Models

This paper demonstrates that the conditional negative binomial model for panel data, proposed by Hausman, Hall and Griliches ( 1984), is not a true fixed- effects method. This method — which has been implemented in both Stata and LIMDEP — does not, in fact, control for all stable covariates. Three alternative methods are explored. A negative multinomial model yields the same estimator as the conditional Poisson estimator and, hence, does not provide any additional leverage for dealing with overdispersion. On the other hand, a simulation study yields good results from applying an unconditional negative binomial regression estimator with dummy variables to represent the fixed effects. There is no evidence for any incidental parameters bias in the coefficients, and downward bias in the standard error estimates can be easily and effectively corrected using the deviance statistic. Finally, an approximate conditional method is found to perform at about the same level as the unconditional estimator.

51791.pdf
大小:(83.35 KB)

 马上下载

FACTORTEST: Stata module to perform tests for appropriateness of factor analysis

Joao Pedro Azevedo (

51793.pdf
大小:(8.93 KB)

 马上下载

[此贴子已经被作者于2006-5-8 6:37:34编辑过]

[此贴子已经被作者于2006-5-8 6:42:20编辑过]

Factor analysis (and its relative, principal components analysis) are performed in Stata using the 4 basic commands (factor, greigen, rotate, score) listed below.

factor [varlist],[algorithm] [modifiers]

[varlist] Choose variables to analyze.


[algorithm] Choose type of analysis (algorithm):


pc - principal components


pcf - principal component factors


pf - principal factors (default)


ipf - iterated principal factors


ml - maximum likelihood

[modifiers] Choose appropriate modifiers:


factors (#) Select maximum # of common factors.


mineigen (#) Set eigenvalue minimum for common factor.


covariance Analyze covariance matrix (in pc only).

greigen Graph eigenvalues for "scree plot".

rotate [,option]

varimax Choose orthogonal rotation.


promax (#) Choose nonorthogonal (oblique) rotation. (# defines correlation)

score "newvarnames" [,option]

bartlett Use (unbiased) Bartlett method rather than default regression
method (smaller MSE).


norotate Score unrotated factors.

STATA Commands:


factor x1 x2 x3 x4 x5 x6 x7 x8 x9, ipf


greigen


factor x1 x2 x3 x4 x5 x6 x7 x8 x9, ipf factor (3)


rotate, varimax


score "factor1" "factor2" "factor 3", bartlett

Notations:



The first command factor analyzes 9 variables (x1 through x9) using iterated principal factor analysis.


Command 2 graphs the eigenvalues by factor number.


Based on the results of the first two command, a decision is made to retain
three common factors with command 3.


The fourth command provides an orthogonally rotated (varimax) solution.


The last command produces factor scores for each of the three factors using Bartlett's algorithm.

[PPT] Introduction to Factor Analysis

[此贴子已经被作者于2006-5-8 6:52:27编辑过]

Implementation of quasi-least squares using xtgee in Stata

Liang and Zeger's original formulation of generalized estimating equations (GEE) has been widely applied since its introduction in 1986 because it extends the application of generalized linear models to clustered data. In this presentation we discuss a method, quasi-least squares (QLS), that is in the framework of GEE and builds on this popular approach by allowing for consideration of correlation matrices that were previously difficult to apply. In particular, we describe how to QLS in a straight-forward fashion by making use of Stata's xtgee procedure. We then discuss some data analysis examples.

http://repec.org/nasug2004/Shults_Stata_2004.ppt

[此贴子已经被作者于2006-5-8 10:44:52编辑过]

XTIVREG2: Stata module to perform extended IV/2SLS, GMM and AC/HAC, LIML and k-class regression for panel data models

Mark E Schaffer

xtivreg2 implements IV/GMM estimation of the fixed-effects and first-differences panel data models with possibly endogenous regressors. It is essentially a wrapper for ivreg2, which must be installed for xtivreg2 to run (version 2.1.15 or above of ivreg2 is required: ssc install ivreg2, replace). xtivreg2 supports all the estimation and reporting options of ivreg2; see help ivreg2 for full descriptions and examples. In particular, all the statistics available with ivreg2 (heteroskedastic, cluster- and autocorrelation-robust covariance matrix and standard errors, overidentification and orthogonality tests, first-stage and weak/underidentification statistics, etc.) are also supported by xtivreg2 and will be reported with any degrees-of-freedom adjustments required for a panel data estimation.

File URL: http://fmwww.bc.edu/repec/bocode/x/xtivreg2.ado

[此贴子已经被作者于2006-5-9 23:40:44编辑过]