"Stata Journal, The"

Hosmer, David W.; Royston, Patrick

doi: 10.1177/1536867X0200200401pmid: N/A

In this paper, we describe a new Stata command, stlh, which estimates and tests for the significance of the time-varying regression coefficients in Aalen's linear hazards model; see Aalen (1989). We see two potential uses for this command. One may use it as an alternative to a proportional hazards or other nonlinear hazards regression model analysis to describe the effects of covariates on survival time. A second application is to use the command to supplement a proportional hazards regression model analysis to assist in detecting and then describing the nature of time-varying effects of covariates through plots of the estimated cumulative regression coefficients, with confidence bands, from Aalen's model. We illustrate the use of the command to perform this supplementary analysis with data from a study of residential treatment programs of different durations that are designed to prevent return to drug use.

Reichenheim, Michael E.

doi: 10.1177/1536867X0200200402pmid: N/A

The command roctg allows visualizing sensitivity (Se)and specificity (Sp) curves according to the range of values of a new diagnostic test, given a “true” state of an event, the reference test. On request, several options for displaying Se and Sp estimates in, or enhancements for, the graphs are also available.

Becker, Sascha O.; Ichino, Andrea

doi: 10.1177/1536867X0200200403pmid: N/A

In this paper, we give a short overview of some propensity score matching estimators suggested in the evaluation literature, and we provide a set of Stata programs, which we illustrate using the National Supported Work (NSW) demonstration widely known in labor economics.

Guan, Weihua; Gutierrez, Roberto G.

doi: 10.1177/1536867X0200200404pmid: N/A

With the release of Stata 7, the glm command for fitting generalized linear models underwent a substantial overhaul. Stata 7 glm contains an expanded array of variance estimators, regression diagnostics, and other enhancements. The overhaul took place to coincide with the release of Hardin and Hilbe (2001). With the new glm came a modular design that enables users to program customized link functions, variance functions, and weight functions to be used if Newey–West covariance estimates are desired. Because cases requiring customized link functions are the more prevalent in the literature, only those are considered here. We give two examples where a nonstandard link function is required: the relative survival model of Hakulinen and Tenkanen (1987)and a logistic model that accounts for natural response as described in Collett (2003). The relative ease (over previous versions of Stata) with which these alternate links can be programmed into glm is demonstrated.

Schonlau, Matthias

doi: 10.1177/1536867X0200200405pmid: N/A

In hierarchical cluster analysis, dendrograms are used to visualize how clusters are formed. I propose an alternative graph called a “clustergram” to examine how cluster members are assigned to clusters as the number of clusters increases. This graph is useful in exploratory analysis for nonhierarchical clustering algorithms such as k-means and for hierarchical cluster algorithms when the number of observations is large enough to make dendrograms impractical. I present the Stata code and give two examples.

Poi, Brian P.

doi: 10.1177/1536867X0200200406pmid: N/A

This article provides an example illustrating how to use Stata to estimate systems of household demand equations. More generally, the techniques developed here can be used to estimate any system of nonlinear equations using Stata's maximum likelihood routines.

Cox, Nicholas J.

doi: 10.1177/1536867X0200200407pmid: N/A

Functions in Stata take two main forms, built-in functions that are part of the executable and egen functions written in Stata's own language. These are surveyed, giving a variety of tips and tricks, and noting the large number of user-written egen functions available for download from the Internet. Two substantial examples, the calculation of percentile ranks and plotting positions, and the calculation of measures summarizing properties of the other members of a group, provide detailed illustrations of egen in action.

Hosmer, David W.

doi: 10.1177/1536867X0200200408pmid: N/A

The new book by Cleves et al. (2002) is reviewed.

doi: 10.1177/1536867X0200200409pmid: N/A