Search
   >> Home >> Resources & support >> User Group meetings >> 2011 Mexican Stata Users Group meeting >> Abstracts

2011 Mexican Stata Users Group meeting: Abstracts

Una Comparación De Los Modelos Poisson Y Binomial Negativa Con Stata: Un Ejercicio Didactico

Noé Becerra Rodríguez
UAM-Xochimilco
Fortino Vela Peón
UAM-Xochimilco
El objetivo de esta presentación es mostrar el uso de modelos econométricos cuando se tienen variables de tipo contable, es decir, en términos de simples conteos. La aplicación se hace a considerando la información proveniente de una muestra de académicos levantada durante 2008-09 en la Universidad Autónoma Metropolitana. Se considera como variable de respuesta al número de artículos publicados que fueron reportados en la encuesta mientr as que cómo variables predictoras se usan características de tip personal y de redes de investigación de los académicos, tales como edad, tipo de actividad profesional (básica, aplicada o desarrollo tecnológico), afiliación al Sistema Nacional de Investigadores, país de obtención del doctorado, entre otras. El trabajo se inicia mostrando por qué razón resulta inapropiado el usar mínimos cuadrados ordinarios para variable s contables. Se continúa realizando diversas estimaciones para explicar algunos temas relacionados con este tipo de modelos. A lo largo de la exposición se señalan las siguientes técnicas de modelación econométri ca utilizando los comandos de Stata: Poisson simple, Poisson exacta, Poisson con efectos mezclados, Poisson truncada al cero, Poisson inflada por el cero, binomial negativa simple, binomial negativa generalizada, binomial negativa truncada al cero, binom ial negativa inflada por el cero y pruebas de bondad de ajuste para estos modelos. La ponencia tiene una motivación didáctica al considerar importante el que los alumnos tanto de las licenciaturas en economía y dministración así como de los posgrados en Ciencias Sociales que se imparten hoy en día instituciones como la UAM Xochimilco, identifiquen las posibilidades que ofrece este tipo de herramienta así como sus posibilidades de fácil implementación en Stata.

Additional information
becerra-vela.pdf

Nonparametric classification of fossil gymnosperm’s foliar area

Erika Lourdes Ortíz-Martínez
Laboratorio de Paleontología
Isaías H. Salgado-Ugarte
Laboratorio de Biometría y Biología Pesquera, Facultad de Estudios Superiores Zaragoza, UNAM
María Patricia Velasco de León
Laboratorio de Paleontología
In Paleobotany, the features of leaf laminae linked to the climate have been used to infer past climatic conditions. However, in spite of the increasing number of paleoecological studies, studies worldwide have been focused on the angiosperms, leaving aside other important groups throughout geological time. In Mexico, for example, the most abundant and widely distributed fossil flora are found in Early and Middle Jurassic deposits and are constituted mainly of gymnosperms. So far, the contributions of investigations of the climate where this vegetal group dominated have only been carried out at high-latitude localities, counting genera not present in Mexico besides the use of an informal leaf-size proposal. It has been demonstrated that this character has a direct relationship with climate. Therefore, this research work was aimed at proposing for gymnosperms leaf-lamina size classification using foliar area of each of the 737 plant fossils collected from Jurassic outcrops at Mexico’s Oaxaca State Mixtec region in the Rosario, Zorrillo, Taberna-Simón, and Tecomazúchl formations and the Tecocoyunca group. To classify the area by its extension we used kernel density estimators (smoothed histograms) with the (empirical suggestion) half of optimal bandwidth (Härdle, 1991, Smoothing Techniques: With Implications in S [Springer]). The resulting multimodal distribution was the base to set up categories defined with a new algorithm denominated “antimodes”. All the estimations were performed with Mosqueda-Romo and Salgado-Ugarte’s programs amodes.ado (a new one), bandw.ado, warpdenm.ado, bhataplt.ado, bhatgauc.ado, and their updated versions for Stata 11. The results allow the demarcation of seven categories, the most abundant being foliar area number three (Microphile I, Zamites, Pterophyllum, and Ptilophyllum genera) with a 58.06% representation, while the group with the largest surface exposed to sunlight radiation (over 57.74%) but the least percentage representation at the study zone (0.53%) is the Pelourdea genus. The gathered data permit the inference that during the Jurassic age in the Mixtec terrain, xeric conditions prevailed, which conditioned the small size of the leaves.

Additional information
salgado1.ppt

Technical tips on time series with Stata

Gustavo A. Sánchez
StataCorp
On a daily basis, we at Stata receive a broad variety of technical questions from users working in different areas with a large number of Stata commands. I selected a few common, interesting questions to help provide a brief view of the tools that are available in Stata for time-series analysis. I will start with a quick, simple introduction to time series in Stata, and I will then illustrate the use of a few commands to perform common tasks that are normally involved in the kind of empirical analysis developed by some of the Stata users who regularly contact us for technical assistance.

Additional information
gsanchez.pdf

Análisis de microdatos de la ENOE con Stata

Juan Francisco Islas Aguirre
CIDE
Fortino Vela Peón
UAM-X
Una de las principales necesidades que involucran el manejo de información proveniente de encuestas periódicas es la generación sistemática de estadísticos y tabulados para el análisis. En esta presentaci&oacut e;n, realizamos un ejercicio comparativo mediante comandos Stata y Mata para la generación de indicadores estratégicos, estadísticos y tabulados en ámbitos determinados, a partir de los microdatos originales con criterios prec odificados de la Encuesta Nacional de Ocupación y Empleo (ENOE) que trimestralmente levanta el Instituto Nacional de Estadística y Geografía (INEGI). La sintaxis de usuario, las matrices de resultados y plantillas que presentamos en este trabajo muestran la versatilidad de Stata como herramienta ideal en el proceso de enseñanza-aprendizaje enfocado particularmente a los temas de economía laboral que requieren aplicación empírica.

Additional information
Islas.pdf

Endogenous treatment effects for count data models with endogenous participation or sample selection

Massimiliano Bratti
University of Milan
Alfonso Miranda
Institute of Education, University of London
We propose an estimator for models in which an endogenous dichotomous treatment affects a count outcome in the presence of either sample selection or endogenous participation using maximum simulated likelihood. We allow for the treatment to have an effect on both the participation or the sample selection rule and on the main outcome. Applications of this model are frequent in—but not limited to—health economics. We show an application of the model using data from Kenkel (2001, Kenkel and Terza, Journal of Applied Econometrics 16: 165–184), who investigated the effect of physician advice on the amount of alcohol consumption. Our estimates suggest that in these data a) neglecting treatment endogeneity leads to a wrongly signed effect of physician advice on drinking intensity, b) accounting for treatment endogeneity but neglecting endogenous participation leads to an upward biased estimate of the treatment effect, and c) advice only affects the drinking-intensive margin but not drinking prevalence.

Additional information
miranda.pdf

Pitfalls in the analysis of complex surveys using Stata

Carlos Guerrero de Lizardi
Tecnológico de Monterrey
The purpose of this presentation is to show the common mistakes in the analysis of complex surveys. In Mexico, we have a significant number of complex surveys available, which cover (among other issues) household income and expenditures, the labor market, consumer confidence, public security perception, and family life. The heart of the matter is the following: if you ignore the sampling design of a complex survey (basically, the probability weights, the clustering, and the stratification), inevitably you will get an erroneous estimation of whatever you are dealing with.

Stata is a fully survey-capable software that takes into account the sampling design. I explore Stata’s survey methods capabilities and, as far as I know, illustrate the best practices in the analysis of complex surveys for the following topics: descriptive statistics, variance estimation methods, hypothesis testing, and econometric models.

Additional information
Guerrero.pptx

Robust inference with clustered data

Colin Cameron
University of California—Davis
This talk considers robust inference for regression models where data are clustered, with correlation of observations in the same cluster (such as state) and independence across clusters. The talk will range from the simplest case of heteroskedastic-ro bust (one individual per cluster) through to complications such as a small number of clusters and two-clustering. The relevant Stata commands and Stata add-ons, where available, will be discussed.

Additional information
cameron.pdf

Uso e importancia de DASP en Stata

Jean-Yves Duclos
Département d’Économique, Université Laval
Abdelkrim Araar
Département d’Économique, Université Laval
Luis Huesca
Departamento de Economía, Centro de Investigación en Alimentación y Desarrollo A. C.
Esta presentación muestra los beneficios que ofrece al usuario el paquete de Análisis Distributivo en Stata (DASP) para la evaluación del bienestar y su distribución, la medición de la pobreza y la desigualdad, y la m anera en que los datos requeridos por DASP se relacionan con temas importantes en la economía del bienestar entre otros tópicos, tales como los efectos impositivos y el impacto de las transferencias. El objetivo es que de forma general, se comprenda la estructura básica del funcionamiento de los módulos de DASP y su fuerte potencial en conjunto con las bondades del programa Stata. Se presentan casos empíricos para mostrar sus aplicaciones.

Additional information
Huesca.pdf (Inglés | English)

Efectos del programa social oportunidades en zonas urbanas

Armando Sánchez Vargas
Instituto de Investigaciones Econoómicas, UNAM
Este trabajo tiene como objetivo examinar las tendencias de mediano plazo del comportamiento de los hogares beneficiarios del Programa Oportunidades en las zonas urbanas. EspecĂ­camente, se estudian las tendencias en los siguientes aspectos sociales: e l consumo alimentario y no alimentario, las condiciones de vivienda, la inversión en activos del hogar, el empleo infantil y juvenil y la toma de decisiones en los hogares desde un enfoque de género. Lo anterior se lleva a cabo mediante el análisis del panel de datos de Oportunidades, con base en el programa Stata, procurando explicar las tendencias observadas en dichos indicadores y sus posibles cambios en el tiempo. Se presenta un análisis econométrico de los impact os de mediano plazo del Programa sobre los indicadores de aspectos sociales destacados y, al mismo tiempo, se busca determinar la existencia de sinergias de Oportunidades con otros programas sociales. Cabe destacar que los métodos utilizados para la estimación de los impactos toman en cuenta la naturaleza no experimental de la información estadística disponible para el análisis.

Additional information
sanchez.ppt

Mixed models: A simulation approach

Isabel Cañette
StataCorp
Simulating data is a powerful tool for understanding statistical models and for spotting identification problems. I will use simulation techniques to explain the building blocks for linear mixed models, and I will also show how to estimate the parameters using the xtmixed command. Using these basic blocks, I will explain how more-complex models can be constructed. Finally, I will explain some nice (but not obvious) applications of xtmixed.

Additional information
icanette.pdf

Updated programs of improved Stata ado-files for nonparametric smoothing

Néstor Aarón Mosqueda-Roma, Isaías Hazarmabeth Salgado-Ugarte
Laboratorio de Biometría y Biología Pesquera, FES Zaragoza, UNAM
In this talk, we present the updated versions of some simple, improved programs for nonparametric smoothing. These updated ado programs remain simple and versatile—mainly with the Stata command style—but now they have Stata’s version 11 syntax, a change that becomes apparent overall in the graphical routines.

The programs include the following:
  • Density traces
    • boxdetr1 (uniform weight functions by direct calculation)
    • boxdt21 (uniform weight functions by discretized calculation)
    • cosdetr1 (cosine weight function)
    • dentrac1 (uniform and cosine weight functions)
  • Kernal density estimators
    • kernsi1 (rectangular weight function)
    • kernep1 (Epanechnikov weight function)
    • kerngau1 (Gaussian weight function)
    • adaker1 (Gaussian weight function with variable bandwidth)
    • kerneld1 (permits choice of the kernel weight function)
    • warpdenm1 (based on the averaged shifted histograms; permits choice of the kernel weight function besides counting and estimating the distribution’s modes)
  • A program for practical rules to smooth parameter (bandwidth) determination in density estimation
    • bandw1 (permits the kernel specification with automatic adjustment)
  • Nonparametric assessment of multimodality
    • with critiband1, a program to find critical bandwidths for a given number of modes
    • with bootsamb (generates bootstrap samples) and silvtest1 (calculates the significance of a modal number), programs to evaluate the significance of a number of modes by means of a smoothed bootstrap procedure
As with previous versions, this collection of simple programs continue to be very useful in the analysis of data distribution (biological and of other kinds) and permits the efficient use of time and effort in the calculations. It must be stressed, too, that these commands make possible the use of Stata’s new graphics syntaxes.

Additional information
salgado2.pptx

Report to users

Gustavo A. Sánchez
StataCorp
The Stata Blog: Not Elsewhere Classified Find us on Facebook Follow us on Twitter LinkedIn Google+ Watch us on YouTube