Los efectos perturbadores en las relaciones entre variables
Archivos
Fecha
2002
Autores
Chaves Esquivel, Edwin
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad Nacional (Costa Rica)
Resumen
Se analiza el efecto perturbador que una característica puede provocar en la relación entre dos o más variables. Se utilizó la información correspondiente a un cuestionario que se aplicó a una muestra de educadores de enseñanza primaria en 1998. Por medio de ejemplos concretos se observa como la relación entre dos variables que aparentan una dependencia estadística importante, desaparece cuando se controla su efecto con una tercera variable. Esto se debe a que las variables utilizadas para explicar un determinado fenómeno, en general, se encuentran muy correlacionados entre sí, y es precisamente esa correlación la que algunas veces provoca un efecto espurio sobre la relación original. Los análisis se efectúan por medio de tablas de contingencia y contrastes Chi-Cuadrado para independencia estadística. Estas herramientas mostraron su utilidad para determinar algunos de los efectos confusores entre las variables.
This article evaluated the importance of analyzing the disturbing effect that a specific characteristic can cause between two or more variables. In order to analyze the problem, it was used information got from a questionnaire that was applied to a sample of primary teachers on 1998. Through concrete examples it is possible to observe how the apparent statistical dependence between two variables disappear when its effect is controlled using a third characteristic. This happens because the variables used to explain a specific phenomenon are strongly correlate among them. This correlation sometimes causes a false effect on the original relation. The analyses are made by contingency tables and Chi-Square test. These tools showed their utility to determine some of the confused effects among the variables.
This article evaluated the importance of analyzing the disturbing effect that a specific characteristic can cause between two or more variables. In order to analyze the problem, it was used information got from a questionnaire that was applied to a sample of primary teachers on 1998. Through concrete examples it is possible to observe how the apparent statistical dependence between two variables disappear when its effect is controlled using a third characteristic. This happens because the variables used to explain a specific phenomenon are strongly correlate among them. This correlation sometimes causes a false effect on the original relation. The analyses are made by contingency tables and Chi-Square test. These tools showed their utility to determine some of the confused effects among the variables.
Descripción
Palabras clave
EFECTO DE CONFUSIÓN, EFECTO SUPRESOR, TABLAS DE CONTINGENCIA, CONFOUNDING EFFECT, CONTINGENCY TABLES