Does anyone know how in R code to view say the top 10 maximum values of a column, alongside their values for another column - ie - county name, with income level, sorted so I only view the top 10 richest counties? _______________________________________________ gov2001-l mailing list gov2001-l(a)lists.fas.harvard.edu http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l

Hi again, Holger has helped me with code suggestions but they are still not running for me on my computer. Can anyone spot my error? object<-cbind(var1,var2) object<-as.matrix(object) object[sort(object$var1, decreasing=T)[1:10],object$var2] The first two lines of code run, and I create the matrix object with 2 columns and 3000 rows - but I can't get the next line to run, what I want is to sort the rows in decreasing values of var1, such that the name for that county (var2) also moves with it. Thanks again, Suzanna On Sat, 29 Apr 2006, ghumphr(a)fas.harvard.edu wrote: _______________________________________________ gov2001-l mailing list gov2001-l(a)lists.fas.harvard.edu http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l

Hi all, Below is our preliminary abstract. We were not sure how to refer to the article/authors in the subheading. We welcome comments and criticism. Thanks, Elena & Omar ################ It?s Not About Race: Confirming Fryer and Levitt's Understanding of the Black-White Test Score Gap Upon School Entry Although more than half a century has passed since the landmark Brown v. Board of Ed. decision, black children continue to lag academically behind their white peers. For decades, research comparing the test scores of black and white students has consistently found a substantial gap, even among four-year olds. Fryer and Levitt (2004), using a new nationally representative data set, found that the test score disparity among Kindergarteners almost entirely disappears after controlling for a small set of socio-economic, family and health related characteristics. If robust, this exceptional finding would suggest that black-white differences in testing are neither innate nor insoluble. In this article we demonstrate that the results are not sensitive to model specification, different treatments of missing data, or weights. The gap evident in later years is not present when children enter school. _______________________________________________ gov2001-l mailing list gov2001-l(a)lists.fas.harvard.edu http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l

Hi, I'm trying to put two graphs into the same pdf file. When I use par(mfrow=c(1,2)) the two graphs become a little distorted (very long and skinny). Is there any way to put two graphs on the same pdf document but keep the graphs as squares, and not turn them into long rectangles? The only way I can figure out how to do this is with par(mfrow=c(2,2)) but then the bottom half of the page is blank, which is not a great use of space. Thanks! Becky _______________________________________________ gov2001-l mailing list gov2001-l(a)lists.fas.harvard.edu http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l

I'm making plots of propensity score distributions - and specifically for the matched data. If TR is my treated subsample of the matched data, do I want TR$propens, or TR$distance, or TR$propscore? The homework solutions for PS7 use TR$propscore, but this doesn't come up under names(TR) for me as an option - I used distance - but I see propens listed under names. Can someone tell me which it is supposed to be? _______________________________________________ gov2001-l mailing list gov2001-l(a)lists.fas.harvard.edu http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l