Zelig August 2007

zelig@lists.gking.harvard.edu

9 participants
7 discussions

by Eduardo Leoni

FYI I am seeing a similar problem with the "by" option in zelig. It repeats the solution for the whole dataset instead of estimating the models by subset. -eduardo On 8/3/07, James Honaker <tercer(a)ucla.edu> wrote: > > > Dear zeligites, zeligistas? (is there a term of art?) > > I ran across curious behaviour in the mi() function in the newest version > of R. I wrote a short code snippet below to demonstrate what is > happening. > Everything works fine in R 2.4.1, but in R 2.5.1, if you have m imputed > datasets, zelig seems to repeatedly estimate the model m times in the > first dataset, rather than moving through them. Thus the combined results > > are simply the results from the first imputed dataset. > > Again, everything looks fine in R 2.4.1, but fails in R 2.5.1 (these > things seem to crop up every new version). > > If it's worth anything, it doesn't appear to be the mi() function itself, > but how zelig handles this object as an argument. > > regards, > james. > > > > # test of the mi() function in zelig > # jH, Aug 2, 2007 > > library(Zelig) > > beta1<- -2:2 # This is a set of true coefficients > # Averaging over them should give beta1=0 > n<-1000 > testdat<-as.list(0) > > for(i in 1:length(beta1)){ # Construct datasets > x<-runif(n) > y<-beta1[i]*x + rnorm(n) > testdat[[i]]<- as.data.frame(cbind(y,x)) > } > > mi.object > <-mi(testdat[[1]],testdat[[2]],testdat[[3]],testdat[[4]],testdat[[5]]) > output<-zelig(y~x,model="ls",data=mi.object) > > print(summary(output)) # should give coefficient on x of 0, not -2. > > # This seems to be five copies of the result from the first dataset in R > v2.5.1 > # But works just fine in R v2.4.1 > > print(summary(output), subset=1:5) > > - > Zelig Mailing List, served by Harvard-MIT Data Center > Send messages: zelig(a)lists.gking.harvard.edu > [un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig > Zelig program information: http://gking.harvard.edu/zelig/ >

16 years, 7 months

twosls problem/bug?

by Andy Harris

Hi, I've run across a problem in the two stage least squares model in Zelig, which may be due to my code or to a bug. I want to use 2SLS for an instrumental variable regression. However, whenever I attempt to do so, I get an error: Error in checkNrReq(modelNumEqn, userNumEqn, modelPar) : The parameter "mu" requires between 2 and Inf equation(s). You have provided 1 See model doc. for more details I've provided some simulated data and the code that produces the error below. Note that I used the same syntax as published in the Zelig manual. Any thoughts on where I/zelig might be going wrong? Doing the procedure with lm() seems to recover the correct parameters. Many thanks, Andy > nn<-1000 #sample size > V<-rnorm(nn) #confounding omitted variable > W<-rnorm(nn) #instrument > X<-rnorm(nn) #exogenous regressor > > Z<-rep(NA, nn) #instrumental var. holder > for (i in 1:nn){ + Z[i]<-rnorm(1, mean=V[i]+W[i]) #instrumental var as a function of instrument and omitted variable. + } > > Y<-rep(NA, nn) > for (i in 1:nn){ + Y[i]<-rnorm(1, mean=V[i]+0.25*Z[i]) #outcome as func of omitted var. and Z. + } > > data<-as.data.frame(cbind(X,W,Z,Y)) > names(data)<-c("X","W", "Z", "Y") > > ### using the 2sls syntax in section 12.48 of the zelig manual > > fml <- list ("mu" = Y ~ X + Z, + "inst" = Z ~ W + X) > z.out <- zelig(formula = fml, model = "twosls", data=data) Error in checkNrReq(modelNumEqn, userNumEqn, modelPar) : The parameter "mu" requires between 2 and Inf equation(s). You have provided 1 See model doc. for more details > > > ### 2sls using lm() > lm.out1<-lm(Z~W+X, data=data) > lm.out2<-lm(Y ~ X + lm.out1$fitted, data=data) > summary(lm.out1) Call: lm(formula = Z ~ W + X, data = data) Residuals: Min 1Q Median 3Q Max -3.7840 -0.9809 0.0200 0.9883 4.2447 Coefficients: Estimate Std. Error t value Pr(>|t|) (Intercept) -0.00241 0.04329 -0.06 0.96 W 0.99100 0.04236 23.39 <2e-16 *** X 0.02788 0.04414 0.63 0.53 --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 1.37 on 997 degrees of freedom Multiple R-Squared: 0.355, Adjusted R-squared: 0.353 F-statistic: 274 on 2 and 997 DF, p-value: <2e-16 > summary(lm.out2) Call: lm(formula = Y ~ X + lm.out1$fitted, data = data) Residuals: Min 1Q Median 3Q Max -4.975 -1.122 -0.055 1.105 5.361 Coefficients: Estimate Std. Error t value Pr(>|t|) (Intercept) -0.00901 0.05170 -0.17 0.86164 X -0.04195 0.05274 -0.80 0.42653 lm.out1$fitted 0.18591 0.05104 3.64 0.00028 *** --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 1.63 on 997 degrees of freedom Multiple R-Squared: 0.0135, Adjusted R-squared: 0.0116 F-statistic: 6.85 on 2 and 997 DF, p-value: 0.00111 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Andy Harris Ph.D. Candidate Department of Government Harvard University

16 years, 8 months

covariates on correlation coefficient using bprobit model

by Michael Braun

First, thanks for creating such a useful R package. I am a new Zelig (2.8.3) user, but I can already see its time-saving potential. However, I am having some trouble using the bprobit model, and I hope someone can help me. I want to model something like: mu1 = x1 %*% b1 mu2 = x2 %*% b2 rho = f(x3 %*% b3) where mu1 and mu2 are the means of the underlying bivariate probit, and f() is the exp(x)-1/exp(x)+1 transformation. x1, x2 and x3 are vectors of variables, and are all different. The zelig command I am using is z.out <- zelig(list(mu1 = y1~x1, mu2 = y2~x2, rho = ~x3), model="bprobit",data=mydata) Note that x1 is really x11+x12+x13+..., and so forth. My problem is that when I do summary(z.out), I get coefficient estimates for b1 and b2, but only a single intercept estimate for b3. If I change x3 variables, I get the same estimate for b3. It is as though zelig is ignoring all of the covariates I put on rho. I didn't see a reference to this in the list archives, but I'm wondering if this is a known issue. From the documentation, it looks like this should be possible. Thanks in advance, Michael -- Michael Braun Assistant Professor of Marketing MIT Sloan School of Management One Amherst St., E40-169 Cambridge, MA 02142 (617) 253-3436 braunm(a)mit.edu - Zelig Mailing List, served by Harvard-MIT Data Center Send messages: zelig(a)lists.gking.harvard.edu [un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig Zelig program information: http://gking.harvard.edu/zelig/

16 years, 9 months

Zelig_3.0-3 : another bug-fix release

by Ferdinand Alimadhi

Hi everyone, The third bug fix release of Zelig version 3.0 (that's Zelig_3.0-3) is online. Two bugs fixed: - "bprobit" model in Zelig was ignoring the covariate (reported by Michael ) - "relogit" model had problems with with large number of variables in the formula (reported by Nathan). To update/install this release: - *nix / windows users: install.packages("Zelig", repos="http://gking.harvard.edu") - mac users: install.packages("Zelig", repos="http://gking.harvard.edu", type="source") Please let us know if you encounter any other problem. thanks, Ferdi - Zelig Mailing List, served by Harvard-MIT Data Center Send messages: zelig(a)lists.gking.harvard.edu [un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig Zelig program information: http://gking.harvard.edu/zelig/

16 years, 9 months

Problem with large numbers of variables in ReLogit?

by nathan.toronto＠us.army.mil

Good day, All. I would like to run ReLogit in Zelig using a large number of time dummy variables (to account for duration dependence), but I get an error when I try to include all of the temporal dummies. Does anyone have any ideas as to why this might be occurring? The script below creates a data set that approximates the data that I'm actually using and replicates the error that I've been experiencing. A note: the Zelig model represented by object "b" below returns an error message that the model does not converge. In the real data it actually does converge. The real problem is with object "c", and I suspect it has something to do with the large number of variables, since when I reduce the number of variables to around 40 or 50, everything if fine (or, at least I have results that need interpreting). Thank you very much, Nathan -----begin script----- ## Creating the mock data set: # time variable: time<-rep(c(1:150),50) # caseid variable: q<-c(1:50) caseid<-rep(q,each=150) # dependent variable: y<-rep(0,times=7500) y[40]<-1 y[209]<-1 y[556]<-1 y[1980]<-1 y[2297]<-1 y[2806]<-1 y[3345]<-1 y[4025]<-1 y[5567]<-1 y[6983]<-1 # independent variable: x1<-rep(0,times=7500) x1[27:150]<-1 x1[101:300]<-1 x1[307:450]<-1 x1[547:600]<-1 x1[1020:1050]<-1 x1[1287:1350]<-1 x1[1355:1500]<-1 x1[2333:2400]<-1 x1[2536:2550]<-1 x1[3051:3150]<-1 x1[3501:3600]<-1 x1[4408:4500]<-1 x1[5361:5400]<-1 x1[6573:6600]<-1 x1[6727:6750]<-1 x1[7159:7200]<-1 x1[7422:7500]<-1 # 150 time dummy variables: r<-as.data.frame(matrix(0,nrow=150,ncol=150)) for(i in 1:150){r[i,i]<-1} r$time<-c(1:150) # joining the variables together: X<-as.data.frame(cbind(time,caseid,y,x1)) X<-merge(X,r,by="time",all.x=T,all.y=F) ## Replicating error in Zelig library(Zelig) a<-zelig(y~x1, model="relogit",robust=F,data=X) b<-zelig(y~x1+ X[,1]+X[,2]+X[,3]+X[,4]+X[,5]+X[,6]+X[,7]+X[,8]+X[,9]+X[,10]+ X[,11]+X[,12]+X[,13]+X[,14]+X[,15]+X[,16]+X[,17]+X[,18]+X[,19]+X[,20]+ X[,21]+X[,22]+X[,23]+X[,24]+X[,25]+X[,26]+X[,27]+X[,28]+X[,29]+X[,30], model="relogit",robust=F,data=X) c<-zelig(y~x1+ X[,1]+X[,2]+X[,3]+X[,4]+X[,5]+X[,6]+X[,7]+X[,8]+X[,9]+X[,10]+ X[,11]+X[,12]+X[,13]+X[,14]+X[,15]+X[,16]+X[,17]+X[,18]+X[,19]+X[,20]+ X[,21]+X[,22]+X[,23]+X[,24]+X[,25]+X[,26]+X[,27]+X[,28]+X[,29]+X[,30]+ X[,31]+X[,32]+X[,33]+X[,34]+X[,35]+X[,36]+X[,37]+X[,38]+X[,39]+X[,40]+ X[,41]+X[,42]+X[,43]+X[,44]+X[,45]+X[,46]+X[,47]+X[,48]+X[,49]+X[,50]+ X[,51]+X[,52]+X[,53]+X[,54]+X[,55]+X[,56]+X[,57]+X[,58]+X[,59]+X[,60]+ X[,61]+X[,62]+X[,63]+X[,64]+X[,65]+X[,66]+X[,67]+X[,68]+X[,69]+X[,70]+ X[,71]+X[,72]+X[,73]+X[,74]+X[,75]+X[,76]+X[,77]+X[,78]+X[,79]+X[,80]+ X[,81]+X[,82]+X[,83]+X[,84]+X[,85]+X[,86]+X[,87]+X[,88]+X[,89]+X[,90]+ X[,91]+X[,92]+X[,93]+X[,94]+X[,95]+X[,96]+X[,97]+X[,98]+X[,99]+X[,100]+ X[,101]+X[,102]+X[,103]+X[,104]+X[,105]+X[,106]+X[,107]+X[,108]+X[,109]+X[,110]+ X[,111]+X[,112]+X[,113]+X[,114]+X[,115]+X[,116]+X[,117]+X[,118]+X[,119]+X[,120]+ X[,121]+X[,122]+X[,123]+X[,124]+X[,125]+X[,126]+X[,127]+X[,128]+X[,129]+X[,130]+ X[,131]+X[,132]+X[,133]+X[,134]+X[,135]+X[,136]+X[,137]+X[,138]+X[,139]+X[,140]+ X[,141]+X[,142]+X[,143]+X[,144]+X[,145]+X[,146]+X[,147]+X[,148]+X[,149]+X[,150], model="relogit",robust=F,data=X) -----end script----- Nathan W. Toronto, Ph.D. Foreign Military Studies Office 731 McClellan Avenue Fort Leavenworth, KS 66027 913-684-5614 (office) nathan.toronto(a)us.army.mil - Zelig Mailing List, served by Harvard-MIT Data Center Send messages: zelig(a)lists.gking.harvard.edu [un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig Zelig program information: http://gking.harvard.edu/zelig/

16 years, 9 months

bug fix

by Gary King

Zelig version 3.0-2, which fixes the mi() bug that James Honaker found, is now on line. To update, just do update.packages() and library(Zelig) and you'll be all set. Gary --- Gary King, Institute for Quantitative Social Science Harvard University, 1737 Cambridge St, Cambridge, MA 02138 http://GKing.Harvard.Edu, King(a)Harvard.Edu Direct 617-495-2027, Assistant 495-9271, eFax 812-8581 - Zelig Mailing List, served by Harvard-MIT Data Center Send messages: zelig(a)lists.gking.harvard.edu [un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig Zelig program information: http://gking.harvard.edu/zelig/

16 years, 9 months

mi() in R 2.5.1

by James Honaker

Dear zeligites, zeligistas? (is there a term of art?) I ran across curious behaviour in the mi() function in the newest version of R. I wrote a short code snippet below to demonstrate what is happening. Everything works fine in R 2.4.1, but in R 2.5.1, if you have m imputed datasets, zelig seems to repeatedly estimate the model m times in the first dataset, rather than moving through them. Thus the combined results are simply the results from the first imputed dataset. Again, everything looks fine in R 2.4.1, but fails in R 2.5.1 (these things seem to crop up every new version). If it's worth anything, it doesn't appear to be the mi() function itself, but how zelig handles this object as an argument. regards, james. # test of the mi() function in zelig # jH, Aug 2, 2007 library(Zelig) beta1<- -2:2 # This is a set of true coefficients # Averaging over them should give beta1=0 n<-1000 testdat<-as.list(0) for(i in 1:length(beta1)){ # Construct datasets x<-runif(n) y<-beta1[i]*x + rnorm(n) testdat[[i]]<-as.data.frame(cbind(y,x)) } mi.object<-mi(testdat[[1]],testdat[[2]],testdat[[3]],testdat[[4]],testdat[[5]]) output<-zelig(y~x,model="ls",data=mi.object) print(summary(output)) # should give coefficient on x of 0, not -2. # This seems to be five copies of the result from the first dataset in R v2.5.1 # But works just fine in R v2.4.1 print(summary(output), subset=1:5) - Zelig Mailing List, served by Harvard-MIT Data Center Send messages: zelig(a)lists.gking.harvard.edu [un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig Zelig program information: http://gking.harvard.edu/zelig/

16 years, 9 months

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

Zelig August 2007