Amelia June 2012

amelia@lists.gking.harvard.edu

11 participants
9 discussions

by Matissa Hollister

Hi, I'm using Amelia on a panel dataset. The main variables that I'm imputing are nominal variables. I'd like to include lags/leads of those variables, since they are the best predictors of the missing values, but I'm not sure if the lags/leads option can handle nominal variables. I ran it and it didn't produce errors but the imputations were pretty far off (this is sort of simulated data, so there is a right answer, and the results here were not much better than random chance). At the moment I've changed the data to what stata calles "wide" format with one observation per individual and a separate variable for each year. This improves the imputations but I wonder whether I'm losing some of the predictive ability because I think the lag and lead are more predictive and I'd have a better estimate of these relationships with the larger sample size in "long" format that pools all of the ages. Also, on a related note, is there a way to include double lags/leads (t-2, t+2)? I'm assuming this process is different from just creating a variable in the data that is the lag/lead and including it as a regular variable. Thanks, Matissa

11 years, 9 months

problem

by Arvind Magesan

Hello: I am using the most recent version of Amelia through R. When I follow the instructions from the user manual: " The easiest way to load AmeliaView is to open an R session and type the following two commands: > library(Amelia) > AmeliaView() This will bring up the AmeliaView window on any platform." I get an error message that says: "Error in structure(.External("dotTclObjv", objv, PACKAGE = "tcltk"), class = "tclObj") : [tcl] couldn't open "C:/Program Files/R/R-2.15.0/library/Amelia/doc/gallery19.gif": no such file or directory." I also can not load it from the desktop. Is there any immediate obvious reason why this is happening? thanks for your help. -- Arvind Magesan Department of Economics University of Calgary Room 430 2500 University Drive NW Calgary AB Canada T2N1N4 tel 1.403.220.5276 fax 1.403.220.5262 arvind.magesan(a)gmail.com anmagesa(a)ucalgary.ca

11 years, 10 months

mi.meld exporting dataset

by Shane Gleason

Hi everyone, With the mi.meld command I get the sense it is geared toward running analysis on the multiply imputed data. I multiply imputed values into a dataset about state attorneys general, mostly things like budget and staff size. I want to take these values and insert them into another dataset about court cases. So, the values I imputed are all controls. When I look at the mi.meld command it seems it is geared toward running actual analysis and seems to require said analysis to create a unified dataset. I'm not sure how to proceed, I need to get a unified dataset out; but I do not have any 'dependent variable' or model in my data. Thanks, I appreciate your input! Shane

11 years, 10 months

Re: [amelia] Error in if (sum(non.vary == 0)) in Amelia version 1.2-17

by Guy Grossman

Hi Matt, I am trying to perform MI with Amelia (R and Amelia are updated to the latest versions). After setting up the variables I received the following error: Error in if (sum(non.vary == 0)) { : argument is not interpretable as logical I searched online but was unable to find a way to correct this error. I have actually relatively few missing variables. any idea what I can do to get rid of the error. Your response would be highly appreciated! Guy # define variables' type ords = c("post_elections", "prev_elections", "election_year") noms = c("type", "region", "split_cat", "splinter", "mother", "breakup", "home_winner", "tribe_winner", "support_winner", "opposed", "irregularity", "home_loser", "tribe_loser", "support_loser", "politics_loser", "mobilize_loser", "ethnic_min", "ethnicity_dec_major", "dom_ethnic02", "turn_majority", "contig_border", "TYPE_2") idvars = c("county_name", "district_11", "regions_67", "district_80", "district_91", "district_96", "district_01", "district_02", "district_06", "district_07") lags = c("split_cat", "mshare_county") logs = c("area", "pop") lgstc = c("mshare", "mshare_county", "cshare_county", "elf02", "eduattain", "literacy", "employee", "pio", "nag_pio", "nag_employee", "pl02", "pi02", "perclit91", "percillit91") a.out <- amelia(DistrictFormation, m = 5, ts = "wave", cs = "id", polytime = 2, intercs = TRUE, idvars=idvars, ords=ords, noms=noms, logs=logs, lgstc=lgstc) Error in if (sum(non.vary == 0)) { : argument is not interpretable as logical > str(DistrictFormation) 'data.frame': 474 obs. of 109 variables: $ id : num 1 1 1 2 2 2 3 3 3 4 ... $ wave : int 1 2 3 1 2 3 1 2 3 1 ... $ county_name : chr "BUJUMBA" "BUJUMBA" "BUJUMBA" "KYAMUSWA" ... $ county_id02 : num 1011 1011 1011 1012 1012 ... $ region : Factor w/ 4 levels "CENTRAL","EASTERN",..: 1 1 1 1 1 1 1 1 1 1 ... $ district_11 : chr "KALANGALA" "KALANGALA" "KALANGALA" "KALANGALA" ... $ regions_67 : chr "MASAKA" "MASAKA" "MASAKA" "MASAKA" ... $ district_80 : chr "MASAKA" "MASAKA" "MASAKA" "MASAKA" ... $ district_91 : chr "KALANGALA" "KALANGALA" "KALANGALA" "KALANGALA" ... $ district_96 : chr "KALANGALA" "KALANGALA" "KALANGALA" "KALANGALA" ... $ district_01 : chr "KALANGALA" "KALANGALA" "KALANGALA" "KALANGALA" ... $ district_02 : chr "KALANGALA" "KALANGALA" "KALANGALA" "KALANGALA" ... $ district_06 : chr "KALANGALA" "KALANGALA" "KALANGALA" "KALANGALA" ... $ district_07 : chr "KALANGALA" "KALANGALA" "KALANGALA" "KALANGALA" ... $ ndist80 : num 18 18 18 18 18 18 18 18 18 18 ... $ ndist91 : num 39 39 39 39 39 39 39 39 39 39 ... $ ndist02 : num 56 56 56 56 56 56 56 56 56 56 ... $ ndist06 : num 72 72 72 72 72 72 72 72 72 72 ... $ ndist07 : num 80 80 80 80 80 80 80 80 80 80 ... $ ndist11 : num 109 109 109 109 109 109 109 109 109 109 ... $ n1991 : num 2 2 2 2 2 2 1 1 1 4 ... $ n1996 : num 2 2 2 2 2 2 1 1 1 4 ... $ n2001 : num 2 2 2 2 2 2 1 1 1 3 ... $ n2006 : num 2 2 2 2 2 2 1 1 1 2 ... $ ncounties : num 2 2 2 2 2 2 1 1 1 4 ... $ type : chr "COUNTY" "COUNTY" "COUNTY" "COUNTY" ... $ post_elections : num 2001 2006 2011 2001 2006 ... $ prev_elections : num 1996 2001 2006 1996 2001 ... $ split_cat : Factor w/ 3 levels "No Split","Splinter",..: 1 1 1 1 1 1 1 1 1 3 ... $ splinter : num 0 0 0 0 0 0 0 0 0 0 ... $ mother : num 0 0 0 0 0 0 0 0 0 1 ... $ breakup : num 0 0 0 0 0 0 0 0 0 1 ... $ election_year : int 2001 2006 2011 2001 2006 2011 2001 2006 2011 2001 ... $ mshare_county : num 0.667 0.539 0.639 0.677 0.535 ... $ cshare_county : num 0.317 0.403 0.296 0.304 0.421 ... $ mmargin_county : num 0.343 0.137 0.342 0.365 0.115 ... $ mshare : num 0.693 0.585 0.689 0.693 0.585 ... $ mshare_cat : num 0.695 0.622 0.706 0.695 0.622 ... $ MP_nrm_share : num -999 1 1 -999 0 1 -999 1 1 -999 ... $ home_winner : int 0 0 0 1 1 1 1 1 1 0 ... $ tribe_winner : Factor w/ 37 levels "ACHOLI","ALUR",..: 6 6 6 6 6 6 6 6 6 6 ... $ support_winner : int 0 0 1 1 1 1 1 1 1 1 ... $ opposed : int 1 1 0 1 1 0 0 1 1 1 ... $ irregularity : int 0 0 0 0 0 0 0 0 0 0 ... $ home_loser : int 1 1 0 0 0 0 1 1 1 0 ... $ tribe_loser : Factor w/ 37 levels "ACHOLI","ALUR",..: 6 6 6 6 6 6 6 13 13 6 ... $ support_loser : int 1 1 0 0 0 0 0 0 0 0 ... $ politics_loser : int 0 1 0 0 1 0 0 1 0 1 ... $ mobilize_loser : int 0 0 0 0 0 0 0 0 0 1 ... $ county_dec_share : num 0.6 0.4 0.6 0.4 0.6 ... $ ethnic_min : num 0 0 0 0 0 0 0 0 0 0 ... $ ethnicity_dec_major : Factor w/ 37 levels "ACHOLI","ALUR",..: 6 6 6 6 6 6 6 6 6 6 ... $ control_dec_ethnicity: num 1 1 0.8 1 1 ... $ pop : num 17271 17271 17271 17428 17428 ... $ dom_ethnic02 : Factor w/ 37 levels "ACHOLI","ALUR",..: 6 6 6 6 6 6 6 6 6 6 ... $ share_dom_ethnic02 : num 67.7 67.7 67.7 60.2 60.2 ... $ turn_majority : num 0 0 0 0 0 0 0 0 0 0 ... $ pop02 : int 17271 17271 17271 17428 17428 17428 229297 229297 229297 137199 ... $ elf02 : num 0.528 0.528 0.528 0.624 0.624 ... $ schools_tot_pop : num 0.811 0.811 0.811 0.689 0.689 ... $ schools_tot : num 14 14 14 12 12 12 260 260 260 170 ... $ other_tot_pop : num 0.116 0.116 0.116 0 0 ... $ other_tot : num 2 2 2 0 0 0 1 1 1 1 ... $ primary_tot_pop : num 0.579 0.579 0.579 0.631 0.631 ... $ primary_tot : num 10 10 10 11 11 11 229 229 229 142 ... $ sec_tot_pop : num 0.1158 0.1158 0.1158 0.0574 0.0574 ... $ sec_tot : num 2 2 2 1 1 1 30 30 30 27 ... $ gov_primary_pop : num 0.521 0.521 0.521 0.459 0.459 ... $ gov_primary : num 9 9 9 8 8 8 188 188 188 81 ... $ gov_sec_pop : num 0.1158 0.1158 0.1158 0.0574 0.0574 ... $ gov_sec : num 2 2 2 1 1 1 7 7 7 6 ... $ eduattain : num 0.222 0.222 0.222 0.241 0.241 ... $ literacy : num 0.818 0.818 0.818 0.826 0.826 ... $ pl02 : num 0.12 0.12 0.12 0.04 0.04 ... $ se_po02 : num 1.89 1.89 1.89 0.93 0.93 ... $ pg02 : int 2 2 2 1 1 1 10 10 10 9 ... $ se_pg02 : num 0.48 0.48 0.48 0.2 0.2 ... $ pi02 : num 0.0033 0.0033 0.0033 0.0036 0.0036 ... $ sepi02 : num 1.14 1.14 1.14 1.63 1.63 ... $ npoor02 : int 2014 2014 2014 772 772 772 79497 79497 79497 45674 ... $ se_npoor02 : int 326 326 326 162 162 162 3164 3164 3164 1962 ... $ elf91 : num 0.43 0.43 0.43 0.458 0.458 ... $ area : num 2356 2356 2356 6672 6672 ... $ pop91 : int 7335 7335 7335 6872 6872 6872 132711 132711 132711 105562 ... $ pl91 : num 0.365 0.365 0.365 0.261 0.261 ... $ pg91 : num 10.1 10.1 10.1 7.1 7.1 ... $ gini91 : num 29.9 29.9 29.9 32.6 32.6 ... $ popdensity91 : num 3.11 3.11 3.11 1.03 1.03 ... $ censuspop91 : int 9192 9192 9192 7179 7179 7179 141607 141607 141607 108098 ... $ perclit91 : num 0.711 0.711 0.711 0.729 0.729 ... $ percillit91 : num 0.289 0.289 0.289 0.271 0.271 ... $ distance_border : num 81855 81855 81855 94097 94097 ... $ contig_border : int 0 0 0 0 0 0 0 0 0 0 ... $ distance_kampala : num 86.6 86.6 86.6 80.2 80.2 ... $ pio : num 0.668 0.668 0.668 0.704 0.704 ... $ employee : num 0.358 0.358 0.358 0.393 0.393 ... $ nag_pio : num 0.249 0.249 0.249 0.15 0.15 ... $ nag_employee : num 0.1628 0.1628 0.1628 0.0701 0.0701 ... $ pop_county18 : num 10257 10257 10257 12185 12185 ... [list output truncated] -- ___________________________ Guy Grossman Postdoctoral Research Associate Princeton University, NJ o: 212-854-4985 m: 917-664-6946 www.guygrossman.com

11 years, 10 months

Imputation model: a trade off?

by Philippe Sulger

Dear all I have a question concerning the inclusion of (auxiliary) variables into the missing data procedure. I understand that a rather "inclusive" strategy can increase efficiency and reduced bias. Now, I also have the feeling that the inclusion of a certain (auxiliary) variable can have an additional cost that depends on the degree of missingness of this (auxiliary) variable itself. If the latter is "too high", couldn't this result in a higher disadvantage of including this variable relative to the advantage (increase in efficiency and/or reduced bias) that the inclusion of the variable could have? If yes, does there exist a measure/rule of thumb to evaluate and judge on this trade off? Thank you for your efforts. Philippe

11 years, 10 months

Overimputation diagnostics

by Karakaya, Suveyda

Hi, I have a question about imputation disgnodtics, the imputed and observed densities of polity 2 are quite different, first one with central density, second one with two hills. Also, I could not figure out how to interpret overimputation graph. There are vertical lines all around with different colors and I dont know which one denotes observed or imputed values. For 90 percent comfidence , should these vertical lines concentrate on the leftside of the graph? My vertical lines is scattered all around, is this a sign of trouble? One last question, since there are five imputations, should I try all of them in stata for robustness check or would using just one of them is fine? Thank you very much Suveyda Karakaya PhD Candidate The University of Tennessee

11 years, 10 months

Imputation too slow , and numerical options for tscs data?

by Karakaya, Suveyda

Dear Prof King, I am new to Amelia. I try to impute a data set with 8400 obs and 32 vars in total, 14 vars have missing obs, some a lot like around 4 000. Since I have many missing obs, I set the ridge prior to 40, is that correct? I also chose time spline with k=1 and interaction with cross section, is that also correct option for a tscs data? My second question is: how long on average does it take to complete imputation? I am waiting for 12 hours and do not see any progress at all yet, the computer is running but there is no imputation yet. Is it because I put ridge option to 40 or because of many logical bounds and log transformations that I set for my vars? Thank you very much Suveyda Karakaya PhD Candidate The University of Tennessee

11 years, 10 months

AmeliaView on Mac

by Ann-Renee Blais

Matt, I did exactly what you suggested on my Mac, but I too get the error message described previously and nothing actually happens... > AmeliaView() Loading required package: tcltk Loading Tcl/Tk interface ... Can anybody help? Thanks! Ann-Renee Blais

11 years, 10 months

ameliaview fails to load on Windows 7

by nazli avdan

Hello, I am somewhat new to Amelia. I had previously successfully run Amelia a few times with Windows XP and just reinstalled R and Amelia on my new device which runs Windows 7. I have successfully installed R but would prefer to use Amelia View as standalone program as I am a novice at R as well. Whenever I try to load Amelia View the black window shows up for a split second before vanishing. I also installed R and Amelia on a Windows Vista machine with the same result. I tried searching through the mailing list for fixes and tried loading Amelia from within R but am getting an error message every time. I also tried running R and Amelia with administrative privileges in case it's an issue with later versionof Windows. More specifically, here's the error message: Loading required package: tcltk Loading Tcl/Tk interface ... done Error in structure(.External("dotTclObjv", objv, PACKAGE = "tcltk"), class = "tclObj") : [tcl] couldn't open "C:/Program Files/R/R-2.15.0/library/Amelia/doc/gallery19.gif": no such file or directory. If anyone knows a fix to this, I'd appreciate a response. Best, N. Avdan

11 years, 10 months

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

Amelia June 2012