Baseline, Placebo, and Treatment: Efficient Estimation for Three-Group Experiments

ISPS Data Archive: Terms of Use

By using, contributing, and/or downloading files associated with scholarly studies available on the ISPS Data Archive, you agree to these terms and conditions. Please read the ISPS Data Archive Terms of Use.

ISPS ID: 
D034
Suggested citation: 

Gerber, Alan S., Donald P. Green, Edward H. Kaplan, Holger L. Kern (2010), Replication Materials for ‘Baseline, Placebo, and Treatment: Efficient Estimation for Three-Group Experiments.’ http://hdl.handle.net/10079/26e4e431-6937-4afd-90c8-13a517d25eb9. ISPS Data Archive.

Keyword(s): 
Author(s): 

Holger Lutz Kern, Donald P. Green, Edward H. Kaplan, Alan S. Gerber

Research design: 
Data type: 
Administrative
Data source(s): 

“Qualified Voter File” (QVF), the official state voter list (as constructed by Mark Grebner at Practical Political Consulting).

Field date: 
August 1, 2008
Field Date: 
2008-08
Location: 
Location details: 
Michigan
Unit of observation: 
Individuals
Sample size: 
369,211 individuals/registered voters
Inclusion/exclusion: 
Households in the state of Michigan. Households with three or more registered voters were eliminated as were precincts with fewer than 100 voters. Households were eliminated that had been shown their vote history in previous field experiments as well as households where addresses listed on the QVF did not appear to be valid. The final sample was restricted to voters who voted in one of the previous two elections (2005 or 2006).
Randomization procedure: 
Voters were grouped into households, which were defined to include everyone with the same last name at the same address. Households were randomly assigned to either the control group or one of three treatment groups described next. Precincts were first given a random identifier and sorted in random order. Within precincts, households were sorted according to an arbitrary household identifier. These households were then randomly assigned into one of three treatment conditions using a skip pattern.
Treatment: 
Mailing; Individuals assigned to treatment groups were sent one mailing 14 days prior to the November general election. There were 3 conditions: Civic Duty, Shown Past Abstention, and Shown Past Vote.
Treatment administration: 
Mail
Outcome measures: 
Turnout rates in 2007 local elections
Archive date: 
2010
Owner: 
Gerber, Alan S., Donald P. Green, Edward H. Kaplan, Holger L. Kern
Owner contact: 

isps(at)yale(dot)com

Terms of use: 

ISPS Data Archive: Terms of Use

Discipline: 
Data file number Descriptionsort ascending File format Size File url
D034F09 Treatment materials .wav 277504 Download file
D034F10 Treatment materials .wav 327680 Download file
D034F11 Supplementary materials Adobe acrobat (8.0) .pdf 200704 Download file
D034F01 ReadMe file .txt 1024 Download file
D034F05 Program file R (2.9.1.) .R 19456 Download file
D034F08 Program file .R 1024 Download file
D034F06 Program file .R 1024 Download file
D034F04 Program file R (2.9.1.) .R 7168 Download file
D034F07 Program file .R 1024 Download file
D034F14 Metadata (DDI 3.2) .xml 131746 Download file
D034F02 Dataset Stata (10.0) .dta 79796633 Download file
D034F03 Dataset Excel (.csv) 73505177 Download file