Victims: 1,181; family: 1,220, Neighbors: 9,632
9/11 victims: Matched a list of victims from public sources, including only victims who were residing in New York and who were killed in the attacks, to voter registration file (1,181 of 1,729 (68%) New York state victims were
identified on the 2001 voter registration file).
Family members: Used a family identification number generated by Labels & Lists that groups individuals together who live in the same household.
Neighbors: Composed a database of all registered voters who, in 2001, lived on the same Census block and on the same street as a victim; For those who live in houses, further restricted the neighbors to the 10 whose house numbers were closest; For those victims identified with an apartment unit, restricted their neighbors to those who lived in the same building or on the same floor as the victim.
All individuals then matched with Catalist database.
For each victim, used a combination of exact matching and distance matching to find up to five “control victims” whose observable traits are as close as possible to the observable traits of the real victims; excluded from the control group the 66 control victims and 132 control family members (2% of all control matches) who were themselves true family members or true neighbors. Using the same procedures as used for the true victims, for each control victim, identified the person’s family members and nearest neighbors.
(i) participation in federal elections, (ii) participation in political party primaries, (iii) campaign contributions, and (iv) party affiliation.