Assigning Contacts to Households

Householding

Exact Household Match
After standardizing street number, name, apartment number
All records with the same street number, name zip code and apartment 
number are assigned to the same household using a key 
[StNumber] & [StName] & [Zip] & [Apt]
Flaws: using this method if voters in the same household have used 
different designations for apartment number, designations that cannot be 
standardized, or omitted an apartment number they will not be 
householded.  Voters living in a private house are not accounted for in this 
method.

Last Name Household Match
For records with no apartment number
After standardizing street number, name, 
All records with the same street number, name, zip code and last name 
are assigned to the same household using a key
[StNumber] & [StName] & [Zip] & [LastName]
Flaws: Voters with different st names living in the same house (private 
house or apartment without apartment number) well not be householded.  
The increase in non married couples living together, or hyphenated last 
names or women keeping their maiden names all detract from the efficacy 
of this method.

Phone number Household Match:
If the datafile is phone matched professionally, many vendors will 
append the phone number to all members of a household (not sure what their 
methodology is, if it goes beyond name/address style household 
matching, or if it comes from their telephone databases.  Phone number can be 
used to identify households, though it does not necessarily identify all 
members of a household.

Points of comparison that can be used to identify households when there 
is no apartment number:

Partial Name Match: a comparison that looks for parts of names, either 
comparing the first few letters of last names or the last few letters 
of last names could help to identify hyphenated names.

Age and gender can be used to identify likely members of a household.
Records at the same address that are within a certain number of years 
(5 for example) in age and of opposite genders could be assumed to be 
members of the same household.
Flaw: this does not account for roommates or same sex couples.
Records where one is between 18 and 24 and the other(s) are 20 to 40 
years older could identify children and their parents as members of the 
same household.


Once the household has been identified there is the complex issue of 
household salutation for mailing purposes.  In many cases a campaign may 
not want to mail to all members of the household (republican man, and 
dem women in the same household).  We have created household salutations 
based upon the number of target voters in a household.  For example:
1 Targeted voter: FirstName LastName
2 Targeted voters with same last name: FirstName1 & FirstName2 LastName
2 Targeted voters with different last name: FirstName1 LastName1 & 
FirstName2 LastName2
3 or more Targeted Voters: The LastName (picked at random) Household
In the 3rd scenario, the last Name could be chosen based upon the most 
prevalent last name, or the last name of the oldest mail voters, or it 
could be a concatenation of last names of voters in the household (eg 
The Smith and Jones Household, or The Smith Jones Wilson 
Household
The Household salutation could be The Smith Family if all the 
last names are Smith.