PDI Mail Ballot Return Scores

Permanently deleted user
Permanently deleted user
  • Updated
PDI Mail Ballot Return (MBR) Score Information
Using PDI (early Late) Mail Ballot Return Scores Options in All Mail Ballot Elections
Splitting a Universe by household for Early and Late mail drops (Three Steps)

 

PDI Mail Ballot Return (MBR) Score Information    

It is no secret that in a typical California election, at least half of the ballots are cast prior to Election Day.  Almost every campaign now implements a strategy for contacting early mail ballot voters as well as late or Election Day voters.   

 

Campaigning to mail ballot voters usually entails identifying which voters will vote early by mail and then timing the communication (direct mail and phone banks) to coincide with when the voters fill out and returns their ballot (by mail or at a drop box).  

 

Since 2004 we have collected information from tens of millions of mail ballots cast and when and where they voted.  These mail ballots reflect the voting behavior of any active voters with a history of voting.  To make this data simple for campaigns to use, we have created a Mail Ballot Return Score (MBR Score) that provides a simple classification for mail ballot voters in California.

 

MBR Score

We have assigned a comprehensive score value reflecting the number of mail ballot votes available, when the ballot(s) were returned, and the consistency or inconsistency of a voter’s behavior.  We identified early votes as ballots returned prior to eight days before an Election Day.  Late votes were returned within eight days before an election.

 

mail_ballot_return_early_vs_late_period_table.png

 

The score values are:

  • Always Returned Their Ballots Early (Always Early / AE) – Multiple mail ballots cast, ALL prior to eight days before Election Day.
  • Usually Returned Their Ballots Early (Usually Early / UE) - Only one mail ballot vote cast prior to eight days before Election Day, or Multiple mail ballots cast, mostly or the most recent being consistently, prior to eight days before Election Day.
  • Returned Ballots Early And Late (Early-Late / EL) – Voters with an equal number of mail ballots cast within both the early and late period, or lacking a consistent pattern of early or late behavior.
  • Usually Returned Their Ballots Late (Usually Late / UL) - Only one mail ballot vote cast within eight days before or on Election Day, or Multiple mail ballots cast, mostly or the most recent being consistently, within eight days before or on Election Day.
  • Always Returned Their Ballots Late (Always Late / AL) – Multiple mail ballots cast, ALL of which were within eight days before or on Election Day.

 

The MBR Score is intended to help campaigns predict the likelihood of an voter casting their ballot early or late in an upcoming election cycle.  The information should only be used for prioritization and does not in any way guarantee behavior in a future election cycle.  Consider the data to be one more piece of a puzzle that provides a more comprehensive look at mail ballot voting behavior.

 

If you have questions or comments about this information, please contact your PDI account representative.

 


Using PDI (early Late) Mail Ballot Return Scores Options in All Mail Ballot Elections

The options below attempt to address the issue of all voters being issued a Vote By Mail (VBM) ballot.  The options are endless, but it really depends on how you are using this data to contact voters. Mail is less forgiving when you split it up into too small of files, because mailing it could drive your data and postage costs up. So if you plan on splitting your mail drops up, it is probably more practical to split you mail in two group or drops, but if you are just calling or emailing voters at different times it could be helpful with scaling  your campaigns work load into smaller buckets or three, four, five or six splits.  For example, your volunteers may not be able to call all of the voters in your target universe at once but you could send mailing that would be mailed out and delivered to everyone at the same time.  However, if you send mail to voters too late after a voter has voted or returned their ballot it is a waste of your print and postage and a missed opportunity.  But the fact is that in regularly scheduled elections, with a long ballot and many choices and issues, most voters wait to vote later in the last week of on election day, that includes voters that drop their ballot in the mail or at a drop box on election day.   If you are working on a special election with just one or two issues or races on the ballot the behavior of voters is different with more voters voting earlier than they would in a normal election, in which case you should use this early and late data with great caution.

 

If you are trying to split a universe between voters with history of always voting late and everyone else. It is safest to assume that any voter who has never voted by mail before could very well vote early (as soon as they have the option to vote by mail or at a vote center or drop off location), so you will want to choose the following split/combination:

 

Early = NOT ALWAYS LATE

&

Late = ALWAYS LATE

 

Another common or more conservative combination is to define Early universe is based on any voters that are NOT (or neither) ALWAYS LATE or USUALLY LATE, which will leave you with those voters that are scored as being in any one of Early groups as well as those voters without any Early or Late score available.

 

Early = NOT (ALWAYS LATE OR USUALLY LATE)

&

Late = ALWAYS LATE OR USUALLY LATE

 

If you are just trying to be more aggressive and limit your early mailings to just those voters that are most likely to vote early, than you will want to target those voters that have typically voted early in one group and then leave all the late and first time mail ballot voters in the second later group using the following split:

 

ANY EARLY (ALWAYS EARLY, USUALLY EARLY, EARLY & LATE)

&

NOT ANY EARLY (ABOVE) OR ALWAYS LATE

 

You can also get more granular and split your list three ways with the “EARLY & LATE” group in the middle with all the first time VBM voters and include the more consistent Early and Late voters as bookends, which could look like the following:

 

ALWAYS EARLY OR USUALLY EARLY

&

EARLY & LATE (THEN ADD), OR NOT (ALWAYS EARLY, USUALLY EARLY, ALWAYS LATE OR USUALLY LATE)

&

ALWAYS LATE OR USUALLY LATE  

 

You could even include the USUALLY scores in the middle and just include the ALWAYS scores as bookends, but the reality is that there just aren’t that many voters that are ALWAYS EARLY. So, a safer granular three part approach would include Any Early in the first group and then exclude those Any Early groups and ALWAYS LATE, which leaves you with those voters without a score plus USUALLY LATE in your middle group and then you have just the ALWAYS late left in the last group as follows:

 

ANY EARLY (ALWAYS EARLY, USUALLY EARLY, EARLY & LATE)

&

NOT (ALWAYS EARLY, USUALLY EARLY, EARLY & LATE, OR ALWAYS LATE)

&

ALWAYS LATE

 

Finally, keep in mind that everyone in an All Mail ballot election is going to get a mail ballot.  There are some voters that have voted by mail before and some that either have never voted or choose to vote in person either because thar is what they prefer or because they lost or can’t find their mail ballot. So, your campaign’s strategy may be to send more or early mailers to those voters that are going to be mailing for the first time to help educate them and get them out to vote by mail or at a vote center or drop off location.  In this case you could do the following split:

 

ANY_VBM  - Voters Who Voted By Mail At Least 1 Time in Prior Election Cycles in Prior Election Cycles

&

NOT ANY_VBM  - Voters Who Voted By Mail At Least 1 Time in Prior Election Cycles in Prior Election Cycles

 

But remember virtually everyone that votes in person does so in the LATE period of the last week or day of voting. So, whether or not that vote in person or by mail matters much less than when they vote.

 


Splitting a Universe by household for Early and Late mail drops (Three Steps)

STEP 1 - Start by Creating a saved universe of your total base target universe that also excludes the dead, moved and already voted.  The example below that base universe is saved as “22G6 Less Dead Moved Voted” 

 

Picture4.pngPicture5.png

 

STEP 2 – Create a saved universe of the voters you want to use to select your Early mail universe. This universe is based on any voters that are in your base universe who are NOT (or neither) ALWAYS LATE or USUALLY LATE, which will leave you with those voters that are scored as being in any one of Early groups as well as those voters without any Early or Late score available. This will require two sets of criteria the first criteria is selected by Voter and the second criteria is selected by Household.

  1. Select your saved base target universe   “22G6 Less Dead Moved Voted” and Add that criteria (by Voters), which is first set of criteria that will apply to all voters.
  2. Select both your saved base target universe   “22G6 Less Dead Moved Voted” AND from the Early Voting tab check both the boxes to Exclude ALWAYS LATE or USUALLY LATE voters.
  3. Once you have both of those Criteria in the pending Criteria box, click on the switch to Add Criteria by Household and then click on the “Intersect Criteria (AND)” button.  This will select any voters in your Base universe Criteria 1 that also are in the same household as any voters that are in your second Criteria.
  4. Now click Save Universe and save this universe as your “22G6 Early Houses” universe.

Picture6.png