By Robert Kibrick

Edited by Ellen Theisen of

May 12, 2004



Precinct-count Optical Scan System for New York



The State of New York should seriously consider deploying statewide a precinct-count optical scan system. While this "interim" solution could be implemented early in 2005, it would actually form the foundation of a permanent solution for New York State that could be fully implemented by 2006.




      The capital cost would be approximately half the cost of deploying DREs.

      It would significantly reduce the number of uncounted votes in New York, especially in races for Senate and Governor

"For example, three percent of voters using hand-counted paper and scanned paper ballots had no vote recorded for Senate or governor, but seven percent of voters using lever machines recorded no vote for Senate or governor." [CalTech/MIT Voting Project, page 8]

"Had the counties using lever machines used optical scanning, we estimate that there would have been 830,000 more votes recorded in Senate and gubernatorial elections." [CalTech/MIT Voting Project, p. 22]

      It would provide protection against over-voting (as lever machines do), and ALSO provide warnings about under-voting (which lever machines do not)

      It would support bilingual ballots for a variety of foreign languages

      It would be accessible to voters in wheel chairs (lever machines are not)

      It would provide in-precinct ballot counts immediately at the close of the polls (just like lever machines do)

      It would provide a voter-verifiable paper trail for all voters in New York State

      With the addition of a modest amount of additional equipment, all of the HAVA accessibility requirements for 2006 could be met on time

      It would provide the foundation of a permanent solution so that NONE of the investments in equipment purchase, election and poll worker training, and public education would be wasted.



The optical scan paper ballots would be printed on large sheets of paper:

      Ballots would be available in different languages as needed.

      Voters would cast their votes by filling in the "bubbles" on optical scan ballots.

      Voters would use lightweight, portable voting booths to afford privacy while voting. (These are available in heights to accommodate standing voters as well as ADA-compliant wheelchair-height versions.)

      Once voters completed their ballots, they would insert them into the precinct- count optical scanner:

- Over-voted ballots would be rejected; the voter would get a fresh ballot

- Blank ballots would produce a warning, and the voter would get the opportunity to correct the ballot.

- Once a ballot was successfully completed and accepted by the optical scanner, the votes on the ballot would be counted into the scanner's memory and the scanner would deposit the ballot into a locked ballot box


- At the close of the polls, the optical scanner would produce a printout of all of the vote totals. Those vote totals would be sent to election central, just the same as the vote totals from lever machines were sent in previous elections. The locked ballot box would be transported to election central so the ballots would be available for any subsequent recount or audit.




- One precinct-count optical scanner Average cost: $5,000

- As many voting booths as needed Average cost: $250 each

- Marking pens to mark optical scan ballots Average cost: $25


Such equipment is readily available from several major manufacturers of voting equipment. Without implying an endorsement of any specific vendor's equipment, the following examples are offered for illustration purposes only:


Precinct-count optical scanner:

Example scanner: ES&S Model 100. Additional information is available at:


Portable, fold-up voting booth:

Example: Vogue Election Systems Vogue-I voting booth



The only variable is the number of voting booths needed for each polling place. The minimum would be one voting booth and the maximum would probably be something like 10. So the range of capital costs per polling place would be:


1-voting booth polling place: ($5000 + $250) = $5250

5-voting booth polling place: ($5000 + $1250) = $6250

10-voting booth polling place:($5000 + $2500) = $7500




This depends on the total number of polling places in New York, which is not a number that I have available. Since New York has approximately 20,000 lever machines, we know that there must be no more than 20,000 polling places. As a very rough estimate, let's assume an average of four lever machines per existing polling place, which would yield 5,000 polling places.


For our new system, let's assume the average polling place will have five voting booths. Since a 5-voting-booth polling place requires $6250 in equipment, if we have 5,000 polling places to equip, the total cost would be approximately $31 million dollars for the entire State of New York. (If the number of polling places is larger, than the total cost would be correspondingly higher, and if the number is lower, the cost would be lower.)


Are these rough estimates reasonably in the ballpark? New York State has about 11,200,000 registered voters. Let's assume a relatively high turnout (i.e., > 70%), and plan for 8,000,000 voters to vote on election day. We have allocated an average of 5 voting booths for each of 5,000 precincts, for a total of 25,000 voting stations (or about 25% more voting stations than the number of existing lever machines). That works out to an average of 320 voters per voting booth, which is well within the typical range for this voting technology.


Add to these capital equipment costs the costs for training election officials, poll workers, and educating the public. I do not have a good way of estimating these costs, but can imagine that these could run in the range of $10 million to $20 million (i.e., roughly $1 to $2 for each registered voter in New York).




The precinct-count optical scan solution described above could be more than just an "interim" paper solution. Rather, it could form the foundation of a permanent and extremely cost-effective solution that could be used for many years to come. As a result, none of the investments in that "interim" solution would be wasted.


To expand this "interim" solution into a permanent one would only require the addition of one additional device per polling place to accommodate the needs of blind, visually-impaired, and language-impaired voters. This device could be one of either:


- A ballot marking device Approx. cost $4,500


- A DRE/touch screen voting machine + VVPB printer Approx. cost $4,500


Assuming 5,000 polling places as in our previous example, the incremental cost of equipping (by 2006) each polling place with one of the above devices to provide disability access would be $22,500,000. Add to that another $5,000,000 for additional training for election and poll workers and another $2,000,000 for an educational campaign for blind, visually-impaired, and reading-impaired voters. Thus, a rough estimate of upgrading the "interim" solution to the permanent solution would be of order $30,000,000 for the entire state.


Such equipment will be readily available from several major manufacturers of voting equipment no later than 2005, well in time to meet the HAVA accessibility requirements of 2006. Without implying an endorsement of any specific vendor's equipment, the following example is offered for illustration purposes only:


Ballot marking device


This device provides a touch screen or audio interface (for blind, visually-, or reading-impaired voters) to a standard optical scan ballot. It also provides over-vote and under-vote protection directly within the marking device, thus ensuring that the optical scan ballot that it completes on behalf of any voter is correctly filled in. Thus, any optical scan ballot completed using such a ballot marking device should be readily accepted by the precinct-count optical scanner, so that there would be little likelihood of it being rejected by that scanner due to over-votes or under-votes, and thus no need for any further intervention by the voter.



Assuming 5,000 polling places in New York State, the solution proposed above is roughly estimated to cost:


- Interim solution:

= $31,250,000 for polling place equipment

= $20,000,000 for election/poll worker training & public education

total $51,250,000

- Cost to add accessibility for blind, vision/reading-impaired by 2006:

= $22,500,000 for polling place equipment

= $ 7,000,000 for election/poll worker training & public education

total $29,500,000


Subtotal for capital costs only (both phases): $53,750,000

Total cost for both phases: $80,750,000


If, instead, New York chose a solution which deployed an equivalent number (i.e., 25,000) of DRE touch screen voting machines and each one of those voting machines included an accessible voter-verified paper ballot printer, the estimated cost for capital equipment alone would be $112,500,000 (as compared to a total capital equipment cost of $53,750,000 for the solution proposed above). Thus, a solution employing a 100% deployment of DRE voting machines with attached VVPB printers is roughly twice as expensive as the proposed solution. The hardware and software maintenance costs for the DRE+VVPB solution are also estimated to be significantly higher due to the significantly larger number of programmable electronic devices.


The proposed solution has the added advantage that it can be deployed incrementally: the "interim" solution is put in place in first and includes voting access for voters in wheelchairs, while access for blind, vision- and reading-impaired voters is added in 2005 and 2006, in time to meet HAVA deadlines.




The State of Illinois and the City of Chicago provide examples of a large urban state and a densely-populated city that have successfully used precinct-count optical scan systems for their elections. For example, 83% of the population of Illinois (10 million) votes using such systems, including Chicago, which is a large, ethnically-diverse city like New York. For additional information from one Illinois County describing their rationale for choosing a precinct-count optical scan solution, please see:


Also, 80% of the State of Arizona also uses precinct-count optical scan, including the city of Phoenix. For details, see:


The Secretary of State of Michigan has recommended that that entire State use precinct-count optical scan. For details, see:,1607,7-127-1640_9150-43906--M_2001_5,00.html


Other states that use mostly precinct-count optical scan systems also include


South Dakota ( ) and

Minnesota ( ).

An example of another large city that uses precinct-count optical scan is Seattle:

In addition to these examples, two major studies of voting systems determined that precinct-count optical scan systems outperformed DRE voting machines in terms of residual voting errors and cost per voter.

CalTech/MIT Voting Project

George Washington University Report on Election Systems Reform





By deploying a precinct-count optical scan system for the entire State of New York soon, the number of New Yorker voters who are now disenfranchised by problems with or limitations posed by lever voting machines could be immediately reduced. In addition, voting access would be provided to voters in wheelchairs.


The equipment needed for such a solution is readily available, certified, and in widespread use in other states, counties, and cities across the U.S., including states and cities with similar population densities and ethnic diversities to New York. The equipment required for this solution not only provides a voter-verified paper ballot solution now (something that most touch screen systems currently can't do) but it also provides a more cost effective solution than would be provided by DRE/touch screen voting machines with attached VVPB printers.


Finally, this solution provides the foundation for a permanent solution, so that none of the initial investment is wasted. While all voters will need to be trained to use any new system that replaces lever machines, most voters would not need any retraining for 2006 and beyond. This proposal provides an incremental solution, delivering a VVPB capacity early, while fully meeting the HAVA accessibility requirements by 2006.


Were New York to embrace this proposed solution, New York would be helping to lead the way towards an accessible, voter-verified voting system, while at the same time increasing access for the disabled and reducing the number of voters who are disenfranchised by lever machine problems. That would establish New York as the role model for other states having large deployments of lever machines (e.g., Connecticut, Pennsylvania, etc.), and give New York national visibility as a leader on this issue that is so important to a large and rapidly growing number of concerned voters.