Monday, November 5, 2007
Absentee ballot processing & Directive 07-21
I am serving as an Election Observer for this Nov. 6 election. In that stead, I was at the Cuyhahoga County BOE yesterday for the opening of absentee ballots - removing ballots (and possibly any other materials) from the ID envelopes/ flattening ballots for upcoming scanning, which will begin Tuesday morning, 11/6 at 7:30a.
What a difference a year and change of leadership makes! I was informed of the times, and was welcomed as an Observer - and I even got to see what was happening! And they had a process! - an amazingly efficient one - a process they've honed and documented over the past number of non-county-wide races this year. I do have observations for a few changes I'd like to see, which I will take to the board, before reporting them here.
To date, as of yesterday, Sunday, and not yet counting Monday and Tuesday arrivals (all except overseas will be due by 7:30p, November 6) there were approx. 23,000 absentees.
(By comparison, in May, 2006, under old guard and few procedures, when the absentees needed to be handcounted because scanners didn't read correctly, and it took them approximately a week to actually count that primary with no handcounting plan, there were about 18,000 ballots.)
Yesterday, there were approximately 80 workers there. They got directions and started at 9:00a, and finished opening at about 2pm (including a one-hour lunch hour.) The ballots, still in their ID envelopes had been verified for HB3 ID, and had been sorted by precinct by the absentee department. They worked one precinct at a time. All flattened ballots, once opened, stayed in precinct groupings apparently to ease otherwise potential computer confusions and memory capacities in the upcoming scanning and tabulation.
Especially in this odd year race, many cities/precincts each have completely different "ballot styles" - very different things on each precinct's ballot - ranging from court races in various suburbs to suburb or precinct-specific zoning issues. The scanner is going to have to recognize which kind of ballot is being scanned, to tabulate correctly, so they do one ballot style at a time.
They also counted all ballots again, to make sure they still had the correct number of ballots that were returned in each opened pile. They have been keeping a count of those requested and those returned. Return rates so far are at about 74%, with highest returns in cities with the most hotly contested races.
The CCBOE is thus, ready to begin scanning and tabulating absentees on Tuesday morning after the polls are open, per the law. The CCBOE estimates that with the number of workers and scanners they have ( a number I still have to check, though I think I remember it's around 20 scanners, each needing a D and an R) that process will take about 4-5 hours. Most of these ballots are only one page (some front and back.)
For comparison, last November's race, an even congressional year, the ballots were 3 and 4 pages, and there were less than 23,000 - less than 20K. It took about 1 1/2 days to scan those multi-pagers, at the Diebold "high speed" rate (not their very expensive - if I remember more than $30K each - highest speed scanners) of- if I remember correctly, about 200 sides of pages per hour.
As is usual, the absentee ballots received by Sunday, yesterday, will be included in the unofficial count. The Monday and Tuesday-received ballots, because they have not been yet verified, opened, sorted etc. and the overseas absentees will then be handled and included in the later official count.
Interestingly, last year because of all the problems (and embarrassments) around scanning previously experienced in May '06, and the length of the ballot in November '06, Cuyahoga, and I believe a few other counties asked Mr. Blackwell for permission to start scanning absentees a day early, so they would be ready for election night counts. As I remember, Blackwell said no, but later a court order by Judge Gaul permitted the early scanning - under certain conditions, like they were supposed to have observers there, they were supposed to have a "security plan" and most of all - they were not allowed to tabulate (count) the votes until the law said - after the polls were open.
They were only allowed to "process them" and have the scanners just read, but not count the absentees early. The concept here is that if any tabulations are known and/or get leaked before the polls close, it could in a number of ways, affect the way people yet to vote, would cast their ballots.
For November 6, SoS Brunner is allowing early processing too - like the opening, flattening, etc. AND she is allowing reading of the ballots before the polls open for all counties using any scanners - EXCEPT those using DIEBOLD SCANNERS/SYSTEMS. I will link Directive 07-21 soon.
Apparently Diebold scanners and/or the GEMS tabulator cannot assuredly just read-scan without counting. In last year's brief audit of the CCBOE November,'06 election, apparently information was found that although the CCBOE had been court ordered not to tabulate when they started scanning a day early, the audit showed they were getting tabulations which they did not let the public know about. This makes one also wonder why Diebold - if a responsible company - when hearing the court order last year, did not explain that their system could not comply, if it apparently cannot. What they did was, they leased the CCBOE more than 20 extra scanners, so while the board staff started to scan early, they could still finish by election close.
So this year the CCBOE and all BOE's using Diebold/Premiere scanners need to start scanning only after the polls open.
Also interesting about SoS Directive 07-21, is that I had to get it from the CCBOE. The SoS website continues to, more often than not, prevent access to her directives, advisories and memos. It errors out. I have advised that office at least four separate times over the past few months of this "glitch." Last time, when they were told they got the problem fixed pretty rapidly. Last time I checked , however, they were only up to Directive 07-18. So again, I know we're missing some.
The SoS office needs to get this frequently occurring problem fixed permanently and soon.
What a difference a year and change of leadership makes! I was informed of the times, and was welcomed as an Observer - and I even got to see what was happening! And they had a process! - an amazingly efficient one - a process they've honed and documented over the past number of non-county-wide races this year. I do have observations for a few changes I'd like to see, which I will take to the board, before reporting them here.
To date, as of yesterday, Sunday, and not yet counting Monday and Tuesday arrivals (all except overseas will be due by 7:30p, November 6) there were approx. 23,000 absentees.
(By comparison, in May, 2006, under old guard and few procedures, when the absentees needed to be handcounted because scanners didn't read correctly, and it took them approximately a week to actually count that primary with no handcounting plan, there were about 18,000 ballots.)
Yesterday, there were approximately 80 workers there. They got directions and started at 9:00a, and finished opening at about 2pm (including a one-hour lunch hour.) The ballots, still in their ID envelopes had been verified for HB3 ID, and had been sorted by precinct by the absentee department. They worked one precinct at a time. All flattened ballots, once opened, stayed in precinct groupings apparently to ease otherwise potential computer confusions and memory capacities in the upcoming scanning and tabulation.
Especially in this odd year race, many cities/precincts each have completely different "ballot styles" - very different things on each precinct's ballot - ranging from court races in various suburbs to suburb or precinct-specific zoning issues. The scanner is going to have to recognize which kind of ballot is being scanned, to tabulate correctly, so they do one ballot style at a time.
They also counted all ballots again, to make sure they still had the correct number of ballots that were returned in each opened pile. They have been keeping a count of those requested and those returned. Return rates so far are at about 74%, with highest returns in cities with the most hotly contested races.
The CCBOE is thus, ready to begin scanning and tabulating absentees on Tuesday morning after the polls are open, per the law. The CCBOE estimates that with the number of workers and scanners they have ( a number I still have to check, though I think I remember it's around 20 scanners, each needing a D and an R) that process will take about 4-5 hours. Most of these ballots are only one page (some front and back.)
For comparison, last November's race, an even congressional year, the ballots were 3 and 4 pages, and there were less than 23,000 - less than 20K. It took about 1 1/2 days to scan those multi-pagers, at the Diebold "high speed" rate (not their very expensive - if I remember more than $30K each - highest speed scanners) of- if I remember correctly, about 200 sides of pages per hour.
As is usual, the absentee ballots received by Sunday, yesterday, will be included in the unofficial count. The Monday and Tuesday-received ballots, because they have not been yet verified, opened, sorted etc. and the overseas absentees will then be handled and included in the later official count.
Interestingly, last year because of all the problems (and embarrassments) around scanning previously experienced in May '06, and the length of the ballot in November '06, Cuyahoga, and I believe a few other counties asked Mr. Blackwell for permission to start scanning absentees a day early, so they would be ready for election night counts. As I remember, Blackwell said no, but later a court order by Judge Gaul permitted the early scanning - under certain conditions, like they were supposed to have observers there, they were supposed to have a "security plan" and most of all - they were not allowed to tabulate (count) the votes until the law said - after the polls were open.
They were only allowed to "process them" and have the scanners just read, but not count the absentees early. The concept here is that if any tabulations are known and/or get leaked before the polls close, it could in a number of ways, affect the way people yet to vote, would cast their ballots.
For November 6, SoS Brunner is allowing early processing too - like the opening, flattening, etc. AND she is allowing reading of the ballots before the polls open for all counties using any scanners - EXCEPT those using DIEBOLD SCANNERS/SYSTEMS. I will link Directive 07-21 soon.
Apparently Diebold scanners and/or the GEMS tabulator cannot assuredly just read-scan without counting. In last year's brief audit of the CCBOE November,'06 election, apparently information was found that although the CCBOE had been court ordered not to tabulate when they started scanning a day early, the audit showed they were getting tabulations which they did not let the public know about. This makes one also wonder why Diebold - if a responsible company - when hearing the court order last year, did not explain that their system could not comply, if it apparently cannot. What they did was, they leased the CCBOE more than 20 extra scanners, so while the board staff started to scan early, they could still finish by election close.
So this year the CCBOE and all BOE's using Diebold/Premiere scanners need to start scanning only after the polls open.
Also interesting about SoS Directive 07-21, is that I had to get it from the CCBOE. The SoS website continues to, more often than not, prevent access to her directives, advisories and memos. It errors out. I have advised that office at least four separate times over the past few months of this "glitch." Last time, when they were told they got the problem fixed pretty rapidly. Last time I checked , however, they were only up to Directive 07-18. So again, I know we're missing some.
The SoS office needs to get this frequently occurring problem fixed permanently and soon.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment