Categories: Capital Read, Inside the Queensway

Vancouver South Recount: Welcome back to elected officialdom, Ujjal Dosanjh!

by kadyomalley on Tuesday, November 4, 2008 3:26pm - 67 Comments

According to ITQ’s tireless source on the ground, the recount is over and after a full ballot-by-ballot recall, Ujjal Dosanjh has held onto his seat with a 20 vote margin. Whoo! Go democracy! Does everybody feel a comforting sense of closure now?

The Vancouver Sun has the details.

UPDATE: Yes, she did concede, but ITQ has been told that every vote was, in fact, counted.

Tags: maybe slightly less exciting than the other vote going, over at last, psychologically comforting, Ujjal Dosanjh, vancouver south recount, wai young

More by kadyomalley

http://bigcitylib.blogspot.com bigcitylib

An estimate is a guess, now, is it?
madeyoulook

Now you need a reading course. I never said that. Samples lead to estimates. Estimates are approximations, not guesses. Better and stronger samples lead to better estimates. Only counts lead to counts. Even then, errors are possible, hence the requirement to recount, with a review of rejected ballots, when the initial measurement is so close to a tie.

The judge’s guess that the result wouldn’t change was proven right. It was even likely to be correct at the time he guessed. And none of that matters, in math or in electoral law.
madeyoulook

Help me out, gang. Is my vocabulary too sophisticated, or is BCL just being a pain? Anyone else who has a fondness for mathematics and stats want to back up or challenge my statements?
T. Thwim

MYL has the point here.

An estimate, a statistical probability, is not a count. And while the judge’s methodology may have been accurate within 1%, 19 times out of 20, an election requires a much higher certainty than that.

The judge does not deserve an apology unless the recount was stopped based on the will of the candidates because the point of a recount is to ensure that both sides agree who won and who lost. The judge failed to do that when he stopped the initial count. As such, he failed in his duty.
Ti-Guy

Yeah, Conbots. A swarm is needed here. Mobs rule, after all.
madeyoulook

One more try, with examples, to enlighten BCL:

Estimate: The results of this random-sample survey is judged to be within 3.5 percentage points (of the true population score), nineteen times out of twenty.

Guess: I don’t understand the question or the five multiple-choice answers, but since there is no penalty for wrong answers I will guess C.
madeyoulook

Are. Results are, not is. Sorry.
KRB

BCL, why do you bother? You know you’d be up in arms if the tables were reversed. You know it, and I know it.

Liberal party HQ was probably hoping Dosanjh would lose. Ha!

G’Obama!
Austin So

What is it with you guys? Where do you think we live? We have the most respected electoral system in the world, except by certain segments of the Canadian population.

The point is that on election night, our ballots are counted by hand. They are monitored by members of each party. This kind of quality control in place on election night is pretty damn high.

The partial recount is geared towards validating that the quality control measures on election night are sound. Based on the partial recounts, it was highly unlikely that something fishy was going on.

And guess what? The full recount validates that premise (that the QC on election night is top notch).

But most of all, it should tell all of you that our electoral system is one of the most accurate in the world, and deserving of our respect.

Austin
Austin So

MYL: the problem with people’s convoluted understanding of stats is that you have to know to what you are applying it to before applying them.

Austin
madeyoulook

Based on the partial recounts, it was highly unlikely that something fishy was going on.

More misunderstanding of mandatory re-count, alas. It is not a smell test for the absence of fish. It is not an audit in which “…according to generally accepted accounting principles: the enclosed figures have not been found to materially misrepresent…” It is a recount.

Not sure who you are accusing of disrespecting Canada’s electoral system: (a) those of us, regardless of partisanship, who wanted the will of the voters of Van-S to carry, or (b) the participant(s) in the mandatory recount who are unclear of the definitions of “mandatory” and “recount.”

As to your Yogi Berra-ish wisdom on statistics. This is not a chi-square or a Mann-Whitney, here. It is a count. Maybe I am trying too hard with the post-secondary math. Maybe a certain muppet from Sesame Street would have been more helpful…
Austin So

Clearly MYL, if that is the level which makes sense to you, then by all means continue, and I will adjust accordingly.

Austin
madeyoulook

…to make it clear enough for certain folk who steadfastly refuse to understand what a count means.

Really, Austin, I always took you for a partisan, but never for as thick-skulled and intellectually challenged as you are attempting to portray here. I will just assume that you are playing dumb to keep picking a fight for the sport of it, because, frankly, any other conclusion makes you look worse.

At least BCL paid attention to the lesson and quietly went away to mull it over a bit.
Mike T.

Mr. Thwim, the judge deserves an apology if he acted as a judge is required to in these circumstances. Certain allegations have been raised, at first glance they are enough to make me want to hear the entire story. I haven’t yet.
Austin So

Thick-skulled! Intellectually challenged!

Oh my my!

Okay MYL, let’s play your game, then.

1. what kind of checks and balances exist on voting night to ensure that a vote is properly tabulated/enumerated?

2. what level of confidence do you place on these checks and balances?

You are essentially arguing that the difference between the counts of Uji and Wai falls within the margin of chance (or, more precisely “human error”). But what is the basis of your “statistics” if you have an n of 1? Not only that, but what about the impact of #1 and #2 on minimizing this error in the first place?

On the other hand, you *can* apply statistics on what measures can increase the accuracy of your results (e.g. the number of poll supervisors on increasing count accuracy), but they would not be a statement of the counts themselves. Does that make sense? Do you understand or appreciate the distinction?

Performing a sampling of ballot boxes to assess the quality of the verification process is fundamentally different than sampling to determine the accuracy of the counts. There is a fine but important distinction that needs to be made here.

Statistical measures only applies to the rejection of the statement you wish to validate, MYL. That is a *fundamental* definition.

This is why this “uproar” was unfounded, and was borne out in the final count.

But don’t mind me…I’m just “a hyper-partisan dumbass”…

Or was that you? I forget…

Austin
madeyoulook

You are essentially arguing that the difference between the counts of Uji and Wai falls within the margin of chance (or, more precisely “human error”). Yup, you really don’t get it. There is no chance here. Only measurement. And all your flowery nonsense about applications of statistics has zero bearing on a mandatory judicial recount. With regret, it makes me inclined to reject with high confidence the hypothesis that you are just playing dumb for the sport of it.

Sigh, bring out the muppet.

There are no statistics. No probabilities. No samples. No estimates. No “18 boxes have no or only small screw-ups, therefore all is probably well.”

The “check and balance” is right there in election law: a difference of less than 1/1000 of votes requires a mandatory judicial recount. We have however many boxes with ballots in them. Many ballots were rejected on election day. Recount all the ballots, and have a look at the rejected ballots so that rejections can be overruled by a judge whenever it is proper to do so; this will allow more ballots to be counted in a close race.

Count. Them. All. It really is that simple.
Jack Mitchell

“Sigh, bring out the muppet.”

That’s pretty funny.
Pingback: Vancouver South-What’s up with you? -OPEN THE BOXES- « Conservative Reporter’s Weblog
dan in van

Unfortunately, the muppet won’t run, you can’t make him.
I’m thinking CONs would rather hear ‘Bring out the Gimp’ (this is a movie reference and not implied to assail anyone real, or pretending to be real)…
Glad that the long wait for democratic justice-assured has been done. Whether we ever know the real intrigue behind the closed door as Mike T hopes, I rather doubt it. An amazing run by Young, a overly dramatic hold by Dosanjh. Now let’s get back in the House!
Karol Karolak

Brossard–La Prairie Recount

http://enr.elections.ca/ElectoralDistricts_e.aspx?ed=1470

Results as Certified by a Judge
Recount date: October 24, 2008

Brossard–La Prairie

Party Candidate Votes % Votes

Conservative Maurice Brossard 11,062 18.9
Marxist-Leninist Normand Chouinard 157 0.3
Bloc Québécois Marcel Lussier 19,034 32.5
NDP-New Democratic Party Hoang Mai 7,452 12.7
Liberal Alexandra Mendes 19,103 32.6
Green Party Sonia Ziadé 1,816 3.1
Total number of valid votes: 58,624
Rejected ballots: 563
Total number of votes: 59,187
=====================================

Results Validated by the Returning Officer
Validation Date: October 15, 2008

Party Candidate Votes % Votes

Conservative Maurice Brossard 11,090 18.8
Marxist-Leninist Normand Chouinard 162 0.3
Bloc Québécois Marcel Lussier 19,202 32.6
NDP-New Democratic Party Hoang Mai 7,477 12.7
Liberal Alexandra Mendes 19,100 32.4
Green Party Sonia Ziadé 1,840 3.1
Total number of valid votes: 58,871
Rejected ballots: 572
Total number of votes: 59,443
=========================================

Preliminary Results

Brossard–La Prairie

Party Candidate Votes % Votes

Conservative Maurice Brossard 11,084 18.8
Marxist-Leninist Normand Chouinard 159 0.3
Bloc Québécois Marcel Lussier 19,200 32.6
NDP-New Democratic Party Hoang Mai 7,530 12.8
Liberal Alexandra Mendes 19,057 32.4
Green Party Sonia Ziadé 1,860 3.2
Total number of valid votes: 58,890

Polls reporting: 258/258 Voter turnout: 58,890 of 91,339 registered electors (64.5%)
The number of registered electors shown in this table does not include electors who registered on election day.
================================================
What do we make out of all this???
Total number of valid votes: goes down from 58,890 in preliminary results to 58,871 Validated by the Returning Officer, to 58,624 Certified by a Judge on October 24, 2008,

Total number of votes: goes down from 59,443 Validated by the Returning Officer, to 59,187 Certified by a Judge on October 24, 2008

Total of votes that went missing between certification and recount 59,443 – 59,187=256.

Total number of Rejected ballots goes down from 572 Validated by the Returning Officer, to 563 Certified by a Judge on October 24, 2008

Conservative Maurice Brossard goes from 11,084 in preliminary count to 11,090 Validated by the Returning Officer, to 11,062 as Certified by a Judge for total loss of 11,084 – 11,062 = 22 votes.

Marxist-Leninist Normand Chouinard goes from 159 in preliminary count to 162 Validated by the Returning Officer , to 157 as Certified by a Judge for total loss of 159 – 157 = 2 votes.

Bloc Québécois Marcel Lussier goes from 19,200 in preliminary count to 19,202 Validated by the Returning Officer, to 19,034 as Certified by a Judge for total loss of 19,200- 19,034 = 166 votes.

NDP-New Democratic Party Hoang Mai goes from 7,530 in preliminary count to 7,477 Validated by the Returning Officer, to 7,452 as Certified by a Judge for total gain of 7,530 – 7,452 = 22 votes

Liberal Alexandra Mendes goes from 19,057 in preliminary count to 19,100 Validated by the Returning Officer, to 19,103 as Certified by a Judge for total gain of 19,057 -19,103 = 46 votes

Green Party Sonia Ziadé goes from 1,860 in preliminary count to 1,840 Validated by the Returning Officer, to 1,840 as Certified by a Judge for total loss of 1,860- 1,840 = 20 votes.

Once again it seems that number of votes that went missing between certification and judicial recount was deciding factor in who have won election in that riding and it explains reversal of fortunes in favour of Liberal candidate.

Most effective method to make sure that the votes go missing is pair of tweezers. Votes can easily be pulled out of ballot box thru the slot that was used to put them in without having to break any seals. Once votes are pulled out of the box than the can be sorted and votes for “winning candidate” can be put back in and vote for loosing candidate spoiled and put back in a box or trashed somewhere.

Above procedure seems like a “wining scenario”.
Cheers,
madeyoulook

Karol, if you are presenting your “winning scenario” as a way to “spoil” opponents’ ballots, that evil scheme would only work if the tweezer caper went on between the provisional count Oct 14-15 and the Judicial Certified Recount Oct 24. Is that what you are suggesting?

Aren’t the sorted and counted ballots placed in signed & sealed envelopes and put back into its own ballot box? I think the tweezer plan needs some refinement in order to be criminally successful.
Karol Karolak

Kitchener–Waterloo Recount
http://enr.elections.ca/ElectoralDistricts_e.aspx?ed=1378
Results as Certified by a Judge
Recount date: October 31, 2008
Kitchener–Waterloo

Party Candidate Votes % Votes

Conservative Peter Braid 21,830 36.1
Independent Mark Corbiere 107 0.2
Libertarian Jason Cousineau 333 0.6
CAP Kyle James Huntingdon 105 0.2
NDP-New Democratic Party Cindy Jacobsen 8,915 14.7
Green Party Cathy MacLellan 7,326 12.1
Communist Ramon Portillo 105 0.2
Liberal Andrew Telegdi 21,813 36.0
Total number of valid votes: 60,534
Rejected ballots: 229
Total number of votes: 60,763
===========================================

Results Validated by the Returning Officer
Validation Date: October 17, 2008

Party Candidate Votes % Votes

Conservative Peter Braid 21,851 36.1
Independent Mark Corbiere 107 0.2
Libertarian Jason Cousineau 337 0.6
CAP Kyle James Huntingdon 104 0.2
NDP-New Democratic Party Cindy Jacobsen 8,920 14.7
Green Party Cathy MacLellan 7,329 12.1
Communist Ramon Portillo 107 0.2
Liberal Andrew Telegdi 21,803 36.0
Total number of valid votes: 60,558
Rejected ballots: 245
Total number of votes: 60,803
=======================================

Preliminary Results

Kitchener–Waterloo

Party Candidate Votes % Votes

Conservative Peter Braid 21,871 36.1
Independent Mark Corbiere 108 0.2
Libertarian Jason Cousineau 337 0.6
CAP Kyle James Huntingdon 107 0.2
NDP-New Democratic Party Cindy Jacobsen 8,920 14.7
Green Party Cathy MacLellan 7,329 12.1
Communist Ramon Portillo 108 0.2
Liberal Andrew Telegdi 21,798 36.0
Total number of valid votes: 60,578

Polls reporting: 265/265 Voter turnout: 60,578 of 94,504 registered electors (64.1%)
The number of registered electors shown in this table does not include electors who registered on election day.
================================================

What do we make out of all this???
Total number of valid votes: goes down from 60,578 in preliminary results to 60,558 (twenty votes go missing) Validated by the Returning Officer, to 60,534 Certified by a Judge on October 31, 2008, for total loss of 60,578- 60,534 = 44 votes

Rejected ballots: go down from 245 Validated by the Returning Officer to 229 Certified by a Judge on October 31, 2008, for total change of 245 – 229 =+16 valid votes

Total number of votes: goes down from 60,803 Validated by the Returning Officer, to 60,763 Certified by a Judge on October 31, 2008, for a total loss of 40 votes.

Now, the interesting part; who lost votes???

Conservative Peter Braid goes from 21,871 in preliminary results to 21,851 Validated by the Returning Officer (twenty votes go missing) to 21,830 for a total loss of 21,871-21,830 = 41 votes

Independent Mark Corbiere goes from 108 in preliminary results to 107 Validated by the Returning Officer to 107 for a total loss of 108-107= 1 vote

Libertarian Jason Cousineau goes from 337 in preliminary results to 337 Validated by the Returning Officer to 333 for a total loss of 337-333 = 4 votes

CAP Kyle James Huntingdon goes from 107 in preliminary results to 104 Validated by the Returning Officer to 105 for a total loss of 107-105 = 2 votes

NDP-New Democratic Party Cindy Jacobsen goes from 8,920 in preliminary results to 8,920 Validated by the Returning Officer to 8,915 for a total loss of 8,920- 8,915 = 5 votes

Green Party Cathy MacLellan goes from 7,329 in preliminary results to 7,329 Validated by the Returning Officer to 7,326 for a total loss of 7,329 – 7,326 = 3 votes

Communist Ramon Portillo goes from 108 in preliminary results to 107 Validated by the Returning Officer to 105 for a total loss of 108-105 = 3 votes

Liberal Andrew Telegdi goes from 21,798 in preliminary results to 21,803 Validated by the Returning Officer to 21,813 for a total gain of 21,798 – 21,813 = +15 votes

Here you have it folks; 40 votes go missing between certification and recount. Leading Conservative candidate lost 41 votes (one more than number of votes that went missing), Liberal candidate gained 15 votes and Conservative candidate won nevertheless.

Funny how the numbers add up 40 missing vots as if somebody pulled them out of the ballot box sorted them out and threw out votes for the conservative candidate.
Karol Karolak

Brampton West Recount

http://enr.elections.ca/ElectoralDistricts_e.aspx?ed=1298

Results Validated by the Returning Officer
Validation Date: October 15, 2008

A judicial recount was requested in this electoral district.
Brampton West

Party Candidate Votes % Votes

Green Party Patti Chmelyk 3,330 6.2
Liberal Andrew Kania 21,739 40.3
Conservative Kyle Seeback 21,516 39.9
NDP-New Democratic Party Jagtar Shergill 7,339 13.6
Total number of valid votes: 53,924
Rejected ballots: 342
Total number of votes: 54,266
================================================

Preliminary Results

Brampton West

Party Candidate Votes % Votes

Green Party Patti Chmelyk 3,330 6.2
Liberal Andrew Kania 21,639 40.2
Conservative Kyle Seeback 21,516 40.0
NDP-New Democratic Party Jagtar Shergill 7,339 13.6
Total number of valid votes: 53,824

Polls reporting: 284/284 Voter turnout: 53,824 of 108,046 registered electors (49.8%)
The number of registered electors shown in this table does not include electors who registered on election day.
================================================
What do we make out of all this???
Total number of valid votes: goes up from 53,824 in preliminary results to 53,924 Validated by the Returning Officer, for total gain of 53,824 – 53,924 = 100 votes

Liberal Andrew Kania goes from 21,639 in preliminary results to 21,739 Validated by the Returning Officer, for total gain of 21,639 – 21,739 =100 votes.

All the other results remain identical.
Liberal candidate lead over Conservative candidate increase from 21,639 – 21,516 =123 to 21,739 – 21,516 =223 with 342 Rejected ballots.

Once again we see a pattern where the change between preliminary results and Validated results benefits Liberal candidate. Would that be the case of stuffing ballot box with votes on behalf of people who did not show up to vote??? It might as well be.
100 votes is a nice round number that could easily be attributed to a counting error.
This time number of rejected votes is larger that margin of victory and that makes fortunate counting mistake quite suspicious to say the least.

So far it seems that Election Canada stole three electoral victories from Conservatives and one electoral victory from Block for the benefit of the Liberal candidates.

Is there a pattern?? I sure think so, do you???
Austin So

There is no chance here. Only measurement.

Really.

If you had to sort and tally 30,000 pieces of paper with 5 different markings, do you think that you would get it exact the first time? The second time? The third time? And do you think that accuracy in your count will depend on how much time you spend perusing each piece of paper and how carefully you tallying each one?

Now, what if there were ambiguous markings on these pieces of paper? Is there a truly objective way of deciding which category to put them in? Or does it boil down to opinion?

If there was yet another recount with different counters and a different judge presiding, there will be a difference yet again.

If you think muppets work, great…it would probably make the process more entertaining (one…ah ah ah…two…ah ah ah…three…ah ah ah…)…but don’t try to pretend that statistics is not useful here or that a full recount is magically infallible.

I’m sorry you don’t appreciate this.

Austin

P.S. in fact I have no problem with the full recount, but I also appreciate that there are very good reasons to believe that a sampling of boxes can be sufficient under certain circumstances. What I don’t like is the presentation of this as some mad conspiracy in some corners on the part of Elections Canada (hi Karol!). Anecdotally, though, the sad thing is that Uji doesn’t realize that his close win reflects how little his constituency believes that he truly represents their interests.
madeyoulook

Austin’s main nugget, supplemented by a long-winded shpiel on the inherent error associated with any measurment attempt:
but don’t try to pretend that statistics [are] not useful here or that a full recount is magically infallible.

It’s almost like you’re trying to not get it. I am not pretending anything. Statistics are NOT useful here. A full judicial recount is not infallible; it is merely a better reflection of the will of the voters than the original election night count (by minimizing the inherent error mentioned above); that’s why the law insists on it in a very tight contest. The law does not say “sample a few and plug the #’s into SPSS, and if you get a halfway decent P-value, call the whole thing off.” The law says re-count. With an impartial judge working to re-enfranchise some voters on the basis of clear intent, because those some voters could make a difference in this particular riding.

Keep exposing your ignorance if you feel so inclined, Austin. There being no more productive ways to express a simple point to someone incapable of grasping it, I am done.