Election predictions

[pmcalk]

Clearly all mid-town people here.  What about Mautino/Troyer?  Patrick/Turner?

I think Perkins will do better than most independents, but 28% is way too high, RM. I think more likely 15%, which is pretty good considering he has 1/3 the money as the other guys.  I like Perkins, but think he is too inexperienced.  I say Adelson 44%; Bartlett 43%. 

While I understand the frustration with the two party system, I still think it is better.  Whomever wins next week, you can pretty much guess that more than half the voters in Tulsa preferred someone else.  Nobody can get much done because, even after the election, the majority of voters will be against him or her.  I think it is much easier for powerful special interests to gain control.  If we had instant runoffs, perhaps I would think differently, but I would worry that too many voters would not understand the process of instant runoffs.

I think city auditor will be interesting to watch.  Phil Wood has been auditor for a very long time, and has good name recognition.  On the other hand, this is the first time I recall a significant contender.  If Mr. Wood wins this time, you can pretty much guess he'll have his job until he decides to quit.

I am surprised at the Barnes/Gomez predictions.  Gomez nearly lost his primary--only capturing 55%, and losing several precincts.  He has a tough road ahead anyway, since District 4 has more democrats.  Not only does he have to capture all of the republicans, he needs a significant number of democrats, and all of the independents.  I guess he might pull it off, but he hasn't been without controversy.

[Bat Bat]

Adelson 47
Bartlett 38
Perkins 15

Wood 53
Doerflinger 47

Barnes 55
Gomez 45

Troyer 58
Mautino 42

Patrick 52
Turner 48

[Wrinkle]

Which brings up the question of why there have been absolutely no polls published. Guess the World doesn't like what it sees in them.

Imagine they'll post one of those 10-day old polls later in the week. Their polling has been pretty inaccurate in recent years anyway. At least, as it relates to contenders.

[waterboy]

Adelson boasted that polls showed him with a double digit lead last week. Most likely a privately conducted poll but I didn't hear Bartlett quote any poll at all. World won't spend the money. They are all on TV now with pretty good ads. Perkins has much appeal and Inhofe has reared his image. Gomez has a good mailer that is going to make the difference.

My guess is these races are very close and my projections could easily flip:

Adelson 47%
Bartlett 42%
Perkins 11%

Barnes 45%
Gomez 55%

Woods 60%
Doerflinger 40%

[akupetsky]

I saw in Urbantulsa yesterday that Rocky Frisco endorsed Maria Barnes.  I think that will have an effect.  My predictions:

Adelson:  44%
Bartlett:  42%
Perkins:  14%

Barnes:  53%
Gomez:  47%

Wood:  55%
Doerflinger:  45%

Troyer:  54%
Mautino:  46%

Patrick:  53%
Turner:  47%

[rwarn17588]

Imagine they'll post one of those 10-day old polls later in the week. Their polling has been pretty inaccurate in recent years anyway. At least, as it relates to contenders.

I'm calling B.S.

Inaccurate compared to what? In city races, Taylor outpolled LaFortune, Bartlett outpolled Medlock and Falling, and LaFortune outpolled Miller and Medlock. Guess who won?

Sure, the numbers shift around, but that's because of the undecided voters. The ultimate outcomes matched what the polls indicated.

The only poll I can recall the World published that was outright wrong was Carson holding a slim lead over Coburn in the Senate race. That one was close to the margin of error, and it was a polling firm that wasn't the usual one being used.

[Wrinkle]

I'm calling B.S.

Inaccurate compared to what? In city races, Taylor outpolled LaFortune, Bartlett outpolled Medlock and Falling, and LaFortune outpolled Miller and Medlock. Guess who won?

Sure, the numbers shift around, but that's because of the undecided voters. The ultimate outcomes matched what the polls indicated.

The only poll I can recall the World published that was outright wrong was Carson holding a slim lead over Coburn in the Senate race. That one was close to the margin of error, and it was a polling firm that wasn't the usual one being used.

Outpolling or winning are terms not necessarily associated with accuracy.
And, I qualified it as to contendors. I'll agree the winners were
often picked correctly, but the polling of others were way off.
For example, World put Medlock's poll at (working from memory only)
around 5-7%. He received about 18% of the vote. If half or less counted,
I'd be great at horse shoes.

And, why are polls not published upon conclusion, rather than waiting 10 days or more? As you know/claim, things can change a lot in that period.
I'd suggest it to depict what was desired.

[rwarn17588]

Outpolling or winning are terms not necessarily associated with accuracy.
And, I qualified it as to contendors. I'll agree the winners were
often picked correctly, but the polling of others were way off.
For example, World put Medlock's poll at (working from memory only)
around 5-7%. He received about 18% of the vote. If half or less counted,
I'd be great at horse shoes.

Again, I'm calling B.S.

Why would you think those polls are inaccurate? After all, I mentioned that there was a sizable percentage of undecided voters during the polls -- as there usually are even a day or two before an election. This is nothing new or unusual. What makes you think Medlock's take of 5-7% (as you claim) a week or two before an election -- with as many as 20-25 percent of the electorate undecided -- was inaccurate?

As for your "contenders" (whatever that means), it still doesn't change the fact that the poll before the primary showed Medlock had a huge hill to climb to even have a shot at winning the race. The results during the primary did nothing to refute that poll.

I think you're equating "inaccurate" to "I don't like what the poll is saying." Those are not the same things.

[MDepr2007]

My favorite "inaccurate" Tulsa World poll was the one they used a Wescott photo instead of a Medlock photo..

[Wrinkle]

Again, I'm calling B.S.

Why would you think those polls are inaccurate? After all, I mentioned that there was a sizable percentage of undecided voters during the polls -- as there usually are even a day or two before an election. This is nothing new or unusual. What makes you think Medlock's take of 5-7% (as you claim) a week or two before an election -- with as many as 20-25 percent of the electorate undecided -- was inaccurate?

As for your "contenders" (whatever that means), it still doesn't change the fact that the poll before the primary showed Medlock had a huge hill to climb to even have a shot at winning the race. The results during the primary did nothing to refute that poll.

I think you're equating "inaccurate" to "I don't like what the poll is saying." Those are not the same things.

I'm not suggesting the methodology of the polling firm was inaccurate. The context was. Waiting 10 days to post obsolete data two days before an election paints an inaccurate picture. And, was intended to do so, which makes it unethical, IMO.

iirc, on that particular poll, there were other oddities to go along with the weird results. Like a polling count of 483 "likely voters" when polls such as this are typically in round numbers like 400 or 500. And, the published margin of error was something like +/-5.35% when these are very normally more like +/-4.0%.

It was like data were grouped and targeted selectively.

[rwarn17588]

I'm not suggesting the methodology of the polling firm was inaccurate. The context was. Waiting 10 days to post obsolete data two days before an election paints an inaccurate picture. And, was intended to do so, which makes it unethical, IMO.

iirc, on that particular poll, there were other oddities to go along with the weird results. Like a polling count of 483 "likely voters" when polls such as this are typically in round numbers like 400 or 500. And, the published margin of error was something like +/-5.35% when these are very normally more like +/-4.0%.

It was like data were grouped and targeted selectively.

OK, so now you admit there's no real evidence the methodology was wrong or inaccurate.

There was no inaccurate picture of the poll, since it was perfectly clear when the poll was taken.

The number polled isn't always in round numbers. They're taken in numbers so that the proportions match the proportions of Republicans and Democrats in an area. And error rates vary according to how many people you poll.

And now you're suggesting that the report about the poll was a conspiracy to suppress the already woefully anemic Medlock vote. And guess what? It was woefully anemic. It's like you're mad that the poll was accurate.

Sheesh. Do you know how ridiculous you sound?

[Wrinkle]

OK, so now you admit there's no real evidence the methodology was wrong or inaccurate.

There was no inaccurate picture of the poll, since it was perfectly clear when the poll was taken.

The number polled isn't always in round numbers. They're taken in numbers so that the proportions match the proportions of Republicans and Democrats in an area. And error rates vary according to how many people you poll.

And now you're suggesting that the report about the poll was a conspiracy to suppress the already woefully anemic Medlock vote. And guess what? It was woefully anemic. It's like you're mad that the poll was accurate.

Sheesh. Do you know how ridiculous you sound?

No, the poll, as presented, was intended to present an inaccurate picture at the time of its posting. The poll itself, in terms of raw data and methodology was probably fine. IOW, it was irrelevant at the time, presented in a way to make it look relevant, and at a time when it could most affect voter impressions.

The World made mention of its intent to publish a Mayoral poll in this Sunday's edition (for the first time), two days pre-election. Let's see how that one presents.

[pmcalk]

Personally, I don't think polls should ever be reported right before an election.  There should be some sort of blackout period, maybe for a week before the election, or at least a few days.

Of course, the media is free to do what it wants.  I just think for the sake of fairness, they just shouldn't post them. 

[rwarn17588]

No, the poll, as presented, was intended to present an inaccurate picture at the time of its posting. The poll itself, in terms of raw data and methodology was probably fine. IOW, it was irrelevant at the time, presented in a way to make it look relevant, and at a time when it could most affect voter impressions.

The World made mention of its intent to publish a Mayoral poll in this Sunday's edition (for the first time), two days pre-election. Let's see how that one presents.

To recap, you now no longer claim the polls were "pretty inaccurate," as you did earlier.

Your main objection now is not based on fact, but an opinion that a poll that's a few days old and shows Bartlett with a humongous lead in the GOP race might sway voters' opinions, even though you've presented no evidence it ever did so.

This is what we call "moving the goalposts."

[Wrinkle]

To recap, you now no longer claim the polls were "pretty inaccurate," as you did earlier.

Your main objection now is not based on fact, but an opinion that a poll that's a few days old and shows Bartlett with a humongous lead in the GOP race might sway voters' opinions, even though you've presented no evidence it ever did so.

This is what we call "moving the goalposts."

My claims have not changed. But, your trying to put words in my mouth never ceases.

That particular poll looks to be manipulated to achieve a goal, even though the underlaying data probably remains valid. The sample and margin of error were facts they presented, not me. And, the timing of the posting was an issue.

So, it was indeed not just pretty inaccurate, it was wholly inaccurate for the timing. Context, but do also question the content as it appeared oddly formulated for presentation, not odd in original methodology.

But, I'm repeating myself and you didn't get it the first time.