Want proof that FOX, CNN and AP rig the Polls? Here it is.

[TruthIsAll]

Just compare their polling numbers against the other 10 polls.

That's why I took these three Bush whores out of my National
Poll Election Forecast Model. 

But we all knew they prop up Bush, didn't we? This just
confirms it.

Watch these muthas. Watch them close.

10-POLL AVERAGE
IBD,ABC, NWK, ARG,NBC,CBS,PEW,LAT,ZOGBY,TIME				
	Current Polling	TIA Projected 		
Month	Kerry	Bush 	Diff	Kerry	Bush 	Diff
Jan 	42.50	50.17	-11.5	46.48	53.52	-7.03
Feb	48.00	45.43	2.57	52.60	47.40	5.20
Mar	47.50	44.75	2.75	52.93	47.08	5.85
Apr	47.38	44.88	2.50	52.80	47.20	5.60
May	47.11	44.22	2.89	53.18	46.82	6.36
June	47.13	45.00	2.13	52.64	47.36	5.28
July	49.57	44.29	5.29	53.87	46.13	7.74
						
AVG	47.0	45.5		52.1	47.9	


FOX, AP,CNN	 (BUSH WHORES)			
	Current Polling	TIA Projected 		
Month	Kerry	Bush 	Diff	Kerry	Bush 	Diff
Jan 	38.7	51.3	-12.7	45.7	54.3	-8.7
Feb	45.5	46.5	-1.0	51.1	48.9	2.2
Mar	46.3	44.3	2.0	52.9	47.1	5.7
Apr	45.3	44.0	1.3	52.8	47.2	5.6
May	44.0	44.0	0.0	52.4	47.6	4.8
June	45.0	47.0	-2.0	50.6	49.4	1.2
July	47.0	47.0	0.0	51.2	48.8	2.4
						
AVG	44.5	46.3		50.9	49.1	

	

[troublemaker]

1. If you feel free to toss unwelcome data why bother with polls at all?

If I saw a RW person tossing three polls because they seem to favor Kerry I wouldn't take the results too seriously, so why is this different?

[Lefta Dissenter]

8. I like this approach

I've already decided (and declared) that I'm only going to believe the polls that I like the results of, so this works for me!

After the whole Iowa debacle, I guess I realized that my own opinions mean just as much as (if not more than) the polls.

[TruthIsAll]

14. I don't toss UNWELCOME data! I just showed what we all knew anyway..

Come back to the Milky Way.

[TruthIsAll]

81. IMPORTANT! HERE ARE THE 13 MONTHLY POLL NUMBERS.

I should have placed the data in my original post, rather than
just comparing the 3 bad guys vs. the other 10.

Source: Pollingreport.com

.....Kerry Bush Spread

IBD			
Feb	44	41	3
Mar	45	43	2
Apr	40	44	-4
May	43	42	1
June	43	44	-1
July	49	44	5
			
			
ABC			
Feb	52	43	9
Mar	53	44	9
Apr	48	49	-1
May	49	47	2
June	53	45	8
			
			
			
AP.....WTF! NEVER HAD KERRY AHEAD OF BUSH!		
Jan 	37	54	-17
Mar	45	46	-1
Apr	44	45	-1
May	43	46	-3
July	45	49	-4
			
			
NWK			
Jan 	41	52	-11
Feb	50	45	5
Mar	48	45	3
Apr	50	43	7
May	46	45	1
July	51	45	6
			
			
ARG			
Jan 	47	46	1
Feb	48	46	2
Mar	50	43	7
Apr	50	44	6
May	47	44	3
June	48	46	2
July	49	45	4
			
			
NBC			
Jan 	35	54	-19
Mar	45	47	-2
May	42	46	-4
June	47	47	0
July	54	43	11
			
			
FOX....WTF! NEVER HAD KERRY AHEAD OF BUSH!		

Jan 	32	54	-22
Feb	43	47	-4
Mar	44	44	0
Apr	42	43	-1
May	42	42	0
June	42	48	-6
			
			
CBS			
Jan 	48	43	5
Feb	47	46	1
Mar	43	46	-3
Apr	48	43	5
May	49	41	8
June	45	44	1
July	49	44	5
			
			
CNN/Gallup.... They won't let Kerry get too far ahead. Oh,
and they said Bush was ahead by 15 points in NC a few days
ago. They wanted to convince us that Edwards is having a
NEGATIVE effect in his OWN state. Oh, the Mason-Dixon poll
has Bush ahead by 3 in NC)		

Jan 	43	55	-12
Feb	48	49	-1
Mar	52	44	8

Apr	46	51	-5 < WTF happened here? Kerry goes from UP 8 to
DOWN 5 in ONE month? Who did Kerry kill in April?

May	49	47	2
June	48	49	-1
July	50	46	4
Aug   na   na	na	 <<<< and guess what will happen
here?
			
PEW			
Jan 	41	52	-11
Feb	47	47	0
Mar	48	44	4
Apr	47	46	1
May	50	45	5
June	46	48	-2
			
			
LA Times			
Apr	49	46	3
June	51	44	7
			
			
			
ZOGBY			
Mar	48	46	2
Apr	47	44	3
May	47	42	5
June	44	42	2
July	46	44	2
			
			
TIME			
Jan	43	54	-11
Feb	48	50	-2
May	51	46	5
July	49	45	4
…………………………………………………………………………………………………….

tia

[TruthIsAll]

83. Feb-July Average Monthly Poll Numbers: Kerry vs. Bush

I did not include January. Bush was ahead by 12 points. Kerry
was not yet the candidate.

I report. You decide.
Are CNN, AP, FOX to be trusted?

FEB-JULY MONTHLY POLLING AVERAGES
Poll	Kerry	Bush 	Diff

CNN   48.00	48.71	-0.71 < Gallup poll. Most volatile.
AP	44.25	46.50	-2.25 < No Kerry lead in ANY month
FOX	42.60	44.80	-2.20 < No Kerry lead in ANY month

AVG	44.95	46.67	-1.72 < Bush average lead: 1.72%
------------------------------------------------------------

IBD	44.00	43.00	1.00
ABC	51.00	45.60	5.40
NWK	49.00	44.60	4.40
ARG	48.67	44.67	4.00
NBC	47.00	45.75	1.25

CBS	46.83	44.00	2.83 
PEW	47.60	46.00	1.60
LAT	50.00	45.00	5.00
ZOGBY	46.40	43.60	2.80 < Most accurate in 2000
TIME	49.33	47.00	2.33

AVG	47.98	44.92	3.06 < Kerry average lead: 3.06%
....................................................................
	

Total		
AVG	47.28	45.33	1.96 < Kerry average lead: 1.96%(13 polls) 

[oldhat]

2. That's amazing.

You've manage to prove absolutely nothing other than prove there are three sets of data you personally don't like.

This means nothing. They don't "rig" the polls.

[RummyTheDummy]

6. Right on

And they're Fair and Balanced too!

[oldhat]

10. Take a statistics class.

[RummyTheDummy]

15. What?

You don't thinkg they're Fair and Balanced?

[TruthIsAll]

19. Oldhat, how many math degrees do you have?

I have three:
BA Mathematics
MS Applied Mathematics
MS Operations Research

I took that statistics course. Did you?

[HEFFA]

20. The real truth isn't even in the numbers; it's in the polling METHOD used

I don't trust any poll that I can't get background info. about. Hence, I rarely trust polls. There is abundant evidence to show that different methods yield results that are more or less trustworthy. Some methods have been discredited altogether.

The bigger question here is: why doesn't any news outlet inform you about the methods, the differences, the nuances of polling? After all, isn't their responsibility to inform the voter rather than to mislead or underinform?

Moral of the story: ALL media are WHORES!!

[Feanorcurufinwe]

40. "Take a statistics class." Is that an attempt to be clever?

If so, you failed.

[TruthIsAll]

12. They don't, huh? Like they never LIED about Kerry and Gore?

And like they told you the truth when they said that Bush got more votes than Gore in Florida?

What makes you think they wouldn't rig the polls? Especially when they (especially FOX) take their marching orders from Rove.

Did you see F9/11? You are OUTFOXED, pal.

Get your head out of the sand, buddy.

[oldhat]

13. Methodology

Read the methodology appendix with each poll report.

Do you know what methodology is? Have you taken a statistics class?

[TruthIsAll]

16. Did you read it? Did Murdoch tell you to?

Bush is TOAST.

Get over it.

[Imperialism Inc.]

18. I don't understand your point.

I'm sure the methodology differs, but that is the point. Whatever methodology they are using is resulting in consistently higher numbers for Bush. Or did you think he meant they were just manually fudging the numbers?

[TruthIsAll]

24. Oh, right. Their methodology is different. Its called BIASED.

God help us all.

[StandUpGuy]

26. Methodology horse shit

Statistics class more horse shit.

1st. they lie about the methodology.
2nd. they poll less than 0.000001% of the population
3rd. they get $7 hr telemarketers to make the calls
4th. they hand them lists of pre ordained phone numbers to call.

Its that simple.

Yes in theory your Methodology and statistics classes are valid.
In practice they are horse shit.
Sorry.

[TruthIsAll]

38. Oldhat, want some more of this? Come back, wherever you are.

B
U
S
H

I
S

T
O
A
S
T

[BillZBubb]

21. "They don't rig the polls" Wow, I almost couldn't stop laughing!

It is very easy to rig polls. It's trivial in fact. And it happens. Skewing results to a preconceived answer is no problem.

Also, rigging has nothing to do with the statistics. It has to do with methodology, presentation, and interpretation.

[TruthIsAll]

49. No, they don't rig polls. Only elections. I feel faint.

I've got to lie down. The insightful logic is truly revelatory.

I am not worthy.

[burythehatchet]

30. That's right oldhat

the networks that tia referenced are objective and independent media outlets. They serve the critical function of truthful observation of facts. A function that makes our democracy work so darn well....they....oops I gotta go, a monkey just flew out of my ass.

[TruthIsAll]

41. ROFLMAO and I can't get up.

Monkeys, no you mean Bushies, out of your ass!

[taxidriver]

3. ...they show kerry winning, mostly?

at least the final numbers do.

[TruthIsAll]

5. Look at the CURRENT poll columns. The projected numbers are mine.

tia

[Mz Pip]

4. Must be hard

for FOX to even put in writing that +2.4 for Kerry. I can just see Hannity turning purple having to even look at it. :-)

MzPip
:dem:

[Raster]

9. don't kid yourself... these guys are having to spend all of their ...

...time defending gee-dubya and $hrubco. They would MUCH prefer to attack. It's more their nature. Besides, that's what's bringing in the bux.

[TruthIsAll]

27. WANT MORE PROOF? LOOK AT THE ACTUAL POLL NUMBERS!

NATIONAL POLLS			


.....Kerry Bush Spread

IBD			
Feb	44	41	3
Mar	45	43	2
Apr	40	44	-4
May	43	42	1
June	43	44	-1
July	49	44	5
			
			
ABC			
Feb	52	43	9
Mar	53	44	9
Apr	48	49	-1
May	49	47	2
June	53	45	8
			
			
			
AP.....WTF! NEVER HAD KERRY AHEAD OF BUSH!		
Jan 	37	54	-17
Mar	45	46	-1
Apr	44	45	-1
May	43	46	-3
July	45	49	-4
			
			
NWK			
Jan 	41	52	-11
Feb	50	45	5
Mar	48	45	3
Apr	50	43	7
May	46	45	1
July	51	45	6
			
			
ARG			
Jan 	47	46	1
Feb	48	46	2
Mar	50	43	7
Apr	50	44	6
May	47	44	3
June	48	46	2
July	49	45	4
			
			
NBC			
Jan 	35	54	-19
Mar	45	47	-2
May	42	46	-4
June	47	47	0
July	54	43	11
			
			
FOX....WTF! NEVER HAD KERRY AHEAD OF BUSH!		

Jan 	32	54	-22
Feb	43	47	-4
Mar	44	44	0
Apr	42	43	-1
May	42	42	0
June	42	48	-6
			
			
CBS			
Jan 	48	43	5
Feb	47	46	1
Mar	43	46	-3
Apr	48	43	5
May	49	41	8
June	45	44	1
July	49	44	5
			
			
CNN/Gallup.... They won't let Kerry get too far ahead. Oh, and
they said Bush was ahead by 15 points in NC a few days ago.
They wanted to convince us that Edwards is having a NEGATIVE
effect in his OWN state. Oh, the Mason-Dixon poll has Bush
ahead by 3 in NC)		

Jan 	43	55	-12
Feb	48	49	-1
Mar	52	44	8

Apr	46	51	-5 < WTF happened here? Kerry goes from UP 8 to
DOWN 5 in ONE month? Who did Kerry kill in April?

May	49	47	2
June	48	49	-1
July	50	46	4
Aug   na   na	na	 <<<< and guess what will happen
here?
			
PEW			
Jan 	41	52	-11
Feb	47	47	0
Mar	48	44	4
Apr	47	46	1
May	50	45	5
June	46	48	-2
			
			
LA Times			
Apr	49	46	3
June	51	44	7
			
			
			
ZOGBY			
Mar	48	46	2
Apr	47	44	3
May	47	42	5
June	44	42	2
July	46	44	2
			
			
TIME			
Jan	43	54	-11
Feb	48	50	-2
May	51	46	5
July	49	45	4
…………………………………………………………………………………………………….			

[Cessna Invesco Palin]

7. Huh?

Is this just a summary? You've not taken margins of error into account. You can't just add them up that way.

[TruthIsAll]

17. Oh, I see we have a mathematician here. Margins of Error, hmm..

Margin this.

[Feanorcurufinwe]

43. Unnecessarily rude comment.

[Cessna Invesco Palin]

84. Deleted.

Not worth the trouble of getting into it again.

[olddem43]

11. I think everyone here agrees that Fox

is slanted to the right in their reporting. Their polls would be unlikely to be fair also. Your numbers bear that out.

[xocolatl]

66. The more interesting conclusion is that AP is as bad as FOX

See analysis below in #53 and #57.

[stellanoir]

22. This may be a silly observation but I'm just wondering. . .

since so many people have replaced their land lines with cellular phones since the last elections, their numbers may not be on many lists and they're generally not in phone books. Could there be a rather huge demographic flying under the radar of even the most methodical of polls? Hope so because none of these polls reflect what I hear. I only know 3 Republicans who intend to vote for *. Granted I live in a deeply blue state but my town is very conservative and predominantly Republican, I talk with people from all over the country with some regularity as well.

Has anyone ever been polled on their cell phone?

[xocolatl]

23. Not a statistically significant difference

Sorry, no significant difference between the two groups. P-value is 0.356; and this is not even taking into account the fact that the monthly numbers are autocorrelated. If you were to take the AC into account the P-value would be even higher.

I don't like Fox much, but you should get your statistical methodology right.

[TruthIsAll]

28. Check the individual poll numbers in post #27, Mr. P-statistic.

I guess I shouldn't believe my lying eyes.

Just a coincidence. NO statistical significance. Right.

You're full of Chebyshev, my Gaussian friend.

[xocolatl]

37. Naah, too much data entry...

... I shouldn't be such a smart-ass anyway. Can't help myself tho, I do this kind of thing for a living.

If you want to exclude FOX from your models, be my guest. They suck anyway.

:) :toast: :)

[TruthIsAll]

39. Just copy and paste into Excel. You do this for a living, right?

All your statistical functions are right there in Excel.

Do your thing. Knock yourself out.

I'll be right here.

[xocolatl]

42. Yah, that's why I object to doing it in my spare time...

... but OK, I'll bite. Of course, the data is never in the format you want it, so it'll take me a while. I'll post the results when I'm done.

[xocolatl]

53. More careful analysis

OK, so it turns out that your claim appears to have merit, given a more careful examination of the raw data. The P-value turns out to be 0.002.

Using a crossed-effects model (with random month and news-outlet effects), and a single fixed effect "Gp" (=1 for FOX, AP, and CNN, =0 otherwise):

Solution for Fixed Effects

Standard
Effect Estimate Error DF t Value Pr > |t|

Intercept 1.4360 1.8789 6 0.76 0.4737
Gp -4.8344 1.4882 52 -3.25 0.0020


Covariance Parameter Estimates

Cov Parm Estimate
News 2.1626
Mon 20.8829
Residual 16.7650


More interesting are the empirical BLPs for the random effects:


Std Err
Effect News Mon Estimate Pred DF t Value Pr > |t|

News ABC 1.0692 1.1872 52 0.90 0.3719
News AP -0.6102 1.2565 52 -0.49 0.6293
News ARG 1.0133 1.1223 52 0.90 0.3708
News CBS 0.8099 1.1223 52 0.72 0.4737
News CNN 1.2737 1.2298 52 1.04 0.3051
News FOX -0.6635 1.2416 52 -0.53 0.5954
News IBD -0.8934 1.1523 52 -0.78 0.4416
News LA 0.6055 1.3281 52 0.46 0.6503
News NBC -1.5324 1.1886 52 -1.29 0.2030
News NWK 0.2028 1.1541 52 0.18 0.8612
News PEW -0.5912 1.1532 52 -0.51 0.6104
News TIME -0.5247 1.2299 52 -0.43 0.6714
News ZOGBY -0.1591 1.1871 52 -0.13 0.8939
Mon Apr 0.8177 2.0586 52 0.40 0.6929
Mon Feb 0.8211 2.1102 52 0.39 0.6988
Mon Jan -9.6705 2.1118 52 -4.58 <.0001
Mon July 3.4853 2.1116 52 1.65 0.1049
Mon June 0.4052 2.0846 52 0.19 0.8466
Mon Mar 2.3544 2.0561 52 1.15 0.2574
Mon May 1.7869 2.0356 52 0.88 0.3841


I can send SAS code and raw data to your DU email if you wish.

[TruthIsAll]

68. Thanks, but no thanks. I prefer an Excel spreadsheet.

The SAS presentation is too cryptic. I realize it's a system you're familiar with, and you wanted to respond quickly.

I must admit, I don't follow the presentation. This is due to my ignorance of SAS and long absence from using the P-statistic.

As I said in the previous post, you should look at the relevant Excel stat functions and present your analysis in layman's terms.

In any case, I am impressed with your rapid response. Oftentimes, whenever I have posted Excel-based probability analysis (generally using the Cum. Normal Distribution or the Poisson function) critics were quick to respond, but offered no analysis of their own.

Perhaps, that was the reason I was a little snippy with you at first, thinking you were like the others - quick to criticize, yet also obviously lacking in mathematical training. These individuals, who are still around, thought they could snow me and others with philosophical arguments and thus avoid the nitty-gritty work of actually creating an alternative model to solve the problem.

[xocolatl]

70. Actually I despise SAS and prefer Excel

but SAS is the only program I know that does crossed-effects models.

Anyway, thanks for collecting the data. It was interesting to analyze.

You might find it interesting to look at the table at the bottom of both posts (inconsistently labelled "Empirical BLPs" or "Random Effects Solution"). The "News" estimates give you relative bias of each organization (within Gp). For example, in the second analysis, ABC and ARG have the largest estimates, indicating the most favorable towards Kerry. Among the non-AP/FOX group, NBC and CNN are the least favorable to Kerry, although NBC appeaers worse than CNN.

There are some hidden assumptions in this model that are probably untrue, but without the sample size for each poll, nothin' can be done about it.

Your timeline was interesting, BTW.

[dsc]

72. As one of the critics

I pointed out at the time it wasn't my job to do your homework. If you are going to state what others said do it accurately. Is that too much to ask? Of you it evidently is.

On edit, It should be noted that one of the points I, and several other critics made, was that NO HONEST ANALYSIS COULD BE DONE. Hence it was totally idiotic and or dishonest TO ASK FOR ONE. Again, if you are going to attribute words to people DO SO HONESTLY.

[TruthIsAll]

77. A revealing, sad comment; it confirms your total lack of understanding

"On edit, It should be noted that one of the points I, and several other critics made, was that NO HONEST ANALYSIS COULD BE DONE. Hence it was totally idiotic and or dishonest TO ASK FOR ONE. Again, if you are going to attribute words to people DO SO HONESTLY."

DSC, are you accusing my analysis, and that of others, as being dishonest? You have criticized my honest efforts to analyze problems for over two years. And you never present a valid argument for your side.

With your post, you have finally overstepped the line of decency. It must be due to desperation on your part.

What the hell is your problem?

[dsc]

80. My problem is people who misrepresent what I say

which you did. Note the post I responded to MENTIONED PAST BEHAVIOR AND NOT THIS THREAD. And WITH THE ONE ANALYSIS OF YOURS I CRITICISED IN THE PAST, the problem was that YOU WERE TREATING AS INDEPENDENT THINGS WHICH WERE DEPENDENT. Thus an honest analysis is impossible. STOP MISREPRESENTING WHAT I SAID AND I WON'T CRITICISE YOU FOR HAVING DONE SO.

[TruthIsAll]

71. My mistake was in not showing the individual poll numbers in the

original post. Instead I focused on the two groups, knowing that CNN, FOX and AP were the bad guys from my long familiarity with the numbers.

I could have better presented my argument if I had provided the detailed monthly polls at the start, rather than at post #27.

Looking at the actual numbers, its pretty obvious what is going on.

[xocolatl]

57. I do think you're a bit hard on CNN...

The random effect EBLP for CNN is rather positive. Therefore, I did another analysis where Gp = 1 only for AP and FOX, and = 0 for everybody else (including CNN).

Results below. Note that the P value is even smaller (0.0004).



Covariance Parameter
Estimates

Cov Parm Estimate

News 1.3016
Mon 20.8486
Residual 16.8125


Solution for Fixed Effects

Standard
Effect Estimate Error DF t Value Pr > |t|

Intercept 1.1999 1.8401 6 0.65 0.5385
Gp -6.1503 1.6160 52 -3.81 0.0004


Solution for Random Effects

Std Err
Effect News Mon Estimate Pred DF t Value Pr > |t|

News ABC 0.8213 0.9878 52 0.83 0.4095
News AP -0.00486 1.0546 52 -0.00 0.9963
News ARG 0.8335 0.9458 52 0.88 0.3822
News CBS 0.6829 0.9458 52 0.72 0.4735
News CNN -0.6728 0.9458 52 -0.71 0.4801
News FOX 0.004858 1.0546 52 0.00 0.9963
News IBD -0.5749 0.9656 52 -0.60 0.5542
News LA 0.4231 1.0695 52 0.40 0.6940
News NBC -1.0215 0.9885 52 -1.03 0.3062
News NWK 0.2208 0.9665 52 0.23 0.8202
News PEW -0.3600 0.9661 52 -0.37 0.7109
News TIME -0.3027 1.0135 52 -0.30 0.7664
News ZOGBY -0.04967 0.9877 52 -0.05 0.9601
Mon Apr 0.9128 2.0569 52 0.44 0.6590
Mon Feb 0.7850 2.1084 52 0.37 0.7112
Mon Jan -9.6752 2.1097 52 -4.59 <.0001
Mon July 3.3896 2.1094 52 1.61 0.1141
Mon June 0.3761 2.0823 52 0.18 0.8574
Mon Mar 2.3904 2.0552 52 1.16 0.2501

[TruthIsAll]

60. Nice work. But you should try to explain the methodology in English.

You obviously know your statistics. I haven't opened a stat book in thirty years. But I know probability theory.

And as for bias in the numbers, I think it is fairly obvious to the unbiased human eye.

1) FOX has never showed Kerry leading.
2) AP has never showed Kerry leading.
3) CNN's March to April switch from Kerry (+8) to Bush (+5) in a month in which Richard Clarke testified before the 9/11 committee, exposing Bush as being totally oblivious to terror threats. And also in a month when the Iraq torture scandal broke.

Look at the numbers. Look at the events.

Is it clear? Crystal.

[xocolatl]

65. In English

I assumed you knew enough stats to read my post, given your background.

Essentially, from my analysis I would conlcude that FOX and AP were biased (either through selection bias or simply fudging the numbers), but CNN is equivocal.

Re: your other comments. Basically, if you're going to argue averages (or other summary stats) then you need to consider random variability.

If you're going to argue based on a time-line, as you do in post #27, then that's another issue entirely.

I still don't believe CNN is as biased as you say. And no I'm not on their payroll or have any unnatural preference for them.

Finally, I'm picking up a bit of hostility from you (tho I could be wrong about that). I have no animosity to your conclusion. It's just think that either you needed to present the time-line up front or support the averages with statistical reasoning.

[StandUpGuy]

29. Grrrr.

My Pvalue (p'd off) is about 10.0 every time I hear people try and defend polls.

[xocolatl]

58. Just trying to be intellectually honest here

... as long as we're criticizing others for dishonesty.

[TruthIsAll]

32. xocolatl? Atlanta? CNN? Naw, can't be.

...

[xocolatl]

36. Huh?

[Beetwasher]

47. Statistical Significance and P-Value is Irrelevant

Now before you get on your statistical high horse, I know all about why it normally would be relevant, but what you're not taking into account is as that as far as public perception is concerned, P-values are irrelevant. For the public and from a PR, marketing and propoganda perxpective it only matters that Bush appears to be doing better than he is, and that would only take a minor adjustment to the data, one in which the p-value would appear to be insignificant, but the effect on perception is not.

[xocolatl]

50. Not true, P value is relevant here

If somebody is going to make a claim that they have proven so-and-so reports biased numbers, then their statistics better support the claim.

Observed differences could be due simply to chance variation.

Public perception is an entirely different matter.

[Beetwasher]

51. No, Public Perception Is THE Matter

TIA may be engaging in a bit of hyperbole by claiming his work is proof, but the point is made.

It would only take a slight tweaking of the data, one that would result in statistically insignificant changes, yet have profound effects as far as public perception is concerned, and that's ALL that would be needed and that's ALL that matters as far as the reasons for doing it are concerned. You would expect some chance variations, but not all in the same direction. Furthermore, the observations TIA made are not the only one's to be made. There's additional evidence that some of the polling orgs are not behaving objectively. The new Gallup NC poll for instance. I wrote about some others here:

http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=104&topic_id=534560

The bottom line is, this type of thing could of course be all due to chance. But it keeps happening and it ALWAYS happens in favor of Bush. If it were chance, we'd see some weird variations that favor Kerry, and that NEVER happens.

[xocolatl]

54. Calm down

I'm no fan of FOX news.

I just want to see statistical evidence when somebody says that they have data that supports a claim.

As it turns out, TIA is right when a more careful statistical analysis is done. See my post above.

[Beetwasher]

55. Calm Down? LOL!

Was I yelling? Did I say you were a fan of Fox? OK, whatever..:shrug:

I just saw your post and appreciate your honesty and your additional analysis. Statistical evidence is a good thing, there's no denying that, but it's not the only evidence.

[xocolatl]

56. Agreed... Statistics aren't everything

:pals:

[patcox2]

82. Never mind, on edit, I mistook the meaning nt

nt

[leyton]

25. Yes, there are three networks with higher-than average numbers for Bush.

In any set of data other than a set of identical numbers, some data will be above average. Just because in this case it's Fox, CNN, and AP proves nothing.

[DemBones DemBones]

63. Well, if you've watched CNN and FOX, and heard

them shill for Bush all during the 2000 campaign and since, and then you see how their polling data differs from all other polling data (except AP) in giving Bush* higher numbers, it certainly suggests their data is suspicious. Unless, of course, you believe in coincidence theory.

[Robin Hood]

31. Good job, great re-search.

I can't believe some of the idiots on this board who would stick up for CNN and FOX. The freaking Saud family owns CNN, and FOX news is a state sanctioned news outlet at this point. BOth of these fake news outlets would make Pravda proud.

[Elidor]

33. Contrary to what some have said

TruthIsAll is not throwing away poll data he doesn't like, but highlighting polls with a strong shift in their data. It's a perfectly valid observation. The interpretation of those facts is somewhat open to question, but the facts themselves are plain for all to see.

[TruthIsAll]

34. You, my friend, are the voice of reason. And TRUTH.

Its sooooo easy to spot those BushCo lurkers.

They all come out of the woodwork as soon as some facts are posted that they don't want the world to know about.

I can spot them faster than you can say "Jackie Robinson".

[spanone]

35. They should call them what they are : 'Spins'

[Beetwasher]

44. Many People Are In Denial About This

They can believe that the Repubs can buy a news station and influence the news, but somehow, a polling org. is immune from their influence? It would make sense that they would TRY to influence the polls if they could wouldn't it? Why is it so hard to believe that they did in fact buy themselves a few of the very best polling outfits. It would fit their M.O. and there is certainly evidence to suggest bias. I don't think we need Sherlock Holmes here...

Now I know all the arguments to the contrary about the differences not being statistically significant, and that's technically true, however, it's irrelevant. What the statisticians here might not realize, because they're statisticians and not experts at PR, marketing and propoganda, is that as far as public perception is concerned it doesn't matter if p>.05 or not, it only matters that Chimpy looks as if he's doing better than he is, and that only takes a slight shift in data. Perhaps a little tweaking of the sample or the weighting, whatever. You know the innumerable ways data can be manipulated if you're unscrupulous.

BTW, thanks, as always TIA for all your work!

[AZCat]

45. Thanks TruthIsAll, but...

My problem with this sort of thing is this: while it is fine and good for those of us already aware of a sickness at the center of this administration to discuss the ways that the media props up the Great Pretender, how can we disperse this to the general public? I know that with my circle of pre-converts, I am miles from addressing this topic. Heck, I haven't even broached PNAC with half of them!

While I haven't taken a close look at your numbers, I don't have any reason to dispute them, or your conclusions about the bias. But can we do anything with this? Do you think the other media organizations that don't skew their polls are interested in a story about this? Should we be emailing journalists, or is this something that they won't write about as a professional courtesy?

[TruthIsAll]

62. OK, send these numbers to CNN, FOX, AP...all over.

Send them to your friends and family. To the newspapers.

Anyone can relate to these discrepancies. They are obvious.
The bias is clear. No fancy statistical analysis is necessary.

We KNOW CNN and FOX are biased. Here is tangible, circumstantial evidence.

[dsc]

46. Since these are within the margin of error

and we don't have the real humbers to compare to, we have no idea if they are skewed toward Bush due to Bush actually doing better or skewed toward Bush due to some other reason. I will give you that it appears they are skewed a bit toward Bush but they would argue that they are accurately reflecting the situation.

[TruthIsAll]

48. dsc, I suggest you take a look at post # 27. Just open your eyes.

Or do you still believe in coincidences?

Kerry has been ahead in the polls since Feb. That's 5 months.
And Fox and AP don't show him ahead in any one month. Does this sound right to you?

And wghat about CNN showing Kerry up 8 ponts in March, then down 5 points in April. That's a 13 point switch. Come on.

[dsc]

59. ABC had a ten point switch in the same time frame

You also don't note when in each month these polls are taken. If they aren't the same time each month the time between polls can be anywhere from 1 to 8 weeks. 13 points in 8 weeks isn't inordinate. Again, I will admit that there appears to be a skewing but without the knowledge of where Kerry really is we have no idea whose polls are right.

[Lerkfish]

52. ALL POLLS ARE INHERENTLY FLAWED....

There are three areas where the best conducted polls are inherently flawed: Sample, methodology, and conclusions.

SAMPLE: When looking at a poll, pay close attention to the SIZE of the sample. For example, a poll of 500 people is not large enough to predict the voting of 200 million people...its statistically insignificant. The smaller the size, the more likely is the sample population unrepresentative. Although even large samples can be artificially skewed (see methodology below).
Also pay attention to the MAKEUP of the sample. Good polls provide a demographic breakdown so you have some idea whether its representative. However, even a sample that appears to represent the demographics of the larger whole, can still be artificially skewed by such factors as is discussed in methodology below.

METHODOLOGY: there are many things in this category, all of them important factors in how polls can be skewed:
-- WILLINGNESS TO BE POLLED. Consider: if you are a republican, and the Democratic Party calls you for your opinion, would you stay on the line? If Fox news calls, someone who watches fox would be more likely to sit still for a poll than one who does not. That automatically skews the sample to those who are first of all predisposed to participate in polls in general, and secondly inclined to participate in a particular poll.
-- ACCESS TO POLLSTERS. Many polls are conducted via telephone. That automatically culls the sample by work shift, access to a phone, whether your listed or not, etc. Also, the methodology of where those phone numbers were obtained is important. For example, is it a subscriber list, a random phone dialing, etc. If the poll were about how people felt about being out of work, and its a telephone interview, how accurate could that sample be if most unemployed people have no phone or even a home?
-- QUESTIONING. This one of the first things I'll mention that has the worst abuse. How a question is asked, and what questions get to be asked is subjective on the part of the pollster and can really reflect a bias in the organization funding the poll. A relatively benign example was a poll that was conducted last year that pointed out that teenagers could not locate certain countries on a map. It made the kids look really stupid until you noticed who sponsered the poll : National Geographic.
Here's an example of how the apparent same question can elicit different responses, depending on how it is asked:
------1. Do you think Bush is doing a good job? yes or no.
------2. Do you think Bush is doing a horrible job? yes or no.
------3. Do you think Bush is performing well as commander in chief? yes or no.
------4. Do you thing Bush is a good president?
------5. Would you prefer to have Bush as president?
------6. Is Bush the best choice for president?
you see, even though this seems to be asking the same question, the same person might not answer the same for #2 as they would for #6.
(there is actually more to this part, but you get the idea)

CONCLUSIONS: Even the most honorably conducted poll can have someone analyze and spin the data in such a way to fit an agenda. Here is an example:
a recent poll of journalists found that, in their opinion, they consider themselves:
---65% Centrist or moderate
---25% Liberal
---10% Conservative
Now, someone who thinks there is a liberal conspiracy will state that there are more than twice as many liberals than Conservatives (which is accurate, as far as it goes, though it neglects to say "of those expressing a side"), OR you could go so far as to say that only 10% of newsrooms are Conservative, and that proves 90% of them are liberal biased.
(this, btw, is a real life example of an argument made by a republican on another board. No matter how much I pointed out that well over half the respondents claimed no affiliation either way, he kept insisting 90% were liberal)
Another real life example is that there was a study to show the "fitness" of teachers to be teaching. Most of the parameters made sense, like their level of education, etc. But one parameter was "does the teacher belong to a union or teacher's organization?". If they didn't , that was one parameter AGAINST them. The study was sponsored by a teacher's union. Their data was accurate, but they were biased in how they analyzed or collated the data, because the union question held as much weight in their 5 point criteria and whether the teacher had an advanced teaching degree.

at any rate, at BEST polls are a rough indicator, but they are never conclusively representative of the whole, only of the limited sample. They should always be taken with a healthy dose of salt, even when they tell you what you want to know...probably ESPECIALLY when they tell you what you want to know. :)

as far as the Fox polls in this thread, I think the sample is automatically skewed by the factors of willingness. Fox doesn't even have to falsely conduct or analyze the data, all that has to happen is that only those people willing to be polled are going to more likely be fox viewers, which are therefore more likely to be right wing.

[West Coast Democrat]

61. The proof was in the 2000 Election...

Nearly all of the polls were wrong. The media tried to make it look like Bush would win easily.

[DemBones DemBones]

64. Here's what Democrats.com had to say

in today's newsletter:


__Kerry Pulls Ahead of Bush So Fast that CNN Shaves Points with Bogus 'Weighting'

When Bush had, a while back, inched ahead of Kerry by one measly point, the mainstream media's headline blared "Bush surges ahead!" Now look at the headline on this CNN story reporting that Kerry has pulled ahead by FOUR points!: "Poll Finds Apparent Kerry Lead in Close Contest." More comical still, this poll was WEIGHTED (for the first time!) so that Kerry's lead was cut down. The actual poll result was 51% to 44%. What a hoot! Who do these media minions figure they are fooling at this point?

http://www.cnn.com/2004/ALLPOLITICS/07/12/presidential/index.html


Of course, the media minions think they're fooling all of us. No doubt they are fooling many.

[moondust]

73. They did the same to Gore.

Strong negative bias against Gore, strong positive bias for Bush*. I'm sure it helped create an election that was close enough to steal.

[ItsThePeopleStupid]

67. something you may be interested in - Professor Pollkatz

Professor Pollkatz did something like this, and Fox came out the most positive for *. On another page, he says he used to believe some polls were fixed but doesn't anymore. Whether Fox is fixing it or not, they are consistently more pro-Bush, so you could take that into account or just throw out the highest and lowest (like when you take a measurement in a science class).

Prof. Pollkatz:
WHICH POLLSTERS LIKE BUSH THE BEST?
pollster average difference from Ipsos-Reid spread (since 9/11/01) t-ratio (t > 2.0 or t < -2.0 means "significantly different from Ipsos")
FOX 6.933 5.093
NBCWSJ 5.267 2.830
GALLUP 4.118 3.362
PEW 4.053 2.252
ABCPOST 3.085 1.880
CNNTIME 2.230 1.049
CBS 1.944 0.679
HARRIS 1.794 1.121
NEWSWEEK 1.171 0.612
TIPP 0.561 0.221
IPSOS-REID 0.000 na
NYT (1.271) (0.403)
ARG (3.093) (1.847)
IBD (6.174) (5.644)
ZOGBY (7.483) (5.383)
RANGE (between highest and lowest) 14.416
http://www.pollkatz.homestead.com/files/bushindexprobushtable.htm

[TruthIsAll]

69. Thanks for the Pollkatz analysis. Seems to confirm my KISS analysis.

FOX and CNN/GALLUP are bad news. But I don't see AP here.

Zogby is the best, as I have always stated.


tia

[xocolatl]

75. Seems consistent with what I just posted...

...here.

Except according to my analysis, Zogby is smack in the middle with ABC being most favorable to Kerry; but the differences are within the margin-of-error.

But clearly, FOX and AP are consistently unfavorable to Kerry.

[TruthIsAll]

76. Zogby has the BEST track record of all the pollsters.

Glad we agree about FOX and AP.

[Phelan]

74. I've done plenty of data resarch in my day

SPSS and I are old friends, its a love hate relationship. it loves to take up space on my harddrive and I hate to use it...but it does get you some good numbers. But one thing that I have learned using SPSS and just following polls is that the saying there are 'lies, damn lies and statistics' since I can pretty much use one data set to 'prove' anything you want me to.

I wouldn't throw out those three polls, I just wouldn't. Seeing Bush having good numbers makes me work harder on getting more Dems and friendlies to vote. If the Bushies don't think Bush has anything to worry about they may not show up come the 2nd. and damn would that make my day

[TruthIsAll]

79. I am not throwing out the polls, just not using them in my database.

I will not however, use what I believe to be biased numbers which would diminish the accuracy of my forecast model.

It's like statistical sampling, where the statistician wants to determine a representative subset of the universe for his anaysis.

A better analogy would be a financial forecast, for which we need accurate financial statements, not cooked numbers.

Likewise, I want a representative sample universe of polls, which form the basis of my forecast.

[KansDem]

78. I don't understand ANY of the above...

...but it still looks good to me! :D

[TruthIsAll]

85. This is an important thread.

TIA