Ron Baiman explains USCV position and how it relates to Febble's work

[eomer]

I get the impression that some people think that Febble's function refutes or invalidates the latest USCV paper. Quite the contrary, USCV was fully aware of Febble's function and what it means before they published this paper.

Here is a recent statement by Ron Baimain:

The point of the latter discussion is to show that UScountvotes has been fully aware of, and has incorported, Liddle's insight into our analysis. It makes our "implausibility of rBr" claim stronger.

Also in the same thread on mysterypollster.com, Ron gives a good summary of what the paper is about:

Dear Mystery Pollster Readers:

I would you to review the most recent study at: www.USCountvotes.org in detail - especially the Appendices.

Our basic point, has been, and continues to be that the E-M data as reported are not consistent with constant mean uniform "bias" across partisan precincts.

In our latest study we have demonstrated this in three different ways. We show that:

a) overall exit response rates, that would be required in “representative” high Bush and high Kerry precincts to generate the within precinct errors, are mathematically infeasible.
b) output simulation of individual precincts shows that Bush voter exit poll reluctance would have to change to by at least 40% across partisan precinct categories to generate the outcomes reported by EM.
c) input simulation of partisan exit poll response rates randomized around mean values of .56 for Kerry voters and .5 for Bush votes – which E-M claims can explain all of the within precinct error in the exit poll samples (see p. 31 of their report) – shows in multiple runs of over 10,000 simulations that mean and median WPE and overall response rates, especially in high Bush and Kerry precincts, cannot be generated.

On the other hand, our analysis shows that EM’s data IS consistent with a non-uniform response bias that would require further explanation, a uniform bias and vote shifting, or vote shifting. The patterns of mean and median WPE and overall response in high Bush and high Kerry precincts in particular cry out for further investigation.

-snip-

These two excerpts can be found toward the end of this page:
http://www.mysterypollster.com/main/2005/05/aapor_exit_poll.html#comments

Ron gives a lot of explanation about the paper, what it means, how it relates to Febble's work and other critiques. Look for Ron's two posts, May 19 at 1:33 PM and May 19 at 5:00 PM.


Edit: add date/time of Ron's posts.

[OnTheOtherHand]

1. it's debatable

As a participant in USCountVotes discussions, I disagree with most of what Ron has to say here. (But it's certainly true that Febble's function, by itself, doesn't refute the USCV paper, which does incorporate it.) The MysteryPollster thread does air the arguments pretty well. I hope to write a detailed explanation of why I disagree with the USCV paper.

Short explanation: as I understand it, Ron thinks that the WPEs are "just too low" on the high-Kerry end and "just too high" on the high-Bush end to be consistent with (in his words above) a constant mean uniform "bias." If he is right about that, then I think the regression line through bias should slope up with partisanship -- but actually it is flat within statistical slop.

After that, the argument gets more complicated, but I don't think he wins it.

But it's worth emphasizing that this isn't a debate about whether there was massive fraud in the 2004 election -- or, at most, it is one small part of that debate.

[LightningFlash]

2. According to that very paper the mean...

Spikes, it does NOT remain flat, and that is the overall point of this analysis.

[eomer]

3. Linear regression masks what is happening at the ends of the ranges

Whatever slope a regression line has cannot tell you what is happening at the ends, especially when the middle has such a larger percentage of the data. The ends are overwhelmed by the middle.

Here is Ron's more technical explanation of what I just said:

The precinct sample was heavily weighted toward more
competitive precincts (1110 observations) with much smaller 40 and 90
samples for the partisan precincts (roughly 10%). The correlation line
will of course mostly reflect what's going on with 90% of the observations
in the competitive precincts, where based on aggregate calcutions, Alpha
(see Appendix C Table 1 or our recent report) has a more or less symmetric
hump shape - and thus a flat linear correlation.

Thus running a flat linear correlation line through a scatter plot of
Ln(Alpha) does not show invariant Alpha. In the most partisan districts in
particular Alpha needs to vary by at least 33% based on (most favorable to
hypothesis) simulations and by about 54% based on aggregate analysis.

http://www.mysterypollster.com/main/2005/05/aapor_exit_poll.html#comments

You are not addressing Ron's actual argument because you continue to talk about what a regression line looks like. Ron's argument clearly has to do with a more detailed look at the shape that cannot be seen through the lens of a regression line. What we need to discuss is his actual argument, which is about the underlying shape. Instead you keep converting the shape into a line, which totally masks the shape we need to be looking at, and then you want to discuss the slope of the line. Ron is not making a claim about the slope of the line - he is making a claim about the more complex shape of the data, which is clearly not linear.

This linear approach that you keep coming back to is starting to look like a strawman. Are you willing to discuss Ron's actual point or are you going to continue to switch the conversation to the slope of a line?

[OnTheOtherHand]

4. not a straw man, actually -- bear with me

It's true, as we've discussed, that a single line doesn't describe the shape of the data.

But Ron's comments above miss something important. Regression lines are actually much more heavily influenced by points that are far from the center of gravity than by points near the center. Therefore, if a bunch of high-Bush precincts have too-high WPEs, and a bunch of high-Kerry precincts have too-low WPEs -- which I think is the point of Baiman's part of the USCV paper -- the regression line really ought to tilt positive.

When I looked at O'Dell's simulation, that's what actually happened -- the regression line had significant positive slope. The same thing was true of the Dopp simulation in the AAPOR version of the paper, although I don't know what it does right now -- I haven't figured out the new modeling of vote shift yet.

I'm pretty sure O'Dell agrees with me about this. Coincidentally, I was talking with him a little while ago, and he offered to post here.

Now, here's the thing. I'm pretty sure that Baiman's response rate arguments are undermined by the actual data. And I don't know about you, but when I look at the scatterplots, I don't see fraud concentrated in Bush strongholds, which was one theory. I do believe the Bush-stronghold theory would entail a positive slope (at least if the fraud were nearly enough to swing the election). But there are lots of other theories that wouldn't entail a positive slope. I've spent most of three weeks now (mostly not on DU) arguing against one pretty weak theory, but it doesn't mean that I'm against all the other theories.

If you see people insisting that folks like me and Febble assume or insist on a linear model, that really is a straw man. I don't think that any line or curve we fit through those points is going to tell the whole story.

[TruthIsAll]

5. OnTheOtherHand, you can try to fit an obfuscation curve through this...

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x369374

Despite your commendable, gentlemanly prose, like all the other naysays and obscurantists, you too shall fail in your attempts to promote the desperate, faith-based fiction that is rBr.

[OnTheOtherHand]

6. thanks for the "gentlemanly" part, TIA!

It's not an accident that "rBr" didn't appear in my posts. (I think I used it once yesterday, but not to promote it.) I'm talking about whether the discrepancies between exit poll results and official results were concentrated in one or two parts of the partisanship spectrum, or not. I conclude that if there was fraud, it wasn't focused in "Bush strongholds."

I followed your link, although I didn't read the whole thread. I distrust touch screens as much as you do. Maybe more. It seems to me that if I were going to hack a touch screen, I wouldn't let the voters see what I was doing. I'm not questioning your facts (although I haven't checked them, either), but it seems to me that they could be the tip of the iceberg.

I'm reasonably convinced that an election could be hacked, and that makes me spitting mad. You too, right?

I don't see any contradiction between your arguments and mine, so I'm not sure what your point is. You're more sure than I am that the election was stolen, but I'm not here to argue that it wasn't. I'm trying to sort out one little corner of what we know and don't know. It's what I do. Sometimes it's helpful, sometimes it isn't, but it shouldn't "obscure" anything from anyone who is paying attention. (And if people aren't paying attention, it probably doesn't matter what we say anyway.)

[TruthIsAll]

58. Your Bush/Kerry partisanship focus is a red herring...

"It's not an accident that "rBr" didn't appear in my posts. (I think I used it once yesterday, but not to promote it.) I'm talking about whether the discrepancies between exit poll results and official results were concentrated in one or two parts of the partisanship spectrum, or not. I conclude that if there was fraud, it wasn't focused in "Bush strongholds."

So that's why you're here, to tell us that if there was fraud, it was pervasive among BLUE and RED partisanships?

Where's the beef? Instead of focusing on the fraud itself, you are creating a straw man, attempting through subtle language to find flaws in USCV's analysis where none exist.

You ignore the physical evidence.
You ignore the exit poll evidence.
You ignore the media lockout of a meaningful fraud discussion.
You ignore the fact that Mitofsky refused to meet with Conyers.
You ignore the fact that the media is complicit by not authorizing Mitofsky to release the data.
You ignore the fact that rBr was created by Mitofsky in 1992 for another Bush in the NH primary.
You ignore the fact that Febble has been used as a dupe.
You ignore the fact that rBr is a contrivance which flies in the face of experience and common sense.
You ignore the fact that science is anathema to RW faith-based fundies who have stolen the House, Senate and WH and who yearn for the rapture.

You ignore the fact that National Exit Poll weightings are mathematically impossible.

You ignore everything that is important.
You are just another prop in a long line of distractions.

You should be posting at the MP site. You will have no shortage of sheep to fleece there.

It would be a lot better than the fleecing YOU are experiencing here.

So take the wool out of your eyes.
Stop insulting us.

[tommcintyre]

59. "just another prop in a long line of distractions". GREAT SUMMARY TIA! n/t

[eomer]

7. Still waiting for a thoughtful argument.

OTOH, you said:

I'm pretty sure that Baiman's response rate arguments are undermined by the actual data.

I'm not sure what this could mean. What Ron is doing is to simulate various response rates to see what they produce and then compare this simulated data with the actual data. His finding was that the uniform rBr response rate hypothesis produces simulated data that does not match the actual data. Thus the rBr hypothesis, by itself, cannot explain the actual data.

How could that argument be undermined by the actual data? The actual data is one of the pillars of the argument.

I am trying to bear with you but it would be helpful if you would stop throwing out careless aspersions while you try to figure out a thoughtful one.


Edit: clarify who the reply is directed to

[OnTheOtherHand]

8. hmm

First of all, as elsewhere, it would be helpful to know whether you found anything else in my post that made sense. It's a bit frustrating to get hammered on one sentence out of context.

Again, my understanding of Baiman's response rate argument is that the WPEs in high-Kerry precincts are too low, and the WPEs in high-Bush precincts too high, to be explained by a constant mean bias. I believe this argument is refuted by the flat regression line for the reasons explained in my preceding post, to which I invite you to respond.

[eomer]

9. You are misstating Ron's argument.

Ron's argument is that the shape of the actual data is not reproduced if you run a simulation with rBr (constant mean bias) as the assumption.

In the first graph on page 7 of the current version of the USCV paper (revision date of 5/17), there is a plot of the shape of the actual data and the shape produced by simulating rBr.

rBr produces a U shape. The actual data produces a shape that is not at all like the U shape of rBr.

The shape of the actual data is the standard that he is measuring against.

[OnTheOtherHand]

11. reproducing shapes

Frankly, the shape of the "Vote Shift Simulation with Bias" (second graph on p. 8) seems even further off than the shape of the "Simulated Response Bias" on p. 7, especially if you consider numbers of precincts.

The one on p. 7 (without vote shift) is off by 0.06 or 6% for mean WPE in the high-Bush precincts, at left. It's off by about 0.01 or 1% for mean WPE in the high-Kerry precincts at right.

The (latter, because I assume this one is supposed to be better) one on p. 8 is still off by almost 0.04 or 4% for mean WPE in the high-Bush precincts, and it is off by close to 0.06 or 6% for mean WPE in the high-Kerry precincts. And bear in mind that there are more than twice as many high-Kerry precincts as high-Bush precincts.

Honestly, I'm still staring at the spreadsheet and trying to figure out where, exactly, these numbers come from in the first place. But while it may be obvious to Baiman that the p. 8 model fits better than the p. 7 model, I don't see it.

[eomer]

10. As far as whether I found anything else in your post #4 that made sense,

You said:

But Ron's comments above miss something important. Regression lines are actually much more heavily influenced by points that are far from the center of gravity than by points near the center. Therefore, if a bunch of high-Bush precincts have too-high WPEs, and a bunch of high-Kerry precincts have too-low WPEs -- which I think is the point of Baiman's part of the USCV paper -- the regression line really ought to tilt positive.

No, Ron was not missing that point. What he said was that data that produces a regression line that has roughly no slope can still have a non-linear shape. You have agreed that this is true. In fact this statement is proved by the actual data. It has a regression line with roughly no slope and it has a non-linear shape.

Ron's work then was to find what assumptions will reproduce as closely as possible that non-linear shape.

Presumably when you get simulated data that approximates the same shape then the regression line of that simulated data will also be roughly no slope. But whether it does or not turns out to be moot for Ron's rBr simulation for reasons I explain below.

Ron has run various simulations and found which ones came close to the shape of the actual data and which ones didn't.

Conspicuously, the rBr simulation did not produce a shape that was similar to the shape of the actual data. This is Ron's main point according to his MP posts. The point is that he rules out rBr as a possible explanation because it does not reproduce the shape.

Once we see that the shape is not reproduced then we don't need to see what the slope of a regression line is because that simulation has already failed.

I don't see where Ron has shown us the slope of the regression line of any of his simulations and, if you are particularly interested in that, then maybe you should ask him over on mystery pollster what the regression line slopes were for his various simulations.

But, whatever the slope of the regression line is for the rBr simulation, even if it is zero, can't rescue that hypothesis from being ruled out because its shape is already enough to disqualify it.

As far as the rest of the things you say in post #4, please do me a favor: please point to a particular claim by USCV in the paper or by Ron in his posts. I can't tell which claim you intend to refute.

[OnTheOtherHand]

12. most of this

I think I've basically answered above. Maybe the fit is substantially better in the midrange, but not much. The fact that the simulation on p. 8 goes a little bit up from category 1 to 2, and then a little bit down from 2 to 3, for me isn't enough to prove very much. You have to imagine error bars around these results.

I'm still a little puzzled by the shape of that second p. 8 simulation line. Actually, it seems to posit a flat (or maybe downward-sloping?) regression line even for WPE -- i.e., lower errors toward the high-Bush precincts -- which is not the result I thought Baiman was looking for.

Most of my post 4 was trying to make sense out of Appendix F on pages 19-20, which I admit I've never really understood (and I swear I have tried), so I am open to instruction there. Or maybe we should leave Appendix F out of it for now. But what I take from Table 1 is that the errors should be lower in high-Kerry precincts (smaller alphas) and higher in high-Bush precincts (much larger alphas). The second simulation on p. 8 -- the one that supposedly shows the telltale zig-zag from category 1 to 2 to 3 -- indicates _virtually no difference_ between the high-Bush and high-Kerry precincts. Even accounting for the distinction between WPE and alpha, that isn't going to fly (which is why, in fact, the simulation line is so very far from the E-M results). So this simulation seems to be at war with Table 1 on page 19. If Ron is responsible for both, I hope he will tell us which one to believe.

[eomer]

13. What Ron points to as his main conclusion

is that a constant mean uniform bias hypothesis cannot be the explanation.

Mitofsky posted on MP a couple of hours ago saying he didn't intend the statement in his paper about rBr to advocate a constant mean bias as the explanation and that in fact there is a great deal of variation in the bias.

Mitofsky and Baiman apparently agree on what Baiman has been describing as his main point. Maybe that will allow the discussion to move forward.

[OnTheOtherHand]

14. but not very far

No, I don't think they agree on the main point, but let's backtrack and compare notes. Here is a snip from Mitofsky:

We told you that it does not say there was a constant response rate, as you insist on claiming we said. It gave an example of a hypothetical response rate in order to show how little it would take to create the within precinct error we cited. We clearly told you there was a huge variance in response rates. That it was not constant.

And indeed, Mitofsky apparently showed a slide that shows response rates all over the place. If the Kerry and Bush response rates varied separately (as, for instance, the USCV simulations assume), then bias likewise will be all over the place. If that were the main point of the USCV paper, well, I don't see why there would be a paper.

But Ron's central conclusion has been, as you I think accurately say, that a constant mean bias can't be the explanation. (I'm not quite sure what adding the word "uniform" adds.) If that claim has any bite, it should imply that the bias must be higher in some parts of the partisanship distribution and lower in others.

* Table 1 in Appendix F implies that it is highest in high-Bush precincts and lowest in high-Kerry precincts.

* The first simulation on p. 8 agrees, but depicts a much smaller range of bias than the observed means.

* The second simulation on p. 8 implies that bias is highest in the high-Bush precincts and mid-range, lowest in the the not-quite-so-high Bush precincts (60-80%), and varies very little across the whole range. (Actually, if we transformed the WPEs depicted in this graph to the bias index, it may indicate that bias is highest in the high-_Kerry_ precincts, but I haven't tried that.)

Asserting agreement that response rates aren't constant doesn't seem to cover the waterfront here.

[eomer]

16. You don't think they agree on Ron's main point?

Repeating Ron's quote:

Our basic point, has been, and continues to be that the E-M data as reported are not consistent with constant mean uniform "bias" across partisan precincts.

That is what Ron says his main point is. Mitofsky agrees with that.

Right?

[OnTheOtherHand]

18. sorry, it's not the same point

You don't see a big difference between "constant bias" and "constant mean bias"? It's the difference between saying, on the one hand, that every precinct has the same response rate or at least exactly the same ratio of B/K response rates, and on the other, that the bias doesn't vary systematically as a function of partisanship.

It's sort of like the difference between asserting that a baseball team scores the same number of runs in every game, and asserting that it averages the same number of runs throughout the season (rather than scoring more in September). It's a considerable difference.

As I pointed out, all of Liddle's, O'Dell's, and Dopp's simulations (at least all the ones I can think of) assume that bias varies from precinct to precinct. I don't think anyone ever contested that assumption, because everyone assumed that it would be true in real life. (People could disagree about how the bias varies, but no one denied the variance.) The issue is whether there is any pattern to the bias related to partisanship. Hence the talk about Bush strongholds and Kerry strongholds and vote shift signatures.

[eomer]

19. Well, then, what is your understanding of what Mitofsky is saying?

Mitofsky is the only one using the term "constant response rate". Clearly USCV knew that the response rate varied from one precinct to another (see Appendix G). Not sure why Mitofsky is popping up to state that if that's all he's saying.

I took him to be saying that there was variation not just from one precinct to another but rather variation across partisanship. Since you have a connection that is at most one level removed from him (through Febble) maybe you can ask him to clarify the point that actually could have used clarifying, not the one that everyone obviously knew.

Which is it that you understand him to be saying? That there is in fact a constant mean that carries across partisanship? And then the variation that is seen is just random variation centered on that mean? Or is it something else?

[Febble]

20. OK - can you point to

where Mitofsky uses the term "constant response rate"? The E-M report certainly proposes a hypothetical pair of response rates (56% and 50%), which I certainly took to mean average, not constant. It clearly isn't constant - in fact the striking thing about the data is the variability, which the E-M report draws attention to.

In fact, I'm not sure that anyone has used the term "constant response rate", although I've seen "constant mean differential response rate", and may have used that one myself. (If I ever left out the "mean" it was definitely a typo!)

Statistics is all about variance, and whether something varies when another thing varies. When it does, you've got a correlation, like age and height in children. It's rarely exact, but it's often close.

If a correlation is "significant", it means that a slope is a better fit to the data than a flat line through the middle (the average). If the correlation is not significant, it means that a flat line through the middle is no worse a fit than the slope.

A non-significant correlation doesn't mean that there's no variability. And it doesn't mean there is not a non-linear relationship. Non-linear relationships are often important, as Ron says.

[eomer]

21. Here is where Mitofsky used it:

Ron:
We read the statement below to you in person in Miami. We told you that it does not say there was a constant response rate, as you insist on claiming we said. It gave an example of a hypothetical response rate in order to show how little it would take to create the within precinct error we cited. We clearly told you there was a huge variance in response rates. That it was not constant.

Below is the original statement from page 31 of our January 19 report. Try reading it slowly and stop when you get to the word "hypothetical" in order for its meaning to penetrate.

"While we cannot measure the completion rate by Democratic and Republican voters, hypothetical completion rates of 56% among Kerry voters and 50% among Bush voters overall would account for the entire Within Precinct Error that we observed in 2004."

Clearly, completion rates were not constant. If they were constant then the bias would not vary from precinct to precinct. And it does vary -- quite a bit. All you have to do is look at the slides Mystery Pollster posted in his article above.

Posted by: Warren Mitofsky | May 20, 2005 04:00 PM
http://www.mysterypollster.com/main/2005/05/aapor_exit_poll.html#comments

Now it's your turn. Where does Baiman say that Mitofsky said there was a "constant response rate". I believe Baiman has been saying that Mitofsky said there was a "constant mean response rate".

And thanks for the lesson on what the terms mean. Believe it or not I know what all those terms mean. I'll return the favor some time and give you an insulting little lesson that implies you don't.

Apparently you didn't understand my words so let me try rewording it for you. Does Mitofsky say there is a correlation between partisanship of the precint and response rate or not?


Edit: provide link for Mitofsky comment

[Febble]

22. I'm sorry, I really did not mean to insult.

I have been accused of obfuscating, so I was trying to be clear. Also it is a public forum, and I thought I'd better not assume knowledge.

I thought you thought Mitofsky said it. I realise now you were referring to his response to Ron on MP. I presume Mitofsky was responding to Ron's comment:

While I'm at it, here's a note that specifially explains why Mitofsky's scatter plots and linear correlations do not support a non-varying rBr hypothesis, and a note on the orgins of "w" and "alpha" analysis.

But in answer to your question, now I understand it:

Mitofsky presented a plot showing no linear correlation between partisanship of the precinct and bias.

[eomer]

23. Thanks.

Are you confounding bias and response rate? Not necessarily the same thing.

The question was whether there is a correlation between partisanship of the precinct and response rate.

Edit to add that I see that Mitofsky was just misconstruing Baiman's words. Anyone who has looked at the USCV report or has read Baiman's posts more widely would know he didn't mean non-varying in the sense of constant rate across all precincts.

[OnTheOtherHand]

24. yes, I think that's right

I have no idea what Baiman said in Miami, but in all likelihood Mitofsky misunderstood him as saying that Mitofsky had argued for actually constant rate across all precincts. Which I think we all agree is silly. Sorry if I seemed to be defending part of Mitofsky's comment that I was actually ignoring! I still think there is a huge gap between Mitofsky's position and Baiman's as I understand them, but not that huge.

Febble may comment on the correlations, but as I understand it, Mitofsky presented evidence at AAPOR that neither bias nor response rate is correlated with partisanship. (Again, there could be a significant non-linear relationship, at least in principle.) Mitofsky further argued that the observed response rates did not vary significantly from the expectation of lower overall response rates in high-Bush precincts. The response rates seem so noisy that I'm not sure we can say much more one way or another about them.

[Febble]

25. No, but I thought the bias was the plot you were

talking about.

I'm clearly getting my wires crossed here.

I THINK what Ron and Mitofsky were sparring about on MP was actually "differential response rates", which can be measured variously, but in the instance of the plot I thought they were discussing, by my bias function, aka "ln(alpha)", or "WPE_bias", or even, in my paper "Bias index".

Mitofsky also showed refusal rates plotted against vote-count margin (or percentage of Bush responses if you like). Again, the variance was very great, and there was no significant correlation.

[eomer]

27. I think we are back where we were before Mitofsky made that comment on MP

A hypothesis of rBr based on constant mean response rate.

The USCV paper says that simulations based on constant mean response rate could not reproduce the WPE means, WPE medians and overall response rates.

For a successful simulation (with those same criteria) they had to assume mean response rates that varied significantly by partisanship.

[Febble]

28. You might have meant to respond to OTOH

but as you commented on my comment, I'll just say that you might like to look at my response to Ron on MP.

Essentially, as far as I understand it, Ron's model assumes that overall response rates have a tight relationship with differential response rates.

On this assumption, I'm sure his conclusion is correct (I haven't read the latest version of paper, but I trust his algebra).

In my model, I let response rates be a free parameter.

[Bill Bored]

15. Fraud in Bush strongholds??? Are we forgetting that

in the data, there were only 40 Bush strongholds?

That's only 3.2% of the 1,250 sample precincts.

How much fraud would it take to account for the exit poll discrepancy of about 6%?

Based on only the Bush strongholds, almost 100%!

In other words, every vote for Kerry in these precincts would have to have been switched to Bush, which would actually have converted them from Kerry strongholds to Bush strongholds! And that assumes there were NO actual votes for Bush in these "strongholds" to begin with (pre fraud). But the Mean WPE in these 40 precincts was -10%, the Median WPE was -5.8% and the Absolute Mean WPE was 12.4%. That's a lot less than 100% WPE, isn't it?

If we are hypothesizing that fraud only took place in Bush strongholds, they had better represent more than 3.2% of the total precincts. Otherwise, the fraud would not have been anywhere near enough to account for the exit poll discrepancy, would it?

So why the fixation on Bush strongholds?

[TruthIsAll]

26. Bill, you are part right, part wrong.

How do you define a RED Bush stronghold?
Well, its rural and conservative, at least 60% of the vote to Bush.
The great majority of rural precincts are in Bush strongholds.
That's where they did the popular vote padding.
There are plenty of RED precincts in BLUE states.

Not to say they didn't swipe in BLUE precincts.
Ballot spoilage and voting hassles gave them a head start there.
That's how they stole OH and FL.

Look at FL.
Optiscan counties were 57-42 Bush.
BBV counties (Palm Beach, Dade, Broward, etc.) were 52-47 Kerry.

[Bill Bored]

29. TIA, I am going by Mitofsky's and USCV's definition of "Bush stronghold"

If you look at the precinct distribution, there were only 40 out of 1250 precincts that they considered to be strong Bush. They define this as 0.00<=Kerry<.20, or in other words, at least an 80% Bush vote.:puke:
For strong Kerry, it's Kerry>=.80, which turned out to be 90 precincts.

Our friend Febble has commented that the 80% is largely an arbitrary classification, and I think we all probably agree on that.

Personally, I'd like to see the 100 strongest precincts for each candidate and do a 2-tailed paired t-test on the WPE in those to calculate p-values. But then I'm not a statistician -- what do I know?

But if USCV is saying that fraud is only evident in Bush strongholds, as defined above, that can't be a significant amount of fraud, since there weren't a significant number of these precincts. So why bother?

I agree we need to think in terms of fraud either in more general terms when in comes to the popular vote, or more specifically from state to state.

[tommcintyre]

17. Thanks for posting this eomer. It helps to undo some of the "spin"...

that the pro-rBr cabal has been pushing around here.

"It makes our "implausibility of rBr" claim stronger."
That's a pretty clear message, isn't it?

Also, according to Kathy Dopp, Febble obfuscated, and even "stole"? ("derived") much of the work:

http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=203&topic_id=369091#369140
"We offered to work with you on our last paper until we could all agree, and you declined. You rushed to publish your derivations of a few new formulas independently, along with your own (in my opinion) unjustified conclusions.

I've shown how simple it is to derive the three formulas that you derived on one page in our work, in Appendix C and D, since you did not reveal how you derived them yourself, and generally obfuscated the math and the issues in your own paper. Everyone else who contributed derivations of new formulas contributed them to the group and wrote up the steps to derive them so we could all easily understand, but not you.

If anyone wants to UNDERSTAND the formulas you are using clearly, they can go to our paper, which shows how easily they were derived from all our previous work.

You also obfuscated by changing all the notation we'd began with and complicating it, by calling precincts defined by percentage of votes for Kerry or Bush, Democratic or Republican precincts, etc."
-----------------

I'll tell ya, the more I look into this, the more it STINKS TO HIGH HEAVEN!

Again, thanks for "clearing the air" some more.

[tommcintyre]

30. A quick summary of what I've seen here, AND A CHALLENGE

Eomer posted a very clear (un-obfuscated) statement from a known (and fully qualified) professional.
"It makes our "implausibility of rBr" claim stronger."

It was immediately "jumped on" by someone with unknown qualifications who has made UNSUBSTIANTED CLAIMS to have been a "participant in USCountVotes discussions", and (IMO), has a CLEAR "agenda". The immediate thrust was "it's debatable", and "I disagree with most of what Ron has to say here." <I believe this is a pattern we have seen many times in the last four+ years - it's called "creating doubt". This alone was enough to "raise my antennae"; but I carefully read on to see what arguments would follow.

Basically, what I saw was a pattern of CLEAR questions being directed to the "implausibility of rBr" doubters (to explain their pro-rBr claims); and vague (obfuscated answers). Statements such as the following became necessary:

"Still waiting for a thoughtful argument."
<snip>
"I am trying to bear with you but it would be helpful if you would stop throwing out careless aspersions while you try to figure out a thoughtful one."

This is only one example of such an exchange. I believe if you take the time to read through the posts, you will also see this clear pattern of: sincere questions asked, followed by vague (obfuscated) answers.

Of course this pattern is NOT new, and (sadly) it should be expected.

The real problem is, the pro rBr cabal doesn't have a solid "leg to stand on".

The "argument" (scam) is based on a conveniently concocted "cover story" by Mitofsky to explain away (cover up) once more a clearly STOLEN ELECTION.

And, unfortunately, since "like begets like", we have even seen a <admitted> pattern of deceit in this forum to "aid and abet" this cover-up.

Before any of the pro-rBr "apologists" object to what I have posted here; I once again ask you to to confront (and deal with) the challenge I have posed to you repeatedly (so far, they have assiduously avoided it ;) ):

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x371523
"Mitofsky invented reluctant Bush responder scam for Bush Sr. in '92!"
<snip>
"Actually, it turns out rBr was invented FOR THE FATHER! (George Herbert Walker Bush). <Thanks for catching this mods.>

So, it gets even MORE amazing! Voter's are not only prouder to vote for candidates BOTH to the right AND left of Bushs' running for president; but, apparently, THE TRAIT IS ALSO INHERITED!

Hey, pro-rBr "number crunchers" <or other "apologists"> - WHAT EXACTLY IS THE PROBABILITY OF THAT HAPPENING? CAN "LIGHTENING" STRIKE ONLY TWICE, AND ONLY TO TWO SUCCEEDING GENERATIONS?"

I sincerely want to know HOW you can explain this (with a straight face) how you can continue to maintain that Mitofsky's "excuse" has ANY SHRED OF CREDIBILITY in light of these facts.

This IS a huge "elephant in your pro-rBr room". It's NOT going away, so you better deal with it!

[OnTheOtherHand]

31. oh, there you are

I'm a college professor. I assume you know my name, actually, but maybe not. How do you propose that I substantiate my involvement with USCV?

If you want to talk about coverups, I can document how USCV's 5/12 working paper showed that there should be a positive correlation between partisanship and bias. And then, after Mitofsky showed that there wasn't, USCV went back to the drawing board and came up with another model that probably doesn't have a significant positive correlation (I'm not sure). Unfortunately, it doesn't fit the results very well either -- just vaguely zig-zags a bit. Is that a coverup? Like the 2004 exit polls, if it is, then it was really poorly done. I just think it was bad work, and that's why I'm criticizing it.

If we're arguing that the 2004 exit polls prove fraud, I think it's a huge problem to start by arguing that Mitofsky has been covering up fraud for years. Then everything that points to fraud is evidence of fraud, and everything that points away from fraud is evidence of the conspiracy. If I tell you I participated in USCV discussions, then I was probably sent by... Glaxo? Rove? right.

[eomer]

32. A USCV cover up, you allege?

What I see is that USCV has continuously published their work fully sourced and has even provided the simulator in a form that you can download and run. Totally open.

Because USCV's work has been totally open, peer review has been rigorous and enthusiastic (as opposed to the type of peer review Mitofsky has used, which is to pick one or two reviewers you already know agree with your position).

As a result of the openness and the real peer review, USCV has been able to respond to any critiques by refining their approach and methods. The newly refined version was then published in the same, fully open manner.

Apparently in your world, if a researcher receives criticism, acknowledges that it was fair and then responds to it by refining their approach, this is called a cover up. I call it professionalism.

[OnTheOtherHand]

33. try me again

As I said, I don't think it was a coverup, I think it was bad work.

Unfortunately, I can't yet characterize it as professionalism or total openness, either, because as far as I know, there is no public explanation anywhere of why the models in the 5/12 paper were replaced with the models in the 5/17 paper. I'm not sure I know the answer myself.

But I hope my post indicated that once we start imputing conspiratorial motives, we can portray honest attempts to do better work as malign attempts to cover our tracks. This is IMHO not a road that DUers should go down.

[eomer]

34. Still obsessing on an alleged flaw in a prior iteration.

On May 17, shortly after the latest USCV paper was released, Febble began a dishonest attack on USCV. Rather than read their latest work and give an honest critique of what is right and what is wrong, Febble started hammering with an argument that applied only to their prior iteration, not the current one. Kathy Dopp (sunshinekathy) asked Febble several times to read the latest paper and give feedback on it.

Here is the thread with that exchange between Febble and sunshinekathy:
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x369091#369112

You arrived not too long after and since then you have remained focused on that same alleged flaw in a previous iteration.

Neither you nor Febble made it clear in your posts that your criticism was not of their latest work. You left the impression that your criticism applied to what they were currently saying. That's why I started this thread.

You've suggested that we shouldn't start imputing conspiratorial motives so I won't even mention that possibility here. ;-) I'll just say that maybe it is time for you and Febble to focus on what USCV is saying now rather than continuing to try to nail them on an alleged flaw in a previous version.

[Febble]

36. To clarify:

Mine was not a dishonest attack - I simply had not read the latest update. I think the paper was updated a number of times over a short space of time, and I had not realised the version Kathy had sent me the day before was not the current version - I had printed it out, and understood that it was the version presented at the conference.

I have not yet digested the latest version, but will try to do so.

[eomer]

37. On May 19 you started a thread:

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x370789

In the OP of that thread you point out a flaw in "USCV's hypothesis that "Bush strongholds have more vote corruption". You neglect to point out that this was a previous position taken by USCV but not part of their current position.

You and OnTheOtherHand have kept the focus on issues related to the previous version and neither of you have ever pointed out that you are taking issue with a prior position of USCV and that it is not part of their current position. I don't agree that this is an honest approach.

[Febble]

38. That thread

was an attempt to give the history of the debate, in chronological order.

Here is what I said:

I wanted to know who won the 2004 US election.

I developed a better tool with which to find out.

It led me to believe that USCV's hypothesis that "Bush strongholds have more vote corruption" was flawed, not because they made an error, but because the information on which it was based, in the Edison-Mitofsky report, was flawed.

It was flawed because E-M used a poor measure of bias (the WPE). I proposed a better measure. Like the WPE, it does not distinguish between bias in the count and bias in the poll. Unlike the WPE it is not confounded with the actual vote-count margin.

The thread was an attempt to give a brief account of the development of the discussion that has been going on since the original USCV paper came out (on which I was even a signatory), between various members of the USCV team, including myself, and an explanation of why I produced my paper. It's a debate that is continuing, fruitfully, right now.

I'm sorry I didn't specify the original Mitteldorf et al USCV paper, and clearly I should have done, although I specified the hypothesis that my paper was addressing, which was the one presented in that paper. It was that hypothesis that my paper addressed. Not that it was wrong, but that we could not infer it because of the problem with the WPE. And my paper, of course, pre-dated the later Baiman, Dopp, Velleman and Sheehan series of working papers. The thread was certainly not a dishonest attempt to misrepresent the working paper. It was simply an attempot to put the debate in context and to try to make it clear that I don't believe the re-analyses done with my variable need be interpreted as an argument against fraud. And I'm sorry that the paper I was referring to wasn't clear from the context.

I realise that this debate has become a wrangle about who said what when, and I'm really sorry that that has happened.

But for the record, now, yes, I believe that the current working paper by does not propose "Bsmvcc". I am not take issue with the current version of the working paper. And I absolutely applaud their recommendations, without reservation.

I am not at war with USCV! Far from it.

[tommcintyre]

44. So YOU agree! "It makes our "implausibility of rBr" claim stronger."

YOU said in the post above:
"I am not take issue with the current version of the working paper. And I absolutely applaud their recommendations, without reservation."

Wonderful Febble! It takes a really BIG PERSON to admit when they've made a mistake (without a hidden agenda, that is).

So, now that you've "come around", and got your senses back, will you please help us with this one?

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x371523
"Mitofsky invented reluctant Bush responder scam for Bush Sr. in '92!"
<snip>
"Actually, it turns out rBr was invented FOR THE FATHER! (George Herbert Walker Bush). <Thanks for catching this mods.>

So, it gets even MORE amazing! Voter's are not only prouder to vote for candidates BOTH to the right AND left of Bushs' running for president; but, apparently, THE TRAIT IS ALSO INHERITED!

Hey, pro-rBr "number crunchers" <or other "apologists"> - WHAT EXACTLY IS THE PROBABILITY OF THAT HAPPENING? CAN "LIGHTENING" STRIKE ONLY TWICE, AND ONLY TO TWO SUCCEEDING GENERATIONS?"

<Part of this post:>
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x371175#371617
--------------

I'm looking forward to your help on this one. I'm sure your boss (Mitofsky) won't mind, since HE doesn't have a hidden agenda (or vested interest), and only wants the truth to come out, right?

Thanks in advance for your help.

[Febble]

48. Here are USCV's recommendations

From their most recent paper:

Recommendations

Few procedures are more important to the strength and health of U.S. democracy than credible and transparent, fair and free, elections. We appeal to our fellow citizens, the media, our courts, and our elected officials, to implement:

· a thorough non-partisan investigation of the 2004 presidential election
· full funding of the National Election Data Archive precinct level database
· election equipment that permits access by non-specialist citizen election judges to recount voter verified paper ballots
· routine 3%, randomly selected, independent audits of all elections
· transparent and publicly accessible exit polling
· election administration by non-partisan public civil servants
· non-proprietary open-source coding for all computerized election equipment
· no wired or wireless network connections to any vote casting or counting equipment

Vote counts in America need to be routinely and independently audited. It is not enough to require voter verified paper records of ballots. These paper records must be easily and "independently" auditable by persons other than the voting machine vendor, preferably without having to hire computer technicians, paper roll advancers, bar code readers, and laptops, as is true with many voting systems on the market today.

In particular, 3% of randomly selected precincts can be recounted, using the paper record, immediately when polls close, in the precinct, before removing ballots from the precinct. If discrepancies are found, a county-wide recount can be automatically triggered. Additional funding may need to be allocated to state and county election offices to routinely perform independent audits of vote counts.

In order to audit their vote counts and monitor the accuracy of vote counting systems, all state and county election offices should set up election data reporting systems to quickly and easily make publicly available, their precinct-level vote totals, broken out by vote type (i.e. election day,absentee, overseas, provisional, early voting, etc.) If vote counts are not reported down to this detailed level, then padded votes in one vote type can easily "cancel out" under-votes in another type. In other words votes can be subtracted from one candidate in one vote type, while being added for another candidate in another vote type, yet these two problems, when added together, may look perfectly normal.

Edison/Mitofsky can materially improve collective understanding of the exit polls -- and whether they are evidence of vote fraud -- by a full release of the data with precinct identifiers and by conducting further tests on the exit poll data, including analysis of variables for voting method and size of place to help resolve this issue. We emphasize that our objective is to ensure that, if election fraud occurred, it never does again in the future.

I whole-heartedly concur with these recommendations, with one minor addendum which is simply that any public release of the data should conform to AAPOR guidelines regarding confidentiality.

I've attempted to address your Buchanan point on a different thread.

And I have no hidden agenda. I never have had a hidden agenda. My agenda is simply that I want to know whether Bush won the election. I produced a mathematical measure that I hoped would make that question easier to answer. It is widely agreed by many, including USCV, that the measure is a useful one. Mitofsky is now using that measure.

[tommcintyre]

55. For the record, do YOU, or do you NOT agree with...

"It makes our "implausibility of rBr" claim stronger." (From Ron Banaim - in the opening post of this thread above. A CLEAR YES or NO please.
-------------
Again, you said in your post above:
"I am not take issue with the current version of the working paper."

Since Ron Banaim clearly states, "It makes our "implausibility of rBr" claim stronger."

You must agree, or you are still "taking issue" (with at least part - THE MAIN PART/THRUST) of Bainam et al's paper.

This IS pretty CLEAR isn't it?

------------

Also, for the record, do you believe that USCV should be allowed to present their case against the rBr hypothesis at the upcoming Carter/Baker Election Reform Hearing on June 30th?

Again, a CLEAR YES or NO answer would be very much appreciated.

I'm only asking you for your opinion, so please don't tell me your not qualified to answer this etc. After-all, just about EVERYONE here has an opinion on this; and you have set yourself up (and/or, been set up) as an "opinion-maker" on the pro/anti rBr issue.

BTW (just in case you're NOT aware of this): Mitofsy's rBr "hypothesis" was given as the reason that the exit poll discrepancy evidence was NOT allowed to be presented at the first C/B hearing in April.

So, you can imagine, your work will only be used, even further, to discredit the exit poll discrepancy evidence, and consequently deny USCV a "seat at the table" - resulting in further hiding the possibility from the American public that THIS ELECTION WAS STOLEN. (Welcome to the "smoke and mirrors" world of American politics Lizzie - a little different than the British system where you count the paper ballots on television, eh? ;) ) And, all the evidence I am literally stumbling over, points to the fact that Mitofskys CHIEF role has been to be an obfuscater - NOT an unbiased pollster.

One final question to illustrate/elucidate this point: Do YOU think there is ANY REALISTIC POSSIBILITY that Mitofsky would've hired (and publicly praised) you if your formula INDICATED A CLEAR EVIDENCE OF FRAUD? If you can be honest (especially with yourself), the answer is pretty clear, isn't it? NO WAY IN HELL!

So, dear Lizzie, whether it was intentional or not, YOU have become a dupe (pawn?) in Mitofsky's "game" to cover up the election fraud.

[Febble]

60. I'm doing my best here, Tom

I think Ron's statement is from his comment on Mystery Pollster, and no, I don't agree with it. Initially (several weeks ago) I did, as we all did (me, USCV) when I figured out the WPE problem (I won't go into that here, it is well documented now elsewhere), as when my (or Ron's version) formula is applied to the mean values given in the E-M report, it does indeed appear to strengthen the claim. My subsequent modelling work established that the formula needs to be applied at the level of the precinct WPE, not the mean WPE for a category of precincts, or you get rather different results. What my paper claimed was that we wouldn't, therefore, know the answer until it was applied, for real, at precinct level.

As for the main thrust of Baiman's new paper (the one first released at AAPOR), it is difficult to comment, as it has undergone a great many revisions. I have been accused here (as has OTOH) of responding to a non-current version before now. I am not sure what the main thrust of the paper now is.

But I certainly believe that USCV should be allowed to make their case anywhere. Whether this would be helpful for the cause of election reform probably depends on who you mean by USCV - do you mean the current authors of the current paper? Or the authors of the original paper? Or do you think that all current members of USCV should attempt to produce a consensus document or view to present at this hearing?

And I would certainly agree that the existence of one hypothesis should never be a reason for denying the case for an alternate to be presented. It's the whole basis of cross-examination. So yes, an unequivocal YES to the principle. However I would have strong reservations as to whether Baiman's case, as currently presented, is the strongest case to be made against Mitofsky's differential non-response hypothesis, which itself requires more evidential support.

And I would also question whether the exit poll case itself is the strongest case to be made for Election Reform - you say you are stumbling over evidence all the time - USCV has also collected an impressive body of evidence (including some written up by myself) and it is on their website. Others, including Bev Harris, Richard Hayes Phillips, Georgia10 at DKos have more. Rep John Conyers, an absolute hero, as far as I am concerned, has assembled a magnificent catalogue, which I am sure has only been added to since January. To me, this is the evidence that needs to be presented. In that company, I think the exit poll evidence remains weak, and so far, I don't think USCV have made a strong case. But that is certainly not to argue that the case, in principle, shouldn't be made.

Regarding your last point - it is difficult for me to answer, because you have not only answered it yourself, but implied that if I give I different answer, it means I am not being honest. But I'll have a go nonetheless. The only reason I know of for Mitofsky asking me to do some further analyses is because, like USCV and a number of others in the political and polling world, he realised that I was right, that the WPE was a bad measure, and particularly, it was a bad measure when the questions themselves you are asking of it concern partisanship. This applies to questions that test for both fraud (as in Baiman's questions) and differential non-response (was one group of voters more or less likely to respond in different political company?). Because much of the E-M report was based on analyses using the WPE, in effect they need to be re-done. I presume it made sense to him to ask the person who figured out the formula.

Now I know (or I am anticipating, as I realise I have become very defensive on this forum) that you may say that Kathy says I stole the formulas, that they were given to me by Mystery Pollster, or whatever, but that fact remains that these assertions are not true - or not true in the way they have been presented here. No insight is ever the result of one person solely (or rarely) - my thinking developed over the course of email conversations with both the USCV team (including Baiman) and others from the blogosphere (initially on DKos). But the fact is that the case for applying the formula to the precinct level data was made by me, in my paper, and as a result, my formulas were used by Mitofsky to do some initial reanalyses, and these were presented at AAPOR. They did not present "clear evidence of fraud", so I have no way of knowing what would have happened if they had. But I do not know what "clear evidence of fraud" would look like. I'd love to know. The new analyses certainly don't contain "clear evidence" AGAINST fraud.

One of the exciting aspects of USCV's recent work has been the development of hypotheses regarding a "fraud fingerprint". The problem remains, however, that even if such a fingerprint is testable for, it may not test positive, not because it isn't there, but because the data is too "noisy". But having a fingerprint to look for is a great step, as noise-reduction techniques may improve its chances of detection. My formula removes one important source of noise.

But this is why I think that even at best exit polls run a poor second to the kind of evidence you are tripping over, and at worst, can actually damage the case for Election Reform, which I agree is vital. I have never believed a poor argument is better than no argument, and at present, I think Baiman's is a poor argument. I hope USCV will come up with a better one.

[TruthIsAll]

61. Febble, your best is not good enough.

"As for the main thrust of Baiman's new paper (the one first released at AAPOR), it is difficult to comment, as it has undergone a great many revisions. I have been accused here (as has OTOH) of responding to a non-current version before now. I am not sure what the main thrust of the paper now is."

The fact that it has undergone "a great many revisions" is irrelevant to the question: Do you agree with it?

We are NOT talking about prior revisions.

And you say you are not sure what the "main thrust" of the paper is now? Well, why don't you read it then. We are not going anywhere.

Could it be that the "thrust" of the paper strengthens the case against rBr, as Ron Baiman claims - and you either fail to understand why or refuse to admit it?

[Febble]

62. I don't know!

And it is very relevant to the question, as I honestly can't keep up with the revisions. And if I cite one revision, someone says I am reading the wrong one. The wording changes every time. How can one say what the current "thrust" is?

But as far as I can see from the last one I downloaded (Sunday), no I don't find their case compelling. I agree with their recommendations however.

If you can cite a direct quote from a current version that you would like me to comment on, I will. But I do not know where to find the "thrust" you want me to comment on.

You have been consistent TIA, I know what you are saying. I honestly no longer have any clear idea what Baiman is saying.

[eomer]

65. The abstract is only 6 short paragraphs.

As one would expect, the abstract explains the thrust of the paper.

DU rules limit me to 4 paragraphs but 4 of the 6 are enough to give you the thrust:

E/M hypothesized that the discrepancy between their exit poll results and the reported vote was due to
different exit poll response rates by Kerry and Bush voters. However, US Count Votes' simulations
show that no plausible Kerry and Bush response rate distributions with constant mean can (with any
realistic chance) reproduce the distribution and values of the Edison/Mitofsky1 data for mean "within
precinct error" (WPE), median WPE, and overall response rates.²

US Count Votes has simulated a variety of exit poll response rate (Gaussian3) distributions for Bush and
Kerry voters and studied the resulting exit poll within precinct error distributions. The simulations thus
far suggest that possible ways to reproduce patterns of mean and median WPEs, and overall response
rates that resemble the distribution of the actual reported E-M exit poll data include:

voter exit poll response rate distributions with means that vary widely with the percentage of
Bush and Kerry votes cast in precincts and seem implausible.

very plausible exit poll response rate distributions by Kerry and Bush voters, accompanied by
vote shifts from Kerry to Bush.

I'll wait while you read the abstract ... .... ... ...

OK, now you know the thrust of the paper.

The thrust is that the USCV simulations were not able to reproduce the patterns of WPE means and medians and the overall response rate when they were based on constant mean response rates and no other assumed bias. The simulations were able to reproduce the patterns of WPE means and medians and the overall response rate with either of two sets of assumptions. One successful set of assumptions used response rates with means that varied according to partisanship with no other bias. The second successful set of assumptions used constant mean response rates combined with vote switching from Kerry to Bush.

The graph on page 10 shows the response rate means for the different levels of partisanship that were used to successfully simulate. The graphs on page 9 show the WPE means and medians that these response rates produced and show how similar to the actual means and medians they were.

That's the thrust. And, by the way, that is basically the same thing that Ron Baiman is quoted as saying in the OP of this thread.

As far as what you should comment on, I would think that USCV's conclusion that the simulations based on constant mean response rates do not reproduce the WPE means and medians and overall response rate would be a good place to start.

[Febble]

68. Yes

But whether I agree with the abstract depends on whether I think that the body of the paper supports its conclusions, and that's the bit that keeps changing. The abstract itself is simply a series of conclusions.

And at present, no, I don't think the abstract is well supported by the current body of the paper.

But I do find it admirable that they have posted the simulator URL and anyone else can decide for themselves whether the conclusions are supported.

It's here if anyone wants the link:

http://uscountvotes.org/ucvAnalysis/US/exit-polls/simulators/dopp/

[OnTheOtherHand]

40. actually, I've criticized the 5/22 report as well

As I pointed out at some length, I believe up-thread here, the new models probably fit better with Mitofsky's correlations, but they don't fit very well at all with the observed WPEs. They also don't fit well, as far as I can tell, with the Baiman appendix. And I don't think it's unfair to point out that the 5/17 model seems to have vanished without an explanation -- the only constant is that "We know there is fraud," even if how we know there is fraud changes radically.

But I should have left the "cover-up" stuff in another nasty thread, so I apologize for that. It had nothing to do with you.

I have some more criticisms of the 5/22 report, but I need some more time to formalize them. Informally, I just think those little zig-zags are unimpressive as a vote fraud "signature."

[OnTheOtherHand]

41. whoops, I'm not nuts, just really confused....

There have been lots of versions of this paper, often released on different dates than their cover indicates, and I don't have a file folder with all of them. So I can easily get confused, especially between home and office.

I think the first release version was 5/13 (although for a while it said 5/5 on the cover), the second release version may have been 5/17, and now there is a 5/21 release (which I called "5/22" because it has a trailing footer dated 5/22 -- I suppose the PDF was generated after midnight).

I don't even know how this latest version differs from the preceding one, except that one of the models moved from p. 8 to p. 7. The null "simulated response bias" is still at the top of p. 7; the first alternative, "vote shift simulation with no bias," is now at the bottom of p. 7; the second alternative (the one I spent more time on in the "reproducing shapes" post) is by itself on p. 8.

On a quick skim, all my gripes about the 5/17(?) revision still apply to the 5/21-22 revision. But I'm open to correction.

[TruthIsAll]

35. IMHO, your role is quite biased. You have said enough already...

to convince me of that.

I challenge you to rebut the math in my latest post.

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x371584

[OnTheOtherHand]

39. may the record reflect

that, to his credit, TIA admits that I actually did rebut the math in that post. He still has an argument -- I honestly think not such a strong one.

Not my place to say whether I'm quite biased.

[TruthIsAll]

49. Yes, but I fixed it right away. THE EAST WAS THE BEAST.

The East WAS the BIG BEAST of rBr aliens.

So were the other regions, only less so.

[OnTheOtherHand]

50. yes, you did. n/t

[tommcintyre]

43. "If you want to talk about coverups, I can document how USCV's..."

And now you say,
"As I said, I don't think it was a coverup, I think it was bad work."

So you started the paragraph with the first line, and then vaguely took it away at the end of the paragraph? And you wonder why I (and others) accuse you of being an obfuscater (a MUCH nicer term than I'm thinking, but this IS a public forum ;) ).

So you try to equate Mitofsky's pattern of "fixing", to Bainam et. al's attempt of OPENLY presenting and improving their work? Man, you just don't give up, do you?

It's enough to make an honest person sick!
:puke:

[OnTheOtherHand]

46. gee guy, thanks for sharing

Baiman handed out a paper at AAPOR that showed a significant relationship between partisanship and bias, and made a big deal about the observed means in Kerry strongholds and Bush strongholds. Then Mitofsky showed some slides that showed no significant relationship between partisanship and bias. Then USCV removed the AAPOR paper and substituted a new paper with a shift-plus-bias model that (probably) shows no significant relationship between partisanship and bias -- but doesn't fit the observed means in Kerry strongholds and Bush strongholds very well.* The AAPOR version is, as far as I can tell, no longer available on the USCV website.

* (EDIT) In fact, it fits these observed means worse than the rBr (response bias only) model that the paper claims to refute. I can hardly believe that myself, but the graphs on pages 7 and 8 are pretty clear -- at least in the 5/21-22 version. If form holds, there might be a 5/23 version by the time you read this.

I do think that if you were one-tenth as suspicious of Baiman and friends as you are of me, you would find these facts disturbing.

As for me, I find them darkly amusing.

[tommcintyre]

56. So which IS it really? USCV cover-up, or not?

Ya just can't seem to make up your mind.

- First, you insinuate a USCV cover-up:
"If you want to talk about coverups, I can document how USCV's ..."

- Then eomer "calls you" on it:
"A USCV cover up, you allege?"
snip
"Apparently in your world, if a researcher receives criticism, acknowledges that it was fair and then responds to it by refining their approach, this is called a cover up. I call it professionalism."

- So then you "back off":
"As I said, I don't think it was a coverup, I think it was bad work."

- So then I call you on your "techniques"/motives on all this:
"And you wonder why I (and others) accuse you of being an obfuscater ..."

- So, NOW you return to insinuating a USCV cover-up - or worse?
"I do think that if you were one-tenth as suspicious of Baiman and friends as you are of me, you would find these facts disturbing."

Jeez! Do you forget we can read all of this? If these were just words that tumbled out of your mouth at different occasions, it would be much harder to "nail you", wouldn't it? Sadly, this is a pattern I've seen from you since you arrived. :(

BTW: When you posted you were a Professor without posting the specialty, some speculation ensued around here. I mused that it must be poly-sci, since this pattern of attempted "slippery" obfuscation, has the "fingerprints" of politician-type all over it. Certainly, this does NOT imply that ALL poly-sci professors would behave in such a manner; but, for reasons just given, your "pattern" could easily be deduced from this origin.

I would find THIS darkly amusing, but the stakes are just way too high. I just can't laugh at any attempted pattern or tactic that contributes to covering up the election theft. :(

[LightningFlash]

69. Sorry, but there is no reason to consider 9 PHD's disturbed.

They are nine PHD's using forward thinking and computational analyses. There would be no reason to suspect the motives of a group of scientists who put the work together, considering they know far more than anyone when it comes to the subject of math.

If the motives of UsCountVotes are considered disturbing, I shudder to think what the motives of Mitofsky's news-media friends are.

Covering up years of research with the so called bush response bias that dates back to 1988, 1992, and is proven to be the official "explanation" for unexplained wins is actually disturbing.

[OnTheOtherHand]

71. I think the obfuscation is mostly on your end

This is not that hard to understand.

Why is, say, CNN.com's replacement of one set of exit poll results with another more suspicious than USCV's replacement of one working paper with another? Whatever else you say about either one, it isn't a coverup.

Why is my reluctance to post my name more suspicious than that of any other DU member? Whatever else you say, it isn't a coverup.

This is not that hard to understand.

[tommcintyre]

42. And there YOU go again. LOL Ignored the challenge and obfuscated

some more.

"I'm a college professor."
OF WHAT!?! (That's a meaningless, ABSOLUTELY USELESS, statement - as usual.) No doubt, you intend it be impressive (and consequently misleading); but, like I said, that kind of attempt at manipulation jus' don't cut it around here.

"How do you propose that I substantiate my involvement with USCV?"
SIMPLE, state your name (instead of trying to obfuscate further by pretending we know it); and we can EASILY confirm your "involvement" with USCV

<It might also be interesting to determine your political/financial "involvement" with other "groups". Might that help us to know your REAL agenda? I wonder.

We now know Febble's since she has admitted (after being "outed" by Kathy, and subsequently trying to cover it up) that she works for Mitofsky. And, we also know (again, by her own admission, only after being "called" on it), that she feigned no foreknowledge of information to be presented at a conference, in an attempt to advance her position, and discredit USCVs'.

So... jus' gotta wonder WHAT exactly are YOU up to, and "all about". On second thought, I retract the former musing: We DO know what YOU are up to, don't we? ;) >

"...arguing that Mitofsky has been covering up fraud for years."
No, not "arguing" this anymore, simply presenting evidence of the fact that Mitofsky has a "fixing" the exit poll for years.

How I came upon this information: The Thursday before last (in Kathy's thread asking for help to promote their new paper), I "smelled" that Febble was using the opportunity to both "spin" her position; and my "nose" also told me that she was faking not knowing what Mitofsky was going to present that weekend (as mentioned above, this has proven to be the case). <I've been told I have a particularly keen sense of "smell" - especially for bullsh*t/manipulation/"spin", etc. - that's why YOU have "stood out" also ;) >

Anyway, Feeble also said that Mitofsky was basically above reproach in the same thread. Well, since Febble's credibility was beginning to "reek" so badly on the other issues, I wondered if I shouldn't check out Mitofsky as well.

Well... "down the rabbit-hole" I slid. I quickly entered a blatant world of election-fixing that simply boggles the imagination! I literally stumbled on his 1992 "invention" of rBr while looking for something else. There is plenty of information about his shenanigans, which I would be GLAD to share with you (and Febble), but I doubt you are really THAT interested in knowing more about this, are you? ;)

Speaking of the invention of the rBr scam in '92; it really STICKS OUT that you ignored my challenge in the previous post (as you, and Febble, have repeatedly done before).

C'MON, INQUIRING MINDS WANT TO KNOW - how DO you attempt to explain this one away?

We are waiting (but not holding our collective breath) .

[Febble]

45. Tom, you know

because I have explained in emails to you, that I did not "attempt to cover up" after I was "outed" as you call it. And I hope I have now made it clear that I was engaged by Mitofsky to do some work 8 days before the AAPOR conference at which he presented his new analysis, and that I agree I should not have posted at DU during that time. He engaged me because it was my work, discussed on DU and elsewhere, that had led to the new analysis.

You also know that far from "attempting to cover it up" I posted a confirmation of the fact as a comment on a thread as soon as the information was public.

You also know that I then deleted my comment. You do not know why I deleted the comment, although you do know it wasn't an "attempt to cover up" the fact, as I actually emailed you personally, in response to your call for information, to confirm the its contents.

In the various comments and threads you have posted on DU about my correspondence with you, I have not noticed any reference to the fact that it was I who voluntarily contacted you. I did this precisely because I DID want to confirm the information, NOT cover it up, and thought, for better or worse, that that was a more appropriate way of doing so.

You also know, because I have told you, that I emailed the USCV team with the information as soon as it was in the public domain. It was in the public domain NOT because I was "outed" at DU, but because Mitofsky "outed" me, much to my relief, at AAPOR.

I should not have posted at DU during those 8 days, and I apologise, once again.

And FWIW, OTOH has posted on another thread that s/he is a professor of political science.

Regarding your question about Mitofsky's "invention" of "rBr": "rBr" is a term used in the context of the 2004 election for what is normally called "non-response bias" and was not "invented" by Mitofsky in 1992. It is a well-researched phenomenon that can invalidate all survey research, and although it can be minimized by good survey design, it can never be eliminated unless response rates are 100%, and then only if you can be absolutely sure that your sampling technique was truly random.

In most surveys, we have no way of "checking" our results to determine the degree of non-response bias - because by definition, the non-responders didn't respond. In exit polls, almost uniquely, we can check - with the vote itself. But, of course, the check is valid only IF YOU CAN RELY ON THE VOTE COUNT. Once you assume that the vote-count itself is rigged, then of course you can no longer assume that any apparent non-response bias you observed was truly in your survey, not in your vote. In the UK, where the vote-count is assumed to be clean (and almost certainly is, as we have public scrutiny of the count) we can measure non-response bias. And it happens. It happened in 1992, and the exit polls called the wrong result.

So it doesn't strike me as prima facie evidence of Mitofsky's "collusion" with a vote-scam that he attributed an exit poll error to differential non-response, although it does indicate that he assumed the count was correct, and that it was the poll that was in error. On this assumption, presumably, he inferred that the non-response bias had been an undersampling of Bush voters relative to Buchanan voters.

So you can certainly fault Mitofsky, as you do, for assuming that the vote-count is a "quality check" on his poll, rather than believing that his poll is a "quality check" on the vote-count. After the Ukraine experience I'd be (and was, and still am) more circumspect about that assumption.

But that is rather different from accusing him of "covering up" vote corruption.

[kster]

63. Febble,You told me

that we should have a full scale investigation into the election of Nov 2,2004, but that you think we should start the investigation in Ohio. Well can you help us get that investigation in Ohio started,instead of trying to discredit the exit polls.

Our kids are dying in Iraq, as we speak, and the longer we have the wrong president in office (which is obvious now) our kids will keep dying. Quit the b*llsh*t already.

[Febble]

64. I'm not trying to discredit the exit polls.

I'm just arguing that they are lousy evidence. I think lousy evidence can hurt a good case. Think proportional spaced font.

And I HAVE helped get the investigation in Ohio started.

And I investigated fraud in Florida, Ohio and New Mexico.

I spent weeks lobbying lawyers in Ohio with my evidence, including the Kerry-Edwards legal team and Cliff Arnebeck. Also sent evidence to John Conyers.

And I am doing my best to get rid of Tony Blair as leader of the Labour Party.

And God knows, I agree you have the wrong president in office.

[tommcintyre]

66. Full disclosure, and clarity is the only way to go

As far as your first six paragraphs, we have quite different interpretations of what happened. I stand by my "raw data" thread, and, with your repeated permission, would be glad to re-post it. ;)

The reason I posted it in the first place was that I noticed your "report and apology" thread was considerably "diluted" as far as the "apology" ("admission") side went, so I figured providing people with the "raw data" (just as Mitofsky should) was a fair and reasonable thing to do. I figured, provide ALL the uncensored correspondence, and let the readers decide for themselves.

Even one of the posters said the following in YOUR thread: "Can you provide a formula to reduce the noise around that apology". ;)
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x370401#370521

This type of personal "bias" (spinning) is pretty common, so you're not alone. ;)

A short lil' story to illustrate the point: A while back, a friend - a French person (renown for their "frankness"), was going to "tell her mother" how unfairly she was treated by someone. Well, the "offending" person asked her, "are you also going to tell her what YOU did"? She responded, "of course not, I want her to be on MY side"!

Well... perhaps, even though you did apologize to a degree, you were biased in wanting the DU readers on YOUR side more than what was truly deserved? ;)
-------------------
As far as Mitofsky's "invention" goes:

"non-response bias" and was not "invented" by Mitofsky in 1992."
Why of course not - but the reason he gave for it sure was.

And, amazingly, "lightening" struck again, twelve years later , and hit another "Bush"! (Can you give me the statistical odds of THAT happening! ;)

"In exit polls, almost uniquely, we can check - with the vote itself. But, of course, the check is valid only IF YOU CAN RELY ON THE VOTE COUNT. Once you assume that the vote-count itself is rigged, then of course you can no longer assume that any apparent non-response bias you observed was truly in your survey, not in your vote." <bolding mine>
Therein lies one of the major flaws in your rBr hypothesis, since we DO NOT have reasonably transparent elections, there is no constant by which to measure. You are left with two variables, so no solid ground to stand on - so to speak.

Combine that with, "we have no way of "checking" our results to determine the degree of non-response bias - because by definition, the non-responders didn't respond" ; and (in this case) the "argument" is rendered a purely academic discussion, isn't it?

"So it doesn't strike me as prima facie evidence of Mitofsky's "collusion" with a vote-scam that he attributed an exit poll error to differential non-response..."
Alone, it may not; but when taken in consideration with his clear pattern/history of exit poll obfuscation (again, I have links if you want them), the "case" becomes much more clear.

"...although it does indicate that he assumed the count was correct, and that it was the poll that was in error."
Good observation Lizzie - why would the man that "invented" exit polling, and had 25 years experience with them at the time, make such an assumption? Could it be that he starts with a certain "answer" (Bush Sr. beats Buchanan) and has to find a "solution" that will get him there? If so, that sounds much more like a "fixer" than an objective exit pollster, don't you think?

And, further, in 2004, couldn't he have ALSO started out with "a certain answer" (GW beat Kerry), and had to find a solution to match that assumption?

"So you can certainly fault Mitofsky, as you do, for assuming that the vote-count is a "quality check" on his poll, rather than believing that his poll is a "quality check" on the vote-count. After the Ukraine experience I'd be (and was, and still am) more circumspect about that assumption."
I'm sure the original Ukrainian "winning candidate" (WC) wishes he had rWCr to "explain away" the exit poll discrepancy. He'd probably still be in power - just as Bush is.

"But that is rather different from accusing him <Mitofsky> of "covering up" vote corruption."
Again, you have to take the history/pattern into account to fairly assess this. And, I believe it shows a clear pattern of "covering up" vote corruption.

[OnTheOtherHand]

47. hmm

I couldn't care less about trying to impress you. My point is that I don't have any financial interest in this issue. (I tried to make that point clearer by editing the post upthread, but I think I put the edit in brackets and it was gobbled up as bad pseudo-HTML.) If there are any other sordid interests you want me to disavow -- say, has Karl Rove threatened to kill my wife if I don't drink the Kool-Aid? -- I am happy to do that too.

Since (1) most thread participants haven't revealed their names (AFAIK), and (2) there is no master list of USCV participants to check my name against, and (3) the undisputed fact that Febble was a USCV participant and still has working papers posted there doesn't cut any ice with you, I don't see how posting my name would resolve anything. But if you would like Bruce O'Dell to come by and vouch for my identity, I bet he would do that.

As far as I'm concerned, you're the one ignoring the challenge to your 1992 argument, as Mitofsky is involved. If Mitofsky is out to spin exit polls for Bush, then why on earth did he let Kerry appear to win the 2004 election in the first place? You think it would be hard for him to hack his own database? (Did Bush steal the 1992 NH primary? I have no idea. Could be -- I don't have any evidence either way.)

[TruthIsAll]

51. The final exit poll data was leaked and downloaded at 12:22am by Simon.

That was not Mitofsky's doing.

It is true that the early national exit timelines of 8349 (4pm) and 11027 (7:33pm) respondents are available online. But we only have a GIF file from the WP/NEP site for the 13047, which was downloaded at 12;22am. Why was it deleted?

The only reason we are even now discussing this is because a few patriots gave us the early exit data. And analyzing the exit poll data is the pathway to the truth.

The exit data up to and including the 12:22am timeline reflects what really happened. That's why CNN et al had to replace quickly suppress the data after midnight with bogus OH and FL exit poll numbers to match the corrupted vote count.

[OnTheOtherHand]

52. well...

Look, I bet there must be thousands of people who did what I did: waited for the polls to close in Ohio, went to the CNN.com exit poll page, hit "refresh" until they posted the results, did a quick calculation that showed Kerry well up, screamed "Yessssssss!" or some variation on that. And then there were tens of thousands of people who had to listen to us screaming....

I've never grasped the idea that something had to be leaked. CNN.com openly posted these results, as I guess WP/NEP did too. And of course before CNN.com openly posted the results, earlier results were being leaked all over the place.

So the idea that the mainstream media was trying to cover up the exit poll results by quickly changing web sites never made much sense to me. It wasn't just a few patriots who knew how the raw exit polls had turned out.

(EDIT: changed 'gotten' to 'grasped' to clarify the meaning)

[TruthIsAll]

53. AND THAT'S WHY WE HAD TO WAIT UNTIL 1:25PM TO

GET THE 13660 FINAL NATIONAL EXIT POLL.

THEY NEEDED TO GET IT IN SHAPE FOR CONSUMPTION BY THE SHEEP.

LOTS OF NUMBERS TO MASSAGE, LOTS OF DEMOGRAPHICS TO FIX, LOTS OF WEIGHTS TO ADJUST, LOTS OF PERCENTAGES TO RIG. ALL HAD TO FIT THE CORRUPTED VOTE COUNT.

43% BUSH 2000 VOTERS? COME ON.
37%/37%/26% PARTY ID? COME ON.
51% FEMALES FOR KERRY? COME ON.

AND ALL THE OTHERS? COME ON.

AS ORWELL SAID, YOU CAN FOOL SOME OF THE SHEEP SOME OF THE TIME...

[OnTheOtherHand]

54. it wouldn't take long, actually

If all they needed was to get X% more Bush voters, then all they had to do was multiply every Bush-respondent weight by X% and rerun the tabs. The numbers might not make much sense, but apparently you don't think they make much sense anyway, so what the heck? And if they planned to do it all along, then they could have done it in October.

Folks, I just have to say: There are some survey researchers out there who are very suspicious about the 2004 election, but there aren't many (if any) who think that E/M or whoever reweighted the exit poll results as part of a coverup. It would have been pointless. Basically everyone in the country who cared about exit polls already knew what they said. The cat was out of the bag.

[TruthIsAll]

57. OTOH, your back-flips are getting old...

"If all they needed was to get X% more Bush voters, then all they had to do was multiply every Bush-respondent weight by X% and rerun the tabs. The numbers might not make much sense, but apparently you don't think they make much sense anyway, so what the heck? And if they planned to do it all along, then they could have done it in October".

OTOH, it was not so simple. They had a dynamic Bush free-fall which was much worse than they could have ever contemplated - Kerry was close to a landslide. They could not just change the weights, they had to change the percentages as well. For some reason, they left the female/male sply constant at 54/46. I guess that would have been too obvious.

They never suspected they would be blown out. They had to scramble; many fingers had to plug the dyke. That takes time.

"Folks, I just have to say: There are some survey researchers out there who are very suspicious about the 2004 election, but there aren't many (if any) who think that E/M or whoever reweighted the exit poll results as part of a coverup. It would have been pointless. Basically everyone in the country who cared about exit polls already knew what they said. The cat was out of the bag".


OTOH, just SOME researchers?

How do YOU explain the "How voted in 2000" 43% Bush/37% Gore weighting?

OTOH, I will enjoy your contortions as you attempt to spin THAT ONE.

[OnTheOtherHand]

70. I'm confused again

They could not just change the weights, they had to change the percentages as well.

Well, which ones? Changing the weights will change all the percentages, too. So I don't get your point at all.

I have no idea what proportion of survey researchers are suspicious about the 2004 election, so I will have to stand by "some." It would be interesting to see a (gulp) poll.

How do YOU explain the "How voted in 2000" 43% Bush/37% Gore weighting?

Well, it fits pretty well with the hypothesis that they basically multiplied each Bush voter by 1.1 or so. Beyond that, I will have to look at the data files. (It's true, as someone pointed out somewhere, that people tend to exaggerate having voted for the winner after the fact, but I can't believe that will get us to 43/37, if that's where we are.)

[TruthIsAll]

74. Let me educate you with the facts..

OTOH, I'm surprised at you.

You are apparently innocent of the FACT that they had to change BOTH weights AND percentages in a number of categories TO MATCH THE CORRUPTED VOTE COUNT. Voted in 2000 is the MOST outrageous example. There are many others.

And where the WEIGHTS were NOT changed, the percentages were- TO THE POINT OF ABSURDITY. KERRY WENT FROM 54% OF FEMALES AT 12:22AM TO 51% AT 1:25PM? DO YOU REALLY EXPECT US TO BELIEVE THAT?

ESPECIALLY WHEN THERE WAS A MASSIVE TURNOUT OF NEW, YOUNG, FEMALE VOTERS WHO WERE SCARED THAT ROE V. WADE WOULD BE IN JEOPARDY IF BUSH WERE RE-SELECTED?

I refer you to the TIMELINE:
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x363938

BTW, you did not anwser my question on the 43/37 IMPOSSIBILITY.

It shows that you are ALSO unaware or dismissive of the FACT that it can be NO HIGHER than 39.8% MAXIMUM, which is itself implausible since it assumes 100% Bush 2000 voter turnout.

Do you know why that statement is a fact?
Do you want me to prove it to you mathematically?

I will save that for a future lesson.

OTOH, your credibiity is not the only thing in question.
So is your knowledge of the facts.

[tommcintyre]

67. I think TIA explained this well, I would only add...

that, from what I've seen, Mitofsky uses a number of methods to "cover-up". It seems to range from computer meltdowns, "found votes", withholding the data, "convenient guesses", etc.

So what may have happened in 2004?

- More scrutiny because of publicity of the 2000 theft.

- Internet leaks

Finally,
"I don't see how posting my name would resolve anything."
WE could check your background for political/financial affiliations; but YOU probably wouldn't want THAT, would you? ;)

[OnTheOtherHand]

72. well, it COULD be

that Mitofsky was somehow forced to release results that showed Kerry ahead, and then all the rest of it.

I certainly appreciate your willingness to investigate my private affairs, but it seems as if you may have your hands full with Mitofsky. Right now your confidence well outstrips your evidence.

[TruthIsAll]

75. I DON'T WANT TO KNOW YOUR NAME.

Neither do I want to know the names of the dozens of naysayers who proceeded you who have vanished (worn out?) when confronted by FACTS and LOGIC.

The combination of their ignorance, straw men and red herring arguments were exposed, just like rBr is exposed by the absurdity of the simulation scenario assumptions required to explain it.

And your straw men are being exposed as we speak.

[OnTheOtherHand]

77. okay, I DON'T WANT TO KNOW YOUR NAME EITHER (grin)

As for the rest, you make a lot of assumptions about what I believe. Hey, whatever. But I still say the USCV working paper is bad. And I still say I don't think Mitofsky is a plotter. As for what really happened in 2004, we still have a lot to learn.

[autorank]

73. OTOH, this is an anonymous forum, you don't need to give your name.

...but this is the second time, at least that you've made a comment like this one:

I'm a college professor. I assume you know my name, actually, but maybe not. How do you propose that I substantiate my involvement with USCV?

If you were going to give your name or think that we knew it, you would at least hint at it in your profile...but your profile is disabled.

If you want to remain anonymous, that's fine and it's what the vast majority of people do. If you want to hint but not tell, do a profile that's accurate. If you want to tell us who you are, and there is no pressure to do this, than just do it.

Or is there a motive behind your coyness?

[OnTheOtherHand]

76. fair enough

I've explained by PM why I expected that tom would know my name -- apparently just a bad guess on my part.

I've updated and exposed my profile so that anyone who is interested in my name can easily google it.