To Give You a Size of the Immense and Growing Size of Illegal Immigration

[Loren Pechtel]

So let's try this rewording:

A census does a full count of a state's population as 3,011,524. (See how this is better? The census isn't rounding to the nearest hundred thousand or estimating.) According to a survey which might be wrong and have systemic issues, an estimate of an undercount for that state is 151,781 which is a midpoint of a confidence interval.

Do you disagree so far with the changes in wording?

And what is that confidence interval,

I only found it by chance browsing over a map that suddenly had a popup with a 90% confidence interval about the %. I don't know even why they would only do a 90% confidence interval and keep it hidden. Maybe it's just because they are trying to appeal to average Joe readers? ETA: I remember seeing std error with the race "undercounts" and "overcounts" and so I think I may have also seen that with the states. So you can probably construct confidence intervals from those, provided they are there.

Ok, that's what my gut was telling me about the numbers. This is a case of lying with statistics, it means nothing other than the whoever authorized it is Republican deep state.

and what p value is used in calculating it?

No idea. That is something I searched and searched for...their threshold of significance. Convention is 5% but that would be inappropriate for 50 states and it's probably a more complex method.

I'm not aware of any more complex method. The precision simply isn't there.

While reporters will almost certainly omit such things any proper poll will have this information. Why was it not listed???

There are a lot of documents there and I haven't gone through them all. It wasn't out in front, but also, I really don't think we need it to see that the original claims were inaccurate of 6 Republican congress critters missing and 6 extra Democrat congress critters.

You've already found enough, their goose is cooked.

 

[ZiprHead]

Worst member of the House of Representatives​

The nominees:

(a) Lauren Boebert, R-Colo.
(b) Matt Gaetz, R-Fla.
(c) Ralph Norman, R-S.C.
(d) Marjorie Taylor Greene, R-Ga.
(e) Jim Jordan, R-Ohio
(f) James R. Comer, R-Ky.
(g) Mike Johnson, R-La.
(h) Thomas Massie, R-Ky.
(i) Bob Good, R-Va.
(j) Andy Biggs, R-Ariz.
(k) Scott Perry, R-Pa.
(l) Nancy Mace, R-S.C.

The winner: Too many to choose from.

You forgot Tommy Tuberville (Yeah, I know he's a senator). By far the most idiotic and downright stupidest person in Congress, bigly.

 

[TomC]

Worst member of the House of Representatives​

The nominees:

(a) Lauren Boebert, R-Colo.
(b) Matt Gaetz, R-Fla.
(c) Ralph Norman, R-S.C.
(d) Marjorie Taylor Greene, R-Ga.
(e) Jim Jordan, R-Ohio
(f) James R. Comer, R-Ky.
(g) Mike Johnson, R-La.
(h) Thomas Massie, R-Ky.
(i) Bob Good, R-Va.
(j) Andy Biggs, R-Ariz.
(k) Scott Perry, R-Pa.
(l) Nancy Mace, R-S.C.

It's kinda refreshing to see a list like that and for once Indiana is not on it.
Tom

 

[Shadowy Man]

Worst member of the House of Representatives​

The nominees:

(a) Lauren Boebert, R-Colo.
(b) Matt Gaetz, R-Fla.
(c) Ralph Norman, R-S.C.
(d) Marjorie Taylor Greene, R-Ga.
(e) Jim Jordan, R-Ohio
(f) James R. Comer, R-Ky.
(g) Mike Johnson, R-La.
(h) Thomas Massie, R-Ky.
(i) Bob Good, R-Va.
(j) Andy Biggs, R-Ariz.
(k) Scott Perry, R-Pa.
(l) Nancy Mace, R-S.C.

It's kinda refreshing to see a list like that and for once Indiana is not on it.
Tom

Being better than that lot isn’t much of a prize.

 

[Don2 (Don1 Revised)]

I'm not aware of any more complex method. The precision simply isn't there.

I finally found how they defined statistical significance:

The PES performed statistical testing for estimates of net coverage error rates and certain components of coverage. Statements of comparison were statistically significant at the 90 percent confidence level (α = 0.10) using a two-sided test. For any PES release, estimated net coverage error rates that were statistically significantly different from zero were identified by an asterisk (*).

https://www2.census.gov/programs-surveys/decennial/coverage-measurement/pes/2020-source-and-accuracy-pes-estimates.pdf

For a two-tailed 90% confidence interval, you will have 5% on each end. 0 will be included in one side's 5% when they are saying it is significant. When 0 is not included in the 90% middle, they are saying it is statistically significant. In a way, it is like saying there is a 95% chance (or confidence) that it is an undercount or overcount because that whole side of the distribution including one tail will be on one side of zero. So it seems equivalent in a sense to a p-value of 5% or less for significance, i.e. the convention.

To add--their goal is not to provide new numbers but to provide insights for quality and potential improvements for the next census. So they go through multiple rounds of finding problems in their own work and then have to stop because the census numbers must be used by 2023. So in 2022 they release their survey results.

They want a reasonable confidence level that will minimize false positives AND false negatives. They might want to err on the side just a little bit of including some things that are not actual problems so that they do not remove too many real problems. They can find out which is which later upon additional work.

From that same paper, check their conclusion:

In conclusion, while every survey must deal with a multitude of errors, and the PES is no different, we implemented many measures to mitigate the impact of these errors on our estimates. The PES estimates provide helpful insights about the quality of the 2020 Census and should be useful when informing plans for the 2030 Census.

 

[TomC]

Worst member of the House of Representatives​

The nominees:

(a) Lauren Boebert, R-Colo.
(b) Matt Gaetz, R-Fla.
(c) Ralph Norman, R-S.C.
(d) Marjorie Taylor Greene, R-Ga.
(e) Jim Jordan, R-Ohio
(f) James R. Comer, R-Ky.
(g) Mike Johnson, R-La.
(h) Thomas Massie, R-Ky.
(i) Bob Good, R-Va.
(j) Andy Biggs, R-Ariz.
(k) Scott Perry, R-Pa.
(l) Nancy Mace, R-S.C.

It's kinda refreshing to see a list like that and for once Indiana is not on it.
Tom

Being better than that lot isn’t much of a prize.

I could walk to Mike Pence's house from here.
It's not a prize.
Tom

 

[Loren Pechtel]

I'm not aware of any more complex method. The precision simply isn't there.

I finally found how they defined statistical significance:

The PES performed statistical testing for estimates of net coverage error rates and certain components of coverage. Statements of comparison were statistically significant at the 90 percent confidence level (α = 0.10) using a two-sided test. For any PES release, estimated net coverage error rates that were statistically significantly different from zero were identified by an asterisk (*).

https://www2.census.gov/programs-surveys/decennial/coverage-measurement/pes/2020-source-and-accuracy-pes-estimates.pdf

For a two-tailed 90% confidence interval, you will have 5% on each end. 0 will be included in one side's 5% when they are saying it is significant. When 0 is not included in the 90% middle, they are saying it is statistically significant. In a way, it is like saying there is a 95% chance (or confidence) that it is an undercount or overcount because that whole side of the distribution including one tail will be on one side of zero. So it seems equivalent in a sense to a p-value of 5% or less for significance, i.e. the convention.

So we would expect to see 5 false positives from this.

However, the number of survey points in low population states (remember, they listed the number of data points for the nation, not a number for each state) says that the low-population states will have wide error bars, I have a hard time picturing them reaching this threshold with the reported error margins.

To add--their goal is not to provide new numbers but to provide insights for quality and potential improvements for the next census. So they go through multiple rounds of finding problems in their own work and then have to stop because the census numbers must be used by 2023. So in 2022 they release their survey results.

Yeah, someone took data that wasn't meant for this purpose and reported it as if it meant something.

Compare the 2020 response data to the 2010 data. There are more than 4x as many people not willing to cooperate in 2020 as in 2010. I would be very surprised if that is random.

 

[Don2 (Don1 Revised)]

Introduction

I went ahead and made a program to run the Hungtinton-Hill method of apportionment. This is the standard that Congress uses and is described in a previous post. The numbers I got back reproduce the delegate allocations that are present in Congress which gives confidence in the programming. This method I call Original Huntington-Hill.​

​

I also ran 3 other methods.​

​

Original Intuitive simply takes the divisor of the census population divided by 435. This is about 760367. I do not know why there are other numbers out there, but this one is correct and it subtracts the DC population as I had attempted to do before but thought it was weird because it did not reproduce other numbers I was seeing at the time. This one, 760367, happens to also be documented which I found and confirms it is the right one. So in this method, if the decimal part of the state's number after dividing by the divisor is .5 or higher, it is rounded up. Originally Congress was also looking for the highest decimal remainders when something went wrong, like not enough delegates were assigned or too many in order to prioritize the last few delegate assignments. I did not bother to implement that.​

​

Next, I took the percentages of just the so-called statistically significant undercounts and overcounts from the report and used those to create an Adjusted Population for those few states. This is also summed up. Strangely, it is quite a bit more than the original population even though the documentation provided by the Census Bureau says they are similar. Therefore, I do not trust their numbers. In any case, I copied the Original Populations of states into this column for all states that had no significant difference. Their new sum, 331508865, was used to create a new divisor (by dividing by 435).​

​

Computing the adjusted populations and summing, and dividing, allowed for two more methods.​

​

Adjusted Huntington-Hill is running the standard Congressional method of apportionment but on the Adjusted Populations. Then, Adjusted Intuitive takes the Adjusted Populations and divides them by the new divisor. If the decimal part is above .5, it rounds up. As before, there is nothing additional.​

​

Any method that is discrepant in the resulting delegate count appears in bold font for the entire row where it is. This way we can look manually at discrepancies and decide what happened. Possibly what ought to be and count up legitimate results.​

Results

Original Population	330759736
Adjusted Population	331508865
Original divisor	760367.209195402
Adjusted divisor	762089.344827586

State	Original Population	Adjusted Population	Original Huntington-Hill	Original Intuitive	Adjusted Huntington Hill	Adjusted Intuitive
California	39538223	39538223	52	52 (51.998853)	52	52 (51.881349)
Texas	29145505	29716053	38	38 (38.330828)	39	39 (38.992873)
Florida	21538187	22314740	28	28 (28.326033)	29	29 (29.281002)
New York	20201249	19529436	26	27 (26.567754)	26	26 (25.626176)
Pennsylvania	13002700	13002700	17	17 (17.100553)	17	17 (17.061910)
Illinois	12812508	13069987	17	17 (16.850422)	17	17 (17.150203)
Ohio	11799448	11626217	15	16 (15.518092)	15	15 (15.255714)
Georgia	10711908	10711908	14	14 (14.087809)	14	14 (14.055974)
North Carolina	10439388	10439388	14	14 (13.729403)	14	14 (13.698378)
Michigan	10077331	10077331	13	13 (13.253242)	13	13 (13.223293)
New Jersey	9288994	9288994	12	12 (12.216458)	12	12 (12.188852)
Virginia	8631393	8631393	11	11 (11.351611)	11	11 (11.325959)
Washington	7705281	7705281	10	10 (10.133631)	10	10 (10.110732)
Arizona	7151502	7151502	9	9 (9.405327)	9	9 (9.384073)
Tennessee	6910840	7257761	9	9 (9.088819)	9	10 (9.523504)
Massachusetts	7029917	6875897	9	9 (9.245424)	9	9 (9.022429)
Indiana	6785528	6785528	9	9 (8.924014)	9	9 (8.903848)
Missouri	6154913	6154913	8	8 (8.094659)	8	8 (8.076367)
Maryland	6177224	6177224	8	8 (8.124001)	8	8 (8.105643)
Wisconsin	5893718	5893718	8	8 (7.751147)	8	8 (7.733631)
Colorado	5773714	5773714	8	8 (7.593323)	8	8 (7.576164)
Minnesota	5706494	5495468	8	8 (7.504919)	7	7 (7.211055)
South Carolina	5118425	5118425	7	7 (6.731517)	7	7 (6.716306)
Alabama	5024279	5024279	7	7 (6.607701)	7	7 (6.592769)
Louisiana	4657757	4657757	6	6 (6.125668)	6	6 (6.111825)
Kentucky	4505836	4505836	6	6 (5.925868)	6	6 (5.912477)
Oregon	4237256	4237256	6	6 (5.572644)	6	6 (5.560051)
Oklahoma	3959353	3959353	5	5 (5.207159)	5	5 (5.195392)
Connecticut	3605944	3605944	5	5 (4.742372)	5	5 (4.731655)
Utah	3271616	3189020	4	4 (4.302679)	4	4 (4.184575)
Iowa	3190369	3190369	4	4 (4.195827)	4	4 (4.186345)
Nevada	3104614	3104614	4	4 (4.083046)	4	4 (4.073819)
Arkansas	3011524	3171361	4	4 (3.960618)	4	4 (4.161403)
Kansas	2937880	2937880	4	4 (3.863765)	4	4 (3.855034)
Mississippi	2961279	3088204	4	4 (3.894538)	4	4 (4.052286)
New Mexico	2117522	2117522	3	3 (2.784868)	3	3 (2.778574)
Nebraska	1961504	1961504	3	3 (2.579680)	3	3 (2.573850)
Idaho	1839106	1839106	2	2 (2.418708)	2	2 (2.413242)
West Virginia	1793716	1793716	2	2 (2.359013)	2	2 (2.353682)
Hawaii	1455271	1362741	2	2 (1.913906)	2	2 (1.788164)
New Hampshire	1377529	1377529	2	2 (1.811663)	2	2 (1.807569)
Maine	1362359	1362359	2	2 (1.791712)	2	2 (1.787663)
Montana	1084225	1084225	2	1 (1.425923)	2	1 (1.422701)
Rhode Island	1097379	1044625	2	1 (1.443222)	1	1 (1.370738)
Delaware	989948	938784	1	1 (1.301934)	1	1 (1.231856)
South Dakota	886667	886667	1	1 (1.166104)	1	1 (1.163469)
North Dakota	779094	779094	1	1 (1.024629)	1	1 (1.022313)
Alaska	733391	733391	1	1 (0.964522)	1	1 (0.962343)
Vermont	643077	643077	1	1 (0.845745)	1	1 (0.843834)
Wyoming	576851	576851	1	1 (0.758648)	1	1 (0.756934)

Discussion

There are 8 rows bolded which indicates across 4 methods there are a maximum 8 discrepancies in comparing any two. However, comparing any of one method against the original may produce much less than 8 differences such as 3 or 4. Here I do 3 comparisons of methods.​

​

1. Original Huntington-Hill (OHH) vs Original Intuitive (OI)​

​

If you focus on the bolded rows but then scan downward along these two columns, you will find 4 discrepancies:​

State	OHH	OI
New York	26	27 (26.567754)
Ohio	15	16 (15.518092)
Montana	2	1 (1.425923)
Rhode Island	2	1 (1.443222)

​

The standard method deducts 1 from both New York and Ohio as compared to the intuitive method. Meanwhile, the standard method adds one to both Montana and Rhode Island as compared to the intuitive method. You may observe that the decimal part of the delegate is in the .4 range for both Montana and Rhode Island, but the decimal part for both New York and Ohio is in the .5 range. This may be surprising, but as noted earlier the standard method we use is biased in favor of smaller states.​

​

The purpose of this comparison was to illustrate how our current process works in contrast to intuition.​

​

2. Original Huntington-Hill (OHH) vs Adjusted Huntington-Hill (AHH)​

​

If you focus on the bolded rows but then scan downward along these two columns, you will find 4 discrepancies:​

State	OHH	AHH
Texas	38	39
Florida	28	29
Minnesota	8	7
Rhode Island	2	1

In the adjusted version, Florida gains 1, and Texas gains 1. Minnesota and Rhode Island both lose 1.​

​

The purpose of this comparison is to illustrate how the actual process would work if the adjusted numbers were actually used in reality.​

​

3. Original Intuitive (OI) vs Adjusted Intuitive (AI)​

​

If you focus on the bolded rows but then scan downward along these two columns, you will find 6 discrepancies:​

State	OI	AI
Texas	38	39
Florida	28	29
New York	27	26
Ohio	16	15
Tennessee	9	10
Minnesota	8	7

This comparison is most interesting because the lack of knowledge about the exact process of apportionment might lead many people to make this comparison. Even in discussions here, we observe people using the divisor as I also initially thought to do. So I believe we can hypothesize that this is how the conspiracy theory started. We can also observe that there are a total of 6 discrepancies which is a number that is in common with the conspiracy theory that there were an extra 6 Democrats and missing 6 Republicans.​

​

What we do observe here is 3 states that are typically red and 3 that may be blue (is Ohio purple?) being impacted, with red states negatively and "blue" states positively. This does not necessarily translate to 3 Republicans and 3 Democrats.​

​

And as noted, the Post-Enumeration Survey results themselves are of questionable quality. Whether they are better than the census is an open question.​

Conclusion

None of the methods show an extra 6 Democrats and missing 6 Republicans.​

​

A comparison of intuitive methods (that are not used in the process) shows 3 typically red state gains, 1 purple/blue state gain, and 2 blue state gains. This comparison may have snowballed into a more dramatic version magnifying the difference and assuming partisan results based on "blue" or "red" states. In reality, following apportionment, there is redistricting and voting and those results may differ from predictions based on state gains or losses.​

​

In any case, the intuitive method is not used, but instead the Huntington-Hill method of apportionment. In that scenario, which is the actual Congressional process, there are fewer discrepancies with 2 red-state gains and 2 blue-state losses.​

​

Those also do not necessarily translate to two Republicans and two Democrats not merely because of the process following apportionment, but also because the survey results that the adjustments are based on are not necessarily improvements to the census.​

 

[Don2 (Don1 Revised)]

I'm not aware of any more complex method. The precision simply isn't there.

I finally found how they defined statistical significance:

The PES performed statistical testing for estimates of net coverage error rates and certain components of coverage. Statements of comparison were statistically significant at the 90 percent confidence level (α = 0.10) using a two-sided test. For any PES release, estimated net coverage error rates that were statistically significantly different from zero were identified by an asterisk (*).

https://www2.census.gov/programs-surveys/decennial/coverage-measurement/pes/2020-source-and-accuracy-pes-estimates.pdf

For a two-tailed 90% confidence interval, you will have 5% on each end. 0 will be included in one side's 5% when they are saying it is significant. When 0 is not included in the 90% middle, they are saying it is statistically significant. In a way, it is like saying there is a 95% chance (or confidence) that it is an undercount or overcount because that whole side of the distribution including one tail will be on one side of zero. So it seems equivalent in a sense to a p-value of 5% or less for significance, i.e. the convention.

So we would expect to see 5 false positives from this.

Right. Offhand, I am unsure if it's 2.5 or 5. But there's an expected average amount that still has a variation about it. And that's just due to randomness, never mind all the systemic problems they identify with surveys of this nature. Measurement error, processing error, Analysis error, Recall Bias, etc etc.

However, the number of survey points in low population states (remember, they listed the number of data points for the nation, not a number for each state) says that the low-population states will have wide error bars, I have a hard time picturing them reaching this threshold with the reported error margins.

And it gets worse. If you set the confidence interval to 98% or 99% to try to get rid of the average 5 (or 2.5) errors merely from randomness, half the things you call significant go away, especially the smaller states.

To add--their goal is not to provide new numbers but to provide insights for quality and potential improvements for the next census. So they go through multiple rounds of finding problems in their own work and then have to stop because the census numbers must be used by 2023. So in 2022 they release their survey results.

Yeah, someone took data that wasn't meant for this purpose and reported it as if it meant something.

Exactly. One of the things I had brought up before is that the dimensions are cherry-picked to create a narrative. No one online who is promoting this idea of adjustments needed because of states is also considering the Census Bureau reports that had a different dimension of race undercounts and overcounts. Suppose there was a different person out there of the left-wing variety who did the mirror image, completely ignored the state statistics and talked only about the race side of it: whites were overcounted, Blacks were undercounted, Hispanics were undercounted...then they could also create a narrative that their side (Democrats) should have more votes.

But anyway, yeah, someone cherry-picked the dimension, then created material online focusing only on states...then someone else came along and changed it to magnify things. No one also looked at the very poor quality of the survey...or at least that it is not of high enough quality to replace the census.

Compare the 2020 response data to the 2010 data. There are more than 4x as many people not willing to cooperate in 2020 as in 2010. I would be very surprised if that is random.

Another thing to look at is the bottom of page 8 and top of page 9: Recall bias and proxies. You will be astonished. So they are interviewing people up to 2022 about where they were on April 1st, 2020. A time when people who have seasonal homes in the Northeast and Florida might usually be thinking of switching but also a time when there was a pandemic and there were lockdowns. How well do people recall that....But it gets worse. If those people are not available, they ask a proxy for that old information. It's already no longer a primary source with respect to time, but now it's also second-hand info, hearsay, about someone else from up to two years before. In 2010, there were only 3.7% proxies who answered the question, but in 2020, 7% were proxies!

From their own report:

Another feature that could affect the accuracy of the data was the length of time between Census Day and the PES Interview Day. Due to the COVID19 pandemic, some activities in the PES schedule— including the Person Interview operation—had to be delayed. The expanded interval between Census Day and Interview Day increased the potential for recall bias. Recall bias is the phenomenon where, as the interview occurs further from the target event, the respondent may be less likely to remember specific details about the event. These could be details such as specific move dates, the place where they were staying most of the time, or even the birth date of a child. If a respondent couldn’t pinpoint whether events like these occurred before, on, or after Census Day, this could create error in the response data.

Further, this phenomenon could become exacerbated when a proxy responded for the household.

Emphasis added.

 

[RVonse]

We can also observe that there are a total of 6 discrepancies which is a number that is in common with the conspiracy theory that there were an extra 6 Democrats and missing 6 Republicans.

​

​

It should be clearly noted that Musk never called this a conspiracy theory. He instead carefully calls this an incentive why members of our political class will not change anything at the US southern border. They simply have an incentive not to fix our borders and to imply that anyone is a nut case is extremely insulting and disingenuous on your part..

 

[RVonse]

What makes the discussion even further laughable:

Facing Nuclear War |

www.paulcraigroberts.org

Because on the one hand, our political institutions have become so paranoid and worried Russia will somehow invade our western hemisphere that we risk Armageddon with a false warning of incoming missiles.

Then our insane political establishment decides we must do absolutely nothing with the certain and continuing invasion on our southern border!


This woman seems to have done thorough research on the subject of nuclear war at present and the current state of this is both frightening and sobering. Because no matter who launches first all life will end.

One really does have to wonder how we become so obsessed with Russia invading way over in EuropeZ when we don't give 2 shits about our own border??

 

[laughing dog]

One wonders why anyone would pay attention to anything Paul Craig Roberts says, let alone give it any credence.

 

[Shadowy Man]

We can also observe that there are a total of 6 discrepancies which is a number that is in common with the conspiracy theory that there were an extra 6 Democrats and missing 6 Republicans.

​

​

It should be clearly noted that Musk never called this a conspiracy theory. He instead carefully calls this an incentive why members of our political class will not change anything at the US southern border. They simply have an incentive not to fix our borders and to imply that anyone is a nut case is extremely insulting and disingenuous on your part..

So, the interesting question would be this: if the Republicans had not killed the bi-partisan border security bill negotiated in the Senate, would the Democrats have killed it? According to your logic, they should have.

 

[Don2 (Don1 Revised)]

Yikes! Okay, so I have to make some minor corrections. This is because there are also people overseas, such as federal employees and certain military personnel and their dependents who may not have been included in the census. Those people's counts are not affected by the net coverage error from the Census Bureau. So, they have to be added to the tables AFTER adjusting by percentages of "undercounts" and "overcounts." Also, the total is only about 350K people across all 50 states which is less than half of one divisor distributed across 50 states. So, the impact is expected to be negligible. It could conceivably make a difference if any state has way more overseas people than another as a ratio interacting with two such states that have very close on-the-edge rounding going on. However, I checked the numbers and it makes no impact on the discussion or results. The counts of discrepancies for each method comparison remain the same.

In the interest of transparency and showing correct numbers, even if the conclusions are the same, I have recomputed the numbers and made Version 2.0 for the previous post.

At the end, you will also find an addendum that has a comparison of similar work.


Version 2.0:

Introduction

I went ahead and made a program to run the Hungtinton-Hill method of apportionment. This is the standard that Congress uses and is described in a previous post. The numbers I got back reproduce the delegate allocations that are present in Congress which gives confidence in the programming. This method I call Original Huntington-Hill.​

​

I also ran 3 other methods.​

​

Original Intuitive simply takes the divisor of the census population divided by 435. This is about 760367 761169. I do not know why there are other numbers out there, but this one is correct and it subtracts the DC population as I had attempted to do before but thought it was weird because it did not reproduce other numbers I was seeing at the time. This one, 760367, happens to also be documented which I found and confirms it is the right one. This is a number that I had encountered before in documentation that I had mentioned in post 716. One must subtract out the DC Population and then add overseas counts for each state to derive this divisor. So in this method, if the decimal part of the state's number after dividing by the divisor is .5 or higher, it is rounded up. Originally Congress was also looking for the highest decimal remainders when something went wrong, like not enough delegates were assigned or too many in order to prioritize the last few delegate assignments. I did not bother to implement that.​

​

Next, I took the percentages of just the so-called statistically significant undercounts and overcounts from the report and used those to create an Adjusted Population for those few states. I also added the overseas counts to the adjusted population. This is also summed up. Strangely, it is quite a bit more than the original population even though the documentation provided by the Census Bureau says they are similar. Therefore, I do not trust their numbers. In any case, I copied the Original Populations of states into this column for all states that had no significant difference. Their new sum, 331508865 331857563, was used to create a new divisor (by dividing by 435).​

​

Computing the adjusted populations and summing, and dividing, allowed for two more methods.​

​

Adjusted Huntington-Hill is running the standard Congressional method of apportionment but on the Adjusted Populations. Then, Adjusted Intuitive takes the Adjusted Populations and divides them by the new divisor. If the decimal part is above .5, it rounds up. As before, there is nothing additional.​

​

Any method that is discrepant in the resulting delegate count appears in bold font for the entire row where it is. This way we can look manually at discrepancies and decide what happened. Possibly what ought to be and count up legitimate results.​

Results

Original Population	331108434
Adjusted Population	331857563
Original divisor	761168.813793103
Adjusted divisor	762890.949425287

State	Original Population	Adjusted Population	Original Huntington-Hill	Original Intuitive	Adjusted Huntington Hill	Adjusted Intuitive
California	39576757	39576757	52	52 (51.994717)	52	52 (51.877345)
Texas	29183290	29753838	38	38 (38.340102)	39	39 (39.001430)
Florida	21570527	22347080	28	28 (28.338690)	29	29 (29.292627)
New York	20215751	19543938	26	27 (26.558827)	26	26 (25.618259)
Pennsylvania	13011844	13011844	17	17 (17.094557)	17	17 (17.055968)
Illinois	12822739	13080218	17	17 (16.846117)	17	17 (17.145593)
Ohio	11808848	11635617	15	16 (15.514099)	15	15 (15.252006)
Georgia	10725274	10725274	14	14 (14.090533)	14	14 (14.058725)
North Carolina	10453948	10453948	14	14 (13.734073)	14	14 (13.703070)
Michigan	10084442	10084442	13	13 (13.248627)	13	13 (13.218720)
New Jersey	9294493	9294493	12	12 (12.210817)	12	12 (12.183252)
Virginia	8654542	8654542	11	11 (11.370069)	11	11 (11.344403)
Washington	7715946	7715946	10	10 (10.136971)	10	10 (10.114088)
Arizona	7158923	7158923	9	9 (9.405171)	9	9 (9.383940)
Tennessee	6916897	7263818	9	9 (9.087205)	9	10 (9.521437)
Massachusetts	7033469	6879449	9	9 (9.240354)	9	9 (9.017605)
Indiana	6790280	6790280	9	9 (8.920859)	9	9 (8.900722)
Missouri	6160281	6160281	8	8 (8.093186)	8	8 (8.074917)
Maryland	6185278	6185278	8	8 (8.126027)	8	8 (8.107683)
Wisconsin	5897473	5897473	8	8 (7.747917)	8	8 (7.730427)
Colorado	5782171	5782171	8	8 (7.596437)	8	8 (7.579289)
Minnesota	5709752	5498726	8	8 (7.501295)	7	7 (7.207748)
South Carolina	5124712	5124712	7	7 (6.732688)	7	7 (6.717490)
Alabama	5030053	5030053	7	7 (6.608328)	7	7 (6.593410)
Louisiana	4661468	4661468	6	6 (6.124092)	6	6 (6.110268)
Kentucky	4509342	4509342	6	6 (5.924234)	6	6 (5.910861)
Oregon	4241500	4241500	6	6 (5.572351)	6	6 (5.559772)
Oklahoma	3963516	3963516	5	5 (5.207144)	5	5 (5.195390)
Connecticut	3608298	3608298	5	5 (4.740470)	5	5 (4.729769)
Utah	3275252	3192656	4	4 (4.302925)	4	4 (4.184944)
Iowa	3192406	3192406	4	4 (4.194084)	4	4 (4.184616)
Nevada	3108462	3108462	4	4 (4.083801)	4	4 (4.074582)
Arkansas	3013756	3173593	4	4 (3.959379)	4	4 (4.159956)
Kansas	2940865	2940865	4	4 (3.863617)	4	4 (3.854896)
Mississippi	2963914	3090839	4	4 (3.893898)	4	4 (4.051482)
New Mexico	2120220	2120220	3	3 (2.785479)	3	3 (2.779191)
Nebraska	1963333	1963333	3	3 (2.579366)	3	3 (2.573543)
Idaho	1841377	1841377	2	2 (2.419144)	2	2 (2.413683)
West Virginia	1795045	1795045	2	2 (2.358274)	2	2 (2.352951)
Hawaii	1460137	1367607	2	2 (1.918283)	2	2 (1.792664)
New Hampshire	1379089	1379089	2	2 (1.811804)	2	2 (1.807714)
Maine	1363582	1363582	2	2 (1.791432)	2	2 (1.787388)
Montana	1085407	1085407	2	1 (1.425974)	2	1 (1.422755)
Rhode Island	1098163	1045409	2	1 (1.442733)	1	1 (1.370326)
Delaware	990837	939673	1	1 (1.301731)	1	1 (1.231727)
South Dakota	887770	887770	1	1 (1.166325)	1	1 (1.163692)
North Dakota	779702	779702	1	1 (1.024348)	1	1 (1.022036)
Alaska	736081	736081	1	1 (0.967040)	1	1 (0.964857)
Vermont	643503	643503	1	1 (0.845414)	1	1 (0.843506)
Wyoming	577719	577719	1	1 (0.758989)	1	1 (0.757276)

Discussion

There are 8 rows bolded which indicates across 4 methods there are a maximum 8 discrepancies in comparing any two. However, comparing any of one method against the original may produce much less than 8 differences such as 3 or 4. Here I do 3 comparisons of methods.​

​

1. Original Huntington-Hill (OHH) vs Original Intuitive (OI)​

​

If you focus on the bolded rows but then scan downward along these two columns, you will find 4 discrepancies:​

State	OHH	OI
New York	26	27 (26.558827)
Ohio	15	16 (15.514099)
Montana	2	1 (1.425974)
Rhode Island	2	1 (1.442733)

​

The standard method deducts 1 from both New York and Ohio as compared to the intuitive method. Meanwhile, the standard method adds one to both Montana and Rhode Island as compared to the intuitive method. You may observe that the decimal part of the delegate is in the .4 range for both Montana and Rhode Island, but the decimal part for both New York and Ohio is in the .5 range. This may be surprising, but as noted earlier the standard method we use is biased in favor of smaller states.​

​

The purpose of this comparison was to illustrate how our current process works in contrast to intuition.​

​

2. Original Huntington-Hill (OHH) vs Adjusted Huntington-Hill (AHH)​

​

If you focus on the bolded rows but then scan downward along these two columns, you will find 4 discrepancies:​

State	OHH	AHH
Texas	38	39
Florida	28	29
Minnesota	8	7
Rhode Island	2	1

In the adjusted version, Florida gains 1, and Texas gains 1. Minnesota and Rhode Island both lose 1.​

​

The purpose of this comparison is to illustrate how the actual process would work if the adjusted numbers were actually used in reality.​

​

3. Original Intuitive (OI) vs Adjusted Intuitive (AI)​

​

If you focus on the bolded rows but then scan downward along these two columns, you will find 6 discrepancies:​

State	OI	AI
Texas	38	39
Florida	28	29
New York	27	26
Ohio	16	15
Tennessee	9	10
Minnesota	8	7

This comparison is most interesting because the lack of knowledge about the exact process of apportionment might lead many people to make this comparison. Even in discussions here, we observe people using the divisor as I also initially thought to do. So I believe we can hypothesize that this is how the conspiracy theory started. We can also observe that there are a total of 6 discrepancies which is a number that is in common with the conspiracy theory that there were an extra 6 Democrats and missing 6 Republicans.​

​

What we do observe here is 3 states that are typically red and 3 that may be blue (is Ohio purple?) being impacted, with red states negatively and "blue" states positively. This does not necessarily translate to 3 Republicans and 3 Democrats.​

​

And as noted, the Post-Enumeration Survey results themselves are of questionable quality. Whether they are better than the census is an open question.​

Conclusion

None of the methods show an extra 6 Democrats and missing 6 Republicans.​

​

A comparison of intuitive methods (that are not used in the process) shows 3 typically red state gains, 1 purple/blue state gain, and 2 blue state gains. This comparison may have snowballed into a more dramatic version magnifying the difference and assuming partisan results based on "blue" or "red" states. In reality, following apportionment, there is redistricting and voting and those results may differ from predictions based on state gains or losses.​

​

In any case, the intuitive method is not used, but instead the Huntington-Hill method of apportionment. In that scenario, which is the actual Congressional process, there are fewer discrepancies with 2 red-state gains and 2 blue-state losses.​

​

Those also do not necessarily translate to two Republicans and two Democrats not merely because of the process following apportionment, but also because the survey results that the adjustments are based on are not necessarily improvements to the census.​

Addendum

Here is a website that also does these types of computations. Its author seems to have political bias, unsure... Some of their results are different with the most egregious difference being Florida which they say would have gained an additional two seats. Therefore, it is worth it to delve into computation differences for this state. We can try to discover flaws in my or their computations.​

​

They report Florida's numbers as the following:​

State	Original Population	Adjusted Population
Florida	21,870,527	22,631,621

​

I have computed Florida's census plus overseas count as 21,570,527. So these numbers differ by 300K. If you perform a Google search of "Florida 21,870,527" you will see they are the only site with this number describing its population, but if you do a Google search of "Florida 21,570,527" you will find many reputable sites listing the value with 5, not 8 as the third digit. Because all the other digits are exactly the same, it appears to be a typo.​

​

My computation comes from the census page which gives 21538187 plus the overseas count of 32,340 which I obtained here.​

​

21538187 + 32,340 = 21,570,527 not 21,870,527​

​

Let us do three more sanity checks: If I had put in the wrong state population for Florida, off by 300K, then I would not have been able to sum up all the state populations, divide by 435, and reproduce the divisor of 761169. Second, all the other counts I computed for these 13 other states in the table in Results are exactly the same as the 13 other states' original populations on their website. Finally, they are claiming that this produces an error of - 761,094 for Florida but that is less than the divisor. So how could a change less than the divisor result in an error of 2 delegates?​

​

I next will try to reproduce their second number (Adjusted Population).​

​

The way that the undercount and overcount rates are imagined intuitively is not really how they are computed. I will first show the simple, straightforward, imagined way they might be used.​

​

Note the undercount rate is 3.48% for Florida. So one may imagine that all one has to do is multiply the original census number by 1.0348 to get an adjusted census number and following that add the overseas count for Florida.​

​

However, they actually added the overseas count (plus the erroneous 300K padding) first and then applied the intuitive correction factor.​

​

21,870,527 x 1.0348 = 22,631,621.3396 ~= 22,631,621​

​

It turns out that this is wrong for 3 reasons: (a). the correction should not be applied after adding overseas counts because this will magnify the effect, (b). they inadvertently padded Florida by 300K people, and (c) the intuitive way to apply the correction factor is incorrect.​

​

I will now show how I computed Florida's numbers.​

​

State	Original Population	Adjusted Population
Florida	21,570,527	22,347,080

​

I have already explained how I computed the original population which is the census count for Florida plus the overseas count:​

21538187 + 32,340 = 21,570,527​

​

Next, note how the Census Bureau defines the so-called error rates that I posted earlier:​

A net coverage error rate is the difference between the census count and the PES estimate of the number of people in the United States expressed as a percentage of the PES estimate.

​

"a net coverage error rate" means an undercount rate or an overcount rate. The mention of PES estimate twice may be confusing, but it's the Adjusted Census count. Let's define some quick variables:​

e = net coverage error rate, the undercount rate or overcount rate​

NP = new population, adjusted census count based on the alleged error​

OP = old population from the census​

​

The wrong, but intuitive, way to think of it was​

NP = OP x (1 - e); so in the case of Florida, there was a correction factor of (1 - (-3.48%)) or (1.0348)​

​

BUT that isn't what they said. They are saying:​

e = (OP - NP) / NP​

​

The new population is the baseline of comparison, not the old population.​

​

Solve this for NP, you get:​

NP = OP / (1 + e)​

​

For Florida, NP = 21538187 / (1 - 3.48%) = 21538187 / .9652 ~= 22,314,739.95 ~= 22,314,740​

​

But lastly, one needs to add the overseas count​

​

22,314,740 + 32,340 = 22,347,080​

​

​

​

 

[Don2 (Don1 Revised)]

We can also observe that there are a total of 6 discrepancies which is a number that is in common with the conspiracy theory that there were an extra 6 Democrats and missing 6 Republicans.

​

​

It should be clearly noted that Musk never called this a conspiracy theory.

The discussion has morphed from illegal immigrants being far too many causing mayhem and crimes in red states to illegal immigrants being used in census counts and looking the other way while republicans send them off in planes to the places they say they don't want them because that will increase the census counts TO FINALLY a discussion of raw census counts that pretty much ignores the whole illegal immigrant issue but is still alleged to be unfair because of GIANT differences in census vs error adjustments that turn out to be smaller differences with very disputable, low-quality things being called errors. So, essentially, what you are talking about is different but the same theme.

He instead carefully calls this an incentive why members of our political class will not change anything at the US southern border. They simply have an incentive not to fix our borders and to imply that anyone is a nut case is extremely insulting and disingenuous on your part..

Do you think math implies people are nutcases?

 

[TomC]

Then our insane political establishment decides we must do absolutely nothing with the certain and continuing invasion on our southern border!

There's a simple solution. Stop giving millions of jobs to undocumented employees.
No budget busting Wall necessary. No massive ramping up of border guard budgets. Start prosecuting illegal employers.
Tom

 

[Don2 (Don1 Revised)]

Sorry, I just kind of dropped off there with the addendum. To continue...

That website states the following summary:

The following map shows how the 2020 Congressional Apportionment would have changed if each State’s total population were accurately reported by the 2020 Census. California, Illinois, Ohio, Michigan, West Virginia, Pennsylvania, and New York would each still have lost a seat, but Minnesota and Rhode Island would each have lost a seat as well. North Carolina, Montana, and Oregon would each still have gained a seat, but Texas and Florida would each have gained three seats instead of two and one, respectively, while Colorado would not have gained any.

Emphasis is added to the portions they are describing as different if adjustments were to be applied. They are counting 5 delegate changes or discrepancies. Minnesota -1, Rhode Island -1, Texas +1, Florida +2, Colorado -1. If you look at the tables, you can see with adjusted numbers Colorado has an edge case of 7.57... delegates. Because they padded Florida numbers with 300K people, Florida ended up winning an extra delegate it ought not to have won and it takes it from the edge case of Colorado. So, if you put that back with Colorado, there will only be 4 discrepancies: Minnesota -1, Rhode Island -1, Texas +1, Florida +1, which is the same result as I had gotten in the AHH vs OHH comparison. Even though they may have run the error correction after adding overseas counts, those did not impact their conclusions. Their primary issue was the padding of 300K.

 

[scombrid]

One wonders why anyone would pay attention to anything Paul Craig Roberts says, let alone give it any credence.

Some people state that Alex Jones knows and states the "real truth".

 

[Emily Lake]

If there is an accurate enough account to show the census undercounted why don’t we use *that* count instead of the census? Am I missing something here? Sorry, I haven’t followed all the posts in this thread

Because the current method is to reapportion based on the US Census. Reapportionment only occurs every 10 years, when a new Census is done.

There's good sense in not having reapportionment done too frequently - it would be very unstable. Hypothetically, we could do a good estimate every year, and apportion as needed... but that would be a bad idea. Apportionment affects the districting within states, and those districts are what the representatives represent (theoretically anyway). The districts are supposed to be drawn such that they're all roughly equal in terms of the number of people being represented. If it happened every year, it would be constantly changing, we'd find ourselves having to fire the occasional congresscritter, or add new ones without sufficient preparation. It would be more problem than it's worth.

On the other hand... given the rate of movement between states that we see in modern times, as well as the impacts of immigration and emigration... I think that every 10 years is probably too long.

 

[Emily Lake]

...
I was under the impression it was 750,000.
...

You might think that you take the total population and divide it by 435, the set number of reps since 1913. This would give about 330M / 435 ~= 761K. However, it doesn't work that way. It might even be close most of the time. One might go ahead and imagine apportionment methods, but it's actually more complicated than intuitive arithmetic and the process has a history that is centuries long.

Problems began to appear early with intuitive methods. Like rounding off the numbers to the nearest whole number. Divide 19 reps among 3 states each having 6.33 delegates. Do you round up, round down/truncate? If you round up, they are all 6, not 19, but there are more complicated scenarios that you can google which gave rise to very insane stuff.

First, you need to meet a Constitutional mandate that each state has to have at least 1 rep. So, all these reps are allocated in the complicated apportionment process. That leaves 385 reps to be distributed among the 50 states in some way based on populations. They do this 1 rep allocation at a time. They assign each state a priority which is the ratio of the state population (P) to the geometric mean of its assigned reps so far in the process and the number it would get if it got this one (n+1). So priority = P / sqt(n * (n+1)). In the 51st round, California will take that spot because all denominators are the same for all states and California has the highest P. Texas might be the next one since California's denominator has increased dramatically in comparison to its last one. The process continues until all 435 spots are assigned.

This method is called the Huntington-Hill method and was alleged to be unbiased when it was implemented in the 40s. However, analysis since then has shown it is biased toward smaller states.

That all said, I'm not going to go through that process to try to prove anything. That would be a bit of work to do.

Suppose we tried to assess this imperfectly by examining the intuitive ways to do apportionment. The 2020 US Census population was 331,464,948. Divided among 435 delegates, we'd get an average of 761,988. I've seen different figures near this floating around the Internet so let's hope we do not observe numbers on the edge.

Suppose we did look at Arkansas. Its 2020 census population was 3,011,524. The midpoint of the undercount confidence interval is 151,781. So add the numbers together and you get 3,163,305. Now, divide each by 761,988. You go from about 3.95 delegates to about 4.15 delegates. Both of these easily round to 4. And if you check the apportionments for Arkansas, you will find that they do have 4 currently. There was no loss of representation. No missing Republican congress critter.

Florida is supposed to be the questionable one. 21,538,187. The midpoint of the confidence interval for undercount is 3.48% which I will convert to people: 749,529. If we add that midpoint of interval to the census population, we get 22,287,716. The original count divided by the 761K gives 28.27 delegates. The adjusted count divided by that gives 29.25. So intuitively we would think Florida missed addition of a delegate, provided that we accept the estimated undercount as a true undercount. Further, if we look up Florida info, we can see they currently only have 28 delegates. So, it's good reason to believe they'd probably have 29 if they had a different census count, though we cannot say if it is valid without understanding the root cause for the discrepancy between census and sampling.

Now, apportionment is only the first step. Because the count of delegates changed in this hypothetical, they'd next have to do redistricting. We know there might be shenanigans there and so we cannot claim that the result of redistricting is necessarily fairly representative of what ought to be so far as addition of an unbiased district.

Given the extreme partisan gerrymandering in Florida, in all fairness the district ought to undo some of that...and likewise the new district ought to be reflective of changing populations in some way.

(assuming that there ought to be a new delegate).

Very informative