The Official 2020 Census Thread

[jimrtex]

WA is now 2% above its final 2010 rate, surpassing MI (1.3% above).

WA has made big gains over the past week, MI not.

But both states are far above the US average.

Soft NRFU in Washington has begun on July 30 or earlier in four ACO (all except Spokane for Washington east of the cascades).

Michigan had NRFU begun in two ACO last Wednesday.

It appears that NRFU has more of an effect in more urban areas. This might have something to do with the the use of PPE (masks and gloves and hand sanitizer). Their publicity videos show the enumerator knocking on a door and then stepping back off the porch. Do houses in Austria have front porches, and are they almost always two steps?

It also appears in the ads that many of the householders are then filling out the form themselves. If no one is home on the first NRFU attempt at contact, a paper form and Internet ID are left. This might also be an option even if someone is home, and a stranger appears at their home. Even though a census worker provoked the response, it is still counted as a self response.

In a smaller town, it is more likely that they are not total strangers. If the enumerator has a familiar last name, the conversation might establish the relationship with someone the enumerator went to school with, or who had dated their sibling, or worked with, or taught at the high school. They might be willing to go ahead and be interviewed.

Publicity might also be concentrated in larger cities. The head of the local ACO might go on the local TV station, or arrange an interview with a public affairs officer. The station might show the TV clips. In a more rural area, the local radio station won't get as much help.

And it might also take longer to get enumerators everywhere in a more widespread ACO. If an ACO is entirely in a single county (and the largest counties have several ACO) then you will have 100s of enumerators who can go anywhere. In a smaller county, the enumerators might be in handfuls, and any hiring shortfalls might be reflected in slower contact rates.

[Tender Branson]

WA is now 2% above its final 2010 rate, surpassing MI (1.3% above).

WA has made big gains over the past week, MI not.

But both states are far above the US average.

Their publicity videos show the enumerator knocking on a door and then stepping back off the porch. Do houses in Austria have front porches, and are they almost always two steps?

Depends.

Our houses are most of the time not comparable with US houses and there are no front porches in most of them like you see from the Southern US for example.

Most just have a smaller or larger driveway in front of the door.

But it doesn't matter anyway in Austria's case: we don't have any enumerators around in our Census, we have a register-based Census.

[Tender Branson]

Door-to-Door Visits Begin Nationwide for 2020 Census

Census Takers to Follow Up With About 56 Million Households Nationwide

AUG. 11, 2020 — This week, the U.S. Census Bureau began following up with households nationwide that have not yet responded to the 2020 Census. Based on the current self-response rate of 63.3%, the Census Bureau estimates it will need to visit about 56 million addresses to collect responses in person. Up to 500,000 census takers across the country will go door to door to assist people in responding to the 2020 Census.

Census takers began following up with households on July 16 in a limited number of areas and added additional areas each week thereafter. Starting Aug. 9, all remaining offices began following up with households nationwide. Census takers have completed training on social distancing and safety protocols, will follow local public health guidelines, and will be required to wear a face mask when conducting follow-up visits.

“America has answered the call and most households responded to the census online, by phone or by mail,” said Census Bureau Director Dr. Steven Dillingham. “To ensure a complete and accurate count, we must now go door to door to count all of the households we have not heard back from. During this phase, you can still self-respond online (at my2020census.gov), by phone (at 844-330-2020), or by mailing your completed questionnaire.”

The Nonresponse Followup (NRFU) operation is the final stage of conducting the once-a-decade population count of everyone living in the United States. Households can still respond now by responding online at 2020census.gov, by phone at 844-330-2020, or by completing and mailing back the paper questionnaire they received. Households can respond online or by phone in one of 13 languages and find assistance in many more. Those that respond will not need to be visited to obtain their census response.

What Households Can Expect

In most cases, census workers will make up to six attempts at each housing unit address to count possible residents. This includes leaving notification of the attempted visit on the door. The notification will include reminder information on how to respond online, by paper or by phone. In addition, census workers may try to reach the household by phone to conduct the interview.

Census takers will go to great lengths to ensure that no one is missed in the census. After exhausting their efforts to do an in-person interview with a resident of an occupied housing unit, they will seek out proxy sources — a neighbor, a rental agent, a building manager or some other knowledgeable person familiar with the housing unit — to obtain as much basic information about the occupants as they can.

Census takers are hired from local communities. All census takers speak English, and many are bilingual. If a census taker does not speak the householder’s language, the household may request a return visit from a census taker who does. Census takers will also have materials on hand to help identify the household’s language.

https://www.census.gov/newsroom/press-releases/2020/door-to-door-visits-begin-nationwide.html

112 households to visit on average for every census worker in the next 50 days, for initial contact.

That's ca. 3 households on average per census worker per work-day (but I guess many will work Saturday and Sunday too).

[jimrtex]

Door-to-Door Visits Begin Nationwide for 2020 Census

Census Takers to Follow Up With About 56 Million Households Nationwide

AUG. 11, 2020 — This week, the U.S. Census Bureau began following up with households nationwide that have not yet responded to the 2020 Census. Based on the current self-response rate of 63.3%, the Census Bureau estimates it will need to visit about 56 million addresses to collect responses in person. Up to 500,000 census takers across the country will go door to door to assist people in responding to the 2020 Census.

Census takers began following up with households on July 16 in a limited number of areas and added additional areas each week thereafter. Starting Aug. 9, all remaining offices began following up with households nationwide. Census takers have completed training on social distancing and safety protocols, will follow local public health guidelines, and will be required to wear a face mask when conducting follow-up visits.

“America has answered the call and most households responded to the census online, by phone or by mail,” said Census Bureau Director Dr. Steven Dillingham. “To ensure a complete and accurate count, we must now go door to door to count all of the households we have not heard back from. During this phase, you can still self-respond online (at my2020census.gov), by phone (at 844-330-2020), or by mailing your completed questionnaire.”

The Nonresponse Followup (NRFU) operation is the final stage of conducting the once-a-decade population count of everyone living in the United States. Households can still respond now by responding online at 2020census.gov, by phone at 844-330-2020, or by completing and mailing back the paper questionnaire they received. Households can respond online or by phone in one of 13 languages and find assistance in many more. Those that respond will not need to be visited to obtain their census response.

What Households Can Expect

In most cases, census workers will make up to six attempts at each housing unit address to count possible residents. This includes leaving notification of the attempted visit on the door. The notification will include reminder information on how to respond online, by paper or by phone. In addition, census workers may try to reach the household by phone to conduct the interview.

Census takers will go to great lengths to ensure that no one is missed in the census. After exhausting their efforts to do an in-person interview with a resident of an occupied housing unit, they will seek out proxy sources — a neighbor, a rental agent, a building manager or some other knowledgeable person familiar with the housing unit — to obtain as much basic information about the occupants as they can.

Census takers are hired from local communities. All census takers speak English, and many are bilingual. If a census taker does not speak the householder’s language, the household may request a return visit from a census taker who does. Census takers will also have materials on hand to help identify the household’s language.

https://www.census.gov/newsroom/press-releases/2020/door-to-door-visits-begin-nationwide.html

112 households to visit on average for every census worker in the next 50 days, for initial contact.

That's ca. 3 households on average per census worker per work-day (but I guess many will work Saturday and Sunday too).

The 500,000 might be the total of temporary workers, and would include supervisors and above. The first-level supervisors also work from home. The Census also does real-time auditing.

Each worker has a hand-held device or laptop, which has their work assignments, travel directions, etc. It also has GPS location. So the Census Bureau can check whether they were really at 123 Oak Street. There are a number of other metrics, such as a large percentage of vacancies, or deletes, or time between visits. They can also check for a match with administrative records. Depending on the apparent quality of the work, a sample will be rechecked. If everything looks pretty reasonable the sampling rate may be quite low. If it turns out that some reported visits were totally bogus, it might result in a complete redo for visits that had been done by that former worker.

Up to 6 visits may be made. Clearly you don't want that first visit to happen on the 50th day. I'd think they would try to get the first visit by perhaps two weeks. Assuming 10 work-days, that is up around 15 per day.

My understanding is that workers request a work assignment each day, they want to (or can work). X% will get COVID-19, have an auto accident, have hip fracture from stepping backwards off a porch, becoming disenchanted, called back to work at their regular employer, etc.

It may also be that the number of temporary workers will taper off. You don't want to plan on your 6th visit being on the 50th day.

Their initial planning assumed a 60% self-response rate.

Of course the delay may make it harder to track down people who might have moved or escaped New York or other large cities.

[jimrtex]

WA is now 2% above its final 2010 rate, surpassing MI (1.3% above).

WA has made big gains over the past week, MI not.

But both states are far above the US average.

Their publicity videos show the enumerator knocking on a door and then stepping back off the porch. Do houses in Austria have front porches, and are they almost always two steps?

Depends.

Our houses are most of the time not comparable with US houses and there are no front porches in most of them like you see from the Southern US for example.

Most just have a smaller or larger driveway in front of the door.

But it doesn't matter anyway in Austria's case: we don't have any enumerators around in our Census, we have a register-based Census.

I was just curious. They were explicitly showing the census takers wearing masks and practicing social distancing. But they had to knock on the door, and then back off. It seemed like they were backing down and standing with an upper foot on the first step. This also permitted the householder to come out onto the porch. This made a good visual of social distancing, while clearly making contact at a door (enumerators are not permitted to enter houses).

I wasn't necessarily thinking of a sitting porch. In many cases it is not desirable to build on grade since that makes you more vulnerable to flooding or just groundwater seeping up through the floor.

[jimrtex]

Twenty-nine-day averages (since 7/13/2010)

Total 1.34% (0.046% per day)
Internet 1.13% (0.039% per day)
Total 0.21% (0.007% per day)

Michigan reached its 2010 response rate on June 15.
Washington reached its 2010 response rate on June 30.
Nevada reached its 2010 response rate on July 20.
Kentucky reached its 2010 response rate on July 27.
Idaho reached its 2010 response rate on August 1.

Colorado is projected to reach its 2010 response rate on August 21 (gap is 0.5%).
Virginia is projected to reach its 2010 response rate on August 31.
Maine is projected to reach its 2010 response rate on September 2.
Oregon is projected to reach its 2010 response rate on September 2.

11 states in remainder of September
10 states in October
11 states in November
6 states + DC in December
3 states in 2010

Green-7: Reached (5): MI, WA, NV, KY, ID
Green-6: Next 14 days (1): CO
Green-5: Next 28 days (3): VA,OR, ME
Red-3: September (11): AZ, NH, UT, MD, OH, AL, CT, IN, MN, IL, KS
Red-4: October (10): NJ, NE, CA, MA, SD, HI, VT, FL, OK, WV
Red-5: November (11): GA, DE, LA, NM, MS, MO, AR, PA, AR, NY, RI
Red-6: December (7): TX, TN, WI, IA, WY, NC, DC
Red-7: 2021 (3): ND, MT, SC

[jimrtex]

Thirty-day averages (since 7/13/2010)

Total 1.41% (0.047% per day)
Internet 1.19% (0.040% per day)
Total 0.22% (0.007% per day)

Michigan reached its 2010 response rate on June 15.
Washington reached its 2010 response rate on June 30.
Nevada reached its 2010 response rate on July 20.
Kentucky reached its 2010 response rate on July 27.
Idaho reached its 2010 response rate on August 1.

Colorado is projected to reach its 2010 response rate on August 21 (gap is 0.4%).
Virginia is projected to reach its 2010 response rate on August 29.
Oregon is projected to reach its 2010 response rate on September 1.
Maine is projected to reach its 2010 response rate on September 1.
Utah is projected to reach its 2010 response rate on September 6.
New Hampshire is projected to reach its 2010 response rate on September 6.
Arizona is projected to reach its 2010 response rate on September 6.

8 states in remainder of September
10 states in October
13 states in November
4 states + DC in December
3 states in 2010

Green-7: Reached (5): MI, WA, NV, KY, ID
Green-6: Next 14 days (1): CO
Green-5: Next 28 days (6): VA,OR, ME, UT, NH, AZ
Red-3: September (8): MD, OH, CT, AL, IN, MN, IL, KS
Red-4: October (10): NJ, MA, NE, CA, SD, HI, VT, FL, OK, LA
Red-5: November (13): WV, GA, AK, DE, NM, MS, MO, PA, AR, NY, WI, TN, TX
Red-6: December (5): DC, RI, IA, WY, NC
Red-7: 2021 (3): ND, MT, SC

[Tender Branson]

The population-rich states are all lagging far behind their final 2010 rates (except MI).

[jimrtex]

The population-rich states are all lagging far behind their final 2010 rates (except MI).

I received an interesting note from my apartment manager.

Under NRFU, enumerators first contact the manager of multi-unit housing to get a better inventory of units. I had thought that it was to help remove units which had been converted to other uses (storage, offices, exercise rooms, laundry rooms, etc.) or out of commission due to major renovations or significant damages (fire, flooding, etc.). They might also have inquired who tenants were as of April 1, and which units were vacant (5% to 10% vacancy is quite normal, as tenants move out, cleaning and updates occur, and new tenants probably aren't ready to move in on the first).

The company that owns my complex and about two dozen others, owns and manages about 10,000 apartment units. The note is technically from the managers of my complex, but it is clear that it was written by company HQ. Perhaps, an enumerator had contacted a complex manager, and the manager had asked the HQ how to respond.

The note says that a Census Bureau representative had visited "some of our apartment communities". It then explains that US Census law requires a multifamily landlord cooperate in one of two ways.

(1) Furnish the enumerator with name of occupants. The note then says "(WHICH WE WILL NOT DO)" Emphasis in the original.

(2) Permit an enumerator free ingress and egress to the premises.

The latter would be most significant when apartments have internal doors off a corridor, with a controlled outside entrance. But in cases where apartments have external entrances, they can control who is permitted on the property.

The note says that they will closely examine credentials of the enumerators and suggest tenants do the same to masked individuals wearing gloves so as to not leave fingerprints as they burglarize apartments.

It also says that enumerators will work between 9 am and 9 pm.

In an apartment complex, they might begin in the afternoon, and begin knocking on doors of non-responding apartments. There will be a few people home. (I suspect self-response in apartments is closer to 50% than 70%), they can then go back around after 5 or 6, and check on apartments where nobody was home earlier.

[jimrtex]

Thirty-one-day averages (since 7/13/2010)

Total 1.46% (0.047% per day)
Internet 1.24% (0.040% per day)
Total 0.22% (0.007% per day)

Michigan reached its 2010 response rate on June 15.
Washington reached its 2010 response rate on June 30.
Nevada reached its 2010 response rate on July 20.
Kentucky reached its 2010 response rate on July 27.
Idaho reached its 2010 response rate on August 1.

Colorado is projected to reach its 2010 response rate on August 18 (gap is 0.3%).
Oregon is projected to reach its 2010 response rate on August 30.
Virginia is projected to reach its 2010 response rate on August 30.
Utah is projected to reach its 2010 response rate on September 3.
Maine is projected to reach its 2010 response rate on September 3.
New Hampshire is projected to reach its 2010 response rate on September 8.
Arizona is projected to reach its 2010 response rate on September 8.

9 states in remainder of September
9 states in October
13 states in November
5 states + DC in December
2 states in 2021

Green-7: Reached (5): MI, WA, NV, KY, ID
Green-6: Next 14 days (1): CO
Green-5: Next 28 days (6): OR, VA,ME, UT, NH, AZ
Red-3: September (9): MD, OH, CT, AL, MN, IN, IL, KS, NJ
Red-4: October (9): CA, MA, NE, SD, HI, VT, FL, OK, WV
Red-5: November (13): DE, GA, LA, AK, NM, MS, MO, PA, AR, NY, TN, TX, RI
Red-6: December (6): WI, WY, DC, IA, NC, ND
Red-7: 2021 (2): SC, MT

[Tender Branson]

The Census Bureau is now going "all-in":

FOR IMMEDIATE RELEASE: FRIDAY, AUGUST 14, 2020

Census Bureau Adapts Operations to Ensure Everyone Is Counted

As the U.S. Census Bureau continues to monitor the impacts COVID-19 has on 2020 Census operations, changes to operations are being deployed to ensure the safety of staff and the public while maximizing the number of households that respond on their own to the 2020 Census.

“We are taking steps and adapting our operations to make sure everyone is counted, while keeping everyone safe,” said Census Bureau Director Dr. Steven Dillingham. “Our commitment to a complete and accurate 2020 Census is absolute. In this challenging environment, we are deploying these tactics to make sure we reach every household in every community. If you haven't responded, the time to respond is now! Responding to the 2020 Census online, on paper, by phone, or in person with a census taker, helps secure vital resources for your community."

As of today, over 63.5% of households have responded to the 2020 Census. People can still respond online, over the phone or by mail — all without having to meet a census taker.

The Census Bureau will mail an additional paper questionnaire to nonresponding households.

To encourage more households to respond on their own to the 2020 Census, the Census Bureau is contacting nonresponding households by mailing an additional paper questionnaire to some households that have yet to respond.

The Census Bureau is sending a seventh mailing, including a paper questionnaire, in late August to early September to the lowest-responding census tracts.

The Census Bureau is emailing households in low-responding areas.

The Census Bureau recently announced that households in low-responding areas would be receiving emails to encourage response to the 2020 Census.

The emails will go to all households that the Census Bureau has contact information for in census block groups with a response rate lower than 50%. This will include households who may have already responded. In total, the Census Bureau expects to email more than 20 million households in these low-responding areas. The email messages will come from 2020census@subscriptions.census.gov and will give recipients the option to opt out of receiving future messages.

The Census Bureau is using email addresses that households have provided in response to another Census Bureau program, or received from states (such as from their WIC, SNAP or TANF programs) or from a commercial list.

https://www.census.gov/newsroom/press-releases/2020/adapts-2020-operations.html

[jimrtex]

Thirty-two-day averages (since 7/13/2010)

Total 1.54% (0.048% per day)
Internet 1.31% (0.041% per day)
Total 0.23% (0.007% per day)

Michigan reached its 2010 response rate on June 15.
Washington reached its 2010 response rate on June 30.
Nevada reached its 2010 response rate on July 20.
Kentucky reached its 2010 response rate on July 27.
Idaho reached its 2010 response rate on August 1.

Colorado is projected to reach its 2010 response rate on August 20 (gap is 0.3%).
Oregon is projected to reach its 2010 response rate on August 26 (gap is 0.7%)
Virginia is projected to reach its 2010 response rate on August 28.
Utah is projected to reach its 2010 response rate on September 1.
Maine is projected to reach its 2010 response rate on September 3.
New Hampshire is projected to reach its 2010 response rate on September 5.
Arizona is projected to reach its 2010 response rate on September 6.
Maryland is projected to reach its 2010 response rate on September 9.

7 states in remainder of September
11 states in October
13 states in November
2 states + DC in December
4 states in 2021

Green-7: Reached (5): MI, WA, NV, KY, ID
Green-6: Next 14 days (1): CO, OR
Green-5: Next 28 days (6): VA,UT, ME, NH, AZ, MD
Red-3: September (7): OH, CT, AL, MN, IL, IN, KS
Red-4: October (11): NJ, CA, HI, MA, NE, SD, VT, FL, OK, DE, AK
Red-5: November (13): WV, GA, PA, LA, NM, MS, MO, AR, NY, TN, TX, WI, RI
Red-6: December (6): WY, DC, NC
Red-7: 2021 (4): IA, ND, MT, SC

[Tender Branson]

48 days left and with an average of 0.05% per day, the final rate will hit 66% once the Census Bureau ends self-response and all other operations on September 30.

So, 2/3 households in the US will have responded on their own - which is not bad considering the falling and falling response rates in political polls over the decades, anti-government sentiment etc. - and in line with the final rates of 2010 (= 66.5%) or 2000 (67.4%).

Obviously, 2010s and 2000s response-deadlines were far earlier in those years (basically in May).

[Tender Branson]

jimrtex has focused on states topping their final 2010 rates so far.

But what about states beating their final 2000 rates ?

There are only a few so far, because 2000 had mostly even better rates than 2010:

WA, ID, HI, KY, AL and D.C.

Note that final 2000 rates are only reported without decimals.

WA, ID (and with rounding KY) are the only states where the rate increased from census to census.

The light-green shaded states will reach their final 2000 rates, using a uniform 2.4 point gain until September 30.

The yellow states have a potential to reach their final 2000 rates, but would need to gain above-average in the final 1.5 months.

The red states have no chance whatsoever to reach their 2000 rates by the end of September.

[Tender Branson]

There is no such thing as a universal trend for the final 1.5 months though.

Some states will do better than the national trend, some worse.

CO for example could still reach their final 2000 rate, because it had pretty good returns lately.

Other states even in light green may not reach them.

WA is fairly interesting, because it could reach 72% by the end of September - up 6% from 2000.

Will be interesting to see if WA will also see its population growth reported above their 10-year estimates or not as a result and states like MT lower ...

From what I can see, lower response rates had no real impact on reported population growth rates: census takers will make sure to count the remaining population or the Bureau will use imputation measures.

[jimrtex]

48 days left and with an average of 0.05% per day, the final rate will hit 66% once the Census Bureau ends self-response and all other operations on September 30.

So, 2/3 households in the US will have responded on their own - which is not bad considering the falling and falling response rates in political polls over the decades, anti-government sentiment etc. - and in line with the final rates of 2010 (= 66.5%) or 2000 (67.4%).

Obviously, 2010s and 2000s response-deadlines were far earlier in those years (basically in May).

In 2010 at least, self responses were recorded through September. Of course, it was more difficult to self-respond since it usually required a paper form (There were a small number of Internet responses in 2010, but that was primarily an experiment).

So you might have a few people who had stashed away the form they received in March, plus a few who turned in a form that had been left in NRFU.

In 2020, it appears that large numbers are being prompted to self-respond by NRFU. This may be concentrated in urban areas where households may be disinclined to interact with an enumerator, but willing to self-respond via the Internet (recent responses are strongly Internet concentrated). Those who would prefer a paper form, might be willing to simply let the enumerator to fill in the form for them. The enumerator may seem less like a stranger who is trying to pry information from them.

It really isn't anti-government sentiment, but disaffection from the community. With smaller families and many people single and/or childless, there is less connection through the community schools. Greater mobility and mass media (TV then Internet) means that people can commute long distances to work, and when they come home close the door and zonk out on television or the Internet.

Previously, there would be more neighbor to neighbor interaction, as people would sit out on the front porch and talk. They were living in smaller worlds. There was no Europe or Africa, let alone Afghanistan. 50 miles away might have been the furthest someone had traveled from home, comparable to visiting Hawaii or Tahiti or Machu Picchu.

You may be right about reaching 66.5% this year. Some of the top counties last week (0.8% or so) were large counties with soft NRFU.

Washington: King, Pierce, Snohomish.
Illinois: Cook, Lake
New York: Queens
Massachusetts: Middlesex
Maryland: Montgomery
Virginia: Fairfax

NRFU is just starting in the three most populous states: California, Texas, and Florida. NYC was just getting started (on August 6). It takes a bit of time to ramp up, and for people to begin to self-respond. There is an extremely large share of multi-family where the first contact was with the manager rather than the tenant.

[jimrtex]

There is no such thing as a universal trend for the final 1.5 months though.

Some states will do better than the national trend, some worse.

CO for example could still reach their final 2000 rate, because it had pretty good returns lately.

Other states even in light green may not reach them.

WA is fairly interesting, because it could reach 72% by the end of September - up 6% from 2000.

Will be interesting to see if WA will also see its population growth reported above their 10-year estimates or not as a result and states like MT lower ...

From what I can see, lower response rates had no real impact on reported population growth rates: census takers will make sure to count the remaining population or the Bureau will use imputation measures.

This article from Pew Research gives a quick and understandable explanation of imputation.

Imputation: Adding People to the Census

The magnitude won't really matter for drawing districts, but could be decisive in deciding the last seat apportionment.

[Frodo]

The Sunbelt (unsurprisingly) is being especially undercounted this year:

2020 Census 'emergency' threatens to leave out communities of color and rural Americans

At first, this outreach effort, postponed from the spring due to COVID-19, gave volunteers until Oct. 31 to contact these groups. But on Aug. 3, that deadline was moved up to Sept. 30 because, according to the census, that was the only way numbers could be tabulated in time to meet "our statutory deadline of Dec. 31, 2020, as required by law and directed by the Secretary of Commerce."

A review of the Census Bureau's Hard to Count map highlights areas of concern, including the many neighborhoods with heavily Latino and Black populations that have yet to respond.

Texas stands out, with a majority of its counties showing fewer than 50% of residents self-reporting their census information. Edwards County, hard by the Rio Grande River, is at just 14.8%. Other heavily undercounted areas include almost all of New Mexico; California's Central Valley; the southern half of Georgia; and the largely Native American Four Corners region where Colorado, Utah, New Mexico and Arizona meet.

"In 2010, there were undercounts of people of color and Native Americans on reservations, and it seems that now with even more households to visit in less time, the undercount risks being much worse," says Steven Romalewski, who is keeping a close eye on the map as part of the City University of New York’s Center for Urban Research at the Graduate Center. "It's worrisome."

[Tender Branson]

The Sunbelt (unsurprisingly) is being especially undercounted this year.

You can also phrase it differently:

The Sunbelt is not undercounted by the Census Bureau, it's the people living in the South who are lazy ass bastards themselves by not responding to multiple letters and questionnaires sent to them by the Census Bureau !

[Tender Branson]

The national rate has ticked up 0.2% to 63.8% today.

That's the biggest "daily" increase since a few months.

CO has reached its final 2010 response rate today, while UT has reached its final 2000 rate.

[jimrtex]

The Sunbelt (unsurprisingly) is being especially undercounted this year:

2020 Census 'emergency' threatens to leave out communities of color and rural Americans

At first, this outreach effort, postponed from the spring due to COVID-19, gave volunteers until Oct. 31 to contact these groups. But on Aug. 3, that deadline was moved up to Sept. 30 because, according to the census, that was the only way numbers could be tabulated in time to meet "our statutory deadline of Dec. 31, 2020, as required by law and directed by the Secretary of Commerce."

A review of the Census Bureau's Hard to Count map highlights areas of concern, including the many neighborhoods with heavily Latino and Black populations that have yet to respond.

Texas stands out, with a majority of its counties showing fewer than 50% of residents self-reporting their census information. Edwards County, hard by the Rio Grande River, is at just 14.8%. Other heavily undercounted areas include almost all of New Mexico; California's Central Valley; the southern half of Georgia; and the largely Native American Four Corners region where Colorado, Utah, New Mexico and Arizona meet.

"In 2010, there were undercounts of people of color and Native Americans on reservations, and it seems that now with even more households to visit in less time, the undercount risks being much worse," says Steven Romalewski, who is keeping a close eye on the map as part of the City University of New York’s Center for Urban Research at the Graduate Center. "It's worrisome."

It is 75 miles from Rocksprings to Del Rio, even further from Barksdale. This is further than from New York City to Trenton, NJ. To describe Edwards County as being hard on the Rio Grande (River. sic), would be like saying Brooklyn is hard on the Delaware River.

Edwards County has a large share of housing units not inhabited by resident households (1689 housing units, 686 households according to 2018 5-yr ACS). Many of the extras are likely seasonal (Hill County retreats) or deer leases. 96% of the county is Update/Leave and I suspect most don't have Internet.

[jimrtex]

Thirty-five-day averages (since 7/13/2010)

Total 1.77% (0.050% per day)
Internet 1.49% (0.043% per day)
Total 0.27% (0.008% per day)

Twenty-eight-day averages (since 7/20/2010)

Total 1.63% (0.058% per day)
Internet 1.39% (0.050% per day)
Total 0.24% (0.009% per day)

The last four weeks have been: 0.38%, 0.38%, 0.38%, and 0.49%. The week of 0.14% for July 20 to July 27 was aged off the four-week running total this week. That was during the last gap of Update/Leave. The last four weeks have been following the final postcard and the beginning of NRFU. NRFU is now nationwide, and we should see continued strength this coming week before it begins to taper off. That doesn't mean that NRFU is not collecting data - it is just that NRFU after a few weeks is no longer triggering self responses. This can be seen in Idaho and Maine, two of the earliest NRFU restarts.

Michigan reached its 2010 response rate on June 15.
Washington reached its 2010 response rate on June 30.
Nevada reached its 2010 response rate on July 20.
Kentucky reached its 2010 response rate on July 27.
Idaho reached its 2010 response rate on August 1.
Colorado reached its 2010 response rate on August 17.

Oregon is projected to reach its 2010 response rate on August 24 (gap is 0.5%)
Virginia is projected to reach its 2010 response rate on August 25 (gap is 0.5%)
Utah is projected to reach its 2010 response rate on August 29.
New Hampshire is projected to reach its 2010 response rate on August 31.
Maine is projected to reach its 2010 response rate on September 1.
Arizona is projected to reach its 2010 response rate on September 2.
Maryland is projected to reach its 2010 response rate on September 3.
Ohio is projected to reach its 2010 response rate on September 6.
Connecticut is projected to reach its 2010 response rate on September 7.
Alabama is projected to reach its 2010 response rate on September 12.
Indiana is projected to reach its 2010 response rate on September 14.
Minnesota is projected to reach its 2010 response rate on September 14.

7 states in remainder of September
17 states in October
3 states + DC in November
3 states in December
2 states in 2021

Green-7: Reached (6): MI, WA, NV, KY, ID, CO
Green-6: Next 14 days (3): OR, VA, UT
Green-5: Next 28 days (9): NH, ME, AZ, MD, OH, CT, AL, IN, MN
Red-3: September (7): IL, KS, NJ, CA, NE, MA, HI
Red-4: October (17): VT, SD, FL, OK, DE, AK, PA, MO, LA, GA, TN, MS,WV, RI, NY, AR, NM
Red-5: November (3): WI, TX, DC, WY
Red-6: December (3): IA, NC, ND
Red-7: 2021 (2): SC, MT

[Tender Branson]

CA, IL, AL and MA reaching their final 2010 rates before data collection ends on September 30 would make up somewhat for the crappy responses in the South, urban areas or (rural) Plains this time.

Also, what's wrong with IA ?

Iowa is one of the few states that is fully covered by the Postal Service, with no update/leave areas (where there are no postal services to households).

So, every household should have gotten several questionnaires and reminder letters.

Iowa had one of the best rates in 1990 and 2000 (ca. 76%), but has dropped a lot since then.

Has the state really gone so hard Trump / anti-government in recent years ?

[jimrtex]

CA, IL, AL and MA reaching their final 2010 rates before data collection ends on September 30 would make up somewhat for the crappy responses in the South, urban areas or (rural) Plains this time.

Also, what's wrong with IA ?

Iowa is one of the few states that is fully covered by the Postal Service, with no update/leave areas (where there are no postal services to households).

So, every household should have gotten several questionnaires and reminder letters.

Iowa had one of the best rates in 1990 and 2000 (ca. 76%), but has dropped a lot since then.

Has the state really gone so hard Trump / anti-government in recent years ?

Iowa is currently 6th overall.

Perhaps the Census Bureau didn't send out the forms soon enough.

[Tender Branson]

OR, UT and VA are now all within 0.5% of their final 2010 rates.