The Official 2020 Census Thread

[jimrtex]

Responses on college campuses are yikes! (I mean, I understand why, but it's just amazing to see on a precinct-by-precinct basis.)

It was planned that there would be early Non-Response Followup (NRFU) around college campuses beginning on April 9. This contrasted with the main NRFU which would run from May 13 to July 23. Obviously, this would miss almost all students who would be gone for the summer.

College students living in rented housing are probably the least likely to use mail. If they ever received mail as a child, it was delivered by their mom. They don't send mail, and they don't need mail for their credit card bill. Any mail before they vamoosed is still sitting in their mailbox at the apartment.

The Early NRFU areas were more extensive than I would expect, with about 7.5 million housing units.

Map of Early NRFU areas (PDF)

The Census Bureau has announced a program where they contact college administrators trying to get information about students living off-campus on April 1 (or before).

Did you have any particular census tracts that you noticed? There are a couple of census tracts in the Houston area which are entirely on the campuses of Rice University and University of Houston, which are apparently entirely group quarters (dorms) since they  have no housing units.

In other cases, there may just be tiny bits of residential housing along a campus within a census tract.

[jimrtex]

What is happening in northern Wisconsin?

It is interesting that there is gradient that seems to follow the Orange-Los Angeles county line.

My maps will gray out any tracts whose 2010 or 2020 self-response rates are 0%. My theory is that Northern Wisconsin was enumerated by Census takers in 2010. Thus, the 2010 final self-response rate was 0%.

Is there any way to know for sure?

Edit: Yes. Here's the 2010 TEA map:
https://www2.census.gov/geo/maps/DC2010/USLCOwTEA/Cen2010_US_LCOwTEA.pdf?#

Northern Wisconsin was Update/Enumerate.

Interesting. I wonder differentiated Wisconsin from UP Michigan or northern Minnesota.

The eastern end of Long Island (Hamptons) and Fire Island, as well as Martha's Vineyard, Nantucket, and the up-arm of Cape Cod were also Update/Enumerate. Must have something to do with large numbers of seasonal residences, as well as some sort of local option. Or perhaps Update/Enumerate wasn't done at the Census Date, but later towards summer.

[ilikeverin]

Did you have any particular census tracts that you noticed?

Even in anemic Hartford, 5038 (University of Hartford) stands out. Ditto 3614.01 (Yale), 8812 (UConn, alongside the entirety of Mansfield for that matter). Student housing for my alma mater, MSU, is clear—obviously, it's tough right around the university, but tract 112 in Clinton County also has a huge student housing development. Basically, anywhere I know there's a college, I can predictably zoom in and find poor response. I don't know Notre Dame well, it's pretty clear where the campus is without even zooming in: 112.03, where the rate is down 30.9 points.

[Tender Branson]

Did you have any particular census tracts that you noticed?

Even in anemic Hartford, 5038 (University of Hartford) stands out. Ditto 3614.01 (Yale), 8812 (UConn, alongside the entirety of Mansfield for that matter). Student housing for my alma mater, MSU, is clear—obviously, it's tough right around the university, but tract 112 in Clinton County also has a huge student housing development. Basically, anywhere I know there's a college, I can predictably zoom in and find poor response. I don't know Notre Dame well, it's pretty clear where the campus is without even zooming in: 112.03, where the rate is down 30.9 points.

Not sure why exactly students should be counted at the city of their university ...

I know, funding for the colleges and the city itself, but ... once they finish their studies, the students are moving out of that college town again to other places anyway. Only a handful remain in that place.

What this 2020 COVID/Census counting dilemma among students shows, is that the US badly needs a National CPR (Central Population Register) and CVSR (Central Vital Statistics Register):

A 10-year Census cannot track in a competent way the number of students living or not living in a college city. The census this year, coupled with the pandemic, will lead to an undercount of students in college cities and an overcount among their parents cities.

A CPR on the other hand will have every person in a state or country registered at their place of residence, at a "live" up-to-date level.

The Census Bureau would be able to do a register-based up-to-date Census every year, not just every 10 years - for a tenth of the current cost with paper forms or online responses.

[jimrtex]

Did you have any particular census tracts that you noticed?

Even in anemic Hartford, 5038 (University of Hartford) stands out. Ditto 3614.01 (Yale), 8812 (UConn, alongside the entirety of Mansfield for that matter). Student housing for my alma mater, MSU, is clear—obviously, it's tough right around the university, but tract 112 in Clinton County also has a huge student housing development. Basically, anywhere I know there's a college, I can predictably zoom in and find poor response. I don't know Notre Dame well, it's pretty clear where the campus is without even zooming in: 112.03, where the rate is down 30.9 points.

CT,Hartford,5038: Pop. 2885, HU 618, Households 170, Response (6/22) 9.6%, 2010 Response 30.5%.

The census tract pretty much matches the campus, but the tomahawk includes a 1950s public housing project that is under redevelopment. If you are familiar with the area it is off Granby Drive along Nahum Drive. The former Bowles Park Public Housing is pretty grim looking 1950's buildings with barred windows on all the first floor windows. Weirdly scattered about as if you were transporting attached slum housing to the suburbs. I've seen more attractive prisons. The area is being redeveloped as Willow Creek Apartments, including subsidized apartments. Another area along Mark Twain Drive, Westbrook Village is being redeveloped as Village at Park River.

That the two areas are being renamed tells you a lot.

Such a large population compared to the number of households means the remainder is Group Quarters - dorms on the campus. Note that Group Quarters are not included in the response rate - it is only for housing units (houses for apartments for a household). The University of Hartford is pretty new for a private school as a result of a merger of other institutions around 1960. It looks like it bought a chunk of suburban land on the Hartford/West Hartford line (the campus address is West Hartford). I don't think there is any off-campus housing adjacent to the campus, so it would be back towards downtown Hartford. I wouldn't be surprised if there wasn't some sort of circulator bus, or maybe city buses run a frequent route out Albany Avenue and up to the campus.

The HU and household figures are from the 2014-2018 ACS. Redevelopment of Bowles Park began in 2017, so the large housing unit count relative to households would reflect that change. Bowles Park is completely demolished so it might not be included in the housing units for the census. Redevlopment of Westbrook Village is just beginning, so it may be contributing a bunch of housing units that have been vacated, but still included in the census bureau denominator.

So oddly enough, the low response rate does not appear to related to college students, but a small area attached to the campus with slum housing under the process of redevelopment.

I was trying to figure out whether your alma mater is in Bozeman, Mankato, Springfield, or Starkville. But the mention of Clinton County gave me a clue. Maybe they should be like TOSU and go by TOOMSU (The other other MSU).

[cinyc]

I haven't checked the more recent map, Jim, but in the May maps, ilikeverin is right. Pick a major college campus - any major college campus - say Penn State. Look at the university precincts - things are usually pretty grim, but your explanation might explain it. But then look at the campus-adjacent precincts. They're usually worse.

Many college students moved home before Census invitations and/or reminders were sent out due to the virus.

Now, Census might be able to fix the on-campus reporting issues by contacting colleges. I'm a little more worried about off-campus housing, though. Colleges may or may not have the relevant info. Even if they do for both, the duplicate count program is going to have to go into overdrive for sure.

[jimrtex]

This map illustrates the net change over the past week.

NM and VT have increased 0.5%; AK, AZ, HI, NH, and WY 0.4%; MT, NY and SD 0.3%.
15 states increased by 0.2%
26 states increased by 0.2%.

The national increase over the past week was 0.2%.

The greatest increase has generally been in states with large numbers of Update/Leave Households, particularly with later restarts. There is a clear west/east bias reflecting more Update/Leave areas in the more rural west.

[jimrtex]

Twenty-nine-day averages* (since 5/31/2010)

New base date reduces daily average by about 10%, and lengthens estimates of when the 2010 response rate will be reached, especially when several months out. Estimates for states with larger Update/Leave shares are optimistic as we were getting responses from these areas over the past month. We are 8 weeks into Update/Leave for the earliest restarts (e.g. WV), and responses are now negligible. We might get a bit of a nibble from postcards now being sent to P.O. Boxes in areas with no home delivery.

Total 1.20%
Internet 0.68%
Paper&Phone 0.52%

Daily average:

Total 0.041%
Internet 0.023%
Paper&Phone 0.018%

Michigan reached its 2010 response rate on June 15.

At the recent rate, those estimated to reach their 2010 response rate in the next 30 days:

Washington on July 1, 2 days from now. Gap is 0.1%
Nevada on July 8
Kentucky on July 11

0 other states in July
8 states in August
9 states in September
11 states in October
8 states in November
7 state in December
3 states +DC  in 2021

Green-7: Reached (1): MI
Green-6: Next 14 days (3): WA, NV, KY
Green-5: Next 28 days (0):
Green-4: July (0):
Green-3: August (8): NH, ID, CO, UT, SD, WY, AZ, HI (month omitted for NH, HI)
Red-3: September (9): VT, AK, VA, ME, WV, MN, OR, OH, NE
Red-4: October (11): NJ, NM, AL, IN, MD, MA, CT, IL, FL, MT, CA
Red-5: November (8): PA, NY, OK, LA, GA, WI, KS, DE
Red-6: December (7): RI, MS, NC, AR, TX, MO, IA
Red-7: 2021 (4): ND, DC, TN, SC

[ilikeverin]

Did you have any particular census tracts that you noticed?

Even in anemic Hartford, 5038 (University of Hartford) stands out. Ditto 3614.01 (Yale), 8812 (UConn, alongside the entirety of Mansfield for that matter). Student housing for my alma mater, MSU, is clear—obviously, it's tough right around the university, but tract 112 in Clinton County also has a huge student housing development. Basically, anywhere I know there's a college, I can predictably zoom in and find poor response. I don't know Notre Dame well, it's pretty clear where the campus is without even zooming in: 112.03, where the rate is down 30.9 points.

CT,Hartford,5038: Pop. 2885, HU 618, Households 170, Response (6/22) 9.6%, 2010 Response 30.5%.

The census tract pretty much matches the campus, but the tomahawk includes a 1950s public housing project that is under redevelopment. If you are familiar with the area it is off Granby Drive along Nahum Drive. The former Bowles Park Public Housing is pretty grim looking 1950's buildings with barred windows on all the first floor windows. Weirdly scattered about as if you were transporting attached slum housing to the suburbs. I've seen more attractive prisons. The area is being redeveloped as Willow Creek Apartments, including subsidized apartments. Another area along Mark Twain Drive, Westbrook Village is being redeveloped as Village at Park River.

That the two areas are being renamed tells you a lot.

Blue Hills is definitely the nicest neighborhood in the North End, but Hartford in general has been subjected to massive disinvestment by the state; rather than distributing affordable housing around the state, Connecticut has instead decided to put it all in neighborhoods with high poverty because suburban NIMBYs get too upset. Compare to, say, Weston, where it is illegal to build anything but single-family homes on at least two acres of land.

Such a large population compared to the number of households means the remainder is Group Quarters - dorms on the campus. Note that Group Quarters are not included in the response rate - it is only for housing units (houses for apartments for a household). The University of Hartford is pretty new for a private school as a result of a merger of other institutions around 1960. It looks like it bought a chunk of suburban land on the Hartford/West Hartford line (the campus address is West Hartford). I don't think there is any off-campus housing adjacent to the campus, so it would be back towards downtown Hartford. I wouldn't be surprised if there wasn't some sort of circulator bus, or maybe city buses run a frequent route out Albany Avenue and up to the campus.

63% of UHart students live on-campus; of the 37% who live off-campus, I suspect the majority are locals who live with their parents. UHart does not have a significant out-of-region population, although the Hartt School (the drama school) has an excellent reputation and does attract people from all over. Albany is indeed a major bus route, but oddly many of the lines sort of splinter just before campus. That said, the 153 bus is branded "Fastrak", which means it comes every 20 minutes. (Why CTtransit decided that was the best way to do things I don't know; Fastrak was a branding introduced for the excellent bus rapid transit service from Hartford to New Britain, but they also decided it should be used for a series of buses that have relatively frequent service but don't have the benefits of the BRT proper. It's confusing enough that people also think the Express buses, which are commuter buses, are also Fastrak.)

The HU and household figures are from the 2014-2018 ACS. Redevelopment of Bowles Park began in 2017, so the large housing unit count relative to households would reflect that change. Bowles Park is completely demolished so it might not be included in the housing units for the census. Redevlopment of Westbrook Village is just beginning, so it may be contributing a bunch of housing units that have been vacated, but still included in the census bureau denominator.

So oddly enough, the low response rate does not appear to related to college students, but a small area attached to the campus with slum housing under the process of redevelopment.

Hmm, that surprises me!

I was trying to figure out whether your alma mater is in Bozeman, Mankato, Springfield, or Starkville. But the mention of Clinton County gave me a clue. Maybe they should be like TOSU and go by TOOMSU (The other other MSU).

We would never pattern ourselves after that disgusting school in Columbus

[Tender Branson]

WA is the 2nd state after MI to reach its final 2010 response rate as of today (actually, as of yesterday).

NV is just 0.3% and KY 0.4% away from reaching their 2010 rates.

[jimrtex]

Blue Hills is definitely the nicest neighborhood in the North End, but Hartford in general has been subjected to massive disinvestment by the state; rather than distributing affordable housing around the state, Connecticut has instead decided to put it all in neighborhoods with high poverty because suburban NIMBYs get too upset. Compare to, say, Weston, where it is illegal to build anything but single-family homes on at least two acres of land.

63% of UHart students live on-campus; of the 37% who live off-campus, I suspect the majority are locals who live with their parents. UHart does not have a significant out-of-region population, although the Hartt School (the drama school) has an excellent reputation and does attract people from all over. Albany is indeed a major bus route, but oddly many of the lines sort of splinter just before campus. That said, the 153 bus is branded "Fastrak", which means it comes every 20 minutes. (Why CTtransit decided that was the best way to do things I don't know; Fastrak was a branding introduced for the excellent bus rapid transit service from Hartford to New Britain, but they also decided it should be used for a series of buses that have relatively frequent service but don't have the benefits of the BRT proper. It's confusing enough that people also think the Express buses, which are commuter buses, are also Fastrak.)

So oddly enough, the low response rate does not appear to related to college students, but a small area attached to the campus with slum housing under the process of redevelopment.

Hmm, that surprises me!

I did some more looking. Census Tract 5038 is in the extreme northwest corner of Hartford, north of Albany. While census tracts may cross town lines, in New England, they won't except in extreme northern New England where towns are much smaller than an ideal tract size of 4000 - or even the minimum of 2000.

The eastern boundary is a railroad track, except at the northern end where the tomahawk jumps across to Granby Avenue. The whole area looks like it missed out on any development. There is a brook through there, perhaps it was a farm, or drainage made it hard to develop (only recently has their been giant earth-moving equipment that permits the terrain to be reshaped. When you were adding streets one by one when the city was expanding in the late 19th century and early 20th street you keep to the level ground (Hartford is below its 1920 population, the peak was in 1950).

The area with the larger public housing project is completely separated from the campus - no roads at all. The other smaller housing project along Mark Twain Drive off of Albany is west of the railroad tracks and south of the campus it is barely connected to the campus.

The main entrance to the campus is off of Bloomfield Avenue. There may be a tiny strip of under 10 residences along Bloomfield that are in Hartford, before the street slides into West Hartford.

Incidentally as part of the Mark Twain Drive redevelopment there is a shopping center being developed on Albany. The developer trying to attract tenants manages to use "affluent" about 5 times in describing the West Hartford/Hartford corridor. "Affluent suburbanites who live in affluent West Hartford will drive by in their affluence-exuding automobiles on the affluent West Hartford/Hartford corridor to their downtown offices where they earn the money to afford their affluent lifestyle." (not a real quote, but the blurb was rich in affluence).

So in 1950, you had this undeveloped patch of land, perhaps the last in Hartford. The University of Hartford is founded to compete with that school in New Haven and they acquire the farm(?) for a campus. Meanwhile, early urban renewal is replacing slums in or near downtown, with soon to be slums further out. The larger project is adjacent to Blue Hills rather than really part of it. 

Because the University of Hartford was founded so late, there was never commercial or residential development around it. If you were an Junior or Senior and wanted to live off campus, and you were not commuting from your parents house, you would probably find a place in Hartford west of downtown (it seems that there is no gentrification in Hartford, to the extent of thesis published on why that is, or op-eds welcoming gentrification.

Note the other remarkable thing about CT 5038 being 20% below its 2010 response rate is that the 2010 response rate was only 30%.

[jimrtex]

Thirty-day averages* (since 5/31/2010)

Total 1.22%
Internet 0.68%
Paper&Phone 0.54%

Daily average:

Total 0.041%
Internet 0.023%
Paper&Phone 0.018%

Michigan reached its 2010 response rate on June 15.
Washington reached its 2010 response rate on June 30.

At the recent rate, those estimated to reach their 2010 response rate in the next 30 days:

Nevada on July 6 (gap is 0.3%)
Kentucky on July 13

0 other states in July
7 states in August
10 states in September
11 states in October
9 states in November
4 states in December
5 states +DC  in 2021

Green-7: Reached (2): MI, WA
Green-6: Next 14 days (3): NV, KY
Green-5: Next 28 days (0):
Green-4: July (0):
Green-3: August (7): NH, ID, CO, UT, SD, WY, AZ (month omitted for NH)
Red-3: September (10): HI, WV, VT, AK, VA, ME, MN, OR, OH, NJ
Red-4: October (11): NE, MD, NM, AL, IN, MA, CT, IL, NY, FL, CA
Red-5: November (9): MT, PA, OK, LA, GA, WI, KS, DE, RI
Red-6: December (4): AR, MS, NC, TX
Red-7: 2021 (6): MO, ND, DC, LA, TN, SC

[jimrtex]

WA is the 2nd state after MI to reach its final 2010 response rate as of today (actually, as of yesterday).

NV is just 0.3% and KY 0.4% away from reaching their 2010 rates.

The data files that I am using use the current data. So I will show WA as reaching on 6/30.

Daily increases are now down around 0.03%, so a 0.1% increase will only register every 3rd day or so. If there is no registered increase, the estimated date of reaching the 2010 rate will be deferred for a day. If there is a registered increase, decreasing the gap by 0.1%, the estimated date will jump forward a couple of days.

As the increase response rate continues to slow, the estimated date slides further away.

Kentucky has been a slower performer lately, increasing only 0.9% over the past month. This is partly because it had done so well early on, and that it has minimal Update/Leave areas (only 3.6% of total). It is estimated to take 13 days to bridge that gap of 0.4%, but will likely be longer.

It would not surprise me if MI, WA, NV, and KY are the only states to reach their 2010 rate.

[ilikeverin]

Incidentally as part of the Mark Twain Drive redevelopment there is a shopping center being developed on Albany. The developer trying to attract tenants manages to use "affluent" about 5 times in describing the West Hartford/Hartford corridor. "Affluent suburbanites who live in affluent West Hartford will drive by in their affluence-exuding automobiles on the affluent West Hartford/Hartford corridor to their downtown offices where they earn the money to afford their affluent lifestyle." (not a real quote, but the blurb was rich in affluence).

Yes, indeed. The only way to drive investment in Hartford is to try to connect it to West Hartford, sadly. Albany is one of the main streets of the North End (along with, well, Main Street), and it's recently seen a substantial street revitalization project, with wider sidewalks, more benches, and otherwise a more-inviting streetscape for pedestrians. Once you cross the town line into West Hartford, the corridor loses its commercial nature till you hit Bishops Corner.

So in 1950, you had this undeveloped patch of land, perhaps the last in Hartford. The University of Hartford is founded to compete with that school in New Haven and they acquire the farm(?) for a campus. Meanwhile, early urban renewal is replacing slums in or near downtown, with soon to be slums further out. The larger project is adjacent to Blue Hills rather than really part of it. 

True. That was right around when I-84 was built through the middle of Hartford's most thriving black community.

Because the University of Hartford was founded so late, there was never commercial or residential development around it. If you were an Junior or Senior and wanted to live off campus, and you were not commuting from your parents house, you would probably find a place in Hartford west of downtown

Maybe. I'm not exactly sure. There are some duplexes and other housing that might be student-appropriate in the West End (say, 5245.02 and 5042), but I'd associate the students there with grad students at UConn's Law School, not UHart. The West End is, by far, the wealthiest neighborhood in Hartford, and it would be out of many undergraduates' price ranges. The closest Hartford neighborhoods that would be in their price range (Asylum Hill and Blue Hills) would probably be considered undesirable by the parents of those students. I suspect more students would rent out houses in West Hartford.

(it seems that there is no gentrification in Hartford, to the extent of thesis published on why that is, or op-eds welcoming gentrification.

That's also correct. Hartford has been trying to encourage housing growth downtown (5021) to mixed success; there are now some higher-end high-rises downtown, usually renovated office spaces, with more slowly opening (spurred by Hartford's abolition of parking minimums in 2016). That has no connection to any sense of "gentrification", though, since no one was really living there previously. Anecdotally, the neighborhood of Hartford I'd most associate with "gentrification" is Parkville (centered on 5041), where a couple of mixed-use loft-style building refurbishments have been completed, but even then there seems to be very little displacement. I'm disappointed that Parkville Market opened about a month before I'm leaving town; it's amazing!

[jimrtex]

Thirty-one-day averages* (since 5/31/2010)

Total 1.25%
Internet 0.70%
Paper&Phone 0.56%

Daily average:

Total 0.040%
Internet 0.022%
Paper&Phone 0.018%

Michigan reached its 2010 response rate on June 15.
Washington reached its 2010 response rate on June 30.

At the recent rate, those estimated to reach their 2010 response rate in the next 30 days:

Nevada on July 8 (gap is 0.3%)
Kentucky on July 14

0 other states in July
5 states in August
11 states in September
10 states in October
10 states in November
5 states in December
5 states + DC  in 2021

Green-7: Reached (2): MI, WA
Green-6: Next 14 days (3): NV, KY
Green-5: Next 28 days (0):
Green-4: July (0):
Green-3: August (5): NH, ID, CO, SD, UT (month omitted for NH)
Red-3: September (11): WY, AZ, VA, HI, WV, VT, AK, ME, MN, OH, NJ
Red-4: October (10): OR, NE, NM, MD, AL, IN, MA, CT, IL, CA
Red-5: November (10): NY, FL, WI, MT, PA, OK, LA, GA, KS, MS
Red-6: December (4): RI, DE, NC, AR, TX
Red-7: 2021 (6): MO, DC, ND, IA, TN, SC

[jimrtex]

Last week (June 22-June 29), the response rate in counties with 0%-5% Update/Leave Housing Units rose by 0.1% from 63.6% to 63.7%. Meanwhile the response rate in counties with 95%-100% Update/Housing Units rose by 1.2% from 22.4% to 23.6%. The increase for these counties dropped from 1.5% to 1.2%.

I suspect that the final self-response rate for these areas will only be around 30%.

The response rate for counties with between 45% and 55% Update/Leave Housing Units rose from 36.0% to 36.5%, an increase of 0.5%. This rate is below the midpoint of the 5% minus, and that 95%+ counties. That is, a mix of the two types in a county has a lower response than two counties, one of each type. This might be due to demographic differences. The towns in a 50/50 county are likely quite small, and semi-rural in nature. Even if they received the forms in the mail, their might be less general publicity, with no radio stations or newspapers or other send-in-the-form campaigns.

This is a list of the counties with the largest increase of 0.6 or greater in each state. There were a total of 189 counties in the country last week, compared to 337 counties the  previous week. 189 (6.0% of all counties) seems high, but these are mostly small counties.

County	State	Change	Total	UL HU%	UL Restart
Dillingham Census Area	AK	5.7%	23.5%	44%	4-May
Apache	AZ	0.8%	10.1%	97%	11-Jun
Greenlee	tie	0.8%	34.9%	46%	18-May
Sierra	CA	2.5%	34.8%	95%	25-May
Saguache	CO	1.1%	36.3%	79%	18-May
Kalawao County	HI	8.0%	14.0%	100%	1-Jun
Hardin	IL	3.5%	29.9%	93%	1-Jun
Haskell	KS	1.8%	44.0%	70%	11-May
Dukes	MA	2.3%	24.5%	77%	25-May
Ontonagon	MI	2.2%	40.8%	48%	25-May
Mahnomen	MN	5.6%	30.8%	100%	25-May
Douglas	MO	0.6%	37.0%	86%	11-May
Oregon	tie	0.6%	36.3%	75%	11-May
Ripley	tie	0.6%	31.3%	85%	11-May
Shannon	tie	0.6%	31.9%	80%	11-May
Glacier	MT	1.4%	29.7%	71%	4-May
Thurston	NE	1.4%	26.2%	100%	18-May
Mineral	NV	1.0%	44.3%	87%	18-May
Carroll	NH	2.1%	32.7%	55%	1-Jun
Mora	NM	4.5%	16.7%	100%	25-May
Hamilton	NY	2.1%	15.1%	95%	25-May
Swain	NC	1.1%	31.0%	40%	11-May
Sioux	ND	0.8%	27.3%	100%	4-May
Osage	OK	0.6%	42.6%	100%	4-May
Buffalo	SD	5.4%	21.4%	87%	25-May
Upton	TX	4.9%	23.1%	99%	8-Jun
Duchesne	UT	1.1%	26.8%	90%	4-May
Caledonia	VT	0.6%	50.1%	24%	4-May
Essex	tie	0.6%	34.8%	63%	4-May
Orange	tie	0.6%	51.9%	14%	4-May
Orleans	tie	0.6%	44.1%	24%	4-May
Ferry	WA	1.6%	37.0%	46%	11-May
Menonimee	WI	3.6%	18.0%	72%	25-May
Fremont	WY	2.1%	47.9%	73%	18-May

States with no counties on list (new states are underlined:

AL, CT, DE, DC, FL, GA, ID, IN, IA, KY, LA, ME, MD, MS, NJ, OH, OR, PA, RI, SC, TN, VA, WV. VT had no counties listed last week, but several this week.

[jimrtex]

Thirty-two-day averages* (since 5/31/2010)

Total 1.27%
Internet 0.74%
Paper&Phone 0.54%

Daily average:

Total 0.040%
Internet 0.023%
Paper&Phone 0.017%

Michigan reached its 2010 response rate on June 15.
Washington reached its 2010 response rate on June 30.

At the recent rate, those estimated to reach their 2010 response rate in the next 30 days:

Nevada on July 9 (gap is 0.3%)
Kentucky on July 16

0 other states in July
5 states in August
9 states in September
13 states in October
6 states in November
8 states in December
5 states + DC  in 2021

Green-7: Reached (2): MI, WA
Green-6: Next 14 days (2): NV, KY
Green-5: Next 28 days (0):
Green-4: July (0):
Green-3: August (5): NH, CO, ID, SD, UT (month omitted for NH)
Red-3: September (9): WY, HI, AZ, VA, WV, VT, AK, ME, MN
Red-4: October (13): OH, NJ, OR, NE, AL, NM, MD, IN, MA, CT, IL, FL, CA
Red-5: November (6): MT, NY, WI, PA, OK, LA
Red-6: December (8): GA, KS, MS, DE, RI, NC, AR, TX
Red-7: 2021 (6): MO, DC, IA, ND, TN, SC

[jimrtex]

Census Takers to Start Follow Up With Nonresponding Households in Limited Areas in July

The Census Bureau announced the 6 Area Census Offices (ACO) where they will begin NRFU on July 16.

Beckley, WV. This is entire state of West Virginia.

Boise, ID. This is entire state of Idaho.

Gardiner, ME. This is entire state of Maine.

Kansas City, MO. This is generally the northern part of the state.

New Orleans, LA. This is extreme eastern Louisiana: Washington, St. Tammany, Orleans, St. Bernard, and Plaquemines parishes. Jefferson (Metairie) is not included.

Oklahoma City, OK. This is northwestern part of state, including Oklahoma, but not Cleveland.

Map of Current Census Operations

[jimrtex]

Did you have any particular census tracts that you noticed?

Even in anemic Hartford, 5038 (University of Hartford) stands out. Ditto 3614.01 (Yale), 8812 (UConn, alongside the entirety of Mansfield for that matter). Student housing for my alma mater, MSU, is clear—obviously, it's tough right around the university, but tract 112 in Clinton County also has a huge student housing development. Basically, anywhere I know there's a college, I can predictably zoom in and find poor response. I don't know Notre Dame well, it's pretty clear where the campus is without even zooming in: 112.03, where the rate is down 30.9 points.

Not sure why exactly students should be counted at the city of their university ...

There is a lengthy federal register article about where residents are counted. Most of the comments this Census were about prisoners.

The Census Bureau is interested in consistency, so that results from one census can be compared with those from the next. They will be quite resistant to changing definitions. They will also want less ambiguous definitions.

Students are counted at the location of their residence. This might or might not be in the city of the university. It probably won't be at the location of the academic campus, since they may live off-campus, or the dorms may be separate from the classroom buildings. Some students commute to campus, perhaps 25 to 50 miles.

Older students may be married, and have children. Even if in family housing on the campus, that is where they are living. Parents of some students will be divorced, sometimes while they are at college. Rather than picking one parent or the other, they may stay at school over summer, or go to Europe. If the mother moved to Florida, can I child be said to live in Florida, even if they have never been to Florida? A student from overseas might remain on campus year round. And if they weren't counted on the campus, where would they be counted?

Anyone who is over 18 is an adult. If they reside with their parents it is by choice.

Whether you could determine the "right answer" to a question that has no "right answer" doesn't matter if it would take a complex set of questions to determine what the answer is, when the census is depending on the self-enumeration.

I know, funding for the colleges and the city itself, but ... once they finish their studies, the students are moving out of that college town again to other places anyway. Only a handful remain in that place.

This is irrelevant to the Census. Some states treat students differently than the census for electoral purposes. Governments are free to dispense money to local governments whatever formula they choose. The amount that is blindly dispensed per capita is likely overestimated.

What this 2020 COVID/Census counting dilemma among students shows, is that the US badly needs a National CPR (Central Population Register) and CVSR (Central Vital Statistics Register):

A 10-year Census cannot track in a competent way the number of students living or not living in a college city. The census this year, coupled with the pandemic, will lead to an undercount of students in college cities and an overcount among their parents cities.

A CPR on the other hand will have every person in a state or country registered at their place of residence, at a "live" up-to-date level.

It really doesn't matter if the student is counted in the "wrong" place, so long as they are counted exactly once.

The Census Bureau is running an experimental census (in select areas) using administrative records.

In the US, people would be resistant to to any central registry.

The Census Bureau would be able to do a register-based up-to-date Census every year, not just every 10 years - for a tenth of the current cost with paper forms or online responses.

That is what the American Community Survey does.

[Tender Branson]

In the US, people would be resistant to any central registry.

True, sadly, but also the main reason why CPR countries have much higher-quality statistics than countries with a traditional paper census like in the US, for only a tenth of the cost.

A CPR would also be the smart "starting point" for 100% fair elections, rather than in the US where 20% of adult citizens are not even registered to vote ...

The Census Bureau would be able to do a register-based up-to-date Census every year, not just every 10 years - for a tenth of the current cost with paper forms or online responses.

That is what the American Community Survey does.

Not exactly:

A CPR allows for a full, annual count of the population.

The ACS is just a sample, which surveys 10% or so of US households each year. It is not a full count, more like a survey to gather information that the 10-year Census doesn't ask.

[jimrtex]

Thirty-two-day averages* (since 5/31/2010)

Total 1.27%
Internet 0.74%
Paper&Phone 0.54%

Daily average:

Total 0.040%
Internet 0.023%
Paper&Phone 0.017%

Michigan reached its 2010 response rate on June 15.
Washington reached its 2010 response rate on June 30.

At the recent rate, those estimated to reach their 2010 response rate in the next 30 days:

Nevada on July 9 (gap is 0.3%)
Kentucky on July 16

0 other states in July
5 states in August
9 states in September
13 states in October
6 states in November
8 states in December
5 states + DC  in 2021

Green-7: Reached (2): MI, WA
Green-6: Next 14 days (2): NV, KY
Green-5: Next 28 days (0):
Green-4: July (0):
Green-3: August (5): NH, CO, ID, SD, UT (month omitted for NH)
Red-3: September (9): WY, HI, AZ, VA, WV, VT, AK, ME, MN
Red-4: October (13): OH, NJ, OR, NE, AL, NM, MD, IN, MA, CT, IL, FL, CA
Red-5: November (6): MT, NY, WI, PA, OK, LA
Red-6: December (8): GA, KS, MS, DE, RI, NC, AR, TX
Red-7: 2021 (6): MO, DC, IA, ND, TN, SC

No daily update because of holiday. 4-day update on Monday will likely reach 62.0%.

[jimrtex]

This map illustrates the net change over the past week.

NM has increased 0.4%; AK, DC, and WY 0.3%; 11 states increased 0.2%; 34 states increased 0.1%; and KS and KY 0.0%.

The national increase over the past week was 0.1%.

We are seeing Update/Leave areas completing that operation. There may be a slightly faster rate in more urban states, but some of that may be rounding error.

[jimrtex]

Thirty-six-day averages* (since 5/31/2010)

Total 1.33% (0.040% per day)
Internet 0.77% (0.021%)
Paper&Phone 0.55% (0.016%)

Twenty-nine-day averages* (since 6/7/2010)

Total 0.94% (0.032% per day)
Internet 0.56% (0.019%)
Paper&Phone 0.38% (0.013%).

0.39% of a five-week gain occurred in the first week.

Michigan reached its 2010 response rate on June 15.
Washington reached its 2010 response rate on June 30.

At the recent rate, those estimated to reach their 2010 response rate in the next 30 days:

Nevada on July 15  (gap is 0.3%)
Kentucky on July 25

0 other states in July
3 states in August
7 states in September
9 states in October
5 states in November
10 states in December
12 states + DC  in 2021

Green-7: Reached (2): MI, WA
Green-6: Next 14 days (1): NV
Green-5: Next 28 days (1): KY
Green-3: August (3): NH, CO, ID (month omitted for NH)
Red-3: September (7): UT, SD, WY, AZ, HI, VA, VT
Red-4: October (9): AK, ME, MN, OH, NJ, OR, WV, NE, NM
Red-5: November (5): AL, IL, CA, IN, FL
Red-6: December (10): NY, CT, MD, MA, PA, GA, WI, MT, OK, LA
Red-7: 2021 (13): RI, TX, DE, KS, MS, MO, AL, AR, IA, NC, ND, TN, SC

[jimrtex]

Thirty-day averages* (since 6/7/2010)

Total 0.94% (0.031% per day)
Internet 0.56% (0.019%)
Paper&Phone 0.38% (0.013%).

This was the slowest day yet, with only five states registering 0.1% gains.

Michigan reached its 2010 response rate on June 15.
Washington reached its 2010 response rate on June 30.

At the recent rate, those estimated to reach their 2010 response rate in the next 30 days:

Nevada on July 13  (gap is 0.2%, Nevada was one of the states to register a gain).
Kentucky on July 27

3 states in August
5 states in September
10 states in October
7 states in November
6 states in December
15 states + DC  in 2021

Green-7: Reached (2): MI, WA
Green-6: Next 14 days (1): NV
Green-5: Next 28 days (1): KY
Green-3: August (3): NH, CO, ID (month omitted for NH)
Red-3: September (5): UT, SD, WY, HI, AZ
Red-4: October (10): VA, VT, AK, ME, MN, OH, NJ, OR, WV, NE
Red-5: November (7): NM, AL, IL, MD, CA, IN, MA
Red-6: December (6): FL, NY, CT, PA, GA, WI
Red-7: 2021 (1): MT, OK, LA, KS, RI, TX, DE, MS, MO, AL, AR, IA, NC, ND, TN, SC

In other news, temporary workers increased by 4000 (from 22,000 to 26,000) likely reflecting gearing up for NRFU.

[jimrtex]

Last week (June 29-July 6), the response rate in counties with 0%-5% Update/Leave Housing Units rose by 0.1% from 63.7% to 63.8%. Meanwhile the response rate in counties with 95%-100% Update/Leave Housing Units rose by 0.8% from 23.6% to 24.4%. The weekly increase for these counties dropped from 1.2% to 0.8%.

I suspect that the final self-response rate for these areas will be less than 30%. The earliest areas have been underway for 9 weeks.

The response rate for counties with between 45% and 55% Update/Leave Housing Units rose from 36.5% to 36.8%, an increase of 0.3%. This rate is below the midpoint of the 5% minus, and that 95%+ counties. That is, a mix of the two types in a county has a lower response than two counties, one of each type. This might be due to demographic differences. The towns in a 50/50 county are likely quite small, and semi-rural in nature. Even if they received the forms in the mail, their might be less general publicity, with no radio stations or newspapers or other send-in-the-form campaigns.

This is a list of the counties with the largest increase of 0.6 or greater in each state. There were a total of 105 such counties in the country last week, compared to 189 counties the previous week. Half of these counties are in TX (29), SD (14), and AK (11).

County	State	Change	Total	UL HU%	UL Restart	Counties > 0.5%
Aleutians West Census Area	AK	8.3%	26.1%	60%*	4-May	11
Apache	AZ	1.0%	11.1%	97%	11-Jun	1
Sierra	CA	1.2%	36.0%	95%	25-May	7
Hinsdale	CO	0.9%	16.4%	99%	18-May	1
Kalawao	HI	4.0%	18.0%	100%	1-Jun	1
Calhoun	IL	0.7%	39.3%	92%	1-Jun	2
Hardin	tie	0.7%	30.6%	93%	1-Jun
Stanton	KS	1.1%	43.0%	80%	11-May	6
Dukes	MA	1.1%	25.6%	77%	25-May	1
Mahnomen	MN	2.1%	32.9%	100%	25-May	3
Sweet Grass	MT	0.8%	34.4%	18%	4-May	2
Thurston	NE	6.3%	32.5%	100%	18-May	3
Carroll	NH	1.1%	33.8%	55%	1-Jun	1
McKinley	NM	2.3%	22.2%	66%	11-Jun	9
Hamilton	NY	0.7%	15.8%	95%	25-May	2
Sherman	OR	1.1%	37.7%	11%	11-May	1
Ziebach	SD	5.3%	11.9%	100%	25-May	14
Loving	TX	3.8%	17.3%	98%	8-Jun	29
Beaver	UT	0.7%	39.3%	93%	4-May	2
Menonimee	WI	1.2%	19.2%	72%	25-May	1
Niobrara	WY	1.2%	41.9%	73%	18-May	5

*The other 40% housing units of the Aleutian West Census Area are Update/Enumerate which will have almost no self-response.

States with no counties on list (new states are underlined:

AL, CT, DE, DC, FL, GA, ID, IN, IA, KY, LA, ME, MD, MI, MS, MO, NJ, NV, NC, ND, OH, OK, PA, RI, SC, TN, VT, VA, WA, WV. OR had no counties listed last week, but one this week.

There were 857 counties (27% of 3142 total) that had no gain (less than 0.1%) this last week.