The Official 2020 Census Thread (user search)
       |           

Welcome, Guest. Please login or register.
Did you miss your activation email?
May 31, 2024, 07:00:12 PM
News: Election Simulator 2.0 Released. Senate/Gubernatorial maps, proportional electoral votes, and more - Read more

  Talk Elections
  General Politics
  U.S. General Discussion (Moderators: The Dowager Mod, Chancellor Tanterterg)
  The Official 2020 Census Thread (search mode)
Pages: 1 2 3 [4] 5 6 7 8 9 ... 22
Poll
Question: Are you taking part ?
#1
Yes
 
#2
No
 
#3
Still undecided
 
#4
Not an American, but I would
 
#5
Not an American, but I would not
 
#6
Not an American & still undecided
 
Show Pie Chart
Partisan results

Total Voters: 200

Author Topic: The Official 2020 Census Thread  (Read 118434 times)
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #75 on: April 28, 2020, 06:44:44 PM »

53.7% (+0.3)

A slight uptick from yesterday, but not enough for my taste ...
Overwhelmingly back to the Internet.

Unweighted average:

0.26% total.
0.22% Internet
0.04% Other

The internet response is the best since last Tuesday. The other (mail) response is 10% of the Thu-Fri-Sat rate.

Possibilities: The last of the mailings of the questionnaire were to areas that had received a questionnaire with their first mailing, and the repetition would only help those who had misplaced their original. About 20% of self-response households received a questionnaire with their first invitation in early March. At least half had likely responded so this represented a tiny portion.

There is something systematic in the processing of paper forms. In 2010, it was the check-in date that was recorded, and not the mail receipt date. Maybe they lay the forms out in the UV sunlight to decontaminate them.

They have to have between 1.5% and 2.0% over the remaining five days of the week to reach by weekly estimate.
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #76 on: April 28, 2020, 06:58:01 PM »

53.7% (+0.3)

A slight uptick from yesterday, but not enough for my taste ...

The lower Sunday and Monday response rates are likely a function of there being no mail delivery or, in many places, pickup on Sundays. It will tick up again the rest of the week.
In 2010, Monday was a heavy day because of delivery of Sunday mail. This was even more true after the Census Bureau stopped processing on Saturday's.

A letter placed in the mail on Thursday or Friday is still going to be moving towards the census facilities over the weekend.

For 2020, there are two mail processing centers, Phoenix, AZ for the west and Jeffersonville, IN for the east.

I'm supposing that when the census bureau does a mass mail-out they are injecting locally, perhaps with mail-route sorted boxes, certainly sorted ZIP+9 so they get the best bulk rates. Shaving 0.1 cents in postage on 150,000,000 pieces saves $150,000.
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #77 on: April 28, 2020, 07:17:23 PM »

Säugling

It is not clear that Americans recognize the meaning of this word.

A Säugling is literally a suckling btw, like a tiny pig.

It's a very German word that doesn't exist in English other than the above-mentioned small pig.

But even Germans use the term baby more often than Säugling and we Austrians have our own words.
The English word is infant, which is from the Latin to not speak. As a legal term, infant means someone who has not reached majority, and can not speak for themselves. I don't think if someone referred to a baby as an infant that they meant that he or she was not speaking yet.

Since typical weaning ages and first-speaking ages are similar, maybe they are the same.

Apparently for the US Census, young children (under 5) are most undercounted. It apparently is due to unstable relationships.
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #78 on: April 29, 2020, 04:36:28 PM »

Average of averages:

Total: 0.64%
Internet: 0.21%
Other (paper and phone): 0.43%

Reporting of paper forms must be delayed. 4/28 likely represents mail delivered on Monday, which would be heavier because of no delivery on Sunday.

0.9% and 1.4% is needed over last 4 days to reach my estimate.

Internet responses may be seeing the effect of the vaguely threatening 5th mailing postcard, which says that only complying will prevent an interviewer being sent to their home. Since everybody has received a paper questionnaire by now, this should begin a gradual decline.
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #79 on: April 29, 2020, 09:57:59 PM »

I now expect the final 2010 rate of 66.5% to be achieved by June 7.

That’s 4x10 days for 12.2%

For every 10 days:

29.04.-08.05.: +4.0%

09.05.-18.05.: +3.2%

19.05.-28.05.: +2.7%

29.05.-07.06.: +2.3%
The decay will be greater.

Using a 7-day period which will better match USPS service, the last 3 seven-day periods have been 2.9%, 2.5%, and 2.7% which are at or below your expectation of 0.4% per day.
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #80 on: April 30, 2020, 03:53:32 PM »

54.6% (+0.3)

Unweighted average:

Total 0.29%
Internet 0.19%
Other (Paper) 0.10%

The paper response appears to be bouncy. There must be operational effects. Perhaps they have to periodically shut down the processing lines and decontaminate the equipment.

While the risk of contamination of the mail may be remote, you might not want the risk of 100s of thousands of forms and envelopes from all over the country. The forms are about a foot from the respondents mouth, and their hands are actually touching. Many will use saliva to seal the envelope. They already have had to shutdown their Jeffersonville, IN center to a deep clean, and modify their phone operations to maintain social distancing (callers can make a call-back appointment).

The internet response is steady. For the past 7 days:

Thu 0.20%
Fri 0.15%
Sat 0.12%
Sun 0.15%
Mon 0.23%
Tue 0.21%
Wed 0.19%

The 5th mailing which states the only way to prevent an interviewer from being sent out began being sent out on Monday (April 27).

This should sustain the internet response for another week.

Between 0.6% and 1.1% is needed over the next three days to reach my weekly estimate.

Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #81 on: April 30, 2020, 05:43:19 PM »

Best Responding Cities

No changes on leader board. Andover, MN becomes largest city (33,072) with over 80% response.

1 to 9McMullen, ALPickens65.2%Small all-black town outside Aliceville, west of Tuscaloosa. Likely 16/23 response.
10 to 18Bonanza, COSaguache4.8%Almost ghost town survived effort to have its government abandoned. At one time had 36 dance halls and 7 saloons. Likely 1/21 response.
20 to 49Harbine, NEJefferson76.9%Southwest of Lincoln, a bit north of the Kansas/Nebraska line. Likely 20 of 26 responses.
50 to 99North River, NDCass100.0%North of Fargo along Red River. All 23 households have responded.
100 to 199Oakland Acres, IAJasper86.6%Exurban Des Moines. Most of town is golf course, but not too exclusive ($20 greens fee on weekends). Likely 58/67 response.
200 to 499Argusville, NDCass88.1%Another Fargo suburb on I-29 north of the city. Had a big boom between 2004 and 2008 when new subdivision was built (141/160)
500 to 999Trout Valley, ILMcHenry86.7%Southeast part of county near Cook panhandle. Was modular home showcase (171/196 response)
1K to 2KSt. Regis Park, KYJefferson86.8%11 miles east of Louisville.
2K to 5KHowards Grove, WISheboygan83.7%Farming community in the northern part of the county, seeing modest suburban growth.
5K to 10KHuntington Woods, MIOakland86.0%Immediately northwest of Pleasant Ridge, includes part of the Detroit Zoo.
10K to 20KDardenne Prairie, MOSt. Charles83.7%Northwest St. Louis suburb, incorporated in 1983, has octupled in population since 1990 Census.
20K to 50KAndover, MNAnoka81.8%Just north of Coon Rapids, with half of the population added since 1990.
50K to 100KMaple Grove, MNHennepin78.2%Northwest of Minneapolis, about 20 minutes in light traffic.
100K to 200KCentennial, COArapahoe77.2%Southeast of Denver, incorporated in 2001 as most populous incorporation at that time.
200K to 500KLincoln, NELancaster68.6%State capital and home of University of Nebraska.
500K to 1MSeattle, WAKing65.2%18th largest city, and fastest growing over 500,000 since 2010, slightly exceeding Austin.
1M to 2MSan Jose, CASanta Clara63.6%High-tech Silicon Valley giant, areas along southern edge are over 75% response rates.
2M to 5MChicago, ILCook, DuPage47.9%Three cities in this range are tightly bunched.
5M to 10MNew York, NYKings, Queens, New York, Bronx, Richmond43.8%USA's largest city made up of Brooklyn, Queens, Manhattan, The Bronx, and Staten Island boroughs.


Largest city to have exceeded certain thresholds.

Andover, MN becomes largest city to exceed 80% response. San Diego, CA becomes largest city to exceed 60%.

100%=North River, ND55
90%+North River, ND55
80%+Andover, MN33,072
70%+Naperville, IL148,304
60%+San Diego, CA1,425,976
50%+Phoenix, AZ1,660,272
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #82 on: May 01, 2020, 10:01:49 AM »

IN and OH are the states with the most evenly distributed responses:

If you look at county maps for each state, almost every state ... no matter the size ... has at least a few (rural) counties with extremely low responses.

Even in the leading states MN, WI, MI etc. you have numerous rural counties in the north that are in the 10s or 20% ...

In IN and OH, there is no county below 41% left on the map.

Even in densely populated states like MA, there are some crappy counties on the map.
Isnt that because the response rate is based off of the number of homes?  Pretty much every place with lots of vacation homes has low response rates.  Granted internet also usually sucks in such places so you wont have as many people responding that way either.

I think it also depends on what the Census Bureau decided to do with rural areas. A lot of the most rural areas were designated for in-person-only so haven't gotten mailings. That's why, for example, Hamilton County, NY has such a dreadfully low response rate (barely 5%). They will presumably either get mailings later this year or maybe in-person visits if/as that becomes more feasible.

The Census Bureau enumerates by location rather than by person. They don't know where you live (and the NSA won't share that with them). The census invitations were sent out to Resident 123 Oak Street, Anytown, USA 00000, with a 12-letter code. When you did an internet response, you would key in the code, and you would be asked to confirm the address.

The number of vacant housing units can be extreme. Hamilton County has 8967 housing units, but only 1124 households. If 5% of the 8967 housing units have responded, then there are returns from almost half of the households. Even if every household had filed a return, the response rate would be around 12%.

An interesting feature of Hamilton County is that it has four census tracts. Ordinarily, census tracts have between 1200 and 8000 population with 4000 considered optimum. But they also have a range of 480 to 3200 housing units, with 1600 considered optimal.

This permits areas with considerable seasonal population to provide statistically valid information for the ACS which is conducted monthly year round. The ACS criteria for inclusion is someone "staying" at the residence for two or months. The ACS form includes the date the form is filled out. Someone with a summer job would count in June, July, August. Someone visiting on a two- or three-week vacation would not.

The Census Bureau intends to count everyone, count them only once, and count them in the correct location. This means that they must look in every location that a person might be residing.

The number of housing units (which is the denominator for the response rate) has increased from 130 million to 150 million since to 2010. That is a 15.4% increase, while the population has only increased by about half that. Smaller households, including widowed and widower baby boomers, and extended-adolescent millennials may require more housing, and there may be vacant housing in northern cities such as Detroit or small declining farm towns that can not be rented, but isn't immediately torn down.

In 2010, the final mail response rate was 66.5%, which was 3% higher than it was at the end of the nominal response period (3% responded before an enumerator could track them down). But the mail return rate was 79.3%. The denominator for the mail response rate was the number of housing units forms were sent or distributed too. The denominator for the return rate omitted deleted and vacant housing units.

During the Non-Response Follow-Up (NRFU, pronounced nerfu), the Census Bureau has to send out an enumerator to get a response, or to determine that a housing unit is (or was) vacant or non-existent.

This is relatively expensive ($57 in 2010) compared to self-response which is just a few dollars for mailing. There is a tension between continuing efforts to get people to self-respond, and delaying field work (by August those who were living in an apartment on April 1 may have moved, along with the neighbors who might have recalled who lived there).

The original intent was for most self-response to be finished now, with NRFU beginning on May 13. NRFU is now delayed until mid-August, with variation possible depending on how locked down an area is.

In about 95% of the country, dubbed Self-Response Type of Enumeration Areas (TEA) the census consisted of multiple mailings to invite people to respond. For about 78% of these, they were strongly encouraged to respond by Internet. It wasn't until the fourth mailing that they received a paper questionnaire. This is just now being completed. There is a final post card being sent now, that is more urgent and threatening that the only way to prevent an enumerator being sent out is to respond.

In the other 22% of the country. A paper questionnaire was sent with the first mailing. These areas were selected based on low self-response to the ACS, and low Internet response, older population, or lack of Internet subscription.

If someone is going to respond it should have happened by now (or a week or two). What percentage of persons who have ignored five mailings are going to respond to a sixth.

In about 5% of the country, dubbed Update/Leave TEA, the intent was for a field worker to hand deliver questionnaires to housing units. The update portion of the effort was also verify existence of housing units and to correctly locate them. This was done primarily in areas without mail delivery to street addresses. They might not have mail delivery, or delivery to a rural route box or postal box.

This was intended to occur in March and April, and was just an alternative delivery mechanism to areas where a simple mailing was not feasible. This did not happen, and has been delayed until June 13. These areas tend to have more vacant housing, since they are in rural areas with more seasonal residences, and perhaps iffy housing (e.g. houses converted to barns, or barns converted to houses). While it is possible for residents in these areas to log onto the Census website, this requires more effort and self-initiative. They might have heard that the Census is delayed in their area, and not be concerned.

This probably is dropping about 2% off the overall response rate, though will be concentrated in certain counties, where the percentage of Update/Leave housing units is much higher than 5%.
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #83 on: May 01, 2020, 08:08:22 PM »

55.6% (+1.0)

The ranking page even says +1.1% (rounded) ... !

Yesterday was the best day for returns in weeks ... Smiley

Internet was 0.2, paper/phone was 0.8 to 0.9 ...
Unweighted average:

Total 1.06%
Internet 0.18%
Other (paper): 0.88%

Average for the 32 states on or east of the Mississippi: 1.02%.

Average for the other 19 states: 0.61%. The best performing western state was 20th.

This suggests that there was a hiccup in Jeffersonville, IN vs. Phoenix, AZ and may be unrealistic projecting forward.
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #84 on: May 01, 2020, 08:13:08 PM »

During the Non-Response Follow-Up (NRFU, pronounced nerfu), the Census Bureau has to send out an enumerator to get a response, or to determine that a housing unit is (or was) vacant or non-existent.

This is relatively expensive ($57 in 2010) compared to self-response which is just a few dollars for mailing. There is a tension between continuing efforts to get people to self-respond, and delaying field work (by August those who were living in an apartment on April 1 may have moved, along with the neighbors who might have recalled who lived there).

The original intent was for most self-response to be finished now, with NRFU beginning on May 13. NRFU is now delayed until mid-August, with variation possible depending on how locked down an area is.

This is probably less of an issue this year, because during pandemics and an economic meltdown we are witnessing right now ... people are less likely to move to a new address.

People are more likely to move in economically good times.

It’s therefore very likely that census enumerators will meet the people at the address they also lived in on April 1st, when they visit in the fall ...

Except for students on campus or homeless, but those are bouncy groups in each census ...
The comparison that should be made is not June 2020 and August 2020 to what would have happened between June 2020 and August 2020 but for coronavirus, but compared to how many people would have moved between June 13 12:00 AM and June 13 12:00 AM (nobody).

And in actuality, the Census Bureau took a risk in delaying the NRFU from May (as in 2010) to June (in 2020), and that has been delayed another month.

If you have lost your job, and your rent is due, you are going to move. If you move in with your parents, they will already have filled in the form in March.
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #85 on: May 01, 2020, 08:51:31 PM »

Internet response so far exceeds Census Bureau goals:

Quote
The Census Bureau, rolling with the punches as the pandemic disrupts field operations for the 2020 population count, has seen its top technology innovation exceed expectations.

With dozens of states and the District of Columbia keeping stay-at-home orders or similar restrictions in place, the bureau has seen more households opt to respond to the 2020 census via the internet than it had expected.

Al Fontenot, the associate director for decennial census programs, told members of the House Oversight and Reform Committee in a briefing last Friday that the internet self-response rate has exceeded the bureau’s expectations.

A senior Democratic Committee aide told Federal News Network that bureau officials anticipated that internet self-response at this point would be around 36%, but so far, more than 45% of households have completed the online questionnaire.

In total, about 54% of households have responded to the 2020 census.

https://federalnewsnetwork.com/it-modernization/2020/04/2020-census-internet-self-response-exceeds-goals-amid-pandemic

The Democratic senior aide may not have understood the Census officials.

The Census would presumably expect a total response similar to 2010 (63.5%).

If they had anticipated 36% Internet, that means they would have expected 27% by phone and paper. It is claimed that they expected 8% by phone (according to the article) and 19% by paper.

But those don't seem to be based on the actual communication with the public.

If you read the actual mailings, the phone number is presented as a help line.

If you consider the 12 languages that have online forms:

Spanish, Chinese, Tagalog, Vietnamese, Arabic, French, Korean, Russian, Haitian, Portuguese, Polish and Japanese; then all the other languages combined who speak English less than "very well" represent only 1% of the population.

Some of those who respond that they speak English "well", particularly when filling out an English-language census form, can read English well enough to fill out a census form. And if Oma is among the 15% of German-speakers who are recorded as not speaking English well, she is not going to be the person filling out the census form.

Someone sent over to train workers at a Mercedes or BMW plant in the South is going to be quite proficient in English, even if their wife isn't. Perhaps they would be confused about whether they are supposed to fill out the US form, but not how to fill it in.
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #86 on: May 02, 2020, 08:03:56 PM »

56.0% (+0.4)

Not bad for a Friday.

I originally estimated +4% for the coming 10-day period (May 1-10), but this period has 2 weekends instead of 1, so my new guess is +3.4% for 59% on May 10.

Weekends have bad returns.
Unweighted Averages:

Total: 0.40%
Internet: 0.16%
Total: 0.24%

Internet weekly:

14.2%, 13.3%, 10.2% (CD) 3.3% 1.9% 1.2% 1.3%

Paper weekly:

2.5%, 1.6%, 1.1% (CD) 0.3%, 0.8%, 1.4%, 1.9%

(CD) marks the first full week after Census Day.

The 5.5% who responded on paper in the first four weeks were from the 20% of respondents who received a paper questionnaire in the initial mailing. The other 80% received an invitation to respond via the Internet, with the hint about a questionnaire buried in the mailing. Some may have given up waiting on the questionnaire, or were tricked into responding online.

The questionnaire did not go out persons until after April 8. Let's assume these mailings were evenly distributed over the 23 days from April 8 to April 30. Let's also assume that most of those inclined to respond to the questionnaire in the Internet Choice group did so in March. So those who waited were:

95% in Self-Response TEA X
80% in Internet Choice group X
50% non-respondent,

and that it would take at least 5 days to respond (mail-out time + response time + mail-back time).

So by April 17, those who were mailed a questionnaire between April 8 and April 12 might have responded.

95% X 80% x 50% x 5/23 = 8%. The mail response for that week was 0.8%

By April 24, those who were mailed a questionnaire between April 8 and April 19 might have responded.

95% x 80% x 50% x 12/23 = 20%. The mail response for those two weeks was 2.2%.

By May 1, those who were mailed a questionnaire between April 8 and April 26 might have responded.

95% x 80% x 50% x 19/23 = 31%. The mail response for those three weeks was 3.7%.

By May 8, all those who were mailed a questionnaire between April 8 and April 30 might have responded.

95% x 80% x 50% x 23/23 = 38%.

The response rate for these questionnaires has edged up from 10% to 11% to 12%. If we assume it will be 13% by the end of next week, then the response rate for this questionnaire will 4.9%, or a weekly response of 1.2% (a drop from this weeks 1.9%).

It will fall off even more after the questionnaire has been sitting under a pizza box for several weeks.

The Internet rate should hold up this week because of the 5th mailing postcard.

Together we should see about 2.5% this next week.
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #87 on: May 03, 2020, 02:16:52 AM »

Best Responding Cities

In 100-199 category Lincolnshire, KY is yet another Louisville suburb to join the list. There are about 80 cities in Jefferson County. When Louisville and Jefferson County merged in 2003, they maintained their separate identities and provide certain services beyond that of the Metro Government.

Broeck Pointe, KY regained its lead in the 200-499 range, and Strathmoor Village, KY took the lead for cities with between 500 and 999 population.

In the 100K-200K category Sterling Heights, MI surged ahead of Centennial, CO. In the 500K-1M range, Louisville/Jefferson County metro government (balance), KY passed Seattle, WA. This is area contained in the city of Louisville and unincorporated portions of Jefferson County prior to the merger. Neither of these areas retain a separate legal status under the Metro government.

1 to 9McMullen, ALPickens65.2%Small all-black town outside Aliceville, west of Tuscaloosa. Likely 16/23 response.
10 to 18Bonanza, COSaguache4.8%Almost ghost town survived effort to have its government abandoned. At one time had 36 dance halls and 7 saloons. Likely 1/21 response.
20 to 49Harbine, NEJefferson76.9%Southwest of Lincoln, a bit north of the Kansas/Nebraska line. Likely 20 of 26 responses.
50 to 99North River, NDCass100.0%North of Fargo along Red River. All 23 households have responded.
100 to 199Lincolnshire, KYJefferson87.3%Another eastern suburb of Louisville (48/55, 55/63, or 62/71)
200 to 499Broeck Pointe, KYJefferson89.7%Eastern suburb of Louisville (93/107 or 100/114)
500 to 999Strathmoor Village, KYJefferson88.3%Eastern suburb of Louisville.
1K to 2KSt. Regis Park, KYJefferson87.8%11 miles east of Louisville (491/559 or 534/608)
2K to 5KHowards Grove, WISheboygan87.3%Farming community in the northern part of the county, seeing modest suburban growth.
5K to 10KHuntington Woods, MIOakland88.1%Immediately northwest of Pleasant Ridge, includes part of the Detroit Zoo.
10K to 20KDardenne Prairie, MOSt. Charles84.6%Northwest St. Louis suburb, incorporated in 1983, has octupled in population since 1990 Census.
20K to 50KAndover, MNAnoka83.0%Just north of Coon Rapids, with half of the population added since 1990.
50K to 100KMaple Grove, MNHennepin79.2%Northwest of Minneapolis, about 20 minutes in light traffic.
100K to 200KSterling Heights, MIMacomb78.0%Detroit suburb, 4th most populous city In Michigan.
200K to 500KLincoln, NELancaster69.7%State capital and home of University of Nebraska.
500K to 1MLouisville/Jefferson County metro government (balance), KYJefferson67.5%Lousville and unincorporated portions of county prior to government merger.
1M to 2MSan Jose, CASanta Clara64.7%High-tech Silicon Valley giant, areas along southern edge are over 75% response rates.
2M to 5MChicago, ILCook, DuPage48.8%Three cities in this range are tightly bunched.
5M to 10MNew York, NYKings, Queens, New York, Bronx, Richmond44.9%USA's largest city made up of Brooklyn, Queens, Manhattan, The Bronx, and Staten Island boroughs.

Largest city to have exceeded certain thresholds.

New Berlin, WI; Elk Grove, CA; and San Diego, CA become largest cities to surpass 80%+, 70%+, and 60%+ thresholds. Several cities are poised to reach 90% in the next few days. North River is fairly safe at 100%. Not only must all households respond, all housing units must be occupied. This is possible in a tiny suburb with 23 families but not much beyond.

100%=North River, ND55
90%+North River, ND55
80%+New Berlin, WI39,733
70%+Elk Grove, CA172,886
60%+San Diego, CA1,425,976
50%+Phoenix, AZ1,660,272
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #88 on: May 03, 2020, 02:59:50 PM »

56.3% (+0.3%)

The ranking page has it rounded as +0.4%
Unweighted average of states:

Total 0.36%
Internet 0.14%
Paper 0.22%

If it was 0.36%, then if would round to 0.4%.

But if the total had been between 55.95% and 55.99%, the new total would be 56.31% and
56.35%, and rounding would show a change from 56.0% to 56.3%.

We should be able to track this. Our initial estimate is that the total is between 56.31% and 56.35%.

Internet responses have declined five consecutive days.

It is hard to tell on the mail because there are operational things going on, such as the eastern glitch on Thursday.
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #89 on: May 03, 2020, 03:46:46 PM »

The Uni of NY also has self-response rates for states from the 2000 Census and compared them to 2010 rates:

Link

US-wide, the 2000 response rate was 67.4% vs. 66.5% in 2010, so it dropped a bit.

But it increased in 17 states+DC, remained the same in FL, and decreased in 32 states.

The verdict for 2020 will only be known in half a year, because people have until Oct. 31 to self-respond, but I guess mid-August will be a better indicator which is when census taker will probably hit the roads.

I think you may be misunderstanding the 3.0% difference between the nominal end of the self-response period and the final number (63.5% and 66.5%).

My neighbor was a salesperson who was often on the road attending conventions and shows. Her apartment was as much an office as a place to sleep.

In (June 2010?) a Census worker appeared. I said I thought my neighbor was around and checked to see if her car was there. It was, so the census worker waited a bit and then left a notice on the door. As she was leaving, my neighbor returned from walking her dog. We hailed the census worker back, and she came back and filled out the form.

It is possible that the notice said to either return the form or call this number. If the resident did return the form in this case it might be counted as a self-response.

Right now 2020 is 9.0% behind 2010 on a May 1 to May 1 basis.
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #90 on: May 04, 2020, 06:26:02 AM »

Best Responding Cities


Broeck Pointe, KY becomes second city to reach 90%. Louisville suburbs may be helped by Census Bureau's main processing facility just across Ohio River in Jeffersonville, IN.

Centennial, CO ties Sterling Heights, MI in 100K-200K range.

1 to 9McMullen, ALPickens65.2%Small all-black town outside Aliceville, west of Tuscaloosa. Likely 16/23 response.
10 to 18Bonanza, COSaguache4.8%Almost ghost town survived effort to have its government abandoned. At one time had 36 dance halls and 7 saloons. Likely 1/21 response.
20 to 49Harbine, NEJefferson76.9%Southwest of Lincoln, a bit north of the Kansas/Nebraska line. Likely 20 of 26 responses.
50 to 99North River, NDCass100.0%North of Fargo along Red River. All 23 households have responded.
100 to 199Lincolnshire, KYJefferson87.3%Another eastern suburb of Louisville (48/55, 55/63, or 62/71)
200 to 499Broeck Pointe, KYJefferson90.7%Eastern suburb of Louisville (88/97)
500 to 999Strathmoor Village, KYJefferson88.6%Eastern suburb of Louisville (234/264)
1K to 2KSt. Regis Park, KYJefferson87.8%11 miles east of Louisville (491/559 or 534/608)
2K to 5KHowards Grove, WISheboygan87.3%Farming community in the northern part of the county, seeing modest suburban growth.
5K to 10KHuntington Woods, MIOakland88.3%Immediately northwest of Pleasant Ridge, includes part of the Detroit Zoo.
10K to 20KDardenne Prairie, MOSt. Charles84.9%Northwest St. Louis suburb, incorporated in 1983, has octupled in population since 1990 Census.
20K to 50KAndover, MNAnoka83.4%Just north of Coon Rapids, with half of the population added since 1990.
50K to 100KMaple Grove, MNHennepin79.5%Northwest of Minneapolis, about 20 minutes in light traffic.
100K to 200KCentennial, COArapahoe78.5%Suburb southeast of Denver, largest city in the US at time of incorporation. Tied with Sterling Heights.
(do)Sterling Heights, MIMacomb78.5%Detroit suburb, 4th most populous city In Michigan
200K to 500KLincoln, NELancaster70.0%State capital and home of University of Nebraska.
500K to 1MLouisville/Jefferson County metro government (balance), KYJefferson67.8%Lousville and unincorporated portions of county prior to government merger.
1M to 2MSan Jose, CASanta Clara65.1%High-tech Silicon Valley giant, areas along southern edge are over 75% response rates.
2M to 5MChicago, ILCook, DuPage49.0%Three cities in this range are tightly bunched.
5M to 10MNew York, NYKings, Queens, New York, Bronx, Richmond45.2%USA's largest city made up of Brooklyn, Queens, Manhattan, The Bronx, and Staten Island boroughs.

Largest city to have exceeded certain thresholds.

Broeck Pointe, KY becomes second, and largest, city to reach 90%+. Lincoln, NE is largest to reach 70%+.


100%=North River, ND55
90%+Broeck Pointe, KY283
80%+New Berlin, WI39,733
70%+Lincoln, NE287,401
60%+San Diego, CA1,425,976
50%+Phoenix, AZ1,660,272
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #91 on: May 04, 2020, 06:01:14 PM »

Unweighted average:

Total: 0.32%
Internet: 0.13%
Paper & Phone: 0.19%

Sixth consecutive day for Internet decline.

Paper returns held up from Saturday. It looks like they are processing on Sunday, perhaps working off a backup from last Thursday, or to maintain social distancing in their processing facility.
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #92 on: May 04, 2020, 09:32:11 PM »

Also, could Jefferson County, KY be growing surprisingly fast?
I think there is a connection to the main Census Bureau National Processing Center being in Jeffersonville, IN as it has been since the 1960s, which is directly across the Ohio River from Louisville. New Albany is opposite Louisville at the falls of the Ohio. Jeffersonville is upstream.   I-65 (Chicago-Indianapolis-Louisville-Nashville-Decatur(Huntsville)-Birmingham-Montgomery-Mobile) is a direct route from the eastern suburbs to Jeffersonville. If you are coming from the west on I-64 you cross the river at New Albany and then reach I-65 to go north to the Census Bureau facility.

I suspect that many of those who work at the facility live in the Louisville suburbs. This might be particularly true during the Census (in 2019, they were hiring to expand from 1642 to 4057 employees, mostly clerical). These would likely fade away going in to 2021 etc.

I found out about the hiring in the Louisville newspaper. So there will an inordinate interest in the operational aspects of the Census, where most people's experience will be a letter in the mail followed by logging in to an app, and after 10 minutes pressing submit. There will be awareness, likely with feature stories on TV and in the newspaper.

High growth would not necessarily lead to high response rates. In high growth areas you will have housing units that are under construction or not yet occupied. These units are guaranteed not to respond.

The highest response rates are in well-settled residential areas, with all or almost all single-family housing, and little or no housing that might be disused. Families of four with a dog, living in suburbia are the most likely to respond.
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #93 on: May 04, 2020, 10:30:18 PM »

Hmm. Thanks for doing these updates, jimrtex. I love the city rankings. I'm rather entertained that I grew up next to (and had my first elementary school in) the town of >50k with the highest response rate in the country, and now live in the town of >100k with the lowest response rate in the country, as of the last time you crunched the numbers. Is that still true of Hartford?
Thanks, I enjoyed researching the cities. I think muon2 grew up in Andover, or perhaps Coon Rapids. Andover is the top city in the 20K-50K category.

Hartford is 3.5% below Providence, RI in its category.

Hartford, Bridgeport, New Haven, and Waterbury are 1st, 4th, 6th, and 7th nationally. The city of Hartford is 36.8%, while the county of Hartford is 61.3%.

This shows the effect of how towns of a certain size were established in 18th and early 19th century, and prevented from growing in the automobile era. In states like Connecticut, there would be established town governments that would provide political resistance to growth of the cities. Even now, there likely is limited regional cooperation.
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #94 on: May 05, 2020, 01:13:55 AM »

Just curious, does anyone know when the US reveals the results of the census categorically? Stuff like total population, ethnic background, age, ect? Canada for example releases a schedule of when they release this sort of info so I'm curious.
The first data product is the total numbers for apportionment of representatives. This is required to be delivered to the President by December 31. The Census Bureau has requested that it be delayed until March 31, 2021. This requires a statute change, but will likely be stuck in a coronavirus bill.

The second data product is the PL94-171 data that is supposed to be delivered within a year of the Census Date (April 1, 2021). PL94-171 was passed in the 94th Congress (1975-1977) and requires the Census Bureau to work with the states to provide data for redistricting (gerrymandering).

This is produced down to the census block level, and includes head counts by racial category and ethnicity (Hispanic or not). The Census regards Hispanic as orthogonal to race.

A respondent may specify one or more of six racial categories:

(1) White
(2) Black
(3) Asian (they actually check Chinese, Asian Indian, etc.), but are aggregated as Asian,
(4) American Indian or Alaska Native (with a fill-in the blank for tribe, with a distinction for enrolled members)
(5) Native Hawaiian or other Pacific Islander (with a fill-in-the blank specification).
(6) Other (with a fill-in the blank).

For 2020, White and Black also have a fill-in the blank. For blacks, they would be wanting to distinguish Jamaican, Haitian, Nigerian, etc, from African-American. For Whites, they appear to be able to isolate people from the Middle East, who may consider themselves to be ethnically distinguished from Europeans. I entered "American".

There are 63 (26 - 1) racial combinations - I don't think non-racial is allowed, though perhaps persons can respond: Other (None),

Counts are also produced by Hispanic or not Hispanic. These are produced for the entire population and the over-18 (voting age population).

The main figure is the total population since that is used for equalizing district populations, and may be used for apportionment in states that apportion representatives to counties.

It is illegal to assign persons to districts on the basis of race, so you have to pretend you aren't by using the racial data.

The Census Bureau has asked for a delay until July of 2021, which will likely put a crimp in redistricting, since filing deadlines for 2022 begin in December 2021.

Other reports will come out later, probably around the end of 2021. There is not a lot of additional information in the Census Data. They will be able to produce reports by single years. They will be able to break down the Asian population as Chinese, Burmese, etc.; and do the same for Blacks, Whites, AIAN, and HOPI.

There is relational information. I think a lot of that is for accounting purposes, so if you report five people living at your house, who are they, and to remind you who else to report (a husband and wife, with three children, might think they are reporting their immediate family, and not the cousin who is living with them between jobs or abusive relationships).

The more detailed information about subjects such as income, and rent, and education, and military etc. comes from the American Community Survey (ACS) which is a very large sample (about 1:8) conducted monthly over a five year period. That is of the 16 million or so households surveyed, 1/60 will be questioned each month over a five-year period. The latest data was reported in September 2019, covering the ACS from 2014-2018. You can sort of think of it as reflecting conditions at the midpoint of the period in 2016, particularly if there is a trend, such as increased Hispanic population in an area. If there was 14% in 2014, 15% in 2015, 16% in 2016, 17% in 2017, 18% in 2018, the aggregate sample will show 16%.

The next report, presumably around September will cover 2015-2019. They will drop the 2014 data, and add in the 2019 data. Data collection for the 2020 ACS continues during the census.

The 2020 Census also serves as establishing the sample frame for the ACS going forward.

The ACS also produces one-year reports. This is statistically valid for areas with 60,000+ population, so could be used for states or congressional districts, or larger cities.

It is the ACS that asks the citizenship question, and is used to produce CVAP estimates.
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #95 on: May 05, 2020, 08:56:28 AM »

The Census Bureau will re-start operations now in 13 states & mostly in rural Update-Leave areas:

Quote
After suspending most of its operations in March due to the coronavirus pandemic, the Census Bureau announced Monday it plans to restart some of its fieldwork in rural areas around the country.

Those operations had been delayed through early June, but agency officials said employees would start to hand-deliver invitations to complete the census to homes later this week in parts of 13 states: Alaska, Alabama, Arkansas, Idaho, Maine, Mississippi, Montana, North Dakota, Oklahoma, Tennessee, Utah, Vermont and West Virginia.

The Census Bureau does not mail invitations to areas that have post office boxes instead of street addresses. Many rural areas also do not have internet access that can be used to respond to the census.

In Monday's announcement, the agency said that field staff would receive "safety training to observe social distancing protocols" and personal protective equipment.

The agency will open about two dozen offices around the country that fall under the "Update/Leave" category of hand delivery. The announcement does not cover Puerto Rico, where operations have also stalled amid the pandemic. 

Al Fontenot, the director of the agency's decennial programs, had previously said the agency may start to phase-in operations throughout the country depending on local conditions.

The Census Bureau said it will spin up both mail delivery and hiring in those local offices, making decisions about each local office on an individual basis going forward.

Response rates in rural areas of the country have lagged behind the national rate, which currently stands at more than 56 percent, over the past two months.

Last month, the Census Bureau said it would pause most operations until at least the start of June. The agency also requested a 120-day deadline extension from Congress, which has not yet been acted on.

https://www.rollcall.com/2020/05/04/census-bureau-to-restart-operations-in-rural-areas-of-country

This is the Census Bureau announcement.

Census Bureau to Resume Some 2020 Census Field Operations in Select Locations

This is a list of the Area Census Offices (ACO) that are re-opening, and the number of Update/Leave (UL) Housing Units in each. The ACO's cover the entirety of the 13 states being re-opened. I suspect the governors of each state could make the final decision on whether to re-start operations in their states.

Status of Current Operations (PDF)

This is a map of Area Census Offices. It is somewhat misleading because it shows the location of the offices rather than the area each covers. It appears the 248 offices were located based on anticipated work load, so that each will manage similar numbers of field workers. The offices are then located in more urban locations, where it would be easier to staff and build-out an office. Field workers will be able to work remotely, receiving work assignments on their Census laptop.

Area Census Office Locations and Pay Rates

For example, in Oklahoma, the Cleveland County (Norman) office covers the southern part of the state from Beckham to Le Flore and counties to the south. The Oklahoma County office covers the northwestern part of the state from Oklahoma County north and westward, and the Tulsa office covers the northeastern part of the state.

These area offices should be able to stage operations concentrating on rural areas with less incidence of COVID-19, though Update/Leave areas are inherently rural - even small towns will have street addresses.
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #96 on: May 05, 2020, 10:48:50 AM »

Estimates of Effect of Update/Leave Resumption

StateResponsesSelf-ResponseUL HUSR HUSR TEA ResponseSR Response RateUL ResponsesEstimate Response RateChange
Alabama140054.6%822482139656.2%3956.0%1.4%
Alaska12037.4%11021111554.3%5151.5%14.1%
Arkansas79051.4%77146078653.8%3553.4%2.0%
Idaho47059.6%7171846665.0%3964.1%4.5%
Maine36048.2%11263535455.8%5354.6%6.4%
Mississippi75052.1%55138574754.0%2553.7%1.6%
Montana28048.2%11346827458.6%5656.9%8.7%
North Dakota23056.6%4236422862.5%2261.6%5.0%
Oklahoma96050.7%188170595155.7%8954.9%4.2%
Tennessee180055.7%93223180055.8%455.8%0.1%
Utah73061.1%89110672665.6%5064.9%3.8%
Vermont17047.8%6229416756.8%3055.4%7.6%
West Virginia43044.1%28968641660.6%14957.9%13.8%

From the May 3 Census Response Rates, we know how many households have responded in each state (in thousands), as well as the self-response rates. We also know the number of Update/Leave housing units from the report of ACO openings.

The response rate is calculated using a denominator based on all housing units in a state in both Self-Response (SR) and Update/Leave (UL) Type of Enumeration Areas (TEA).

Self-Response Rate = Total Self-Responses / Total Housing-Units

Total Housing-Units = (SR Housing-Units + UL Housing-Units)

We can solve for number of SR Housing Units in a state. Note it does not matter if some responses have come from UL TEA.

For example, in Alabama there have been 1400K responses, for a response rate of 54.6%, and there are 82K ULHU.

Thus 54.6% = 1400K / (SRHU + 82K)

Solving for SRHU = 2482K

Assumption 5% of housing units in UL TEA have responded. It is possible to respond to the census without an ID (or invitation), and the Census Bureau may have done a mailing to RR and postal boxes.

We can remove these UL responses from the total responses for a state, and calculate a self-response rate for self-response TEA (that is, the percentage of housing units that were mailed invitations and questionnaires beginning in March and have responded.

For Alabama, 5% of 82K UL HU = 4K estimated responses from UL TEA. This leaves 1396K self-responses from SR TEA, or an effective Self-Response rate in those areas of 1396K / 2482K = 56.2%. That is 56.2% of housing units in SR TEA have responded.

Assumption Response rates in UL TEA will be similar to that in SR TEA, with a fudge factor of 0.85 applied. In 2010, Update/Leave TEA had substantially lower response rates, but return rates that were only a few percentage points lower. The response rate is the percentage of housing units that respond. The return rate is the percentage of households that return a form. Vacant or non-existent housing units can not respond. There are more vacant, second residences, etc. in UL TEA. The Update portion of Update/Leave is in part to determine these vacant housing units so that they can be excluded from the NRFU. There also appears to be a tiny differential in returns. Not having a street address may correlate with economic status and other demographic factors (e.g. the UL TEA in Alabama are concentrated in the Black Belt. Only 3% of housing units in Alabama are UL, but up to half in some rural counties are).

With this fudge factor applied, we can estimate the total number of UL responses (This includes persons who have already responded in these areas. Just because there are some self-starters should not effect the overall response).

In Alabama, we estimate a SR Rate for UL TEA = 85% of SR TEA rate. = 85% x 56.2% = 47.8%. Most of this lower performance is due to greater numbers of vacant and deleted housing in UL TEA. There was a tiny bit (around 3% due to a smaller share of households responding).

Now that we have the total number of housing units and the state, and the expected number of responses, we can calculate an estimate for the response rate IF the questionnaires had been delivered to the UL TEA in March and April.

In Alabama we have (1400K total responses - 4K estimated from UL TEA + 39K estimated UL responses) / (2482K + 82K) = 56.0%. This is an increase of 1.4% from the current 55.6% response rate.

Under ordinary circumstances, we would preparing for the beginning of NRFU in mid-May.
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #97 on: May 05, 2020, 05:06:06 PM »

Here’s the CB press release about re-starting the Update/Leave process in 13 states:

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

Key aspects:

Quote
As part of the phased restart of operations, the Census Bureau will resume dropping off 2020 Census invitation packets at front doors of households in areas where the majority of households do not receive mail at their home. This operation is also known as Update Leave. About 5% of households are counted in the Update Leave operation, where census workers will confirm or update a household’s physical location address and then leave a census questionnaire packet.

Quote
The Update Leave operation does not require interaction between households and a Census Bureau employee and follows the most current federal health and safety guidelines.

Quote
The health and safety of Census Bureau staff and the public is of the utmost importance. All returning staff will receive safety training to observe social distancing protocols in the COVID-19 environment. For their safety and the safety of the public, the Census Bureau has ordered Personal Protective Equipment (PPE) for all field staff, including those that work in a field office. These materials will be secured and provided to staff prior to restarting operations.

Quote
Households that receive the 2020 Census invitation packets are strongly encouraged to respond promptly to the 2020 Census using the census ID included in the questionnaire packet. People can respond online, by phone or by completing and returning the paper form by mail. Responding with the census ID or by completing and returning the paper questionnaire helps ensure the best count of their community.

Important to note that WV (southern part) is included in this first re-start.

It’s also the biggest region to „re-open“ with ca. 300.000 mostly rural households to be visited (incl. the poor McDowell County).

Interesting also that rural TX is not among the first re-starters ...
I doubt that they have given up on knocking on doors under Update/Leave it is downright uncivil, if not dangerous, to be walking around houses in a rural area without greeting people you see. The field worker might be mistook for a revenuer.

Perhaps they will knock on the door and then stand back or knock with a 6-foot pole. If I were leaving a questionnaire packet, I'd want to use a door hanger, and not just lay it on a porch or doorstep.

It is not clear how many field offices that there are. It appears that field workers work from home and get their work assignments from an app on their Census Bureau Tablet. Supervision is done remotely.

A field worker would be assigned a Basic Collection Unit (BCU), and would be routed to that area and directed to each housing unit of interest. During an Update/Leave operation they would find each housing unit, verify whether or not it exists, is habitable, and vacant or not. They knock on the door, attempt to get alternate contact information, and leave a census questionnaire (along with an ID that would permit Internet response). It was never intended that they would actually interview anyone (as is done in a few remote locations in remote Alaska and Maine).

All this activity would be logged. If they finished a BCU, they could log in and be assigned another BCU. Or at the end of the day, their progress would be recorded. They might be assigned the remainder of a BCU the next day, or perhaps not. My understanding is that there is flexibility in hours for workers, and AI is driving the assignment.

Since an ACO might cover an entire state, there has to be a way to distribute physical materials such as questionnaires, tablets, and now, PPE. It would be too far to drive to the ACO on a daily basis. It's possible that field offices are little more than supply depots. At least some training is done online.

The Beckley ACO covers the entire state of WV. That it is in Beckley rather than say Charleston is probably some pork barrel dating back to Robert Byrd. Or perhaps it recognizes that more field work will be needed in the southern part of the state.

If you go to the map of ACO offices and click through to tabular information, you can enter a state and county, and find out which ACO covers the county. Jefferson and Ohio counties in the panhandles point back to Beckley.

Scroll down to Find Pay Rates Near You, and enter a state and county name

I think that every ACO corresponds to an area within a single state, with the exception of the ACO in Window Rock, AZ which covers counties in Arizona and New Mexico that include most of the Navajo Nation (San Juan, UT is under the Orem, UT office).

The ACO appear to be designed to balance the overall workload, particularly during Non-Response-Followup (NRFU). In Texas, there are half-a-dozen ACO in Harris Counties. The Fort Bend and Montgomery ACO might appear to be suburban offices, but the Montgomery ACO also covers Huntsville, Lufkin, and the Golden Triangle. The Fort Bend ACO includes Fort Bend, Brazoria, and Galveston, but also down the coast to just past Victoria.

The ACO selected cover entire states. I suspect strongly that governors could give a go/nogo on reopening (11 of the 13 have Republican governors). They probably wanted to have a gradual re-opening (28 of 248 ACO) that had widespread footprint (13 states).

Texas has 26 ACO. Within ACO, they may be able to stage work to avoid current hotspots, but they'd rather have a widespread effort. The field workers in McDowell, WV aren't going to be doing the field work in Randolph, WV. They might not even be going into the next county.

Notice that in Tennessee, that two of the ACO have no UL HU (statewide only 0.5% are UL TEA). They'll be able to knock those out quickly.

It is possible that later on they will selectively open ACO. But for example, in Texas much of rural west Texas is under the Lubbock ACO. Lubbock has a higher infection rate than much of the state, including the major metro areas. Better choices might be the San Antonio ACO's or Williamson ACO, but they probably have fewer UL housing units.

Consider the 289K UL Housing units in West Virginia. Let's guess that it takes 15 minutes per Housing Unit. A field worker will probably have to travel by car. She will have find a place to park her car, or decide to drive into a driveway. She is going to have to walk up to the house. She may have to wait for a child to fetch her Mom, or wait for the householder who was tinkering on his tractor to amble down from the barn. She will have to record the location of the house, and prepare the packet,

289K / 4 per hour / 6 hours per day (excluding prep-time, close-out-time, travel, and breaks) / 25 days per work month = 482 field worker-months, which we will double because I'm just making this all up, back-of-the-envelope. So 1000 field workers. You really want to get this done sooner rather than later, because all of the Update/Leave households become potential targets for repeated mailings to non-responders.
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #98 on: May 05, 2020, 09:35:59 PM »

Under ordinary circumstances, we would preparing for the beginning of NRFU in mid-May.

It is much, much better to let households self-respond longer than usual this year.

Health plays a role (fewer census takers have to interview householders, therefore limiting personal contacts).

But also efficiency: the more households respond on their own, preferably online with their code, the easier it is for the Census Bureau to verify the entries and process them.

A census taker visiting a household is much more costly, if they don't respond.

Therefore, the Bureau should launch much more advertising and even more reminder cards or another round of forms. Because this would still be much cheaper than sending someone.

With these measures, self-response could be 75-80% by August. A ton of money could be saved.

I have highlighted the portions that are hopeful.

What if it doesn't happen?

You seem to forget that the denominator for the response rate is the number of housing units. About 65.5% of households have responded (in 2010 this approached 80%).

If someone has not responded after 2 letters, 2 postcards, and a questionnaire, what makes you think that they will respond to a commercial on the television or another postcard?
Logged
jimrtex
Atlas Icon
*****
Posts: 11,817
Marshall Islands


« Reply #99 on: May 05, 2020, 10:38:26 PM »

JimboRTX

What is your educated guess on when we’ll see counties like McDowell shoot up from the current 4% response rate to 30-40% ?

Let’s assume they start visiting places next week ... maybe in one month ? Two ?
I think that they will do the update/leave in about a month. The way it is set up, it can be done piecework. The Update/Leave actually began on March 15, before it was shut down three days later. They would have hired the field workers, and they will still be available unless they are elderly. They could hire additional field workers. School teachers are usually paid on a 12-month year, but they often work summer jobs for some extra cash, unless they go back to school for advanced degrees (a M.A. may pay extra, and also provide advancement opportunities into management) or some do travel, but not in 2020.

Wearing a mask and gloves should not really slow down the work if you don't have to talk to people. McDowell has only had 6 COVID-19 cases.
Logged
Pages: 1 2 3 [4] 5 6 7 8 9 ... 22  
Jump to:  


Login with username, password and session length

Terms of Service - DMCA Agent and Policy - Privacy Policy and Cookies

Powered by SMF 1.1.21 | SMF © 2015, Simple Machines

Page created in 0.113 seconds with 13 queries.