Census Bureau Error (Estimates vs. Actual Census Count)

[Tender Branson]

I was bored and created an Excel chart to compare Census Bureau errors with the actual 2000 Census count. For this, I used the Mid-1999 Census Bureau estimates and allocated growth from 1998-1999 to April 1, 2000. This means I split the total growth in 4 quarters and multiplied it by 3 to go from July '99 to April 2000 (3 quarters). Here´s what I got:

Percentage Error (Estimates vs. Census Count by state)



As you can see, the CB was really off in predicting the results in DC and was best in West Virginia.

Numerical Error (Estimates vs. Census Count by state)



The Census Bureau predicted that New York would have 752.000 fewer people on Census Day than it actually had, but was really close to the end results in Alaska.

[Cubby]

Very interesting. The fast growing states generally seem to be at the top, which is understandable, since its hard to get a good read on places that are seeing lots of in-migration. No excuses for some slow growing states to be up there too.

Even in 1999 they still had CT as a decline over 1990. They really over-estimated the impact of the '91 recession on the state. It was severe, but it had more of an effect on jobs and finances than population.

I thought it was great that no state declined in population between 1990 and 2000, that last happened in the 1870's I believe.

[jimrtex]

Very interesting. The fast growing states generally seem to be at the top, which is understandable, since its hard to get a good read on places that are seeing lots of in-migration. No excuses for some slow growing states to be up there too.

Even in 1999 they still had CT as a decline over 1990. They really over-estimated the impact of the '91 recession on the state. It was severe, but it had more of an effect on jobs and finances than population.

I thought it was great that no state declined in population between 1990 and 2000, that last happened in the 1870's I believe.

The Census Bureau bases its migration rates on income tax data (the IRS strips out everything but the county) and Social Security data.  And it bases the estimates for the each decade on the previous census.  So if they do a better coverage in a census it will show up as an error in their estimates.

The error in the estimates is highly concentrated in certain age groups.

[Bo]

I thought it was great that no state declined in population between 1990 and 2000, that last happened in the 1870's I believe.

I think you mean 1970s, since I know for a fact West Virginia's population declined in some parts of the 20th century.

[Хahar 🤔]

I thought it was great that no state declined in population between 1990 and 2000, that last happened in the 1870's I believe.

I think you mean 1970s, since I know for a fact West Virginia's population declined in some parts of the 20th century.

Read what he said again.

[Cubby]

States with Population declines between decennial censuses (censi?)  This list may be incomplete:

1880's: Nevada
1890's: Nevada
1900's: Iowa
1910's: Nevada, Vermont
1920's: Montana
1930's: North Dakota, South Dakota, Nebraska, Kansas, Oklahoma, Vermont
1940's: North Dakota, Mississippi, Arkansas, Oklahoma
1950's: West Virginia, Mississippi, Arkansas
1960's: West Virginia, North Dakota, South Dakota
1970's: New York, Rhode Island
1980's: West Virginia, Iowa, North Dakota, Wyoming

Washington DC declined every decade from the 50's to the 90's, but if the estimates are right, this past decade will show an increase for the first time since the 1940's.

[muon2]

The Census Bureau claimed to have learned from their estimation errors last decade, and aimed to do better this last decade. One tool to improve their estimates has been the American Community Survey with its rolling three year samples since mid decade. However, if they were prone to the same type of underestimate in fast growing areas, the recession may be their biggest help at getting it right, since it has generally slowed migration.

[Bo]

I thought it was great that no state declined in population between 1990 and 2000, that last happened in the 1870's I believe.

I think you mean 1970s, since I know for a fact West Virginia's population declined in some parts of the 20th century.

Read what he said again.

I did.

Also, do you have the error data for the 1990 U.S. census?

[Obnoxiously Slutty Girly Girl]

States with Population declines between decennial censuses (censi?)  This list may be incomplete:

1880's: Nevada
1890's: Nevada
1900's: Iowa
1910's: Nevada, Vermont
1920's: Montana
1930's: North Dakota, South Dakota, Nebraska, Kansas, Oklahoma, Vermont
1940's: North Dakota, Mississippi, Arkansas, Oklahoma
1950's: West Virginia, Mississippi, Arkansas
1960's: West Virginia, North Dakota, South Dakota
1970's: New York, Rhode Island
1980's: West Virginia, Iowa, North Dakota, Wyoming

Washington DC declined every decade from the 50's to the 90's, but if the estimates are right, this past decade will show an increase for the first time since the 1940's.

Well, as the government destroyed the economy in the rest of the country, people moved to D.C. to do parasitic government jobs.

[Хahar 🤔]

I thought it was great that no state declined in population between 1990 and 2000, that last happened in the 1870's I believe.

I think you mean 1970s, since I know for a fact West Virginia's population declined in some parts of the 20th century.

Read what he said again.

I did.

Also, do you have the error data for the 1990 U.S. census?

And did you understand what he said this time?

[Bo]

I thought it was great that no state declined in population between 1990 and 2000, that last happened in the 1870's I believe.

I think you mean 1970s, since I know for a fact West Virginia's population declined in some parts of the 20th century.

Read what he said again.

I did.

Also, do you have the error data for the 1990 U.S. census?

And did you understand what he said this time?

Yes. That the last time the population grew in all the U.S. states at the same time was in the 1870s.

[jimrtex]

States with Population declines between decennial censuses (censi?)  This list may be incomplete:

1830s: Virginia (if WV counted separately)
1860's: Maine, New Hampshire
1880's: Nevada, Alaska territory
1890's: Nevada
1900's: Iowa
1910's: Nevada, Vermont,Mississippi.Alaska territory
1920's: Montana
1930's: North Dakota, South Dakota, Nebraska, Kansas, Oklahoma, Vermont
1940's: North Dakota, Mississippi, Arkansas, Oklahoma
1950's: West Virginia, Mississippi, Arkansas
1960's: West Virginia, North Dakota, South Dakota
1970's: New York, Rhode Island
1980's: West Virginia, Iowa, North Dakota, Wyoming

Washington DC declined every decade from the 50's to the 90's, but if the estimates are right, this past decade will show an increase for the first time since the 1940's.

Longest period of recovery.

North Dakota: 1930 to present.
West Virginia: 1950 to present
South Dakota: 1930 to 1990 (ie 1990 population surpassed 1930)
Oklahoma: 1930 to 1970
Arkansas: 1940 to 1980
Nevada: 1880 to 1910
Nebraska: 1930 to 1960
Mississippi: 1940 to 1970
New York: 1970 to 2000

Slowest growing states: 1900 to 2000

IA 1.31
NE 1.60
VT 1.77
MO 1.80
KS 1.83
MS 1.83
ME 1.84
SD 1.88
KY 1.88
WV 1.89
PA 1.95

[Queen Mum Inks.LWC]

Very interesting.  Although your second image isn't showing up.

[Tender Branson]

Very interesting.  Although your second image isn't showing up.

Now it should be working ...