Tennessee Census Estimates are Understated

[Hankster]

As has been stated on this forum, many have expressed doubt that census estimates are very accurate. Today I stumbled upon a series of census estimates for Tennessee for 1990 to 1999 that were revised (one could say corrected) after the 2000 Census results were determined. I decided to compare these revised estimates with the Census 1998 Estmates which were published shortly before the 2000 Census to see how accurate they really were. Just as I suspected, they were way off as the data below shows. Some counties were off by as much as 10%! Davidson County was underestimated by more than 32,000 and Tennessee was underestimated by nearly 140,000. Wow!

County..Original '98 Est...Revised '98 Est...Revision Amt...% Revision

Anderson 71,116 71,321 205 0.29%

Bedford 34,533 36,275 1,742 5.04%

Benton 16,328 16,353 25 0.15%

Bledsoe 10,795 11,834 1,039 9.62%

Blount 101,295 102,631 1,336 1.32%

Bradley 83,292 86,120 2,828 3.40%

Campbell 38,241 39,107 866 2.26%

Cannon 12,139 12,367 228 1.88%

Carroll 29,115 29,281 166 0.57%

Carter 53,323 55,920 2,597 4.87%

Cheatham 35,344 34,591 -753 -2.13%

Chester 14,700 15,123 423 2.88%

Claiborne 29,529 29,365 -164 -0.56%

Clay 7,255 7,816 561 7.73%

Cocke 31,968 32,791 823 2.57%

Coffee 45,767 46,863 1,096 2.39%

Crockett 13,959 14,364 405 2.90%

Cumberland44,291 44,788 497 1.12%

Davidson 533,967 566,150 32,183 6.03%

Decatur 10,807 11,588 781 7.23%

Dekalb 15,943 17,027 1,084 6.80%

Dickson 42,254 41,958 -296 -0.70%

Dyer 36,782 36,881 99 0.27%

Fayette 30,457 27,465 -2,992 -9.82%

Fentress 16,184 16,301 117 0.72%

Franklin 37,465 38,667 1,202 3.21%

Gibson 48,186 48,186 0 0.00%

Giles 28,925 29,090 165 0.57%

Grainger 19,829 19,976 147 0.74%

Greene 60,502 61,357 855 1.41%

Grundy 14,138 14,225 87 0.62%

Hamblen 54,050 57,112 3,062 5.67%

Hamilton 294,745 305,616 10,871 3.69%

Hancock 6,778 6,811 33 0.49%

Hardeman 24,895 27,254 2,359 9.48%

Hardin 24,961 25,143 182 0.73%

Hawkins 49,719 51,954 2,235 4.50%

Haywood 19,525 19,828 303 1.55%

Henderson 24,424 24,794 370 1.51%

Henry 30,066 30,614 548 1.82%

Hickman 20,553 21,178 625 3.04%

Houston 7,853 8,079 226 2.88%

Humphreys17,059 17,635 576 3.38%

Jackson 9,629 10,669 1,040 10.80%

Jefferson 43,663 42,391 -1,272 -2.91%

Johnson 16,755 17,229 474 2.83%

Knox 366,846 378,319 11,473 3.13%

Lake 8,171 8,076 -95 -1.16%

Lauderdale 24,206 26,631 2,425 10.02%

Lawrence 39,358 39,365 7 0.02%

Lewis 10,868 11,124 256 2.36%

Lincoln 29,761 30,809 1,048 3.52%

Loudon 39,052 38,068 -984 -2.52%

McMinn 46,283 47,855 1,572 3.40%

McNairy 24,048 24,149 101 0.42%

Macon 18,181 19,375 1,194 6.57%

Madison 85,954 89,526 3,572 4.16%

Marion 26,851 27,296 445 1.66%

Marshall 26,302 26,240 -62 -0.24%

Maury 69,633 68,417 -1,216 -1.75%

Meigs 9,955 10,639 684 6.87%

Monroe 34,830 37,011 2,181 6.26%

Montgomery127,265 129,865 2,600 2.04%

Moore 5,196 5,565 369 7.10%

Morgan 18,775 19,472 697 3.71%

Obion 32,219 32,249 30 0.09%

Overton 19,557 19,697 140 0.72%

Perry 7,508 7,545 37 0.49%

Pickett 4,629 4,810 181 3.91%

Polk 14,883 15,635 752 5.05%

Putnam 59,143 60,978 1,835 3.10%

Rhea 27,836 27,828 -8 -0.03%

Roane 50,026 51,462 1,436 2.87%

Robertson 53,077 52,142 -935 -1.76%

Rutherford 166,035 171,404 5,369 3.23%

Scott 20,044 20,669 625 3.12%

Sequatchie 10,367 10,819 452 4.36%

Sevier 64,505 67,359 2,854 4.42%

Shelby 868,825 886,007 17,182 1.98%

Smith 16,368 16,818 450 2.75%

Stewart 11,545 11,787 242 2.10%

Sullivan 150,617 152,908 2,291 1.52%

Sumner 124,056 126,068 2,012 1.62%

Tipton 47,343 48,864 1,521 3.21%

Trousdale 6,844 6,929 85 1.24%

Unicoi 17,216 17,457 241 1.40%

Union 16,260 16,999 739 4.54%

Van

[tnwalker]

I'm glad someone shares my distrust of the census. Here in Shelbyville we are conducting a special census because the government says we have a population 16,105 while local estimates put it at around 20,000 consertively, and at over 100 bucks a head that's alot of funding we're missing out on.

[Rural King]

^I just read that the other week, in Business TN I think. Bedford County had a nice piece in that publication last month. You should check it out if you are able to, I think you would be impressed by it.

[tnwalker]

Thank you for pointing me to that Rural. There was lots of good info. It's hard to get development news in this town. For some reason the local media isn't really that interested in reporting it.

[Hankster]

Bedford County was underestimated by almost 2,000 before the last census, so it wouldn't surprise me at all if the Census Bureau's estimates for Bedford are way low again.

[nashvol85]

Davidson 533,967 566,150 32,183 6.03%

I'm sorry, but if you do this for a living, how can you be 32,000 off? Something is SERIOUSLY wrong if you are making that kind of error.

[Hankster]

Davidson 533,967 566,150 32,183 6.03%

I'm sorry, but if you do this for a living, how can you be 32,000 off? Something is SERIOUSLY wrong if you are making that kind of error.

That's exactly my point. How can they be off this much, and yet such tremendous credence is given these estimates because the are the "official" government estimates. Frankly, they're not really worth much, and the estimates developed at the local level are considerably more accurate. Unfortunately, nobody give these more accurate estimates any credence. It's all so frustrating to me.

[sleepy]

The estimates throughout the 90's were off because they used as a benchmark the 1990 census, which even the Census department agreed was an undercount. For the 2000 census, the Census had in place a format for a statistical adjustment if the statisticians felt that the 2000 census was also an undercount. The statisticians agreed however that the 2000 census was accurate.

That's one of the reasons many places had unexpected population increases from 1990 to 2000. Those places didn't actually grow as much as people not counted in 1990 were counted in 2000.

The estimates throughout this decade are based on the accurate 2000 census benchmark.

It's unfair to compare the inaccuracies of the 1990's estimates with this decade's estimates, because the reason for the prior inaccuracies don't exist.