Official US 2010 Census Results

[Kevinstat]

I'm not sure if anyone here has pointed it out yet, but if the major fractions method (which was used after the 1910 and 1930 censuses) was still used today, Rhode Island would lose it's second U.S. House seat and North Carolina would gain a 14th U.S. House seat.

[jimrtex]

I'm not sure if anyone here has pointed it out yet, but if the major fractions method (which was used after the 1910 and 1930 censuses) was still used today, Rhode Island would lose it's second U.S. House seat and North Carolina would gain a 14th U.S. House seat.

After the 1920 snafu, Congress under prodding from President Hoover agreed to automate the process, and specified that the president (or the census bureau on his behalf) would calculate the apportionment under both major fractions and equal proportions.  In 1930, both methods gave the same result.  In 1940, they gave different results, with equal proportions favoring Arkansas, and major fractions favoring Michigan.  The decision to use equal proportions was made on a party line vote, with all Democrats except from Michigan voting for equal proportions.

So if the decision were being made now, would the vote be for North Carolina a swing state in presidential elections, and where there is a reasonable chance of a Republican seat being drawn to avoid pushing a marginals into the Republican column, or for Rhode Island which is a certain Democratic presidential vote and representatives?

Would Speaker Pelosi tried to ram the bill through in the lame duck session?  Or would there be an effort to include representation for DC and RI (this assumes that Rhode Island is 436 in priority) while switching to major fractions.

[minionofmidas]

Hmm? What is the major fractions method? I did a few calculations this morning, actually, though with rounded figures because I didn't have access to the internet, and I got Hare-Niemeyer and the US method agreeing this year. Sainte-Lague shifts two seats to larger states; I forget which but Rhode Island wasn't one IIRC. I also did D'Hondt (single seat not just for RI but for ME and NH as well... HI narrowly double-member. About ten or so seat transfers. Oh, and you have to amend it to include an exception so every state gets at least one seat), and the method used by Turkey (one seat per state automatic, remaining seats distributed by D'Hondt with that one seat not taken into account, which in Turkey has the nice effect of making every province/constitutency at least double-member. Similar number of seat transfers but in the other direction this time. SD the smallest double-member state).

[Antonio the Sixth]

Hmm? What is the major fractions method?

Major franctions is the same as Sainte-Lagüe.

[jimrtex]

Hmm? What is the major fractions method? I did a few calculations this morning, actually, though with rounded figures because I didn't have access to the internet, and I got Hare-Niemeyer and the US method agreeing this year. Sainte-Lague shifts two seats to larger states; I forget which but Rhode Island wasn't one IIRC. I also did D'Hondt (single seat not just for RI but for ME and NH as well... HI narrowly double-member. About ten or so seat transfers. Oh, and you have to amend it to include an exception so every state gets at least one seat), and the method used by Turkey (one seat per state automatic, remaining seats distributed by D'Hondt with that one seat not taken into account, which in Turkey has the nice effect of making every province/constitutency at least double-member. Similar number of seat transfers but in the other direction this time. SD the smallest double-member state).

Differences are relative to Huntington Hill

Jefferson's method

+ CA(2), IL, NJ, NY, NC, OH, TX
- ME, MN, NE, NH, RI, SC, WA, WV

Adams' method

+ DE, ID, IA, LA, MO, MT, OK, OR, SD
- CA(3), FL, GA, NY, PA, TX(2)

Webster's method

+ NC
- RI

Dean's method

+ MT
- CA

[Kevinstat]

Hmm? What is the major fractions method? I did a few calculations this morning, actually, though with rounded figures because I didn't have access to the internet, and I got Hare-Niemeyer and the US method agreeing this year. Sainte-Lague shifts two seats to larger states; I forget which but Rhode Island wasn't one IIRC. I also did D'Hondt (single seat not just for RI but for ME and NH as well... HI narrowly double-member. About ten or so seat transfers. Oh, and you have to amend it to include an exception so every state gets at least one seat), and the method used by Turkey (one seat per state automatic, remaining seats distributed by D'Hondt with that one seat not taken into account, which in Turkey has the nice effect of making every province/constitutency at least double-member. Similar number of seat transfers but in the other direction this time. SD the smallest double-member state).

Differences are relative to Huntington Hill

Jefferson's method

+ CA(2), IL, NJ, NY, NC, OH, TX
- ME, MN, NE, NH, RI, SC, WA, WV

Adams' method

+ DE, ID, IA, LA, MO, MT, OK, OR, SD
- CA(3), FL, GA, NY, PA, TX(2)

Webster's method

+ NC
- RI

Dean's method

+ MT
- CA

One to add:

Hamilton's method (a.k.a. largest remainders with the Hare Quota (1 Hare Quota = total votes (or total apportionment population here)/total number of seats) a.k.a. Hare-Niemeyer

No difference (from Huntington-Hill a.k.a. the method of equal proportions)

Jefferson's method is also known as the method of greatest divisiors or (typically as a seat allocation method in proportional representation systems) the d'Hondt method.

Adams' method is also known as the method of smallest divisors.

Webster's method is also known as the method of major fractions or (typically as a seat allocation method in proportional representation systems) the Sainte-Lagüe method.

Dean's method is also known as the method of the harmonic mean.

[jimrtex]

This is the 2010 apportionment, based on each state's share of the total population. but assuming that share is the geometric mean of  x+1/2 and x-1/2, where x is the actual number of seats that a state is entitled to.

So for example, we divide Alabama's apportionment population of 4,802,982 by 710,767 which is the US apportionment population of 710,767, which gives a quotient of 6.757.  IOW, Alabama has 6.757 / 435 of the US population.

6.757 is the geometric mean of x+1/2 and x-1/2 or sqrt( (x+1/2)(x -1/2) ).

Solving for x = sqrt (6.757^2 +1/4) = 6.776 which would be Alabama's apportionment based on its share of the US population.
 
But if he add all these up, we find that we have apportioned a total of 436.838 seats.  So we adjust our original divisor upward to 713,769 which gives us an apportionment of 435.015, with Alabama being apportioned 6.748 seats.  Repeat another iteration using a divisor of  713,794, gives a total apportionment of 435.000 seats, and Alabama's apportionment of 6.747.

We could use independent rounding (rounding fractions greater than 1/2 upward) and each state would have it correct apportionment based on its relative share of the US population.

But in the case of 2010, this would leave us 5 seats short, as FL, WA, TX, CA, and MN would lose a seat.  The extra 5 seats are apportioned on the basis of the conventional ranking method (not shown here), which illustrates an interesting effect.  When there are extra seats to be apportioned, the ranking favors larger states, and disfavors smaller states.  The reason is that large states can spread their shortfall among a large number of districts.  So in this case California and Texas are favored, even though they are somewhat short of the population needed for 52.5 or 35.5 seats.  Montana is close, but if it did receive a 2nd seat they would be quit small.

It is happenstance that extra seats have to apportioned.  It could be the other way around and more than 435 seats apportioned based on independent rounding.  In that case large states would be disfavored.  We could take a district from them and the other districts would have fairly small increases.

In the following table, the first column is the 2010 fractional apportionment.  The second column is the change from 2000.  We can use this to make a rough projection to 2020 as to which states will gain and lose:

Winners:  AZ, CO(?maybe), FL, GA, ID(?), NV(?), OR(?), TX(+3), VA(1)

Losers: CA(?), IL, MI(1, 2?), MN, NE(?), OH, PA, RI, WV(?)

The 3rd column is the 2010 apportionment.  The 4th is the 2000-2010 relative increase or decrease in population.

The 5th and 6th column shows either the relative increase and absolute increase (in 1000s) for a state to take Minnesota's 8th representative, or relative decrease and absolute decrease for a state to forfeit a seat to North Carolina (Minnesota received the 435th seat, while North Carolina would have received a 436th were the House expanded).

The previous column can be combined.  So for example if Texas had grown at 20.5% rather than 20.9% it would have only gained 3 seats.  It also shows that it is now necessary for a state to have more than a million population before it gains a second seat (Montana has 994K and would need to gain 10K).


Alabama               6.747   -0.137   7    7.7%  -4.3%  -208
Alaska                1.128    0.038   1   14.7%  39.2%   283
Arizona               8.998    1.070   9   24.7%  -6.2%  -396
Arkansas              4.130   -0.025   4    9.2%   8.5%   250
California           52.317    0.090  53   10.1%  -0.3%  -118
Colorado              7.085    0.430   7   17.0%   5.4%   270
Connecticut           5.043   -0.229   5    5.0%   8.6%   308
Delaware              1.358    0.050   1   14.8%  11.5%   104
Florida              26.484    1.808  27   17.9%  -0.6%  -114
Georgia              13.637    0.995  14   18.5%  -1.7%  -162
Hawaii                1.979    0.041   2   12.3% -26.6%  -364
Idaho                 2.260    0.202   2   21.3%  10.6%   166
Illinois             18.029   -1.123  18    3.4%   2.1%   270
Indiana               9.122   -0.266   9    6.7%   3.6%   236
Iowa                  4.307   -0.233   4    4.2%   4.0%   122
Kansas                4.043   -0.133   4    6.3%  10.9%   312
Kentucky              6.116   -0.137   6    7.4%   5.8%   252
Louisiana             6.399   -0.514   6    1.6%   1.1%    49
Maine                 1.933   -0.096   2    4.3% -24.8%  -330
Maryland              8.127   -0.058   8    9.1%   4.1%   237
Massachusetts         9.203   -0.592   9    3.2%   2.7%   178
Michigan             13.895   -1.437  14   -0.4%  -3.5%  -346
Minnesota             7.463   -0.135   8    7.9%  -0.2%    -9
Mississippi           4.202   -0.217   4    4.4%   6.6%   198
Missouri              8.437   -0.207   8    7.2%   0.2%    15
Montana               1.480   -0.000   1    9.8%   1.0%    10
Nebraska              2.615   -0.073   3    6.8%  -5.2%   -95
Nevada                3.829    0.707   4   35.3%  -9.3%  -253
New Hampshire         1.918   -0.053   2    6.7% -24.1%  -319
New Jersey           12.349   -0.627  12    4.5%   0.7%    63
New Mexico            2.939    0.088   3   13.3% -16.0%  -330
New York             27.213   -2.043  27    2.2%   0.6%   107
North Carolina       13.411    0.983  13   18.6%   0.2%    16
North Dakota          1.071   -0.039   1    5.0%  48.6%   329
Ohio                 16.215   -1.300  16    1.7%   1.3%   145
Oklahoma              5.298   -0.049   5    8.8%   3.3%   125
Oregon                5.415    0.114   5   12.3%   1.1%    41
Pennsylvania         17.848   -1.091  18    3.5%  -2.6%  -331
Rhode Island          1.561   -0.131   2    0.5%  -5.0%   -52
South Carolina        6.528    0.313   7   15.4%  -1.1%   -51
South Dakota          1.253   -0.015   1    8.3%  22.5%   185
Tennessee             8.946    0.158   9   11.8%  -5.6%  -359
Texas                35.404    3.225  36   20.9%  -0.4%   -99
Utah                  3.914    0.435   4   23.9% -11.4%  -314
Vermont               1.015   -0.049   1    3.4%  59.3%   374
Virginia             11.272    0.331  11   13.2%   1.5%   122
Washington            9.474    0.366  10   14.3%  -0.4%   -27
West Virginia         2.653   -0.182   3    2.6%  -6.6%  -123
Wisconsin             7.999   -0.284   8    6.1%  -6.9%  -392
Wyoming               0.940    0.028   1   14.7%  76.7%   436

[minionofmidas]

Yeah, RI was one of my two Sainte Lague losers after all; I've checked. The other one (Minnesota; though who was the other gainer? New York probably but I'm not sure right now) may be a rounding issue.
The rounding made no difference to the Huntington/Hill method - it so happens that the gap between 435th and 436th was something of a mini natural break this year.

Seat #431: California 53 (nc)         Priority: 709,631
Seat #432: Florida 27 (+2)            Priority: 709,610
Seat #433: Washington 10 (+1)   Priority: 708,829
Seat #434: Minnesota 8 (nc)         Priority: 708,767
Seat #435: Texas 36 (+4)             Priority: 708,396
Seat #436: North Carolina 14 (+1)  Priority: 706,817
Seat #437: Missouri 9 (nc)               Priority: 705,802
Seat #438: New York 28 (-1)           Priority: 704,775
Seat #439: New Jersey 13 (nc)       Priority: 703,915

Biggest gap in the 431 to 439 run. (Much larger gaps just beyond, though.)

[jimrtex]

RI is the only one.

The difference between Webster's method and Huntington-Hill are pretty minor except for the smallest states.  The square root of  5 * 6 is 5.47, vs the average of 5 and 6 which is 5.5.  But more important is the relative difference between the divisors.

sqrt (9 * 10 ) / sqrt (5 * 6) = 1.732
9.5 / 5.5 = 1.727

The ratio of the two numbers is 1.0027.  Which means that the two methods would make a difference in the relative ranking of an Washington-sized state and an Oregon-sized state only if the quotients were within 0.27% of each other.  And it would make a different in apportionment if they happened to just be at the cutoff line for the membership of the House.

If we use only the quotients that determine whether a state has N or N+1 seats in a 435-member House, while Rhode Island drops 12 places, the 12 states that followed it for the most part simply move up one position in the ranking.

The exceptions are that AL and MI move up two, but that is because NE, a small state, also drops down a couple of places; and TX and CA, the two largest states, move up two place past WA.

The quotient for the State that receives the Nth representative, should be very close to 1/Nth of the total population (the national average representatives/population).  So the expected difference is 1630, vs. the observed difference between NC and MO for Webster's method is slightly closer than expected.

Yeah, RI was one of my two Sainte Lague losers after all; I've checked. The other one (Minnesota; though who was the other gainer? New York probably but I'm not sure right now) may be a rounding issue.
The rounding made no difference to the Huntington/Hill method - it so happens that the gap between 435th and 436th was something of a mini natural break this year.

Seat #431: California 53 (nc)         Priority: 709,631
Seat #432: Florida 27 (+2)            Priority: 709,610
Seat #433: Washington 10 (+1)   Priority: 708,829
Seat #434: Minnesota 8 (nc)         Priority: 708,767
Seat #435: Texas 36 (+4)             Priority: 708,396
Seat #436: North Carolina 14 (+1)  Priority: 706,817
Seat #437: Missouri 9 (nc)               Priority: 705,802
Seat #438: New York 28 (-1)           Priority: 704,775
Seat #439: New Jersey 13 (nc)       Priority: 703,915

Biggest gap in the 431 to 439 run. (Much larger gaps just beyond, though.)

[Antonio the Sixth]

Great list Jim. Would you like to sort it by the percent change necessary to win/lose a seat, please ?

[nclib]

What information will be released in the 2010 census that was not released in the 2000 census and visa versa?

Anybody know?

[jimrtex]

What information will be released in the 2010 census that was not released in the 2000 census and visa versa?

Anybody know?

The 2000 census included only the short form which asked the following questions:

Number Persons in Household
Housing: Owned; Mortgaged; Rented; Squatting.

For each person:

Family relationship to person filling out form
Sex
Age
Hispanicity
Race

The information that was formerly collected by the long form, is now collected in the American Community Survey.

Housing: Type (house, apartment, cave, boat, etc.), age, length of tenancy, acreage, farm income, business use, number of rooms, number of bedrooms, running water, flush toilet, bathtub or shower, sink with faucet, stove or range, refrigerator, telephone, number of vehicles, heating fuel, cost of electricity, gas, water&sewer, and other fuel; food stamps, condominium (and fee), rent (and whether board included), value, taxes, insurance, mortgage payments,

Population: Citizenship, and basis thereof; entry into USA, attending school and level, highest level of education completed,  ancestry, language spoken at home, residence location 1 year earlier, health insurance, deaf, blind, physical disabilities, marital status, recent motherhood, caring for grandchildren, military service, employment, commuting, laid off, productive or government worker, income, social security.

The long form was distributed to a large sample of households (10% to 20%) and was intended to ask all the annoying questions that would have made census participation plummet if asked of everyone.  It was intended to be a large enough sample to provide statistically reliable information for small areas (down to the block group, which has around 1000 persons).

The American Community Survey is administered on an ongoing basis, with a much smaller sample each month, but which when aggregated over 5 years produces a comparable sample size and accuracy to the long form.  The Census Bureau earlier this month released the ACS data for 2005 to 2009.   The Census Bureau also releases ACS data on 3-year and 1-year basis for larger areas (3-year data is statistically reliable for areas with population greater than 20,000; and the 1-year data for areas with population greater than 65,000).

Because the data is collected on a continuing basis, next year the ACS will be released for 2006 to 2010, with the oldest year of the sample being dropped and a new year added in.  So the ACS will be better for trends, while the long form census data, while clearer and more concise because it all is for a single data, won't show changes between decades, and because of unlucky timing might even be misleading (a 2010 long form would show higher levels of unemployment, and more persons in their mid-20s living with their parents).

[Antonio the Sixth]

Was the official list of priority values released ?

[cinyc]

The Census Director is scheduled to hold a press conference at 1PM on Wednesday, January 12 to "brief the media on next 2010 Census data releases, including state redistricting data."  Will we get some lower-level data tomorrow afternoon?

[jimrtex]

The Census Director is scheduled to hold a press conference at 1PM on Wednesday, January 12 to "brief the media on next 2010 Census data releases, including state redistricting data."  Will we get some lower-level data tomorrow afternoon?

I'd be really surprised, since it isn't expected until February for the earliest states.  The new version of American Fact Finder is supposed to go on line.  They have just released the 3-year (2007-2009) data from ACS for area with population more than 20,000, and have been rolling out the geography for the 2010 census (they have 32 states plus DC, done so far).

[minionofmidas]

We might get some more info on the exact timeline, though. That would be very welcome.

[jimrtex]

We might get some more info on the exact timeline, though. That would be very welcome.

I tuned in to late for the press conference.  It says it is being archived.  Anyone catch it?

[cinyc]

We might get some more info on the exact timeline, though. That would be very welcome.

I tuned in to late for the press conference.  It says it is being archived.  Anyone catch it?

I didn't catch it, but there doesn't seem to be much of interest in the press kit for the press conference.  There's some internal migration data for Texas, California and Michigan - but that may or may not have been previously released.

[Torie]

We might get some more info on the exact timeline, though. That would be very welcome.

I tuned in to late for the press conference.  It says it is being archived.  Anyone catch it?

I didn't catch it, but there doesn't seem to be much of interest in the press kit for the press conference.  There's some internal migration data for Texas, California and Michigan - but that may or may not have been previously released.

Yes, in a few places, internal migration could move the numbers by a discernible amount, and I try to factor that in, in my maps, when I am cutting it close to what I think is the point, where the odds of the seat being at risk in a semi wave, start to go up exponentially. I don't draw for a wave quite as strong as 2008. I don't think the odds are high that it will be replicated soon. And even then, the maps I draw should enable competent incumbents to survive. My sense of it is, is that a competent incumbent who is not a partisan attack dog, or becomes demonized by the opposition, can generate about 3 PVI points in his or her favor, over what would otherwise happen; with a superstar, like Gerlach in PA, or Ryan in WI, that is worth perhaps about 5 PVI points, in all cases assuming competent and reasonably funded opposition.

[cinyc]

We might get some more info on the exact timeline, though. That would be very welcome.

I tuned in to late for the press conference.  It says it is being archived.  Anyone catch it?

I didn't catch it, but there doesn't seem to be much of interest in the press kit for the press conference.  There's some internal migration data for Texas, California and Michigan - but that may or may not have been previously released.

Yes, in a few places, internal migration could move the numbers by a discernible amount, and I try to factor that in, in my maps, when I am cutting it close to what I think is the point, where the odds of the seat being at risk in a semi wave, start to go up exponentially. I don't draw for a wave quite as strong as 2008. I don't think the odds are high that it will be replicated soon. And even then, the maps I draw should enable competent incumbents to survive. My sense of it is, is that a competent incumbent who is not a partisan attack dog, or becomes demonized by the opposition, can generate about 3 PVI points in his or her favor, over what would otherwise happen; with a superstar, like Gerlach in PA, or Ryan in WI, that is worth perhaps about 5 PVI points, in all cases assuming competent and reasonably funded opposition.

Just to be clear, Torie, by internal migration, I meant raw statistics regarding people already in the US moving in or out of the state.  I don't think the analysis was strictly based on Census 2010 data, and like I said, it might have been previously released.

[Torie]

Oh sorry, cinyc, I did indeed misunderstand. When I am drawing my maps, I am obsessed with intra county migrations, which can matter a lot in bigger, and politically variegated counties. I don't want my maps to hit the dumpster because I didn't try to anticipate for that.

[cinyc]

The press conference transcript is here.  There's not much there there.  As we knew, the state redistricting data will trickle out by state starting the first week of February.  Census will tell us which states will be released next about a week ahead of time - so we'll start knowing by the end of January.  The release schedule is sensitive to state redistricting needs - so I'd imagine those with 2011 elections might come first, if they have to redistrict beforehand.   State officials will get the data before the rest of us, but we will get it at the same time as the press.

New Tiger shapefiles will be put up on Census' website soon.

There's going to be a webinar on the new American Factfinder on January 18 and on redistricting on January 24.

Oh sorry, cinyc, I did indeed misunderstand. When I am drawing my maps, I am obsessed with intra county migrations, which can matter a lot in bigger, and politically variegated counties. I don't want my maps to hit the dumpster because I didn't try to anticipate for that.

Yeah, my initial phrasing was misleading.  While it certainly wasn't clear from just the handouts, from the transcript, it sounds like the Census Bureau was using Pew and Heritage inter-state migration estimates, among other things, to try to judge the accuracy of the 2010 Census, which was in line with the mid-level population estimates.  Obviously, not everyone responds to the census, and in some cases, census workers had to ask proxies like neighbors or building management.   It also sounds like the full accuracy analysis won't be ready until 2012.  The final accuracy analysis for the 2000 census indicated a slight overcount.

[jimrtex]

The press conference transcript is here.  There's not much there there.  As we knew, the state redistricting data will trickle out by state starting the first week of February.  Census will tell us which states will be released next about a week ahead of time - so we'll start knowing by the end of January.  The release schedule is sensitive to state redistricting needs - so I'd imagine those with 2011 elections might come first, if they have to redistrict beforehand.   State officials will get the data before the rest of us, but we will get it at the same time as the press.

New Tiger shapefiles will be put up on Census' website soon.

There's going to be a webinar on the new American Factfinder on January 18 and on redistricting on January 24.

Oh sorry, cinyc, I did indeed misunderstand. When I am drawing my maps, I am obsessed with intra county migrations, which can matter a lot in bigger, and politically variegated counties. I don't want my maps to hit the dumpster because I didn't try to anticipate for that.

Yeah, my initial phrasing was misleading.  While it certainly wasn't clear from just the handouts, from the transcript, it sounds like the Census Bureau was using Pew and Heritage inter-state migration estimates, among other things, to try to judge the accuracy of the 2010 Census, which was in line with the mid-level population estimates.  Obviously, not everyone responds to the census, and in some cases, census workers had to ask proxies like neighbors or building management.   It also sounds like the full accuracy analysis won't be ready until 2012.  The final accuracy analysis for the 2000 census indicated a slight overcount.

New shapefiles are up for most states now.  It was something like 38 states a couple of days ago.

The first part of the conference was the census director explaining their preliminary evaluation of the 2010 Census.

The mail back rate was slightly higher in 2010 (74%) but that was because everyone got the short form.  The mail back rate for the short form in 2000 was also 74%, with the long form response rate dragging down the overall rate.  I'd guess that the long form took too much effort to complete, and some people would have decided it was too prying.

Of their effort to contact the remaining 26%, 22% was by contact with a building manager or neighbor (so 5.7% overall). vs. 17% in 2000.  But I'd speculate that this might not be too much worse than in 2000, because of the long form.  Some of the people who didn't turn in the long form, did so for other reasons than unwillingness to be contacted, moving, or working 3 jobs.  It was just too much hassle.  But it was easy to find them.

Overall contact rate, by mail, or followup with the household or a proxy was 99.60%, up from 99.57% in 2000.

They then compared the actual count with estimates.  One was based on birth and death data, and guesses about immigration.  They were within a couple 100,000 of the middle estimate, but off by 3 to 4 million on the outer estimates.  So that could just be luck.

The other was to compare their estimates, which are based on the 2000 census, then updated based on births, deaths, migration data.  They were within a few 100,000 on that as well.  Which could say that they are no worse than 2000.  If they had missed 5 million in 2000, their estimate for 2010 would also be off by 5 million or so, and if there count was as (in)accurate they would miss 5 million.

The 3rd party slides were really pretty goofy.  They were based on the ACS data, which includes information on previous residence (perhaps in the PUMS).  They showed a lot more people leaving Michigan than moving in, and the Census showed a population decline.  Texas showed a lot more people moving in than moving out, and the Census showed a 20% gain, plus 4 new seats.  And California showed more people moving out than moving in but not in the ratio of Michigan, and California showed a growth rate about the same as the US and no congressional seats.  There wasn't any analysis of whether the net migration was of the correct magnitude to match the census population.

Washington Post gets the award for most inane questions (plural).  One was whether the redistricting data would include data counts on same sex couples.  And the second was whether in light of the shooting in Tucson, the Census Bureau had beefed up its security.

Amusing moment was when the census director Robert Groves said that the 3-year (2007-9) ACS data would be released later this month.  Someone interjected, "yesterday".  Groves then said that the 3-year (2007-9) ACS data had been released earlier this month.

Most of the press questions were about when would they be able to write an article with little research and little substance about the population in their area with a few quotes of reaction from a government official or demographer at the local university, and would the press get to see the data before the public.  Groves said that the data would be released to the public (and press) via American Factfinder.

[jimrtex]

Media Advisory — Census Bureau to Release Local 2010 Census Data for Louisiana, Mississippi, New Jersey and Virginia

[minionofmidas]

Odd choice of states.