COH Redistricting: CVAP By Council District
Below are the Citizen Voting Age Population totals for each Houston City Council District.
A note on my methodology is in order before that, though. There are better ways to arrive at these numbers than the method I’ve done. My method is simply one that allows for a relatively quick means of determining the percentages for demographic groups listed below. I can’t overstate enough that I’m doing this as a one-man show, without benefit of interns, students, or other minions willing to do work for me. Peer review is a good thing and if anyone else wants to take a crack at it, I’d love to see the numbers that result from that effort.
The 2005-09 CVAP counts are offered at the Census Block Group level at it’s most granular data point. My means of listing which block groups are counted is obtained visually by overlaying the council district map on top of the CVAP by block group map. I then de-select and transcribe each individual block group for each district and then dump those lists into a database where a custom script does the counting.
The code has been checked and re-checked. The CVAP numbers, as previously noted, have several issues that should be kept in mind – they are estimates, they have a margin of error, they are based on smaller samples than in previous years, and there are return rate issues since they come from the American Community Survey rather than the Census forms. Many of those issues get massaged in putting together the estimates. But I don’t pretend to offer the sum totals below as authoritative. The percentages that follow from those numbers, however, I argue are much more reliable.
In the process of visually determining which block groups should be counted, there are inevitably a few judgment calls that result. Block groups that straddle a council district are generally counted as follows: if the district being counted is drawn to favor a demographic group for VRA purposes, I tend to include block groups that have a majority of that demographic group and not count Anglo block groups that are split in order to determine what the best possible showing for that demographic group may be. In some cases, block groups are split fairly close to 50-50 by a border. In most cases, I alternate whether or not to count a block group on that basis. In cases where a block group only covers Limited Purpose Annexation, I err on the side of not counting that block group even though there may be a small amount of population within the LPA portion of the block group.
In many of the cases where there are splits, we’re typically talking about 5 or so block groups out of 100 or more where a judgment call is needed. A significantly higher number were split in areas with a lot of LPA turf like District A and B. District E also contained a larger than normal number of judgment calls. In each case, there is no mistaking that District B is at or near the 70.8% African-American share listed below. Likewise, District E is unmistakable in its proximity to the 69.6% Anglo population listed below. District A, however, might be worth a closer look to see if it is precisely above or below 50% Anglo. It might be worth suggesting that District K could use a closer look to determine if the African-American population there is over 50% based on a more precise analysis. But given the lower number of judgment calls made on split block groups in the district, I’m a lot more confident in standing by the fact that District K is slightly over 50% African-American.
For a comparison, here are the totals that the Census Bureau provides for the entire City of Houston along with two of the most relevant counts from the report below.
Totals (Census Bureau)
Total CVAP … 1,206,360
Hispanic CVAP … 281,235
Totals (Greg’s Count)
Total CVAP … 1,139,280
Hispanic CVAP … 263,684
That means my counts are 5.56% short of the total CVAP count and 6.24% short on the Hispanic CVAP count when compared to the definitive count done by the Census Bureau. In general, the shares for each demographic group don’t change terribly much. If I get an opening in time sometime soon, as well as a block assignment file for the city, I may work with that to get a more refined count. But I’d argue that the percentages you see below are informative enough to draw any conclusions with.
Dist CVAP Anglo Hisp Afr-Am Asian
-----------------------------------------
A 85,005 41,584 21,480 17,160 4,084
48.9% 25.3% 20.2% 4.8%
-----------------------------------------
B 97,105 9,089 17,649 68,735 1,120
9.4% 18.2% 70.8% 1.2%
-----------------------------------------
C 153,105 108,295 24,838 11,896 6,398
70.7% 16.2% 7.8% 4.2%
-----------------------------------------
D 117,505 19,637 16,996 76,385 3,928
16.7% 14.5% 65.0% 3.3%
-----------------------------------------
E 128,970 88,330 24,445 8,170 6,600
69.6% 19.3% 6.4% 5.2%
-----------------------------------------
F 85,490 24,860 18,310 28,720 12,745
29.1% 21.4% 33.6% 14.9%
-----------------------------------------
G 140,945 106,770 13,688 11,930 7,250
78.4% 10.1% 8.8% 5.3%
-----------------------------------------
H 91,360 23,630 47,655 18,923 784
25.9% 52.2% 20.7% 0.9%
-----------------------------------------
I 77,235 13,985 45,914 14,970 1,813
18.1% 59.4% 19.4% 2.3%
-----------------------------------------
J 55,150 16,995 13,939 18,545 5,245
30.8% 25.3% 33.6% 9.5%
-----------------------------------------
K 107,410 28,165 18,770 54,860 4,655
26.2% 17.5% 51.1% 4.3%