District 3 is Still as Majority Latino CVAP as It Ever Was: A Close Look at Block Groups 93 & 106

That .

.

.

(1) BIG NEWS That Probably Isn’t News at All

Kleptocracy-associated blogs are all atwitter over news, reportedly from the Anaheim City Council’s hired demographer Justin Levitt, suggesting that “ha-ha, the latest figures show that District 3 doesn’t have a Latino majority!”  Both suggest, in a screwy way, that Jordan Brandman deserves an apology for suggesting delaying the mapping process until after new Census data came in at the end of January.  (Of course, this wan’t the basis for Brandman’s objection.  You can check; it’s on video!)  Prior to that new census data, District 3 had a 50.9% Latino majority; Levitt’s new figure for it is supposedly 49.1%.

This is especially odd because the difference between these figures and the previous ones were the result of only two changes: dropping the 2009 estimates and adding 2014 into the rolling five-year — average of changes.  That means that the estimates for years 2010-2013 continued to account for a stable 80% of the data — and Latino CVAP went UP in Districts 4, 5, and even 6.  So to the extent that this change matters at all — which it probably doesn’t — if it’s true then this would be a piece of bad luck for proponents of the Recommended Plan.  (Probably not legally consequential, though, but that belongs in another post.)

I and certain anonymous others have done some research and I’m confident that, based on the “kit” that mappers like me and Oscar Reyes and Levitt himself  had to work with, it’s not true.

And yet, it is true that under current calculations, District 3 has 49.1% Latino CVAP (where Majority Latino CVAP is anything over 50%.

And yet, that won’t matter.  There won’t be a lawsuit — at least given that the percentage of registered voters in District 3 is over 50%.

So why did District 3 lose Latino population when Districts 4, 5, and 6 gained it?

Simple: IT DIDN’T!  At least, not enough to move the Latino CVAP even a tenth of a percent.

Confused?  You should be!

I’ll explain why — but there will be a big buildup before we get to the payoff.  (Otherwise, you might not understand the payoff.)  And the conclusion will be that everyone acted in good faith, no real harm was done, and the source of the error should not be held too much responsible, because it was hard to spot.

(2) Many Many Maps

First, take another look at the Recommended Plan.

District 3 is in yellow, the one with the smokestack — “Population Unit 106,” as we’ll discuss — jutting up into Fullerton.  (Note that everything east of District 5 (the purple section) is in District 6, which is  mostly Anaheim Hills.  It’s left off here because including it would reduce the size of the above map by almost 50%.)  The part of District 3 that we’re most interested in is the very bottom part that (literally, as it contains Disneyland Drive) leads into Disneyland.  That part is in “Population Unit 93” — also known as “Census Block Group” or “CBG 93.”  (I’ll generally use the term CBG, for brevity.)

The interesting thing about CBG 93 is that it’s one of the relatively few Population Units in District 3 that is split between two Districts.  Here’s a closeup look at the division of District 93, which is a short distance west of where Districts 2, (blue border line), 3 (purple border line) and 4 (red border line) come together.  It’s that short vertical line, bottom center, that is red on its left and purple on its right. The part of it just west (or left) of that red and purple line is in District 4.

What street is that, you may wonder?  We can help you out there!

CBG 93 is (rather crudely) outlined in red.  That dashed red line between Districts 3 and 4 is Walnut Street.  Let’s call the part colored purple — which, again, is in District 4 — “West CBG 93.”  Its north and south are Crone Ave. and Ball Road.  Its east border is, again, Walnut and its west border is railroad tracks.  West CBG 93 is the key to understanding why District 3 is probably OK.  It’s due to the likelihood that that square was counted in District 3 though it belongs in District 4 — and that it has a proportionately large population that is largely non-Latino.

Let’s first take a look at population density.  You can see that the green territory to the east within CBG 93 (“East CBG 93”) is at least twice as large as West CBG 93.  However, West CBG looks like it has at least as much population.  In the population density map below, more red means more population.

In this map, CBG 93 is bordered in green.  West and East CBG 93 are divided by that purple line.  West CBG 93 is mostly dark orange (15,000-20,000 people per square mile, per the legend at the left) with a stripe of medium orange (10,000-20,000 people per square mile).  East CBG 93 includes an non-residential area east of Disneyland Drive and an area about 1.5 times the size of the part west of the line in District 4.  The larger, less dense area looks like it is about the same population as the smaller, denser area.  So the question will be: what happen when we exclude denser and browner West CBG 93 from District 3 — where it does not belong?

Hint: We solve the problem.  But let’s look at the actual data!

(3) Tables Stocked with Data!

All right, here’s what a (selected) view of the (relatively) raw data table looks like.

One thing to keep in mind: if you don’t understand this table, even after I explain it, you probably don’t need to.  This is mostly to benefit the data geeks; you’ll do just fine without it.  For those of you who want to take the trip — buckle in.

The first three column headers, in orange, are just labels.  Column A shows Block Group, B is the District number, and C is the Census Block Group — of which only 93, 106, and 132 are shown here.  These are the three CBGs that exist in one than one district.  I’ve colored the rows in District 3 in shades of green and the rows in District 4 in shades of purple.  You may notice that there’s no purple shown for CBG 106.  We’ll get to that near the end.

The next four columns show each CBG’s total and then Hispanic (or Latino) population, both overall and for adults.

Now we get to estimating how much of those adult populations are citizens.  The next four columns, in light blue, show what those estimates were in 2013 (shorthand for “2009-2013”), using two different measures known as “ST” and “ACS.”  The Demographer prefers ACS, which is fine by me.

The next six columns are in darker blue.  The medium blue ones are the same measures for 2014 as the light blue are for 2013, although for some reason they are entered into the table in reverse order.  After each pair of measures, in the darker blue columns, are calculation of the “CVAP” statistic — “citizens of voting age population,” largely equivalent to “eligible voting pool” (but for the inclusion of felons, mental incompetents, etc.)  That — and associated measures like “percent of registered voters” — are what you use to see whether or not ethnic and racial minorities are being treated fairly by the map.

As for the rows: the light purple and green at the top are for CBG 93.  Then comes a set in darker green for CBG 106.  (As I noted, there’s no purple here.)  And finally, a shade darker, are the purple and green rows for CBG 132 (which is the very southeast corner of the Colony District.)

Look at the fourth column from the right, Column “GI”.  If you follow this column down, you’ll see that you get a “CVAP” score for each individual Census Block comprising CBG 93, followed by a summary score — in bold — for the entire CBG.  (CBGs 106 and 132 only have the summary score.)  As you’ll see, the “purple” part of CBG 93, which you know as “West CBG 93,” has a CVAP of 16.82% — that’s the percent of its voting pool estimated to be Latino.  Continue down and you’ll see that the overall CVAP for “East CBG 93” is 40.73%.  It turns out that East is a little more populous than West despite its lower density; it has 1,233 residents compared to 972.  (You’ll find those figures in Column M.)

So you can see that including those 972 residents, only about 1/6 of which are Latino, would drag down District 3’s numbers substantially.  Did the Demographer do this?  I’m not sure — at first I thought so, but there turns out to be another explanation.

Why do I think that these might have accidentally been included in the District 3 tally?  Here’s a table of calculations sent out by the Demographer.

The coloring is mine, because at first I thought that the discrepancy might have been due to an error in combining “partial” statuses.  (It wasn’t.)  The first set of 13 rows is for CBG 93, then 106, then 132.  Each set of 13 rows starts with the total for a Census Block Group (in black) and ends with the number of Latinos (in orange).  The columns to attend to are the four to the right.  From the left, they are the total number of estimated citizens and its margin of error; then the total number of Latino citizens and its margin of error.

If you look at the top row (under the headers), you’ll see that this was a calculation of 1385 Latino Citizens of Voting Age  out of 1705 Total Citizens of Voting Age for District 93.  Looking at the total population of CBG 93 from the first chart, it seems likelier that this comes from a total population of 2205 (adding 972 and 1233) than a total population of 1233 alone — mostly because the latter would be impossible.  The same appears to be true of CBGs 106 and 132.  CBG 106 only shows a population of 291 in the upper chart, rather than 1495; CBG 132 only has ONE Census Block in District 3 — containing 398 people of whom 385 are Latino — so the population estimate of 1665 there must include lots of people from District 4.

Then again, maybe the Consultant sent out the wrong table and has another one that (correctly) does not use these figures.  I don’t know.

Now let’s look at CVAPs:

All the rows in yellow are the Census Block Groups fully within District 3.  The total population, Latino population, and CVAP of these districts are shown on Row 29 (in orange.)  That’s at 49.37%, so if District 3 is to push above 50% it will have to come from the other three districts.

Let’s look at Rows 31 and 32 first (in kelly green.)  The numbers for only the portions of them within District 3 DO appear somewhere in what has been released — whether they were calculated before or after the first maps came out is unclear — so I have here substituted them in.  Using the calculation for CBG 106 in Row 30, we come up in Row 33 (blue) with a total CVAP for these final three districts of 72.86%.  adding this to Row 29 gives us a total of 51.48%.  This is almost entirely due to the effect of RBG 106 — as can be seen in Row 35 (light blue) it by itself raises the District 3 CVAP to 51.42%.

Are we done?  Happy ending?  At one point I thought so….

What about that figure in the first table showing only a total population of 291 — rather than 1705 — from CBG 106?  How can we explain that?  Where’s the rest of CBG 106?  Well, remember when it first came up, way above?

“District 3 is in yellow, the one with the smokestack — “Population Unit 106,” as we’ll discuss — jutting up into Fullerton.”

Without having confirmed this, I suspect that I am quoting Justin Levitt verbatim when I first saw this and said: “OH, SHIT!”

Can a census block group in an urban area cover more than one city?  I hadn’t thought about it before now, but I now know the answer: yes it can.  The missing part of CBG 106 — the 1414 citizens of voting age of whom about 1355 are Latino — are in FULLERTON!

I’ve said for years that I thought that it was an abomination for Anaheim to extend north of the 91 between Lemon and Raymond, allowing the city both to foist a Walmart on its northern neighbor and take away its (See’s) candy, but I never expect to see the day when Anaheim had to pay the price.  Yet here we are.  Census Block Group 106 does indeed bust past the latitude of Kimberly Avenue and end up who knows where — well, I guess that at this point Justin Levitt probably very well knows where — and so we have to redo our calculations.  (What a price to have to pay!)

You’ve seen the topmost part already.  In the second block, we substitute in the miserable 10.71% CVAP for the portion of CBG 106 that is actually in Anaheim and redo them.  I come up with a total District 3 CVAP of 49.07%.  Yes, that rounds off to the 49.1% reported by the Demographer.

Was this due to the shrinking of the adult Latino Citizen population of Anaheim?  I decided to check.  Remember, this problem was entirely due to the “partial districts,” so I started the reanalysis using 2013 numbers by pretending that the total CVAP for “full districts” was the same.  What I found was that in 2013, CBG 106 was slightly less Latino, CBG 132 about the same amount more so, and district 93 almost a full 5% more so (44.93% to 40.36%.)  Put them all together and the overall Latino CVAP goes up from 40.07% to … 40.12%.  OK, not such a big change after all.

(4) Implications and Conclusions

Let’s consider some of the outstanding issues that Anaheim now faces.

(Q1) Will Anaheim Be Sued If It Approves This Map on Tuesday?

Nah.  District 3 still have a Latino CVAP majority based on the “Voter Registration” measure, and especially at this late date that will be good enough for the likes of MALDEF, the ACLU, and (at least Statewide) LULAC.  Because of the increases in Districts 4 and 5, Anaheim now has three strong Latino plurality districts and three strong White plurality districts (one of which is a majority), and that’s pretty good for a city approaching equal numbers of Latinos and Non-Latino  Whites.  The damage to Latinos is negligible — and moreover the culpability is negligible.  The City acted in good faith reliance on its Consultant.  (This presumes that there was no City order to Levitt that he should mess with the numbers — and that’s close to inconceivable.  For one: WHY DO THAT?)

(Q2) Will Anaheim Be Sued If It DOESN’T Approve This Map on Tuesday?

That’s actually more likely.  The City’s actions last fall froze challengers to the City’s incumbents into place while the incumbents could continue to raise money.  Using a situation like this to slow down the process further, when no litigation is credibly threatened (or likely even possible), could raise questions about whether incumbents and their allies are simply throwing obstacles in front of their potential opponents.  But as Lucille Kring is happy with these districts — frankly she has good reason to be — the three votes needed to pass it should be there, so we won’t find out the answer to this question.

(Q3) Is Justin Levitt in Some Kind of Trouble?

I certainly hope not.  And I expect not.  I’m not sure how unusual this cross-city CBG sort of thing is, which is what I’d want to know to see how negligent it was, but from decades in Social Science I can tell you that Things Happen.  I think that I personally caused him as much grief as anyone else with my perfectly reasonable requests, but I thought that he did a good job overall.  This stuff looks so easy, but the opportunities to stumble are everywhere, and the slip up regarding CBG 106 was pretty fluky.  Be Kind, City of Kindness!

(Q4) Does Jordan Brandman Deserve an Apology?

That Jordan deserves an apology is one of the nuttiest assertions I’ve seen in years.  Delaying approval of the maps to await the new Census numbers wouldn’t have made a difference because the Census numbers were not the problem!  The drop to 49.1% wasn’t because the Census numbers changed, but because Levitt finally noticed (or was informed) that CBG 106 was mostly in Fullerton!  It could (and should) have happened back in April!  And if Levitt mistakenly counted parts of District 4 in District 3, that has nothing to do with the changing census numbers either.  Brandman’s problem with the Recommended Plan was not that it might swing out of balance, but that he thought it was “divisive” — and it had led to his getting his butt kicked by the DPOC.

That two bloggers have written — without doing any of the work that you see here to figure out what happened — that this somehow vindicates Brandman’s desire for delay just shows you how totally nutty certain “deep thinkers” are.  If you got more from their trying to capitalize on this situation than my (and others) investigating what really went wrong, then just spend your time in their echo chambers.  In that case, this blog is clearly not meant for you — and thank God for that!

Here’s to good sense prevailing on Tuesday night — and then let the Games begin!