What a Fine Predicament This Is: CA 48th and the Top Two Ballot
I spent a lot of time and posts analyzing the 2016 Presidential race, partly for its bizarro arithmetic. I once pointed out that the effect of the undecided could be crucial, and it turned out to be. I promised myself not to waste time on another election, but the dilemma of the California 48th again presents another bizarre arithmetical challenge. The root of the problem is that politicians and voters lay out a set of arithmetic rules that sound fair, but lack the foresight or flexibility to deal with actual situations. Recall college football’s BCC championship games selection dilemmas, almost every year.
California’s 48th district is what I call a typical 60-40 district, where, in this case, the voting goes 58.3%-41.7% for Dana Rohrabacher, the Republican stalwart, since 1988, for 15 terms or 30 years. Redistricting has given Rohrabacher a gold-plated part of the California Golden Coast including the rich Orange County enclaves of Laguna Beach, Corona del Mar, and Newport Beach. If anybody is surfing in money, it is Rohrabacher. Oddly, he backs offshore drilling, which should be anathema to all those in his district who are there for the beach, as he is. Even if some constituents never have a view or get sand in their toes, they are all aware that their expensive property values depend on clean, safe, and attractive beaches. Many local businesses and jobs depend on unspoiled beaches, also. I don’t even understand why oil companies would want to do more drilling off of our expensive shores, since the Deepwater Horizon catastrophe cost $65 billion in damages.
The arithmetic dilemma is posed by California’s fairly new primary law, that only the top two vote getters will appear on the deciding November district ballot. This puts parties second to voter preferences, and in lopsided party districts, still gives voters a choice. But many don’t bother to vote in primaries, who will vote in the general election. Many, of course, don’t vote in off-year elections anyway. This off-year election, where Democrats certainly have the most justified justification ever among all US elections, or any elections anywhere in the world, ever, (to speak in egomaniacal Trump style), they are being hampered by a surfeit of qualified and dedicated candidates.
Here are the dilemmas. Rohrabacher (R) is leading with only 30%!!!! Are we going to elect a Representative who is the first choice of only 30%? Next are two Democrats Harley Rouda (D) and Hans Keirstead (D) both tied at 13% each, and the Republican co-runner, Scott Baugh (R), ALSO at 13%! Besides 6 other candidates at 4% or lower, is a yawning abyss of 18% Undecided! This sounds like a final arithmetic problem posed by a very mean second grade teacher. It is now actually a Ph. D level problem for political science experts. If Scott Baugh gets a smattering of extra votes, Democrats are completely shut out of the November election in this district, while they have 35% of the decided voters. Republicans only have 47% of the decided voters. So a Democratic candidate is justified instead of Baugh, by 35 to 13, or a ratio of 2.7!
Let’s blame the dilemma on the law, and here, the faulty poll. 18% Undecided is a vast gap. Why is there no data on the party preferences of the Undecided? Are they all Republicans just having a tough time choosing between two main Republicans? (I am not getting into the histories or qualifications of the candidates.). What percent might be undecided Democrats? What percent are neither party? What percent won’t decide at all and not vote for a Representative, or not vote at all? Are they favoring one or another of their party’s candidates? How many will change their preference as they become better informed or influenced by friends or ubiquitous annoying lawn signs by the June 5 primary? This was answered by the poll takers. How many will be swept away with the pomp and circumstance of Trump’s meeting with Kim of North Korea just before the primary? How many will suffer the Bernie Sander’s Effect, of not backing their party’s candidate if their primary hopeful does not triumph.
I have a naive suggestion. The Democratic candidates can come together after the election by noting that a House Representative needs a Chief of Staff, and top advisors. They also need someone to handle the home office and community relations. Plenty of posts for the two Democratic campaigns to merge or cooperate to form a strengthened campaign.
Probably, this will sort itself out before June 5, and voters can switch, but new polls will be needed, and voters would have to hold back on early voting. They might even have to show up at the polling booths on the 5th, heaven forbid.
The California Democratic Party has endorsed one candidate, Keirstead, while the Democratic Congressional Campaign Committee (DCCC) has endorsed Rouda. What is the old quote: “A house divided does not get into the House”? As much as there is motivation to battle the maximally corrupt Trump administration, Democrats have to plan much better to succeed.
The way to estimate the error of the sampling of each candidate is to take the square root of their voters as the fluctuation. The percent error is to then divide the fluctuation by the specific voters. The poll was taken by Tulchen Research on May 1-5, by interviewing 400 likely voters. The leading candidate at 30% had 120 voters. It’s square root or standard deviation is 10.95. Then you add the factor √(1-p) where p is the candidate’s percentage. For p = 0.30, this factor is 0.837. Taking 10.95 x 0.837/400 = 2.5% fluctuation. The lead result is then 30% ±2.5% of the total. For the three subleading 13% candidates, out of 400 that is 52 votes each. The square root of 52 or 7.21. The factor √(1-p) = 0.933. The error is then 7.21 x 0.933/400 = 1.7%. The fraction of total voters of the three runner-ups is then 13% ± 1.7%. We thank UCI Physics Professor Jonas Schultz for showing us how to do these calculations.
The errors cited here are called 1-sigma errors, which contain 68% or about 2/3 of the statistical results. Doubling the errors gives 2-sigma errors, or about 95% of cases.
But don’t forget the 800 pound gorilla in the room. The 18% Undecided is the clear number two runner up! Tulchin Research then informed the voters of the stances of the candidates, but only reduced the undecided to 12%! They also remarkably increased Harley Rouda’s vote from 13% to 22%, while leaving Keirstead’s still at 13%. Having an 18% or 12% Undecided in a vote with runner up at only 13% shows why so many election predictors bit the dust after Trump’s seemingly unlikely 2016 victory. If only 9% or half of the undecided voters vote Democratic, and the other 9% vote Republican, than Democrats are still behind by 56% Republican to 44% Democratic, close to the 2016 result of 58% to 42%. Be Better, Democrats.