# Camden: Primary numbers don’t always add up in general election

By

Published:

The week after the primary seems like a good time for a lesson in primary numbers, which is offered by the poli sci department, not the math department.

Election Math 101 teaches us that the most important number is who has the most votes. But other numbers matter, particularly in a Washington primary, which is brought to you by the number 2 — as in the top two vote-getters go to the general election, no matter what.

If a particular office has more than two candidates, the rest fall by the wayside. On primary election night, one critic questioned how a candidate with the most votes could be described as “leading,” when the person with the second-most votes would also advance to the general. The answer is simple: The person with the most votes at any given time is leading the folks who have fewer, regardless of what comes next. (And people say reporters are bad at math.)

If an incumbent in a primary gets less than 50 percent of the votes, the person finishing second will almost certainly claim some type of political victory by saying more than half the voters rejected the incumbent. This may be technically defensible but it is not necessarily politically true, because the opponents might represent a wide range of political flavors, and the second-place candidate may not be to some voters’ taste.

In a large district such as Eastern Washington’s congressional district, it’s important to look at the whole district, not just one portion of it. Some Democrats have been saying more than half the voters rejected Republican Cathy McMorris Rodgers, which is true in Spokane but not in the 10-county district as a whole, where she currently has about 52 percent of the vote. Even in political math, 48 percent is not more than half.

Not all opponent votes are fungible. The 11,000 or so folks who voted for extra-conservative Republican Tom Horne might not vote for Democrat Joe Pakootas — or any Democrat — in November. And even with those votes, Democrats have to add in the 2,000 or so voters who didn’t mark anyone for that office, something technically known as an undervote, to get more votes on the “not McMorris Rodgers” side of the ledger.

Undervotes are probably the least predictable numbers in any election. These represent people who marked the ballot for other offices or issues, but weren’t sold on anyone in that particular race and left all circles blank. Conventional wisdom says they are easier for a challenger to capture, because the voter already has decided not to vote for the incumbent.

A large number of undervotes in a race can be a good signal for a challenger, although large is a relative number in political math. In Spokane County, the congressional race has about 1,200 undervotes, which seems like a lot until one considers the county assessor race has 15,800. Generally speaking, it’s a significant primary number if the undervotes plus the challenger’s total puts him or her close to the incumbent, and the challenger comes up with a plan to reel them in.

### Strong campaign is key

If only two candidates file for an office, their primary results are essentially a bit of free polling, and the numbers are instructive but not necessarily predictive. Some analysts contend that no one can overcome a gap of more than 10 percentage points in the primary, but that’s not always true. In 1980, Democratic Rep. Jerry Hughes got 42 percent of the primary vote against Republican Sen. Bob Lewis’ 58 percent in what was then northwest Spokane’s 5th Legislative District. Hughes won in November with 52 percent of the vote through a strong campaign.

Even small numbers in the primary are not a guarantee of loss in the general. Four years ago, Rob Chase got 2 percent of the primary vote, through a write-in campaign, for county treasurer but won in the general election with 51 percent.

The lesson is, too much attention to primary numbers is a bad idea. Weak ones can be overcome with hard work, and strong ones can be frittered away with a poor campaign. Class dismissed.