We are two states: Seattle and Washington
If anyone were to doubt that there really are two Washingtons — that the Seattle metropolitan core (and its playgrounds) are another world from most rural to small-city Washington, and especially from east of the Cascade crest — a look at the vote on Referendum 71 should be persuasive.
These are not subtle, marginal differences, but indisputable polarization in what we might call the modernist-traditional divide. Referendum 71 passed by a 53 to 47 percent vote, revealing the power of the King County electorate, which alone provided a margin of 104,000 of the statewide margin of 113,000. It points again to what a statistical outlier Seattle is among cities, what I called in an earlier article, "Extreme Seattle."
To overcome the problem of variable size of precincts, and to suppress too-small numbers, I aggregated precincts to census tracts, which have the added advantage of permitting comparison of electoral results with social and economic data from the census.
Looking at the statewide map, one of the first things one notices is that about 85 percent of thegeographic territory of the state (95 percent in eastern Washington, 70 percent in western Washington) voted NO on R-71, a measure to expand domestic partner rights for same-sex couples. But that strong no vote was overwhelmingly rural and small town. The only core metropolitan census tracts that voted a majority no were in the Tri-Cities, Yakima, and Longview.
The heart of traditionalist, and arguably, of anti-gay sentiment,`is the farm country of eastern Washington, especially wheat and ranching areas in Adams, Douglas, Garfield, Lincoln, Walla Walla, and Whitman counties, but also the rich irrigated farmlands of Grant, Franklin, Benton, and Yakima counties. The highest no votes in western Washington were far rural Clark, exurban Chehalis, the Graham area south of Tacoma, and Lynden, home to many Dutch descendants and members of the conservative Christian Reformed Church.
Not surprisingly the census tracts in eastern Washington that supported Referendum 71 were the tracts dominated by Washington State University in Pullman, around Central Washington University in Ellensburg, the mountain resorts tracts in western Okanogan (Mazama, Twisp), in Skamania county, and a few tracts in the core of the city of Spokane.
Across western Washington majorities against Ref 71 prevailed over a sizeable contiguous southeastern area, from northern Clark and Skamania through urban as well as rural Lewis county (reinforcing that county'ês contrary-to-Seattle reputation) and on northward into much of southeastern Pierce county; this region opposed R-71 by at least 58 percent. Still opposed, but less so were voters in much of the rest of rural small town western Washington, including most of rural Snohomish county.
The zones of strong support, over 60 percent, are essentially Seattle and its inner commuting zone; its spillover playgrounds and retirement areas of Port Townsend and the San Juans; college and university dominated tracts (around Western Washington in Bellingham, the Evergreen State College in Olympia); plus the downtown cores of Vancouver, Tacoma, and Everett. Weaker but still supportive were rural spillover, retirement, and resort tracts, often in coastal or mountain areas of Pacific, Grays Harbor, Jefferson, Clallam, Skagit, and Whatcom counties.
Looking next at the detailed map of central Puget Sound, we find contrasts between support and opposition are also starkly revealed. Support over 75 percent almost coincides with the city of Seattle (not quite so high in the far south end), and its professional-workforce commuting outliers of Bainbridge and Vashon, plus the downtown government core of Olympia and tracts around the University of Puget Sound and the UW Tacoma. Moderately high support (60-75 percent) surrounds the core areas of highest support, most dominantly in the more affluent and professional areas north of Seattle through Edmonds (and south Whidbey Island) and east to Redmond, Issaquah, and Sammamish. Weak but still positive votes occurred in the next tier of tracts, around Olympia, north and west of inner Tacoma, most of urban southwest Snohomish county and much of exurban and rural King county (quite unlike other rural areas of the state).
On the contrary, the shift to opposition is remarkably quick and strong in southeastern King and especially in Pierce county, in northern and eastern Snohomish county, including Marysville, and, not surprisingly, in military dominated parts of Kitsap county (e.g., Bangor) and Pierce (Fort Lewis).
I also compared the voting behavior in census tracts with social and economic conditions of those tracts. The results are highly representative of the Red-Blue division of the American electorate, in both the large metropolitan versus rural and small town and city aspect, and in the modern versus traditional dimension (socially liberal or conservative).
The strongest single variable for those favoring R-71 is transit use (.75 correlation), which is a surrogate for the metropolitan resident. The critical social characteristic determining which way voters chose is the nature of households — the traditional family (tending to oppose 71) versus non-families (partners, roommates, singles), where the positive correlation ranges from .48-.65. This is a powerful tendency, and useful to describe differences in areas. But of course many in families, often more educated and professional, support Ref 71; and many singles, often elderly, or opposite sex partners, opposed it, especially in more conservative parts of the state.
The next strongest variables reflect the strong split of the electorate according to the predominant educational level of the tracts. The tendency of the more educated to support R-71 is the key statistic of the 'êmodern'ê vs 'êtraditional'ê dimension, and is closely related to the differences by occupation and industry. Managers and professionals and those working in finance and information sectors tended to be supportive of 71, while those in laboring and craft occupations and in manufacturing, transport and utilities, tended to oppose. (South King county and much of Pierce county have high shares of blue collar jobs). Finally differences by race exist, but are not so strong as, say, in the presidential election in 2008. (Interestingly, the correlation of the percent for Ref 71 and for Obama was .90.)
Here are the strongest positive correlations for people who voted for R-71: users of transit, non-traditional households, singles, same-sex households, college graduates, managerial professionals, aged 20-39, and Asian. The strongest negative correlations (opposing 71) are: drivers of SOVs, traditional families, aged under 20, born in Washington, high-school education only, Hispanics, laboring and craft occupations.
Yes, greater Seattle is indeed different!