I have started getting texts asking my opinion about the Arizona
election. (Three just last week.) That and the barrage of TV commercials
prove that I live in a swing state.

We need your opinion because
“our future is at risk,” says a pitch from a conservative text message.
Another “presidential straw poll” asks if I am voting for Kamala
Harris. (Pro tip: Look for clues, including the source, to see if this
is a push poll … a marketing technique rather than a real question
looking for an answer). A third text offered me a $2 gift card if I
answered a long set of questions. I was curious so I took the bait. It
was legit, a balanced quiz, asking many of the same questions in a
different form to try and uncover an honest answer.

What we all want answered is “who is winning?” That’s where the polls can be really misleading.

FiveThirtyEight’s
polling average shows Harris ahead by 3.5 percent with 67 days to go.
“But there’s considerable uncertainty in those numbers,” says
FiveThirtyEight in its methodology tab explaining why it uses a
color-coded thicker line above the polling graph. This is “kind of like
the range of possible precipitation reported in a weather forecast,
showing you could get anywhere from, say, 1 to 3 inches of rain in an
upcoming storm.”

Speaking of storms, we see snow in Phoenix more often than we get an
accurate reading of the Native American vote. Pew Research explains this
error: It’s because polls are “excluding parts of the population.” That
would be us.

A new Arizona poll by Noble Predictive Insights
surveyed 1,003 registered voters. The methodology: “The sample
demographics were weighted to accurately reflect the registered voter
population by gender, region, age, party affiliation, race/ethnicity,
and education according to recent voter file data, the Arizona Secretary
of State’s Office, and recent Census data. The margin of error was +/-
3.2 percent for likely voters and +/- 3.09 percent for registered
voters. Totals may not equal to 100 percent due to rounding.”

So much there.

First, always remember that plus or minus 3.09
percent means the tally could be off by a factor of 6.18. That’s huge in
a tight race.

Second, and more important, just who are those
“weighted to accurately reflect the registered voter population?” Excuse
me, it’s not weighted. The demographic crosstab does not include Native
Americans or Alaska Natives as a category.

This is the norm. Noble Predictive Insights is not the outlier. And
it’s a reason why people are surprised when Arizona defies the polls and
elects a Democratic governor (not just the current one, but virtually
every Democrat who has won in the state over the last couple of
decades.) What has happened for decades is that the Navajo vote comes in
later than Phoenix and Tucson and is overwhelmingly Democratic. But
that’s not reflected in the polls.

The folks who do the polls counter, saying that a small demographic
group is difficult to track (if 1,000 people were surveyed, less than
five would represent us all). But what if all five vote?

And
that’s the key problem of polling: How does a poll calculate (or really
guess) who will vote? Pew Research says: “Surprisingly, polls
historically have not shown any particular pattern of favoring one side
or the other. The errors that favored Democratic candidates in the past
eight years may be a result of the growth of political polarization,
along with declining trust among conservatives in news organizations and
other institutions that conduct polls.”

That doesn’t make the idea of “not counting” sting any less.

There
is another way to read polls. What’s the trend? How has that number
changed over time, even when the numbers are close? That story is also
interesting because Harris has led since she first entered the race and
that was not the case for President Joe Biden.

Of course a
national poll doesn’t matter anyway. The United States elects its
presidents via the Electoral College – state by state. Most of the
states are pre-determined by demographics, red states, blue states.
(That contest would be one-sided: The last Republican to win a majority
of national voters was George W. Bush in his 2004 re-election bid.) The
ones that do matter are the swing states: Arizona, Nevada, Georgia,
Wisconsin, Michigan, Pennsylvania, North Carolina, and possibly
Florida.

And it’s in these state races where the Native vote matters – and it’s part of a larger story.

Take
Nevada: In 2012 when Barack Obama was re-elected as president, that
state was 65 percent White. This is what I wrote in 2012: “ So it will
be possible to build a winning coalition made up of some white voters (a
third or so) plus significant majorities from Latino, African American,
Asian American and Native Americans.” That’s exactly what happened.

In this year’s cycle the voting demographics will only be about 45
percent White. So a winning coalition in Nevada looks a lot different – a
combination of White voters, African Americans, Asian Americans and
Native Americans. Another key group in this coalition is young people:
Millennials are the largest voter bloc.

This essentially was the
Obama coalition in 2008 and 2012 – and one thing to watch will be to see
if this coalition has been reforged by Harris and Tim Walz.

And if you are waiting for that poll result? Real numbers will start
coming in about a month from now when early voting begins across the
country. Yes, in October. In much of the country Election Day is more
like Election Month.