On July 1, I sat down to talk to Scott Keeter and Hannah Hartig from Pew Research about their 2024 Validated Voter Survey.

We talked about the challenges of analyzing elections using panel data, and about the shifts in makeup of the electorate between 2020 and 2024, and what that means for how campaigns think about balancing turnout and persuasion strategies. More than anything, they tell us, mobilization is a result of campaigning. You gotta play to win.

Links:

How Changes in Turnout and Vote Choice Powered Trump’s Victory in 2024

Behind Trump’s 2024 Victory, a More Racially and Ethnically Diverse Voter Coalition

Commercial Voter Files and the Study of U.S. Politics

My interview with L2 about their approach to assembling the voter file: https://podcasts.apple.com/us/podcast/cross-tabs/id1725891109?i=1000651891510

My interview with Michael McDonald discussing turnout models and his Election Project: https://podcasts.apple.com/us/podcast/cross-tabs/id1725891109?i=1000666055702

The Red Shift Maps from NYT: https://www.nytimes.com/interactive/2024/11/06/us/politics/presidential-election-2024-red-shift.html

Hank Green's response video to the red shift maps: https://www.youtube.com/watch?v=kC9u7NZbGlQ)

David Shor on Ezra Klein talking about changing demographics in the MAGA coalition: https://podcasts.apple.com/us/podcast/democrats-need-to-face-why-trump-won/id1548604447?i=1000699618199

Our Guests:

Hannah Hartig is a senior researcher at Pew Research Center, where she primarily studies U.S. political attitudes and voting behavior. She has authored analyses on topics including domestic opinions of the U.S., voter turnout in 2020 and views of abortion. Prior to joining the Center, she was director of research at the Penn Program on Opinion Research and Election Studies at the University of Pennsylvania. She regularly discusses the Center’s political research with the news media and has served as an election night exit poll analyst for NBC News since 2014. Hartig received her bachelor’s in foreign affairs from the University of Virginia and master’s degree in quantitative politics from the University of Pennsylvania.

Scott Keeter is a senior survey advisor at Pew Research Center. In this role, he provides methodological guidance to all of Pew Research Center’s research areas. An expert on American public opinion and political behavior, he is co-author of four books, including What Americans Know about Politics and Why It Matters (Yale University Press), A New Engagement? Political Participation, Civic Life, and the Changing American Citizen (Oxford University Press), The Diminishing Divide: Religion’s Changing Role in American Politics (Brookings Institution Press), ** and Uninformed Choice: The Failure of the New Presidential Nominating System (Praeger). He has also published numerous articles on survey methodology. Prior to joining Pew Research Center, he taught at George Mason University, Rutgers University and Virginia Commonwealth University, where he also directed a survey research center. Keeter is a graduate of Davidson College and received his doctorate in political science from the University of North Carolina. He is a past president of the American Association for Public Opinion Research (AAPOR). In 2016, Keeter won AAPOR’s highest honor, the AAPOR Award for Lifetime Achievement, for “outstanding contributions to the field of public opinion research.”

You can follow their work, and even donate to support Pew’s Research, at pewresearch.org

----

📬 Subscribe to our newsletter at crosstabspodcast.com for new episodes, insights, and behind-the-scenes content.

📹 Watch video episodes on YouTube @CrosstabsPodcast

💬 Follow us on BlueSky: @crosstabspod.bsky.social and @farrahbostic.bsky.social

📍 Produced by The Difference Engine

Welcome back to Crosstabs, A show about people data and power.

2

I'm your host, Farrah Bostic.

3

I meant to get this episode to

4

But instead, I got caught up watching

the various live streams of the

5

votes being taken in Congress on the

budget reconciliation bill known in

6

the Senate as the one big beautiful

bill act before a minority leader.

7

Schumer had the name change to

something generic and then sent

8

it on to the house overnight.

9

A few Republican holdouts stopped holding

out and after Hakeem Jeffries took the

10

floor in what is known as a magic minute.

11

Uh, as an aside, the Congress has a

lot of supernatural time on its hands.

12

Perpetual days, magic minutes.

13

Anyway, they held the

vote and the bill passed.

14

The Immigration and Customs Enforcement

Division of the Department of Homeland

15

Security now has a budget that exceeds the

combined budgets of basically all of our

16

domestic federal law enforcement agencies.

17

A budget for one agency that some

are saying exceeds , that of most

18

countries, total military spending.

19

On the same day today, July 3rd, the

day before Independence Day, the US

20

Supreme Court ruled on whether the United

States government can move without due

21

process to deport people to countries

they have never set foot in, where

22

they may be tortured and beaten, where

they may never again know liberty, and

23

it found that it can do exactly that.

24

I bring this up because the conversation

I'm trying to bring to you today is not

25

really about that, but in a way it is.

26

I spoke with the excellent researchers,

Scott Keter and Hannah Hardig at the

27

Pew Research Center who worked on the

28

is one of a handful of high quality

post-election surveys designed to

29

tell us who the electorate was in

30

previous electorates who votes and who

doesn't matters in this country still.

31

It is why we fight over how to

draw congressional districts and

32

what identification is required

to register to vote and how much

33

money there should be in politics.

34

It is also in a way why we are fighting

so hard right now over who deserves to

35

be an American at all, and it appears

whether you can have your Americanness

36

stripped from you, if the government

chooses or have it contested depending

37

on which state you are in at the time.

38

Tomorrow, maybe today as you

listened to this, or a few days ago,

39

or whenever is Independence Day.

40

We count this as the nation's

birthday, more or less.

41

It will be 249 years old, who

the electorate is at least in

42

part, is who the nation is.

43

I wonder if the people who made up

this electorate really had in mind

44

ringing in the 250th birthday of this

nation by becoming something terrible.

45

Maybe they did . This conversation

explains how hard it is to

46

know both who the electorate is

and why they do what they do.

47

This research isn't designed to

tell us why they did what they did,

48

though I think that is for us all to

consider now as their choices to vote

49

or not to vote, have consequences.

50

We will all have to reckon with Happy

4th of July to those who celebrate.

51

Here's my conversation with Senior Survey

advisor Scott Keter and senior researcher

52

Hannah Hardig of the Pew Research Center

53

So Hannah and Scott, thank you for being

willing to talk to me about the analysis

54

Pew has done about the 2024 electorate.

55

And I kind of just wanna start by getting

to know each of you a little bit better.

56

I have joked before that this is the

unofficial podcast of the Pew Research

57

Center, uh, because I've had a few

different teams on at this point.

58

But, uh, I haven't had a chance to,

to meet with the two of you before.

59

I'd love to hear a little bit about your

roles at Pew and, and kind of your areas

60

of expertise and, and how you came to

be in the position that you're in now.

61

Hannah Hartig: So I've been on

the politics team at Pew Research

62

Center for almost eight years now.

63

And more broadly, I've been

studying the American public and

64

how they view politics, how they.

65

think about political news and how

they vote for a number of years.

66

I got my start at NBC News during

67

decision desk and, and getting

immersed in the exit poll world.

68

Um, and then from there I sort of,

69

helping them with some horse race

coverage of the primaries for the

70

Democratic and Republican parties

71

And I made my way over to Pew a

few years ago and have been happily

72

studying the public ever since.

73

Farrah Bostic: Excellent.

74

That must be,, is it a nice change

to go from the horse race coverage

75

to, to this well, I guess less, uh,

pitched version of a political polling

76

Hannah Hartig: For me, yes, it was

a really great transition because,

77

you know, election is a mark in time

and so I really wanted to get a, a

78

deeper understanding of, the shifts

over time, how people were viewing.

79

Politics and non-election years,

that's, it's very crucial to

80

understanding our society.

81

It's not just how people vote, it's

how they view a variety of events.

82

And as we'll talk about

later, not everyone votes.

83

And we wanna reflect how everyone

thinks in our country today.

84

And so, yes, it was a welcome change.

85

Farrah Bostic: That's great.

86

Well, thank you Hannah.

87

How about you, Scott?

88

Scott Keeter: Well, I've been

with Pew for a long time.

89

, This is over 20 years for me.

90

I was an academic for much of

my career over 20 years as a

91

college professor of political

science, studying public opinion.

92

I.

93

Got interested in doing surveys

and had some experience with it,

94

and, uh, was fortunate enough to

then be able to transition over to

95

Pew in 2002 on a full-time basis.

96

But for the last, um, almost 10 years

or so, I've been a part-time advisor

97

to the center, working mostly with our

methodology team, uh, sometimes with

98

the politics team on, particularly

on this particular project, uh,

99

validating our panelists, , votes

and writing about the election.

100

Uh, but I too worked at NBC for

a while as a, as a consultant.

101

And I met Hannah in 2014

when she was working there.

102

So

103

Farrah Bostic: Well,

104

polling is a small world.

105

Scott Keeter: Yeah.

106

Farrah Bostic: Yes.

107

That's great.

108

Um, so maybe let's start just

with what this project is.

109

Every, every election cycle.

110

For as long as I've been paying

attention to Pew, I feel like

111

there has been something like this.

112

How, how far back does this type of

this post-election, uh, analysis of the

113

makeup of the electorate go with few?

114

How long has it been

doing that, that project?

115

Hannah Hartig: Yeah, so Scott actually

had a hand in the original validated

116

voter analysis, and that was first

117

And so we do those every two

years after the general election

118

starting in 2016 and the midterms.

119

And so this was our, our fifth edition

of the Validated Voters Report.

120

Farrah Bostic: And I think one of

the things a lot of people who maybe

121

are not in the polling world don't.

122

I don't understand is why, why it's

seven months after the election

123

that we find out kind of what

the makeup of the electorate is.

124

Maybe you can talk a little bit

about the, the challenges of even

125

getting to the question of who voted.

126

Scott Keeter: Yeah, it's,

um, it, it is frustrating.

127

'Cause people want to know pretty soon

after the election, well, why did this

128

happen and what, what were the forces that

drove it and what could have been done

129

differently if you're on the losing side.

130

But the fact is that we are

depending on obtaining the actual

131

state voter turnout records.

132

Farrah Bostic: Mm-hmm.

133

Scott Keeter: this, you know, as

you know, elections in the United

134

States are very decentralized.

135

Uh, each state has control of its own

elections and follows its own calendar.

136

In terms of updating its

records and making them public.

137

So we have to wait until all of the

states have reported to where we have a

138

complete picture of of the electorate.

139

Now that doesn't mean

we can't start sooner.

140

We, we actually get a lot of these data

within a month or so of the election.

141

A lot of states are very, very

fast in turning theirs around.

142

And so we're able to work on the, on the

project and, and to get a good sense of

143

what, what the story is, well before the

time that we've released it last week.

144

But we, we just need to wait because

that's the, the way the wheels of

145

the administration of elections goes.

146

Farrah Bostic: Yeah, it's, it's, uh,

something we've addressed on, um, last

147

year on a couple of episodes, talking to

Michael McDonald at the Elections Project.

148

And, and then also to some folks from

some of the, even just like the voter file

149

companies, about the complexity of getting

these records from different kinds of

150

governmental units, let's say, depending

on the part of the country that you're in,

151

it may be down to like the village or the

township that is holding these records.

152

And so it, it does take time

to, to gather all of that.

153

Um, so you're, it sounds like you're

kind of incrementally collecting

154

the, the data as it is available.

155

Maybe then talk about how you go

about, obviously one of the other

156

things we know from those conversations

is that there are, it's, it's very

157

different state to state how much.

158

Kind of demographic data, or even

for that matter, voter, uh, party

159

affiliation data is on the record

about the electorate itself, um, just

160

from like the voter registration file.

161

Um, and obviously when people are

filling out their ballots, they're

162

not also filling out a demographic

screener prior to doing so.

163

So talk a bit about how you kind of,

uh, make these correlations between what

164

we know about the demographics of an

area or about your, from the, from the

165

panel that Pew has, uh, and projecting

that onto the actual results that we

166

obtained on the kind of just the ballot

167

Hannah Hartig: So , this is a

really unique analysis in that we

168

have a massive panel, our American

Trends panel of thousands of adults.

169

And so we have a, a variety of

information both about, you know, the

170

demographics of the people on the panel.

171

So, their race, their age you

know, their relative income status,

172

their educational attainment.

173

And really what the value of matching that

panel to these voter file records is, is

174

that we're able to confirm their turnout.

175

And so, as you're well aware, one

of the major factors when, when

176

people are responding to surveys is

that they might feel social desire.

177

Ability, you know, to say that they

voted, they might feel like they should

178

say that it's part of their civic

responsibility to say they voted.

179

And so part of what this match

allows us to do, and it's three

180

different vendors, is to confirm

that they actually did turn out.

181

Um, and so it's validating their

turnout, hence validated voters.

182

And so most of the other information that

we're getting about how they, they voted

183

or their political preferences or some of

these other demographic things that you're

184

pointing to are actually from our panel.

185

But that's, you know, the unique value

add of this particular project is that

186

we have really a really rich portrait

of the people who turned out and the

187

people who didn't, , but we're able

to validate whether they actually did

188

turn out in one of three voter file

189

vendors that we use.

190

Scott Keeter: I was just gonna

throw in a plug for a report that

191

we wrote in 2018 where, uh, when we

192

actually had had, uh, relations with

five different commercial vendors.

193

And in addition to validating votes

194

opportunity to actually try to

evaluate the quality of the demographic

195

information and other data that you can

find in these commercial voter files.

196

And of course, as your audience

might know if they saw these earlier

197

podcasts where you talk to some

of these vendors, the value of

198

these is largely, I think for the.

199

Campaigns themselves, they use these

to help them reach, uh, their voters or

200

to try to convert people, to mobilize

people that aren't regular voters.

201

And you know, not so much for research

purposes as we, as we use them, but

202

we thought that it would be helpful

to kind of demystify the voter files.

203

And so we took a look at, you know, how

good is the demographic information in

204

it and sort of what, what does it tell us

beyond just whether somebody voted or not?

205

And that that report's on our website.

206

And I think it's still,

still pretty interesting.

207

Farrah Bostic: That, that

is really interesting.

208

'cause I know that as I was starting to

get into the question of, you know, I'll

209

fess up that like thinking about voter

file companies was not really a thing.

210

I had done a ton of prior,

prior to really, I think prior

211

to the New York Times is.

212

Polling operation.

213

And, you know, I'm a,

I'm a footnote reader.

214

If you write a methodology section,

I read your methodology section.

215

. They mentioned that they were using

L two as as a voter file provider

216

for, for some of the surveys that

they were doing, which, , got me to

217

start to dig in a little bit on who

are all of these different companies.

218

And we did in fact talk to someone

from L two in the end, but now I

219

get emails from all of them, um,

which is exciting, uh, for me.

220

But I will, we will refer

people to that piece.

221

'cause I think that was a really

interesting look at the, the kind of level

222

of quality and, and, um, verifiability

of what's in each of those data sets.

223

So how did you ultimately narrow

down, you started with five, you're,

224

you're with the three, you picked sort

of the three that you felt had the

225

highest quality, uh, data sets to begin

with, or, or were there other kind

226

of factors that drove your choices?

227

Scott Keeter: It was, It was, uh,

a little less systematic than that.

228

Hannah Hartig: Yeah, cost.

229

Scott Keeter: five was just

too many to, to manage.

230

It was just, uh, the logistics

of, you know, the contracting

231

with them and all of that.

232

It was just more of a burden.

233

We didn't feel like the value add

of, that many justified, the effort.

234

So, but we did, we did want to do one

thing, which was we wanted to make sure

235

that we had a vendor that was nonpartisan

and we wanted to have a vendor that

236

represented sort of the conservative

side of the spectrum Republican vendor.

237

And we wanted to have one that

was liberal or a democratic

238

vendor, progressive vendor.

239

And that way it's not that, it's not

that the information in, in each of

240

these is gonna be radically different.

241

I mean, after all, the conservatives

want to turn out people that are not

242

conservative, and the liberals want to

turn out people that are not liberals.

243

So it's in their interest to

have it as accurate as possible.

244

But we wanted to avoid any, appearance

of favoritism from one side or

245

the other, as well as just taking

advantage of the, you know, the, the,

246

the extra information that you get.

247

And we do find unique voters in each

of the three files that we used.

248

Farrah Bostic: Interesting.

249

Yeah.

250

And so, so you go through this

process, you have your own panel, so

251

you have a lot of information there.

252

You're able to verify them against the

voter file that they did in fact vote in,

253

you know, the, the last three elections

or in this, this current election.

254

Um, you, you mentioned I think right

before we started recording that the

255

questions that you ask really are about.

256

Their vote choice.

257

And then obviously you have the, the

demographic makeup of, of the respondents.

258

But maybe talk a bit about what sorts of

things you ask on this survey and, and

259

what things you don't ask on this survey.

260

, Hannah Hartig: So the main

vehicle that we're matching to

261

is our post-election survey.

262

And the main thing we're asking

about is, you know, did you turn

263

out, did you face a long line?

264

If you didn't vote, do

you wish you'd voted?

265

And one of the key things I'm guessing

we'll talk about is if you didn't vote.

266

Who would you have

supported in the election?

267

And so we're asking a variety

of things about voting behavior.

268

You know, we take that opportunity to

ask a couple different issue questions.

269

But again, it is a panel and so we, you

know, we've been tracking pre-election

270

attitudes throughout the 2024 cycle.

271

But yes, that post-election survey where

we take the full panel and we ask them

272

about their voting behavior and, and the

273

we ask about those things like your

vote choice and, and how did you vote?

274

Did you vote by mail?

275

Did you vote in person?

276

Those kinds of things.

277

Farrah Bostic: by, yeah.

278

I think there was a particular just , on

that topic that I noticed in the report

279

was, uh, that the number of people who are

doing in-person early voting is growing.

280

And obviously it's a relatively, I

mean, depending on where you're, I'm

281

originally from Oregon where it's

been vote by mail for like, ever.

282

Um, and, uh, and now I live in

New York where we just got early

283

voting in, in the last, I think in

those last election cycle actually.

284

Um, or maybe the last two.

285

And so it's interesting to

see that rate of growth.

286

Were there any kind of interesting

trends that you saw in terms of things

287

like wait time or method of voting?

288

Across the, across the sample.

289

Hannah Hartig: I think you highlighted

the main, the main things that we

290

found too far, it was just a growth

among people who supported Democrat

291

or Harris or Trump in the election.

292

And so there's been, you know, a

modest uptick in, in Republicans

293

who are voting early as well.

294

Um, and they tend to favor

in person early voting.

295

But, you know, 2020 was obviously a

massive year for, for male voting.

296

Um, but we're seeing a continuation

of people voting before election

297

day among both Democratic supporters

and Republican supporters.

298

Farrah Bostic: Yes, I'm old enough

to remember that when Oregon went

299

to mail-in ballots, everyone was

very worried that this was going

300

to create a Republican advantage.

301

And not really, it just turned

out to be the way it is.

302

Like no one, no one minded.

303

And obviously it's now

just effectively mandatory.

304

I think my mother fills out

her ballot and drops it off at

305

the library across the street.

306

But, but that's 'cause she

waits till the last minute.

307

So, um, sorry mom outed you.

308

But it's interesting to have that

kind of have that kind of data.

309

It's also, you know, heartening to

see that people are, are adopting

310

these, new methods of, voting since

obviously making it easier to vote.

311

I think it's a good thing in general.

312

Any other kind of things that,

that surprised you or that stood

313

out in the, in this research?

314

Were there kind of things that

you weren't anticipating seeing?

315

I know that, you know, you're, you're

doing ongoing research among the

316

electorates, so I'm just curious

if there were, if there were any

317

big surprises demographically or

behaviorally across this analysis.

318

Scott Keeter: Well in, in terms of

the substance of, of what happened

319

I think that a lot of it was

pretty well signaled in advance.

320

Particularly, uh, Donald

Trump's gains among.

321

Non-white voters.

322

And the fact that he was, uh, making

explicit appeals to people who were

323

, sort of disengaged or irregular voters.

324

The thing of course that you don't know

is, uh, that saying that you're doing

325

these things, doing a lot of social media

or whatever the channels were, you don't

326

know whether they're going to pay off.

327

And I think what, what the report

did was it established quite clearly

328

that his efforts to mobilize people

who were, disengaged from the process

329

four years ago, uh, were successful.

330

And while it wasn't a massive amount

of, , you know, sort of net gains

331

for him, it was a close election.

332

And I think it was probably

enough to spell the difference

333

between success and failure in

some of the battleground states.

334

Farrah Bostic: There were definitely

some folks trying to maybe get ahead

335

of the full analysis, a little bit of,

of what the electorate looked like.

336

But nevertheless, the story that has,

I think, sort of begun to, you know,

337

maybe it's not as extreme as it was in

the first instance, but it nevertheless

338

seems to be fairly consistent, is this

sense that two things turnout was down

339

from 2020 still very high historically

340

there were some advantages in gains

from both non voters in switching that

341

Trump enjoyed, that Harris did not.

342

And that the conclusion to some degree

is even if all the non-voters had

343

voted, he still would have won that,

that the, that that sort of taken

344

for granted narrative of non-voters

tend to vote for Democrats is.

345

Declining or, or, or, or has switched?

346

I, I think the interesting thing

that I noticed in, in the report was

347

actually that number has been, or that

348

Um, so I wonder how much this has

fallen off a cliff or whatever.

349

But maybe you can kind of talk a bit

about how you arrive at any kind of

350

analysis around what would non-voters

have done had they, had, they

351

participated, and obviously even if they

had, not all of them would have, but

352

nevertheless, um, maybe talk a bit about

how you arrive at, at that conclusion.

353

, Hannah Hartig: So it's something

that we've been thinking

354

very deeply and hard about.

355

Right.

356

Backing into some of the psych

electoral math is, is really tricky.

357

And of course we don't have the full

counterfactual to just rewind in time

358

and, and simulate a hundred percent

turnout on what that would've looked like.

359

One of the most straightforward ways

we have , to go about that is what

360

we just discussed, is just asking

non-voters who they would've supported.

361

And we found that they

were pretty evenly divided.

362

You know, Trump kind of enjoyed like a,

a modest advantage over Harris among,

363

among this group, but very close and.

364

Ha, Democrats have long enjoyed, I

think , we see that in a variety of

365

ways, a turnout advantage, right?

366

, When turnout, is higher,

Democrats do better.

367

There are kind of like

interrelated themes here, right?

368

It's the people who don't turn

out consistently they're more

369

likely to be voters of color.

370

They're less likely to have

formal levels of education.

371

They're more likely to be young.

372

All of those things compound, right?

373

If, if that relationship between the

democratic advantage among these groups is

374

weakening and they're also less likely to

turn out, that suggests that when turnouts

375

higher Republicans are starting to,

even the playing field among Democrats.

376

And this is something

I think , we've been.

377

Acutely attuned to in the Trump era.

378

So I think there's evidence of

this relationship weakening,

379

beginning potentially in 2012.

380

Again, it's, it's hard to arrive

at some of these estimates.

381

But certainly I would say

382

I think we are getting more

evidence from these variety of

383

different analyses to suggest that

that relationship is weakening.

384

But yeah.

385

Sky, I don't know if there's

anything else you wanna add to that.

386

Scott Keeter: No, I think there

are, there are a couple of.

387

Things that you could do

to sort of bracket this?

388

I mean, one of, one of them is that,

that we don't address this in the report.

389

The report's very factual and focused.

390

But you know, this, movement of less

educated voters non-white voters and

391

so forth, in the direction of the

populist conservative party is not.

392

Confined to the United States.

393

This is something that we're

seeing all around the world.

394

You know, that, these changes

are, are not happening in a vacuum

395

here in, in the United States.

396

Trump is, obviously a factor,

but , he's not the only reason

397

that , this is happening.

398

At the same time, we are kind of

reluctant to even make a statement quite

399

like that , because we know that this

election was held in the context of a

400

very specific set of conditions that

Joe Biden was an unpopular incumbent,

401

that he had presided over the country

during the COVID, uh, years with a lot of

402

inflation, which of course was happening

in lots of other places in the world.

403

And people were in a very

anti-establishment mood.

404

It's hard to know whether any

incumbent could have weathered that.

405

And ultimately he didn't.

406

Uh, but his successor

, couldn't get the job done.

407

So I, I think that if you're looking

for a, a sort of big picture bottom line

408

from this, it's very hard to say what

the implications of the election were.

409

but it's very intriguing to see.

410

As Hannah says that, you know

something that's been true for a long

411

time, just wasn't true last year.

412

Hannah Hartig: This.

413

Turnout birth persuasion conversation

is really interesting because

414

inherent in it is this idea that

views are fixed in a funny way.

415

And we know that views change,

particularly among people who

416

don't turn out consistently.

417

, I, I love to point this out, but

it's important for our community.

418

I think in, in the polling world, it's

like we are some of the weird ones.

419

We study what people think about

politics, and a lot of the country

420

doesn't think about politics.

421

And so again, you know, Scott's suggesting

we don't wanna be too declarative

422

about this, but views change, right?

423

A lot's already happened

in the past six months.

424

Things will happen in the next,

over the course of the next few

425

months or the next few years.

426

And so of course that's our job

is to track how people feel.

427

But that's part of the weird tension in

some of these conversations is, you know,

428

turnout is, you're either this or this,

and you either turn out or you don't.

429

And we know that that's not how

elections work or politics work.

430

Farrah Bostic: Well, I love to hear that

because, um, as a, my, my background is

431

as a qualitative researcher, and so I find

these kind of mechanistic explanations

432

for people's behavior to be not what it

looks like in meet space with real people.

433

And I think that kind of, I, I'll confess,

it wasn't until, I dunno, a couple months

434

ago, I I, I hadn't realized that this

was a ideological idea, that turnout and

435

persuasion, like, I didn't understand

that that had like a, a valence to it.

436

That was anything other than just like.

437

Wouldn't you do both?

438

But apparently, at least

amongst Democrats, there's

439

like a real theory about that.

440

, One thing makes you go far to the left.

441

One thing makes you who to the center odd.

442

As a marketer it's like, hmm,

I think I need both, but okay.

443

But you, you raise a really

interesting question as well, which

444

is, and here I'm thinking about last

week's election in New York City.

445

Uh, I lived in New York for 20 plus

years and I'm now out in Long Island.

446

, But a lot of the analysis that we're

seeing is essentially like, well this was

447

just a different electorate in 2025 than

448

Like there were different.

449

Different types of voters

voted than voted last time.

450

And this is kind of the ongoing challenge

of any sort of attempt to predict

451

the outcome of an election is who do

you think the electorate's gonna be?

452

And then the other thing that I think you,

you do take pains to note in the report

453

is we are living in this period of very.

454

It's very specific, partisan polarization

and deep antipathy, uh, amongst partisans

455

on either side, towards the other party.

456

And so subtle shifts can make a

huge difference in the outcome.

457

And a lot of these shifts in certainly

on the turnout side of things.

458

And, you know, how many Biden voters

voted for Harris versus Trump versus

459

the couch as people like to say.

460

Those, those shifts are subtle.

461

Like it's, it's a two points, 1.3

462

points here and there.

463

And just on that basic sort of, did

you vote, did you switch from one party

464

to another when you chose to vote?

465

All of those things can just like

lead to real interesting results at

466

the same time after the election.

467

Again, I hear, I'm thinking about the

Northeast, there were a lot of those arrow

468

maps that the, that the New York Times

liked to do where it was like the, the

469

Redshift idea, but it was hard to parse

like Redshift plus decline in turnout.

470

I'm curious about whether anybody is

attempting to kind of model this state

471

by state since obviously we don't

have a national popular election.

472

It is an electoral college process.

473

You know, the UK loves to do their MRP

studies and hardly anybody does them here.

474

And I'm just curious about like how

much anyone's trying to model, sort of

475

state by state how turnout worked or

how how that affected, um, outcomes.

476

Scott Keeter: We did take a look

internally at our data, our, our

477

panel's not really designed to do

state level analysis, and so we've

478

always been very reluctant even to

do battleground states versus every.

479

All the others,

480

but we're all very curious people.

481

So I, I did take a look at that and

what you see is really not surprising

482

because the campaign was so heavily

focused on the battleground states.

483

Turnout in the battleground states

was quite a bit higher than it

484

was in the rest of the country.

485

Now, I don't know if you, you know,

you have to take into account which

486

are the battleground states and what's

their baseline level of turnout.

487

But the, the fact is that, uh, there was

a lot of mobilization going on there,

488

and so you didn't have as much fall

off in support, let's say differential

489

falloff that hurt Harris because

Biden's voters didn't, didn't turn out.

490

At a greater rate than, than

Trump's voters didn't turn out.

491

Not saying that very well, but

you know, so the mobilization in

492

the battleground states helped.

493

So what happened there is that

whatever change happened had

494

to come from other sources.

495

And our data suggests, and this is,

this solved, very tentative, and we

496

don't put it in the report 'cause it

is, it is tentative, but our, , our data

497

suggests that it was mobilization of

the non-voters, the new and returning

498

voters, uh, in the battleground states.

499

It probably was more important even

than, as a factor than it was nationwide.

500

But this is slicing the

salami very, very thin.

501

And so, it, it's kind of speculative.

502

But, , I do think it would be fun

if you had the data to be able to

503

model it on a state by state basis

because it, it still is a little

504

true that all politics is local.

505

Farrah Bostic: Yeah, I mean, you, you

get right to the thing that I was,

506

the initial reason I reached out to

the team about talking to someone

507

about these results is I have been.

508

Particularly, I mean, look, I work

on the commercial side in market

509

research, and we also tend not to study

people who outright reject a category.

510

Like, you know, if you're, if you're

just, you're, you're not a soda drinker,

511

there's not a lot of point in me

asking you which soda you'd like best.

512

You don't drink it, there's no point.

513

But I think because there is a

churn in people voting this time,

514

not voting the next time people

sitting at a couple of elections

515

and then, and then getting involved,

new voters coming into the system.

516

It has been a real kind of question

mark for me about what do we know.

517

About non-voters and about that

kind of rate of churn that, that we

518

see between people who vote in one

cycle and then sit out the next one.

519

Uh, new people coming into the electorate.

520

And, and, you know, and obviously a

lot of this comes down to is it even a

521

campaign strategy to try to get those

folks who tend to choose, and I, I

522

really hate this metaphor, but they

tend to choose the couch, um, tend to

523

choose childcare and, uh, commuting to

work and all the other responsibilities

524

they have in their lives over voting.

525

But, uh, what do you know,

what do you know about non

526

voters from, from this study?

527

, Hannah Hartig: So we already highlighted

one of the major findings, which

528

is that they're, they're fairly

closely divided in their preferences.

529

And, and as we noted , they

tend to be less likely to have

530

formal levels of education.

531

They're younger than the people who vote.

532

They're closely divided in their

533

Of course, that changes

from election to election.

534

They're paying slightly less

attention to politics depending on

535

the metric, uh, that you're using.

536

But yeah, I mean, some of the

characteristics at held in

537

previous presidential elections

538

Scott Keeter: Your question also

just sort of goes to, goes to the

539

issue of like, how big is this

group of people that, that are, um.

540

Sort of , up in the air.

541

Like not the chronic voters, the people

who are always gonna show up, but the

542

people who are in play as, as it were.

543

And I mean, our estimate in this election

is, is that at defining this as who did

544

something different between 2024 and

545

of, of the total eligible public who

could vote did something different.

546

They either turned out when they

hadn't before, they didn't turn

547

out when they had before or they

voted in both elections, but they

548

switched their candidate preference.

549

And we haven't gone back and done a

long look at this, although our panel

550

does make it possible for us to at

least go back, you know, four years.

551

And the sheriff voters that were.

552

In that category of, of changers

of one sort or another was

553

about the same four years ago.

554

It just may be that that's the nature of

American politics right now is that about

555

three quarters seem to be locked into

something and one quarters , is available.

556

If you look at it that way, if

you're a campaign to either be

557

demobilized, mobilized, or flipped.

558

Farrah Bostic: Mm-hmm.

559

Hannah Hartig: And I'll add, that's

one of the, the value adds of

560

this particular analysis, right?

561

Is using that panel data as it

helps us kind of triangulate

562

this churn in the electorate.

563

'cause the, you know, the exits are,

are really the exit polls, sorry,

564

are really focused on the people that

showed up in that particular election.

565

And that's a really rich

data source in and of itself.

566

But using this panel data allows us

to walk out that sort of churn who's

567

showing up it, and it matters for some

of these subgroups when people are

568

extrapolating their narratives out, right?

569

It's like, oh, Hispanic

voters shifted by X amount.

570

Well, is that because different

types of Hispanic voters showed up?

571

Is that because the same exact, you know,

I'm, I'm just using an example here.

572

The same exact voters that showed

573

There was a wholesale change

in how they viewed politics.

574

I think it's a really, it's just

another opportunity to showcase the

575

value of panel data here and for

this particular look at the election.

576

Farrah Bostic: Yeah.

577

And then, uh, we'll link to it

as well, one of the reports that

578

you, uh, have on the website.

579

It has some nice animations as you kind

of scroll down of how the different

580

cohorts kind of moved around , and

where where the non-voters and where

581

the switchers funneled themselves to.

582

I was talking to, , a Shankar

Osorio last week and she made the

583

good point of like, actually it's

usually like a four-way race.

584

It's the two parties, some third party,

and then not voting are the choices that

585

Americans actually have when they, when

they decide about what they're gonna do

586

it does look like there were some kind

of in, I mean, just some interesting sort

587

of structural shifts that a lot of the,

I mean, I have two books on my shelf.

588

One is the emerging Republican

majority and one is the emerging

589

Democratic majority, uh,

written about 30 years apart.

590

Um, but um, but like this sort of

demographics as destiny thing just seems

591

like not so much and that there's some

really interesting kind of comparisons

592

even between 2016 and 2024 when you have

a woman democrat running for president

593

and like women dropped off and some

of the ethnic cohorts dropped off.

594

And, um, and there were

some other changes as well.

595

And I'm, I'm curious about, you know,

did you see kind of corresponding

596

gains for Trump among those

groups or did those groups tend

597

to funnel more into the non-voter

side of things or, or can you tell

598

Hannah Hartig: Yeah.

599

I think this points to something you

asked earlier, which is like one of the

600

key questions, and it's that it's, I

almost think of it as a spectrum, right?

601

If you're turning out people who

haven't supported you in past,

602

that's a form of persuasion.

603

You're convincing someone to turn out.

604

And so it's like, you know, it

depends on your starting place.

605

Of course, if you're locked into politics,

, are you sort of souring on whatever

606

political party you've been attached to,

or are you not liking the candidate that's

607

representing that party at that time?

608

And so maybe you decide, sit out this

selection, or as you pointed to, there

609

are a variety of just daily life things

that occur that might prevent someone

610

, from casting a ballot when they plan to.

611

And then.

612

You might make your way all the way

over to the other side and decide

613

to, to switch your allegiances.

614

But obviously if you're someone who

hasn't been engaged with the political

615

process before depending on where

you are in life or that, that is kind

616

of a different calculus, you know,

you're getting different messages from

617

different parties and so they just

might be persuading you to turn out for

618

the first time if you haven't before.

619

And so that's where it all

gets tricky because it's,

620

yeah, it's not black and white.

621

You either turned out or you're

persuaded, to vote for someone different.

622

I think this alludes to something you

were pointing to earlier, which is that

623

if you're seeing a change in the types

of people who are turning out, that's

624

potentially a form of persuasion, right?

625

Um, you're not convinced to switch

sides, but maybe you decided you

626

wanted to sit this, this round out.

627

Farrah Bostic: you know, Scott, you've,

you've been using the word mobilization

628

as part of this equation as well,

which is just, it does seem like

629

there's, you know, there's multiple

layers to get through here, right?

630

You've gotta, you've gotta convince

people to vote and then you've gotta

631

convince people to vote for you.

632

And yet on the other hand, there are all

sorts of factors that demobilize people.

633

And I'm, I'm wondering if either of you

can speak to, do, do you have a sense

634

of broadly what motivates or demotivates

people to participate in elections?

635

Scott Keeter: Well, I don't think any

of our data , can speak directly to it.

636

You know, we have done work in the past

that that shows that at least people who

637

recall being contacted by campaigns and

so forth are more likely to turn out.

638

That's kind of a duh finding maybe.

639

But,

640

That it actually, the

campaigns really matter.

641

They may matter in terms of

persuasion, uh, but they definitely

642

matter in terms of turnout.

643

If you don't, uh, reach out and ask

somebody, there are just enough.

644

Things in our society, if you're not

already an intrinsically, highly motivated

645

voter type of person, there's just enough

going on in people's lives, enough, uh,

646

friction in the process that, um, you

know, it's easy for people not to vote.

647

It's just not as important to, to

a lot of people as it as it might

648

be to the three of us, for example,

because we study politics and, , you

649

know, , keep on top of this stuff.

650

So because of that campaigning,

really does, does make a difference.

651

And we can see it, for example, in

the turnout, in the battleground

652

states as compared with the, with

the rest of the, of the country.

653

The question of persuasion

is a, is a different story.

654

And there, I think what's,

what is interesting is that.

655

We know that, uh, people who are

really highly politically engaged tend

656

not to be very persuadable, right?

657

I mean, they are paying attention

to messages, so, you know, at

658

least theoretically you might

be able to persuade them.

659

But because they tend to be ideologically

inclined, they, are not really open to

660

persuasion, to the same degree as people

who are not as politically engaged.

661

And so , the consequence of that

is that , what moves them tends to

662

be things like unhappiness with the

status quo, inflation, um, having

663

trouble making ends meet, , trying

to figure out who to blame for.

664

The perception of job competition

from recent immigrants or other

665

kinds of, uh, things that we think

were at play in, in this election.

666

And if, the less engaged folks are moved

by those economic factors or, or social

667

factors, uh, and then the campaigns

can come in and mobilize them, then

668

you have the potential for some real

change from one election to the next.

669

And that might be one part of the

explanation of, you know, how Trump was

670

able to come back and make, you know, make

up a fairly wide difference between , his

671

margin of defeat to Joe Biden and the

fact that he was able to, to actually

672

win a plurality of the popular vote.

673

. Farrah Bostic: I'm also curious about some

of the findings that you had around age.

674

Uh, another kind of emergent narrative

from this year is that younger voters,

675

particularly younger male voters.

676

Shifted towards the Republican

side of the spectrum.

677

And I'm, I'm curious about what you

found when you looked at, at those

678

kind of age and gender breakdowns.

679

, Hannah Hartig: We did see, um, a bit of

a shift among young men in particular.

680

If you zoom out, it's, you know,

it, it depends on if you're looking

681

at 2020 compared with 2024, if you

682

understanding of, of where young men are.

683

But our data does suggest between

684

towards Trump than among either

younger women or older men, for

685

example.

686

Farrah Bostic: And, and one of the other

charts that you have in there that, I

687

really encourage people to take a look at

is a like generation, well, I don't wanna

688

say generational, but a like decades based

birth cohort shift, because I think that's

689

one of the other things that sometimes.

690

You know, it's, it's not like,

uh, everyone's 21 forever, sadly.

691

And so, and so, um, you know, the

692

are gonna be more or less in the 30

to, you know, 30 plus category now.

693

and there was also some drop off

among those voters supporting

694

the democratic tickets.

695

So there's sort of a, a shift to the

right, which you know, that's an old

696

narrative that the older you get,

the more conservative you become.

697

And there was an idea that millennials

were gonna break that trend.

698

And, and maybe they are

ish, but also maybe not.

699

And yet this, you know, each success of

year, that kind of 18 to 29 group for

700

young men has like, there, there has

been a lot of hand wringing about , what

701

exactly explains the, the shift there.

702

Some analyses I've seen

have shown sort of a.

703

It's all kind of highly correlated

with a lot of factors that young men

704

are attending college less so, they're

not in the higher education cohort.

705

They may be earning less, they may

be in the less affluent cohort.

706

They may be not as sort of

economically mobile, and so they're

707

less likely to be in urban areas.

708

I mean, there's all these kinds

of, , factors that may correlate there.

709

What, what do you make of any of

those kinds of explanations of

710

what's going on with younger voters?

711

Scott Keeter: And the first thing I

would say is that, that it's just, it's

712

unfortunate, but I just don't think.

713

We have good data to really unpack this

question to the degree we would like.

714

It, it's, it's just a case that

that young adults are very difficult

715

to find and survey to get good

representative samples of young adults

716

these days, especially young adults

that are not engaged in politics.

717

You know, it's just,

it's not a dirty secret.

718

It's hardly a secret at all that,

uh, surveys tend to overrepresent

719

politically engaged people.

720

And that problem is, is really worse when

you're talking about younger demographics.

721

And so, I'm, I'm not saying this to

excuse our inability to give you a

722

definitive answer, but simply to say

that there's a lot of disagreement

723

among the surveys that we saw over

the over the course of the past year.

724

There were some surveys, including

our own, that showed really rather

725

massive shifts in a Republican direction

and party affiliation among young.

726

Men.

727

But then there were other surveys

that we did that were bigger and

728

possibly better, that , showed the

shifts that we're talking about.

729

But they weren't, as dramatic as one

of the studies was, and as some of the

730

other surveys that you've seen published.

731

And so I think, we have to be cautious

in not over reacting to the notion

732

of the manosphere is, is driving,

you know, young men to the, into

733

the arms of the Republican party.

734

I think that's, I think that's

premature, both because we

735

don't have the data for it.

736

And also because there still is the

simple explanation that, you know,

737

the last four years were really tough

on a lot of people economically,

738

especially people who are economically

marginal to begin with, who may not be

739

established in jobs, who sort of more.

740

You know, subject to the, to the bad

winds of the economy and so forth.

741

And so there might be simpler

explanations for why , young people

742

and young men in particular were more

receptive to trump's messages, uh,

743

than they might've been in the past.

744

Um, and so I'd, I'd like to see

another couple of years worth of data

745

before we draw any firm conclusions

about whether there's a, like a real

746

generational shift happening here.

747

Hannah Hartig: I'm glad you did point

out our age cohort data though, but that

748

was, um, something we included in this

year's edition and it, I, I love it for

749

the exact reason that you just said, it

allows you to understand how people are.

750

Changing over time.

751

And as you know, we're not 21 forever.

752

And what our data suggested, again, not

to over interpret this turnout versus

753

persuasion point, but , the people

who were born in the nineties and two

754

thousands the change in the margin

was largely the result of a different

755

mix of people showing up in 2024.

756

Compared with 2020 in the 1980s.

757

It's a slightly different story.

758

We, we saw a higher conversion from

Biden to Trump among this, birth

759

cohort , than the other way around.

760

So Trump did make some inroads among this

group and he effectively changed more

761

of, of those people's minds than uh, uh,

Harris did for Trump to, to convert them.

762

And so that's just a little bit

more specific, you know, within

763

that group when you're kind of

parsing them with a fine tooth comb.

764

Farrah Bostic: Well, and, and this is

why this, this is the analysis I was

765

wa the Pew analysis was, the analysis

I was waiting for is, I think the level

766

of granularity here is actually really

useful because it is so easy for these.

767

Kind of broad demographic cohorts to

get sliced and diced in such a way that

768

a narrative emerges really quickly.

769

That is like, uh, young men are all

turning towards the Republican party.

770

And that is not true.

771

And it is also not explained by

the fact that they're young men.

772

Like that, there, there's all

these things that come with being

773

a young man in, in today's economy

and today's internet and today's

774

everything else that may be influencing

their, their choices in this way.

775

And I think the same thing is true as

we go through kind of, life stages.

776

I mean, this, years ago used to be

a bigger fight I had to have with my

777

clients about we would do focus groups

and they would wanna have like, here's

778

the 18 to 28 group and here's the,

you know, 29 to 34 group and whatever.

779

And I would be like, but do we

want parents in the 18 to 28

780

group because they're gonna be

different than the non-parents.

781

And do you want the college

students together with the not

782

in college people and like.

783

Life stage matters, context matters.

784

They will not sound

like each other at all.

785

Um, and sure enough, uh, they never did.

786

And so thankfully we no

longer have this fight.

787

Those pieces of context do really matter.

788

And you know, there were some amusing

memes coming into the election about

789

how like Gen X was betraying everyone

by tilting Republican, or maybe it's

790

just like when you have a house, and

a mortgage and you know, you're paying

791

your kids' tuition, , you start to

look for breaks, you start to look for

792

ways to, uh, pay less in taxes or get

more of a benefit from those things.

793

And that just shifts your,

your political orientations.

794

And it has nothing to do with your age

specifically or the year that you were

795

born, except in so far as it makes it more

likely that you're in those life stages.

796

The thing I do think about is

these types of pat explanations for

797

things or pat descriptions really

of things take hold really quickly.

798

And so by the time we get these more

detailed analyses of what happened

799

in the election, it does feel like

some narratives have been baked.

800

And I'm curious about are, are there

any others that you think maybe get

801

unbaked by the, by this, this set

of data or this set of analysis?

802

Are there other things that you think

like, actually this shed a little

803

bit of light or cast a little bit of

doubt on some things that had started

804

to kind of emerge as explanations

for what's been happening over the

805

last two, three election cycles?

806

Hannah Hartig: I think you've

pointed to one of the big ones,

807

which is, I'm, I'm really glad Scott

pointed to this, which is just that

808

we need a little bit more data on

809

this, , especially the youngest

cohort of people who are turning out.

810

, So we need more data around that.

811

But I mean, all of these analyses,

right, just help us triangulate,

812

like the exit serve a purpose too.

813

And, and you're a hundred percent right

that sometimes narratives get formed and

814

then it's, it's hard to unbaked them.

815

I like that analogy a lot.

816

But what this offers us is just a chance

to, further confirm something that

817

another analysis might have been saying.

818

For example, um, the exits

suggested, , a large shift in support

819

among Hispanic voters, and sure

enough, we, we found that as well.

820

And I know, you know, catalyst is

another great firm that put out a, a very

821

extensive analysis of this, uh, election.

822

And , they saw something similar in

terms of things that made me jump

823

back and say like, wait a second.

824

That's not what our data says at all.

825

Nothing's immediately coming to mind, but.

826

Looking at Scott to.

827

Scott Keeter: No, I, I, same reaction.

828

Um, and I, I'm glad you mentioned.

829

The Hispanic shifts Hannah.

830

'cause you know, I think that that's

been the subject of intense interest

831

since really before the election,

uh, looking at pre-election polling.

832

But, you know, it was clear both from

precinct analyses that the New York Times

833

did and other organizations from the exit

polls and others that, Hispanic voters

834

have, you know, had shifted quite a bit.

835

The magnitude of the shifts

is, I think, debatable.

836

And while I, I like our data and.

837

Think we've done a good job of building a

good Hispanic sub-sample into our panel.

838

The reality is that surveying Hispanics

is very difficult to, to do well.

839

And so the fact that there are

competing estimates of, how much of

840

a shift there's been or out there

doesn't, doesn't really surprise me.

841

It's just the, the nature of the beast.

842

But, you know, we could also toss into

the mix here, shifts among black voters.

843

There's a, you know, 15 percent

black voters supporting Trump.

844

If, if I remember my number right, up

from 8%, I believe four years, uh, ago.

845

That's not a massive shift.

846

But given how reliable African American

voters have been for the Democratic Party,

847

it was a no, it was a notable shift.

848

And then, you know, we we're

not really able to do a great

849

job with Asian Americans of.

850

Fastest growing minority

group in the country.

851

But, uh, we, we documented about a

10 point shift, I believe, towards

852

Trump, among Asian American voters.

853

Now in our panel, these are English

speaking Asian American voters,

854

and that is not all, um, Asian

Americans probably, you know, misses,

855

misses some important political

subgroups of Asian American voters.

856

But that's a, you know, if you're

thinking about the future of American

857

politics, , these groups are really

important to pay attention to.

858

And I think we're getting a better handle

on them now than we did in the past.

859

We have a large enough sub-sample of them

to be able to report separately on Asian

860

American voters.

861

But I think those are, those are

groups that, you know, I was very

862

keenly interested in seeing what

our numbers were going to be.

863

And I think we want to keep an eye

864

Farrah Bostic: Yeah.

865

Scott Keeter: them going forward.

866

Farrah Bostic: Yeah.

867

I mean, again, thinking about the, some

of the early data coming out of the.

868

New York City mayoral race.

869

There was sort of this assumption

that Queens has long been a kind of

870

conservative leaning borough of New

York City, and it went from Mom Donny.

871

And part of me is like, yeah, it went from

mom Donny like the, the aunties love him,

872

like it's like the most diverse borough in

New York, city , of a very diverse city.

873

And so, you have a lot of a

lot of languages spoken and a

874

lot of, uh, different kinds of

people's origin stories there.

875

And that also leads me to another finding

that you had about naturalized citizens

876

and their, their kind of shifts in their

voting behavior, which I think is another

877

thing that the parties have to adapt to

because there was a, there was a story

878

about that, and that story looks like

it's it's shifting at the very least.

879

Hannah Hartig: that's, that

was one another interesting

880

addition in this year's report.

881

And we were able to look back

882

real, I think, uh, I first heard

it on, uh, David Shores podcast.

883

And we're, you know, already

looking at that in our data as well.

884

And, and we saw it reflected as, in 2024.

885

So Trump seemed to make gains and, and

again, these are mostly a different mix

886

of voters turning out, but Trump did,

uh, improve his performance compared with

887

2020 among, those naturalized citizens.

888

So people who are eligible to vote

but not born in the United States,

889

which is another important data

point for understanding the broader

890

context of the, the whole election.

891

I think.

892

Farrah Bostic: Right.

893

I think the other thing that was useful,

because I do feel like there's sort

894

of a, a public narrative conception

of who immigrants are from a racial

895

makeup, and I think you have, you have

some cuts there as well of, naturalized

896

citizens , by race and ethnicity.

897

And so that, it's, you know,

it's not one particular cohort.

898

It looked like, it looked like kind

of across the board a shift towards,

899

towards the republican side of the aisle.

900

Is that fair to say?

901

Yeah.

902

Yeah.

903

Yeah.

904

, What's the data you wish you

had for an analysis like this?

905

If I

906

could wave a

907

magic wand, give you anything you

wanted, what would be on your list?

908

Scott Keeter: I like about 5,000 more

cases of people who are 18 to 25,

909

because

910

Farrah Bostic: Yes.

911

Scott Keeter: I really want to know

what's going on with that group.

912

And you know, it's, uh, it is just, just,

uh, we're, we're frustrated that we don't,

913

don't have as, as many of them as we want.

914

You know, it's a, it's a great group

because it's, the future of the country.

915

It's more racially and ethnically

diverse than, you know, other cohorts.

916

And so it's interesting in that respect

because it tells you, tells you something

917

about what the country is becoming.

918

But uh, we don't have it

919

so.

920

Hannah Hartig: I second that more, more

younger adults, um, more people who

921

are, are not as politically engaged, who

aren't voting in every single election.

922

Just more, more of them.

923

Farrah Bostic: I went looking last

summer for anybody who just studied

924

infrequent or non-voters, people who

are eligible to vote but don't vote

925

frequently or, or don't vote at all.

926

and I'm curious about

why this seems to be.

927

not more studied, I guess

is the, is the question.

928

Uh, what do you think accounts for us

knowing so little about the non-voter,

929

apart from, you know, what we can

sort of surmise from demographics?

930

Hannah Hartig: It's directly

related to who responds to surveys.

931

And so it's as money is a factor, right?

932

It takes money to, to find people, to

convince them to respond to your survey.

933

it's hard to convince someone to sign up

for a panel, let alone a one-off survey.

934

Um, and so I think that's one of , the

core issues at hand is just getting people

935

to respond to surveys, particularly those

who aren't paying attention to politics.

936

We don't do phone surveys

anymore, but in the days of phone

937

Farrah Bostic: It's only a

few days before the election.

938

Now as I'm writing this, it's Halloween.

939

The election's on Tuesday.

940

After I record this, I'm gonna go

cast my ballot with my in-person early

941

voting friends down in my neighborhood.

942

I hope you have either gone to vote in

an early voting station or submitted

943

your mail-in ballot already, or you know

where you're gonna take your mail-in

944

ballot if you prefer to drop it off or

you've made a plan to vote in person.

945

It's easy to believe that

your vote doesn't matter.

946

You may think you live in a safe state

or that your vote will be drowned

947

out by all those neighbors who don't

agree with you, but your vote does

948

count because it is your voice.

949

And in a close election, every

vote matters in our deeply

950

polarized political environment.

951

It matters for a very particular reason.

952

The electoral college will ultimately

decide the election that's in the

953

Constitution, and so it's reasonable

to think, as Mike Pdoa told us, people

954

don't vote places due, but when the

last two Republican presidents have

955

lost the popular vote, but won the

electoral college, when it seems like

956

it takes at least a three point margin.

957

To guarantee a Democrat

wins the electoral college.

958

The popular vote takes on a salience that

I think goes kind of underappreciated.

959

The popular vote conveys legitimacy on

the decision of the electoral college.

960

So even if you're in a safe district

or you're sure you're outnumbered,

961

nevertheless your ballot will be

counted and your vote can help run up

962

the score for your preferred candidate.

963

It may not help your choice win

electoral college votes, but it

964

can help them win an election.

965

The rest of us can believe in.

966

Maybe we'll see.

967

On Tuesday as we round the corner on the

election season, I wanted to explore the

968

application of what we know about the

electorate from polling and political

969

science with someone who works in

politics and who has a strong point of

970

view, someone who can stare directly

at the world in front of her and see it

971

as it is, not as she wishes it to be.

972

I don't know about you, but

that has been in short supply

973

for me this election season.

974

This season seems to wanna avoid that kind

of clarity as much as it possibly can.

975

So enough of that.

976

My guest today is the inimitable Dr.

977

Rachel Bit coffer.

978

She is a political analyst,

strategist and author of Hit Ware.

979

It Hurts How to Save Democracy by

Beating Republicans at their own

980

game after a career in academia

where she taught political science

981

and ran a survey research center.

982

Bit coffer shifted to focus on

reforming democratic campaign strategy.

983

She is known for her theory of

negative partisanship, which we

984

discuss in this conversation, and for

her accurate electoral predictions.

985

Even if election Twitter

is not ready to admit that.

986

Biter has worked with the Doffer, has

worked with the Democratic National

987

Committee to implement more effective

messaging strategies, and she regularly

988

provides commentary on political campaigns

and electoral dynamics, and she very

989

graciously joined me on cross tabs.

990

Here's our conversation.

991

With only a few days left to election

day, I want you to know that already

992

more than 65 million people have

cast their ballots while this trails,

993

the COVID era 2020 early vote, it

is still a really robust early vote.

994

So if you have already voted,

I wanna extend my deepest.

995

Thanks for your attention to

making your own voice heard.

996

It's important.

997

We're gonna continue the

conversation next week.

998

After the results are in, I have

some great interviews coming up so we

999

can continue to explore the ways we

understand each other as fellow citizens

00:56:39,694 --> 00:56:44,254

at scale, how we forecast outcomes,

imagine and plan for the future, and

00:56:44,284 --> 00:56:48,874

mobilize each other as we continue to

pursue, I hope the Democratic project.

00:56:49,714 --> 00:56:52,354

I don't think it's a secret

who I'm voting for, it's.

00:56:52,744 --> 00:56:56,614

Probably pretty obvious that a highly

educated woman in New York is a Democrat.

00:56:57,034 --> 00:57:00,994

But I wanna tell you that I am voting

for Kamala Harris and Tim Walls

00:57:01,024 --> 00:57:05,254

precisely because I believe in the

Democratic project and because I love

00:57:05,254 --> 00:57:07,864

this country for its unending potential.

00:57:07,864 --> 00:57:07,924

I.

00:57:08,884 --> 00:57:12,634

And to those of you who listen all the

way to the end of each episode, I want

00:57:12,634 --> 00:57:17,434

you to know that I'm grateful for your

interest in this subject and for many of

00:57:17,434 --> 00:57:22,144

you, the work you do to help us all better

understand each other as a body politic.

00:57:22,414 --> 00:57:25,714

It's important not just to

my show, but to all of us.

00:57:26,344 --> 00:57:30,364

So take care of yourselves and

I'll see you on the other side.

00:57:32,145 --> 00:57:35,375

Hannah Hartig: polling, if you're

not interested in political news

00:57:35,375 --> 00:57:37,805

or following politics, and you

pick up the phone and someone says,

00:57:37,805 --> 00:57:39,785

can I talk to you about politics?

00:57:39,785 --> 00:57:40,205

You're like,

00:57:40,685 --> 00:57:40,895

I'm

00:57:40,895 --> 00:57:41,255

good.

00:57:42,005 --> 00:57:43,385

And you know, that's true.

00:57:43,385 --> 00:57:46,715

However, you're surveying them over

the phone or online, if you get a

00:57:46,715 --> 00:57:50,785

mailer that's asking you to respond to

a survey those things are correlated.

00:57:50,885 --> 00:57:56,345

, You just not as willing to respond to a

survey and fill out a long questionnaire.

00:57:56,395 --> 00:58:00,505

And so yeah, that's plagues us

in our, in our panel work is

00:58:00,510 --> 00:58:02,875

in, in our survey work as well.

00:58:02,955 --> 00:58:06,935

Farrah Bostic: I always wonder about you

know, just slipping in political questions

00:58:06,935 --> 00:58:10,865

and what is otherwise like a consumer

questionnaire and just like they don't

00:58:10,865 --> 00:58:12,095

know that they're coming in for politics.

00:58:12,095 --> 00:58:13,535

But we're just gonna quickly ask, uh,

00:58:14,450 --> 00:58:14,740

Yeah.

00:58:14,960 --> 00:58:15,420

By the way.

00:58:16,200 --> 00:58:18,395

by the way, are you, are

you registered to vote?

00:58:18,975 --> 00:58:19,995

Hannah Hartig: Yeah, exactly.

00:58:20,865 --> 00:58:23,205

Farrah Bostic: Tack it on at the end after

we've asked, you know, a bunch of other

00:58:23,205 --> 00:58:26,115

questions about, you know, their favorite

soda, whether they drink it or not.

00:58:26,475 --> 00:58:29,295

I guess my final question for you is, what

should I have asked you about this study?

00:58:29,295 --> 00:58:31,785

Are there, are there particular things

that you're proud of or interested in

00:58:31,785 --> 00:58:34,665

or wanna pull threads on more that, that

we didn't get a chance to talk about?

00:58:35,017 --> 00:58:38,167

Scott Keeter: Well, I've really enjoyed

the, the conversation because we did talk

00:58:38,167 --> 00:58:42,527

a lot about the intricacies of putting

it together and the difficulties of,

00:58:43,037 --> 00:58:47,777

of, of trying to deal with a panel where

you, it allows you to look at, look over

00:58:47,777 --> 00:58:51,047

time change in the same individuals.

00:58:51,047 --> 00:58:55,667

And so I appreciated the opportunity to,

to hold forth about that a little bit.

00:58:56,927 --> 00:58:58,637

Hannah Hartig: Yeah, I really appreciate.

00:58:58,637 --> 00:59:00,317

yeah, yeah, exactly.

00:59:00,317 --> 00:59:04,067

And I really appreciated being able

to just kind of opine on this idea

00:59:04,067 --> 00:59:05,597

of like turnout versus persuasion.

00:59:05,597 --> 00:59:08,867

It's something Scott and I have, you

know, been thinking really hard about

00:59:08,907 --> 00:59:10,707

and so really important questions and I.

00:59:10,947 --> 00:59:13,197

Yeah, you identified them out of the gate.

00:59:13,977 --> 00:59:14,267

Farrah Bostic: Yeah.

00:59:14,367 --> 00:59:19,487

It has become a thing that, . Again, as

a marketer, we can't expect to persuade

00:59:19,487 --> 00:59:24,137

people to buy stuff if they don't hear

from us and have never heard of us before.

00:59:24,407 --> 00:59:26,957

And so this kind of, it is always both.

00:59:26,957 --> 00:59:29,627

It is always like, we have

to let you know we exist.

00:59:29,897 --> 00:59:32,087

We have to give you a reason

to think that's a good thing.

00:59:32,117 --> 00:59:34,277

Uh, we have to entice you

to wanna give us a try.

00:59:34,577 --> 00:59:35,957

All of those things are true.

00:59:35,957 --> 00:59:37,067

And obviously the stakes are different.

00:59:37,067 --> 00:59:40,637

I don't have to get you to do it

like for a two month period, once

00:59:40,637 --> 00:59:41,897

every four years or two years.

00:59:42,257 --> 00:59:43,187

I can do it every day.

00:59:43,637 --> 00:59:46,157

But, um, but the, the.

00:59:46,562 --> 00:59:51,872

The thinking about campaigns

matter and mobilization matters,

00:59:51,872 --> 00:59:55,592

and, and like where you put your

effort does pay off some dividends.

00:59:55,592 --> 01:00:00,052

And so if the Trump campaign has

reported spent more time looking at and

01:00:00,052 --> 01:00:04,912

targeting infrequent or non voters or

new voters, then the Harris campaign did.

01:00:04,912 --> 01:00:08,302

Or if one, if both of the campaigns

spent most of their time and effort in

01:00:08,302 --> 01:00:11,842

battleground states, then that's going

to shape the turnout and it's going to

01:00:11,842 --> 01:00:13,792

shape what what the electorate looks like.

01:00:14,072 --> 01:00:18,002

And I, you know, wrote something over the

weekend about, about the kind of Mom Donny

01:00:18,002 --> 01:00:20,912

case, which I look at as market making.

01:00:20,972 --> 01:00:23,462

Like he went out and tried to

assemble an electorate that

01:00:23,462 --> 01:00:25,232

01:00:25,472 --> 01:00:26,822

And it appears like.

01:00:27,632 --> 01:00:30,782

The five candidates who ran for mayor

on the Democratic ticket did that.

01:00:30,782 --> 01:00:32,852

They created a different

electorate this year.

01:00:33,162 --> 01:00:35,772

, And it makes prediction really hard to

do, but it makes these kinds of studies

01:00:35,772 --> 01:00:40,152

really important because we can find out

what kind of electorate they made, um,

01:00:40,212 --> 01:00:44,052

and, uh, and start to think about what

that means for any future electorate.

01:00:44,052 --> 01:00:46,422

So I really appreciate the work that

you've done and your willingness

01:00:46,422 --> 01:00:47,592

to come on and talk to me about it.

01:00:47,592 --> 01:00:52,092

It's, it's always fun to, for me and I

hope for these listeners to, uh, to get

01:00:52,092 --> 01:00:55,842

into the weeds of how it's done and, and

what we can glean from and what we can't.

01:00:56,072 --> 01:00:56,312

Scott Keeter: you.

01:00:56,477 --> 01:00:56,837

Hannah Hartig: Yeah,

01:00:56,947 --> 01:00:57,617

Farrah Bostic: thank you so much.

01:00:57,957 --> 01:01:00,522

The last thing I like to

do though is how can people

01:01:00,522 --> 01:01:03,642

support Pew and follow your work

01:01:04,001 --> 01:01:06,431

Hannah Hartig: You can

visit pew research.org.

01:01:06,491 --> 01:01:09,071

We study a lot of things,

not just politics.

01:01:09,071 --> 01:01:13,451

So if you're interested, for example, on

how people are viewing AI or a variety

01:01:13,451 --> 01:01:17,411

of different things happening in our

world today, we study a lot of it.

01:01:18,166 --> 01:01:18,436

Farrah Bostic: Yes.

01:01:18,436 --> 01:01:22,156

Anybody who works in my industry of

market research or in, uh, marketing

01:01:22,156 --> 01:01:26,026

in general, if you are not looking

at the Pew Research website on the

01:01:26,026 --> 01:01:30,346

regular, you are missing all this

amazing publicly available data that

01:01:30,346 --> 01:01:31,666

you can't afford to go get on your own.

01:01:31,666 --> 01:01:32,716

So you should go do it.

01:01:32,996 --> 01:01:39,156

, I think the, the site first came on my

radar because of all of the research

01:01:39,156 --> 01:01:42,936

into attitudes towards tech and the

internet, um, because there's a, a ton

01:01:42,936 --> 01:01:44,436

of great research there about that.

01:01:44,826 --> 01:01:47,976

Um, so if you wanna know about

AI adoption or crypto adoption or

01:01:47,976 --> 01:01:50,316

any of those things, pew is great

for that, as well as obviously the

01:01:50,316 --> 01:01:52,386

incredibly robust political coverage.

01:01:52,386 --> 01:01:55,056

So, uh, I thank you for your work

and, and thank you for coming

01:01:55,056 --> 01:01:56,136

on to spend the time with me.

01:01:56,136 --> 01:01:57,966

I really appreciate it, both

of you, Scott and Hannah.

01:01:58,731 --> 01:01:59,181

Scott Keeter: Thank you.

01:01:59,716 --> 01:01:59,956

Hannah Hartig: Thanks.

01:02:02,772 --> 01:02:05,262

Farrah Bostic: Crosstabs is a

production of the Difference Engine.

01:02:05,352 --> 01:02:06,882

It is edited and hosted by me.

01:02:06,942 --> 01:02:10,542

Farrah Bostick music is from

Audio Jungle by S Audio.

01:02:10,872 --> 01:02:15,432

You can subscribe to our weekly

newsletter for free@crosstabspodcast.com.

01:02:16,122 --> 01:02:19,692

You can also follow the show

on Blue sky@crosstabspod.blue

01:02:19,692 --> 01:02:20,592

sky.social,

01:02:20,652 --> 01:02:24,222

and on LinkedIn where we share links

to new episodes and newsletters.

01:02:24,702 --> 01:02:28,062

We also share these episodes via

video, and you can like and subscribe

01:02:28,062 --> 01:02:31,482

to each week's video episode on

YouTube at Crosstabs podcast.

01:02:32,022 --> 01:02:33,942

Please follow us on

your favorite platform.

01:02:33,942 --> 01:02:37,332

Tell your friends about the show, and

don't forget to subscribe on whatever your

01:02:37,332 --> 01:02:39,162

favorite podcast service happens to be.

01:02:39,612 --> 01:02:42,582

If you wanna learn more about what

I do, you can find me on all the

01:02:42,582 --> 01:02:45,972

socials at Fara Bostic, though I

am mostly on Blue Sky these days.

01:02:46,002 --> 01:02:48,582

Or get in touch through

the difference engine.co.

01:02:49,062 --> 01:02:49,722

And that's it.

01:02:49,812 --> 01:02:50,562

See you next time.