December 28, 2015--Social-Desirability Bias

Over coffee at Balthazar, a friend looked around furtively, leaned close, and whispered, "You promise not to tell anyone?"

"Of course. Always. About what?"

"I can't believe I'm about to say this." She was unusually agitated.

I wondered what might be on her mind. "It's OK," I tried to assure her.

"But I think I'm going to vote for," she lowered her voice still further, "Him."

Now I knew where this was headed. It was not the first time I heard similar things from some of my most liberal friends. Including, most surprisingly, middle-age women.

"Trump."

"This surprises me, but . . ."

"Me too. I'm the most surprised of all. I should hate him. And in many ways I do. I'm a women, a feminist, and Hillary has the best chance to become the first female president during whatever is left of my lifetime. But . . ."

She trailed off, looking blankly across the banquette where we had so often met for breakfast.

"To balance things a bit," I said, "I know lifelong Republicans of the old, now-almost-obsolete moderate kind--the ones who voted for Javits and Eisenhower--who tell me that if he is nominated they're going to vote for Hillary. They can't stand her but hate her less than him."

"That's no comfort to me."

"So why are you thinking about abandoning Clinton?"

"I'm not thinking of it that way. So here's the thing I want to confess and ask you to not pass along to anyone we know."

"Again, I promise."

"I think he'd make a better president." She said that so softly I could barely follow her.

"I've heard this from others who one would think would be enthusiastic about Hillary."

"I don't know anyone who feels enthusiastic about her," she said, "And for the most part they're lifelong Democrats, always vote that way, and though they too don't like her that much or think she'd make a very good president--too much personal and ideological baggage--they're resigned to vote for her. Above all, to tell the truth, because she's a woman. That should be enough for me too, especially the woman part, but it isn't. Like I just confessed, I think he'd . . ."

It was as if she couldn't utter those words again.

Trying to shift the subject somewhat from what was clearly painful for her, I asked, "Did you see the piece the other day on the Internet about social-desirability bias?"

"Not really." She sighed, collapsing in her seat, but seemed pleased that I had changed the subject.

"You remember in the old days when they gathered TV ratings by asking people directly what they watched or asked their sample viewers to record their viewing habits in a diary? They still use diaries to a certain extent, but more-and-more they're doing it directly via electronic boxes attached to people's sets that automatically record what's being watched. No way too lie that way."

"I remember that. I remember being called once or twice and being interviewed."

"Well, the problem with gathering data that way was it was easy not to tell the truth. For example, Public Broadcasting always was over-represented because some people didn't want to fess up that they were really not watching Alistair Cooke but Uncle Miltie They didn't want to appear to be lowbrow."

"This is interesting--sort of--but what does it have to do with what we were talking about?"

"The same thing appears to be happening now in the political campaigns. Especially the Republican one. At the moment, he is leading in the polls with about 34 percent or so saying they support him."

So?"

"So a research organization called Morning Consult just did an interesting triple-blind experiment. They surveyed voters three ways--the traditional way by interviewing people over the phone, another sample group via interactive dialing where potential GOP voters were surveyed via an automated program, and  a statistically-equivalent group on line where those being questioned were not asked to identify or in other ways describe themselves."

"And," my friend said, "they got the Masterpiece Theater response."

"Now I'm not following you."

"The percentage of the people who plan to vote for him was higher when people were surveyed anonymously and lowest when interviewed by a live person."

"Correct. Across the board by as much as 6 percent. But he did much better among college-educated Republicans with whom it is said he is not polling well. From them he got 9 percent more confessing that they, like you, plan to vote for him."

"Wow. But I need to correct one thing--I'm not planning to do so but, in the way pollsters describe people, I'm leaning that way."

"Gotcha.

"How do they explain this disparity--between what people tell interviewers and those who respond on line?"

"They refer to it as social-desirability bias, the tendency of people to hesitate to confess unpopular views to a pollster. If, I may say so, sort of like you."

"Interesting. But it is still making me crazy that after all the stupid and offensive and bigoted things he's said, at the moment I'm even leaning . . ." She trailed off again.

"There are clearly quite a few people like you."

"As I said, I think he'll do a better job. Like with Russia. She, when Secretary of State, hit the reset button and what do we have? Cold War II. I think we need a deal maker. I hate to say this--a larger-than-life figure for our larger-than-life problems. A . . ."

"I think I understand."

"I wish I did," my friend said, shrugging and finally smiling.

"Have you talked to your therapist about this?"

"I plan to. But in the meantime do you think it's too early to move on from coffee to Jack Daniels?"

Uncle Miltie