Race and gender affect who is considered “electable.”
In a recent experiment, I presented a nationally diverse sample of 1,947 American adults with profiles of generic, hypothetical candidates. (Though not a random sample, the implementing vendor, Lucid, constructed the subject pool to match the U.S. census on key characteristics.)
After viewing each profile, the subjects were asked, “If this candidate ran for governor in your state, how electable would [he/she] be?” In each profile, I randomly varied the race (black or white) and gender (male or female) of the candidates.
On average, the subjects rated the white male candidates “very electable” 37 percent of the time. Black men were rated “very electable” 35 percent of the time, white women 33 percent of the time and black women 30 percent of the time. As compared with white men, black women were about one-fifth less likely to be rated as “very electable.”
These results suggest that when pundits, donors and primary voters judge candidates based on their “electability,” they are using a biased metric that disadvantages women and people of color.
Race and gender affect which 2020 Democratic presidential candidates are believed to have the best chances of beating Donald Trump.
In a second experiment, I used Amazon’s Mechanical Turk (mTurk) to recruit 1,702 U.S. adults who said they voted against Trump in 2016, and who do not support his reelection in 2020. The subjects were not representative of the national Democratic primary electorate; most notably, they were younger than typical primary voters.
At the start of the experiment, the subjects saw photos of the top eight Democratic presidential candidates. Then they were randomly divided into three groups. Two groups saw different messages highlighting the need to appeal to either white or male voters. A control group did not see any message.
Subjects who read a message emphasizing the strategic importance of white voters were less than half as likely to say that a black candidate had the best chance of beating Trump. This effect was driven almost entirely by a large decline in Sen. Kamala D. Harris’s (Calif.) perceived competitiveness.
Similarly, when subjects read a message emphasizing the strategic importance of male voters, they were less than half as likely to say a female candidate had the best chance of beating Trump. Both Sens. Elizabeth Warren (Mass.) and Harris were significantly penalized in this scenario, while Biden and to a lesser extent former congressman Beto O’Rourke (Tex.) benefited.
To be perceived as more electable, candidates need to show they have a path to victory.
My research shows that strategic discrimination is a real problem. So how can candidates overcome it?
To find out, I used mTurk to recruit a fresh pool of 2,219 Americans who did not vote for Trump in 2016 and do not want to see him reelected. Then I repeated my second experiment, but with different messages. One message told subjects that “electability” is a biased concept that gives white male candidates an unfair advantage. That had no meaningful effect. Neither did a message stressing that most Americans are willing to vote for female and black candidates.
Here’s what worked: emphasizing black candidates’ strategic advantages. When subjects were told that high African American voter turnout is the key to Democratic victory in 2020, they were three times more likely to say that a black candidate had the best chance of beating Trump.
Of course, social science experiments are abstracted from the back and forth of real-world campaigning. But this experiment suggests that diverse candidates can boost their perceived electability by showing that they can win. And they have a strong case to make. After all, when the Democrats took back the House in 2018, more than 60 percent of the Democratic challengers who unseated GOP incumbents were women and people of color.
Regina Bateson (@regina_bateson) is a political scientist and former congressional candidate who will be a visiting professor at the University of Ottawa this year.