Showing posts with label evaluation bias. Show all posts
Showing posts with label evaluation bias. Show all posts

Monday, April 21, 2014

Should women be encouraged to compete?

An article on “The Confidence Gap” between men and women has been making the rounds this week. In it, Katty Kay and Claire Shipman argue that women too often doubt their own abilities as compared to men, leading to an imbalance in women’s representation in most fields. This may have to do with socialization, evolutionary biology, the structure of our social systems, or some combination.

In his seminar on whether “boys and girls respond differently to academic competition,” Prof. Robert Jensen of the University of Pennsylvania explored how this carries over into the realm of competitiveness. He and his co-authors used a real-life experiment in which a math and verbal prep technology suddenly introduced a peer competition in the form of a “leader board.” Prior to the leader board, students would simply answer a series of questions and be told, individually, how well they’d done. After the “leader board”, students were given points for correct answers and the names of the top-three point-earners were displayed for all the participants to see.
Before the points system, girls tended to perform better in both English and math. But after the introduction of the competitive system, girls performed worse than they previously had, and also worse than boys, particularly in math.

Whether it had to do with social stigma of being publicly seen as a “nerd” or just the aversion to and stress associated with competition is unclear. But Professor Jensen concludes that a competitive system simply wasn’t conducive to better learning outcomes for women in this education technology.

So should we reduce competition in how we raise and educate girls? As one seminar participant remarked, “we live in a society of competition in every sphere; to discourage that is to encourage girls to opt-out of success. Instead, perhaps we should raise our daughters and sons the same way so that they can both learn to compete effectively.” Indeed, as Elizabeth Plank writes, instead of telling women to change their personalities, maybe it's time we take a look at the entire system and adjust all of the structures that hold them back.

To this, WAPPP Executive Director Victoria Budson responds that, “Whenever the frame and context for any competition is set in today’s world, it will necessarily be biased---by gendered components, racial components. So we need to understand  what choices are made and how those choices impact outcomes. It’s not that one shouldn’t compete…but to create a new competitive frame.

“When you understand what the mechanisms are and what they produce, you can then guide how institutions create structures. Because whenever we set up structures, we’re really creating pathways toward outcomes that we can predict when we study them effectively. So rather than telling us how we should feel about this, all of these studies are just data that can help us create a world where the majority of our talent is effectively utilized.”


WAPPP Director Iris Bohnet adds that we should do both: “we should enable people to be competitive in the world that we live in, but we also have to change the world to make it easier for everyone, based on whatever preferences they have, to survive and compete in that world.”

Monday, April 14, 2014

Are Women Punished for Seeking Power?

One of the catch-22s of gender relations these days is that women are hemmed by both realistic power structures that do exist, as well as by perceptions of what ‘should’ exist.

Specifically regarding gender stereotypes, many people expect not only that women are more modest in their presentation and interactions, but that they should be more modest.

So what happens when women violate these stereotypes?

That was the question that Professor Victoria Briscoll of Yale University posed in her seminar on “Women and Power: Hard to Earn, Difficult to Signal, and Easy to Lose.” She broke her answer into three parts.

First, women often have to manage people’s impressions of their rise to power. Their intention of seeking power and authority appear inconsistent with people’s perceptions that women should be communal and not dominating. So even female politicians like House Speaker Nancy Pelosi and Senator Patty Murray, who are essentially in the business of power, often downplay the fact that they are there, insisting that they “never expected to run for office.”

Second, once in power, men and women often communicate differently to continue this impression management. According to a great deal of social psychological research, ‘powerful’ people are often given a license to talk more than people with less power, who signal deference. Moreover, women tend to lead in more democratic, non-hierarchical fashions than men. So in spaces like the US Senate floor, men talk to display power, while women tend to talk to establish and maintain relationships and advocate for communal rather than personal causes. This is often in the effort to avoid backlash.



Finally, women’s power is often more fragile and easily lost than that of men. In the case of expressing anger, women are almost always penalized for this, while angry white men are sometimes rewarded for being assertive. But when women can explain their anger away to an external source, women are rewarded.

So clearly there’s a lot of work for society to do. To get there, do women need to keep on adjusting what they do? How can we get societal expectations to change in the long run?

Photo Source



Thursday, March 20, 2014

How can mothers win? Ending the bias against caretakers.

As we’ve seen, both women and men face challenges reconciling work-life integrity. Usually, individuals are forced to adjust their own behavior: share household burdens, manipulate supervisors’ perceptions, or just opt out entirely.

What makes this whole balancing process harder is that caretakers suffer a particular bias at work. In her seminar last week on “Reducing the Caretaker Penalty: Norms, Laws, and Organizational Policy,” Stanford University’s Shelley Correll demonstrated how mothers, for example, are paid less than both fathers and childless women---nearly 5-7% less per child, in fact. But it’s difficult to overcome this because of two absurd and paradoxical societal perceptions:
  • The assumption that mothers are less committed to their office work
  • The normative view that mothers should be more committed to their children than their office work-and if they’re not, they’re bad people.
If mothers try to work their way out of the caretaker bias, they’re seen as selfish, arrogant, and dominant, and are penalized accordingly. Basically, mothers can’t win!


So can laws change societal norms? Professor Correll explains how, yes, they can not only provide punitive protections but also create more symbolic social consensus that implies what’s right and wrong.

The Family and Medical Leave Act (FMLA) gives employees that have worked for over twelve months in an organization with more than 50 people the right to 12 weeks of unpaid leave per year. Just knowing about FMLA and other organizational leave policies positively affects how colleagues view mothers and other caretakers who take short leaves. Even a limited law that’s weakly enforced can promote gender equity.

If we enable workplace leave more reflexively, we can prep society for more openness, namely to the reality that all people---men, women, rich, poor---have and must honor responsibilities outside of the workplace. As Professor Correll put it, “work should be a verb, not a place,” and Best Buy’s management has pioneered Results Only Work Environments (ROWE), where employees are paid for results and output rather than the number of hours worked.

Convincing employers and supervisors of the merits of this kind of flexibility might be more difficult, because they may have legitimate concerns about their workforce. But these employers ought to keep two things in mind. One is that their own expectations of different groups---women caretakers, African Americans, etc---are often incorrectly biased, and these are biases that impede a fair and efficient workplace. Another is that more flexible work environments will mean longer-term retention of good workers.

Low- and hourly-wage workers and their employers face another challenge: their work is inherently based on time commitment, and already feels risky. Even if they can afford it financially, these workers don’t want to take FMLA because it will prejudice their employers upon return.

Accordingly, good laws are even more important in these cases, setting the norm for what is right.





Monday, October 28, 2013

If you can’t change the player, change the game

How many times have you heard people try to deny that they are prejudiced by saying that “some of my best friends are” black, gay, Republican, or women? I’ll bet it’s all too often.

Try as we might, most of us have some kind of latent bias that keeps us from being completely objective, even when we really want to be. Of course, too many of us are also pretty openly prejudiced. Both of these views get in the way of two things: equity and efficiency.

By equity, I mean that women in particular are underrepresented in most places of power---corporate boards, elected office, etc---in proportion to their population. By efficiency, I mean that there are tangible economic benefits to including different people, and women in particular, in decision-making.

According to Professor Iris Bohnet, Director of WAPPP and Academic dean at the Kennedy School, who presented her work on “Gender and Decision-Making” this week, women’s smaller appetites for risk-taking, more tempered competitiveness, and complementary perspectives may account for these contributions.

So how do we overcome these unintended biases and maximize these benefits?

One answer may lie in understanding and fixing “evaluation bias”. We often assess things based on factors completely unrelated to the thing itself. Our evaluations are often driven by the context around the subject, or our other assumptions or preconceived notions.


The Chubb Illusion: We think the center squares are different colors because of their surroundings
http://www.optical-illusion-pictures.com/famous.html 

For example, in many Indian villages people didn’t typically associate ‘leadership’ with women largely because they had rarely seen female leaders. But in 1993, women were required by a constitutional amendment to serve as the head of village councils, or Panchayat Raj, in one-third of all Indian villages---a change that familiarized the electorate with female leadership, and paved the way for further elections of women. (In 2011, 50% of all seats had to be filled by women).

This legislative change was a “nudge” in the right direction. By changing the context a little bit, people’s perceptions changed a little bit too. So if we adjust the process of hiring in different firms so that, for instance, candidates are referred to as “people” rather than as a “man” or “woman,” decision-makers might not be primed by their gender-based association, and instead choose based on merits alone. Similarly, if a school has to hire 5 professors in a year, it makes sense to hire them all at once, so any gender imbalances are immediately obvious when they’re seen in a single group---rather than letting gender slip past the radar in one-off hirings.

To tweak a phrase used by the rapper Ice T, if we can’t change the player, change the game.