On January 1, Virginia Rometty will become the first female CEO of International Business Machines Corp. Articles about her have lauded her ability to blend enthusiasm, charisma, clear communication, strategic thinking, and “cool-minded” decision making. But one New York Times story placed the emphasis on the role self-confidence may have played into her success:

Early in her career, Virginia M. Rometty, I.B.M.’s next chief executive, was offered a big job, but she felt she did not have enough experience. So she told the recruiter she needed time to think about it.

That night, her husband asked her,”‘Do you think a man would have ever answered that question that way?”

“What it taught me was you have to be very confident, even though you’re so self-critical inside about what it is you may or may not know,” she said at Fortune‘s Most Powerful Women Summit this month. “And that, to me, leads to taking risks.”

Self-confidence is one of the elements of active coping, a set of behaviors central to executive success that I’ve identified in my work assessing executives for senior leadership roles. In the first half of 1990s, I headed research at the University of Chicago Booth School of Business investigating the longer-term personality predictors of leadership. Among the relationships examined were those among gender, coping, and motivation in the evaluation of leadership effectiveness.

Among the particularly striking findings of this research were the differences between men and women on measures of active coping. If you are interested in the conceptual and empirical approach underlying the research, an academic article reporting the findings can be found here. In a nutshell, we took measures of coping, motivation, and intelligence at the beginning of the study. At the end of the study, we assessed the ability of these measures to predict leadership effectiveness as evaluated by peers, superiors, and subordinates.

We found that the only measure that predicted leadership for men and women alike was an overall measure of active coping that indicates the ability to respond adaptively to stress and to grow. But this measure was far more strongly associated with perceptions of women’s leadership effectiveness than it was for men’s. And when we broke down the elements of active coping, we found even stronger gender differences.

Category 2, a measure of the readiness to articulate sources of frustration and difficulties in terms of the external environment (as opposed to within the self) was positively and significantly correlated with leadership effectiveness for women but not men. Rejection 3, a measure of defensive vagueness and ambiguity, was negatively and significantly correlated with leadership effectiveness for women but not men. Category 4, a measure of self-confidence and self-esteem, was significantly correlated with leadership for women but not men.

Gender-based expectations for behavior influence the styles and evaluations of leaders. Women are expected to display high levels of social (communal) qualities, including needs for affiliation, a tendency to be self-sacrificing, concern with others, spontaneity, and emotional expressiveness. Men are expected to display high levels of agentic qualities, those associated with acting or exerting power, including independence, assertiveness, self-confidence, and instrumental competence.

Applied to leadership, gender role stereotypes suggest that female-stereotypical forms of leadership are interpersonally oriented and collaborative, whereas male-stereotypical forms of leadership are task oriented and dominating. To the extent that women who are leaders exhibit a masculine style, they amplify their role conflict and increase the chances of receiving unfairly negative evaluations.

Male leaders do not face a basic role conflict analogous to the conflict that female leaders face because expectations about behavior that is appropriate for a leader coincide largely with beliefs about the behavior that is appropriate for men. Men are freer to carry out leadership in a variety of styles without encountering negative reactions because their leadership is ordinarily perceived as legitimate. We expect men to display self-confidence, and true to form, men showed little variance in the measure of self-confidence. As a result, self-confidence did not predict judgments of men’s leadership.

For women, the ability to identify and face difficulties in the external world openly and non-defensively predicted leadership beyond any chance occurrence. The correlation between self-confidence and leadership effectiveness was also overwhelmingly statistically significant.

As a whole, these findings indicate that women have to have high self esteem and high self confidence while leading in a communal style in order to be perceived as effective leaders. In short, they must be stronger copers in order to transcend the constraints placed on their leadership style. Virginia Rometty appears to demonstrate these characteristics.