The most-cited statistic in popular discussions of women and negotiation comes from research by Linda Babcock and Sara Laschever, published in their 2003 book Women Don’t Ask: in an MBA graduating class, they found that male graduates were significantly more likely to have negotiated their starting salary than female graduates, and that those who negotiated received substantially higher salaries as a result. The conclusion drawn was that women should negotiate more.
This advice is not wrong. But it is incomplete in a way that has caused its fair share of frustration. The literature that has developed in the 20 years since is more nuanced, more complex, and ultimately more useful — because it acknowledges what women have always known: that negotiating more is not simply a matter of confidence, and that the social costs of negotiating are not imaginary.
The Backlash Problem
The evidence that women face social penalties for negotiating that men do not face is robust and has been replicated across multiple studies, methods, and cultural contexts.
Hannah Riley Bowles, Linda Babcock, and Lei Lai published a series of studies in 2007 that are among the most rigorous in this area. In experimental settings, they found that female job candidates who negotiated salary were rated as less likable and less hireable by evaluators — male and female — even when the negotiation behavior was identical to that of male candidates who were rated more positively. The penalty applied whether evaluators were men or women, ruling out the hypothesis that the problem is simply male evaluator bias.
The backlash effect operates through specific mechanisms. Negotiating — particularly for oneself, for higher pay, or for more resources — is culturally coded as self-interested, assertive, and communally indifferent. Women are expected to be communally oriented, relational, and other-serving. When they violate this expectation by negotiating on their own behalf, the violation is read as a character flaw. The same behavior in men is consistent with masculine role expectations, so no violation is registered.
This is not a small effect. Studies find that the backlash is significant enough that women’s rational assessment of the costs and benefits of negotiating — in many situations — should lead them to negotiate less than men with identical confidence and knowledge. The women who don’t negotiate are not necessarily making an error. They may be making an accurate calculation.
The Gender Pay Gap and Negotiation
The connection between the gender pay gap and women’s negotiation behavior is real but more complicated than the “women don’t ask” frame suggests.
Research by Claudia Goldin has established that the pay gap is primarily a gap between people who can offer flexible hours and those who cannot — a “flexibility penalty” that falls on women because of caregiving responsibilities. This is a structural explanation, not a behavioral one. Negotiating harder does not resolve a structural problem.
But there is also a behavioral component. Several studies have found that in contexts where salary ranges are not disclosed, women are more likely to anchor low — to accept the first offer, to ask for less than the maximum available, to assume that what is offered is what is possible. In contexts where salary ranges are disclosed (as more companies are now required to do in some jurisdictions), the gap in negotiated outcomes narrows.
The practical implication of salary transparency research is significant. In US states and cities that have required salary range disclosure, early data suggest the gender pay gap narrows. This is not primarily because women suddenly become better negotiators; it is because transparent information enables more accurate anchoring, removing the information asymmetry that disadvantages negotiators who are less likely to have access to informal salary information through social networks.
What Works: Context-Contingent Strategies
The research on what negotiation strategies work for women has evolved considerably since the early “women should just ask” framing. The most developed framework is what Bowles and colleagues call “relational accounts” — framing negotiation requests in terms that are consistent with female communal role expectations rather than in the direct self-interested terms that trigger backlash.
The approach sounds like this: rather than “I want a higher salary because my market value is X,” the relational account might be “I want to make sure we start this relationship in the right place” or “I want to contribute at the level that matches the compensation.” The framing emphasizes relationship, mutual benefit, and the employer’s interests alongside the negotiator’s.
Research finds that this framing reduces the backlash penalty for women in many contexts. It is not eliminating an unfair system; it is working within it while the system is being changed.
Other strategies with research support include:
Negotiating on behalf of others. Research by Emily Amanatullah and Michael Morris found that women negotiate more effectively and with less backlash when they are negotiating on behalf of someone else (a team, a client, a family) than when negotiating for themselves. The communal framing is authentic in this context rather than strategic, which may explain its effectiveness.
Asking questions rather than making demands. Negotiation researchers have found that asking questions — “Is there flexibility on this?” “What do you typically see in this range?” — triggers less backlash than direct statements of what you want, because questions are relational rather than assertive.
Preparing thoroughly. The negotiators who achieve the best outcomes in research studies are those who come with market research — specific, documented evidence of what comparable roles pay. The information itself is protective: it frames the negotiation as reality-checking rather than self-interested demand.
Choosing contexts carefully. Not all negotiation contexts are equivalent in terms of backlash risk. Research suggests that context-specific norms — whether negotiating is expected and normal in a particular situation — affect how negotiating behavior is read. In contexts where negotiating is expected (explicit invitation to negotiate, explicit flexibility statements from employers), the backlash is reduced.
The Pipeline of Negotiation Skill
There is a genuine developmental dimension to negotiation that is less discussed than the backlash problem. Negotiation is a skill. Like most skills, it improves with practice. And women, on average, have had fewer practice opportunities throughout their lives — fewer low-stakes negotiations from which to build skill and calibrate strategy.
Research by Linda Babcock and colleagues on girls’ negotiation development found that adolescent girls receive less encouragement than boys to advocate for themselves, are rewarded less for direct assertion in competitive contexts, and enter adulthood with less negotiation experience. This is not simply a confidence gap; it is a skill gap that reflects different developmental histories.
This has implications for how negotiation training is most usefully structured. Training that addresses only mindset (“believe in your worth,” “be confident”) without addressing skill-building and the specific social context of female negotiation tends to produce disappointment when women attempt to apply it and encounter the backlash that the training didn’t prepare them for. The most effective training programs include practice, feedback, and explicit attention to context-specific strategy.
Negotiating Beyond Salary
Salary negotiation receives the most attention, but it is one of many forms of negotiation women engage in that have significant outcomes.
Negotiating for flexible work arrangements. Negotiating for access to high-visibility projects. Negotiating for training and development resources. Negotiating for team size and budget. Negotiating in relationships — for equitable distribution of domestic labor, for financial arrangements that protect individual interests within partnerships. Each of these negotiations has its own context-specific dynamics.
Research on domestic labor negotiation — the negotiation within relationships about who does what — finds that the backlash pattern applies here too. Women who negotiate explicitly for more equitable distribution of household labor face social costs (being perceived as demanding, difficult, or unloving) that are specific to gender role violation. The result is that many women avoid the negotiation, and the unequal distribution of domestic labor persists.
This matters economically because domestic labor reduces women’s paid work capacity, driving the caregiving penalty that is one of the primary drivers of the gender pay and wealth gaps. The failure to negotiate domestically feeds into the financial outcomes of paid work.
The System Problem
The individual-focused frame of “women should negotiate better” has a significant limitation: it locates the problem in individual women’s behavior and ignores the system that produces the backlash, the information asymmetry, the skill development gap, and the structural conditions that make negotiating more costly for women.
This limitation matters for policy. Salary transparency legislation, which is spreading across US states and increasingly in Europe, addresses the information asymmetry directly. Policies that increase caregiving support reduce the flexibility penalty. Organizational norms that formalize negotiation processes — making clear when and how negotiation is expected — reduce the backlash because negotiation is no longer a gender-role violation but a role-consistent behavior.
Individual women navigating the current system should know the research — both what the research says about negotiating more, and what it says about the costs. They should prepare thoroughly, choose framing strategically, and not blame themselves when the system produces unfair outcomes for the same behavior in different bodies.
And they should support the structural changes that would make the individual-level navigation less necessary.
Related reading on Vanity-X: