A new PNAS study explains why salary negotiation advice works for some people and backfires for others. The difference comes down to social class, not skill
Salary negotiation has been one of the most consistent pieces of career advice for the past two decades. Studies show that workers who negotiate their starting pay earn substantially more over the course of their careers than those who accept initial offers. The data is real. Negotiating works. The advice has been translated into books, podcasts, workshops, and corporate training programs. Just ask.
A new study published in PNAS, the Proceedings of the National Academy of Sciences, has examined a question that career advice literature has been slow to address: works for whom?
The research team, led by Ying Lin at the Hong Kong University of Science and Technology alongside Jackson G. Lu at Columbia Business School and Michele J. Gelfand at Stanford, ran five separate studies involving a combined total of thousands of participants across Australia, the United States, and controlled experimental settings. Their central finding is specific and uncomfortable: the professional backlash from salary negotiation falls unevenly depending on a worker’s social class background. The same negotiation behavior that advances an upper-class employee’s career damages a working-class employee’s professional reputation.
“There’s a large body of literature that documents a gender gap in negotiation,” Lin said. “However, we know very little about social class differences. Social class often remains invisible in organizations. At the same time, we know that negotiation has a long-lasting impact on economic outcomes and career advancement, which is important for upward mobility. So we wonder whether there’s a social-class difference in people’s propensity to initiate negotiations, which may contribute to the class ceiling phenomenon that prior research documents.”
What five studies found
The research built its case across five separate investigations, each targeting a different aspect of the negotiation gap.
The first used a nationally representative survey of 7,883 Australian employees, measuring social class through educational attainment and controlling for age, gender, hours worked, and employment characteristics. Higher-class individuals had 1.15 times greater odds of negotiating their starting pay. The failure to negotiate was associated with measurably lower hourly wages.
The second tracked 634 MBA students at a prominent US business school immediately after they accepted full-time job offers. Class was measured by the educational attainment of students’ parents. Lower-class backgrounds predicted a lower likelihood of negotiating starting salary, controlling for grades, age, gender, and race. Students who did not negotiate started with an average salary $6,450 lower than students who did.
The third was a controlled online experiment with 699 adults. Participants completed a short task and were told they would receive a 50-cent bonus, but were explicitly informed the payment was negotiable up to three dollars. Higher-class individuals had 1.65 times greater odds of explicitly requesting more money, a pattern that held even after controlling for task performance and participant confidence.
The fourth surveyed 995 full-time US workers selected to match the country’s income distribution. Higher social class predicted greater likelihood of negotiating across all work domains, including starting salary, benefits, promotions, and job content. Those who negotiated their starting salary secured an average gain of over $3,800 per year. Among participants who asked for a promotion, 95% received one. Among those who did not ask, 48% received one.
The fifth study is the most directly revealing. The researchers recruited 1,133 HR professionals and showed them a fictional male job candidate’s resume and interview transcript. The candidate was identical in every substantive way across conditions. The only variation was two sets of cues: whether the candidate’s background suggested lower or upper-class origins, signaled through resume details such as listing a first-generation student club versus a sailing club. The other variable was whether the candidate negotiated the salary offer or accepted it.
When the candidate negotiated, HR professionals across the board rated him as less cooperative. But this penalty was substantially larger for the lower-class candidate than for the upper-class candidate. The perception of being uncooperative reduced the likelihood that HR professionals would recommend hiring the candidate. The same words, the same request, the same behavior produced a different professional outcome based entirely on class signals.
Why working-class employees are more penalized for the same behavior
The mechanism the researchers identified is not about individual psychology in isolation. It is about social norms and the assumptions evaluators bring to a negotiation.
In higher-class environments, assertiveness and self-advocacy are treated as signs of confidence and competence. When an upper-class candidate negotiates, evaluators interpret it as appropriate behavior for someone of that profile. The negotiation fits the script.
In lower-class environments, social and economic survival has historically depended more on cooperation and collective norms. Asserting individual interests in a setting where cooperation is the cultural expectation can be read as a violation of that norm. When a lower-class candidate negotiates, evaluators interpret the same behavior through a different frame and flag it as a sign of being difficult or uncooperative.
This is the structural trap Lin calls a double bind. Staying silent perpetuates the financial disadvantage that negotiation is supposed to address. Speaking up exposes working-class employees to a professional penalty that upper-class employees doing the same thing do not face.
“If you’re a person from the lower class, it helps to recognize that hesitation to ask is not a personal or personality failing,” Lin said. “It’s completely normal and also a rational response to potential backlash. But it doesn’t mean don’t negotiate, as avoiding negotiation comes with a cost. It means recognizing that negotiation is normal, preparing for the difficulties, and it helps to frame negotiation in a more cooperative or communal term to help offset the social cost.”
The researchers found that framing a request using communal language, emphasizing shared goals and using we rather than I, reduced but did not eliminate the backlash penalty for lower-class candidates. The advice is useful but it does not solve the underlying problem. It places additional burden on the people already carrying more of it.
The class ceiling it produces
Economic inequality in the workplace has been extensively documented. Even controlling for education and credentials, people from lower social class backgrounds tend to earn less and advance more slowly than those from higher social class backgrounds. This pattern is sometimes called the class ceiling, an invisible barrier to upward mobility that operates inside organizations rather than through formal policy.
What this study contributes is a specific mechanism: the negotiation gap. The data show that working-class employees are less likely to initiate negotiations and more severely penalized when they do. These two facts compound across a career.
In study four, workers who negotiated their starting salary earned over $3,800 more per year than those who did not. Across a 30-year career, even without accounting for raises or promotions calculated from that higher base, the cumulative gap is substantial. And the study found that negotiating for a promotion is associated with a 95% success rate compared to 48% for those who stay silent. The employees who ask are the ones who advance. The employees who face the highest social cost for asking are disproportionately from working-class backgrounds.
The research also identified two psychological mechanisms that partially explain why working-class employees are less likely to initiate negotiations in the first place. The first is a diminished sense of power. Higher-class environments provide more material resources and more experience exercising control over outcomes, which builds a psychological orientation toward assertiveness. Lower-class environments produce the opposite orientation. The second is a heightened fear of social backlash, which the study shows is not irrational. The fear is accurate. The backlash is real and documented in the experimental data.
What organizations are not measuring
One of the study’s most pointed observations concerns what corporate diversity and equity efforts typically track and what they do not. Gender and race are increasingly measured in pay equity analyses, promotion rates, and hiring decisions. Social class background almost never appears in those analyses.
This means that the class ceiling operates largely unmeasured and therefore largely unaddressed. Organizations that audit hiring decisions for gender or racial disparities are not looking for the pattern this study identified: that identical negotiation behavior produces different evaluations depending on class signals, and that those evaluations feed into hiring recommendations.
“Starting to include social class in equity analyses is important. Social class mostly remains invisible,” Lin said. She recommended that organizations periodically check for discrepancies in negotiation patterns, hiring recommendations, and performance notes that flag candidates as not a team player, a phrase that the experimental data suggest is applied unevenly based on class background.
Lin also pointed to structural changes beyond individual coaching: reducing how much career outcomes depend on an employee’s willingness to initiate negotiations, making the rules and opportunities for negotiation explicit and transparent, and proactively distributing information about what is negotiable. The problem is not that working-class employees do not know how to negotiate. The problem is that the same negotiation gets evaluated differently depending on who is doing it.
“The employees who actively ask for more compensation are not always the ones who are most deserving of it,” Lin said.
What the study cannot establish
The research has real limitations. The experimental findings in study five focused on one professional role, one type of compensation request, and US-based evaluators. Whether the pattern holds across different countries, industries, genders of both the candidate and the evaluator, and types of negotiation remains to be tested.
The studies also focused exclusively on the decision to initiate a negotiation, not the strategies people use during the negotiation itself or how evaluators respond to different approaches within a negotiation. The backlash finding may differ depending on how the request is framed, how much is asked for, and the specific relationship between the employee and evaluator.
And while the research documents a clear social class gap across five studies with thousands of participants, the mechanisms proposed, sense of power and fear of backlash, explain the pattern only partially. Other factors the study did not measure may also be driving the gap.
What the study establishes clearly is that the standard advice to negotiate your salary is not wrong. It is incomplete. It describes a strategy that produces reliable returns for one group of employees and reliable costs for another, and treats those two groups as if the same advice applies equally to both.
The study, “The social-class gap in negotiation: Lower-class individuals negotiate less and face more backlash”, was authored by Ying Lin, Jackson G. Lu, and Michele J. Gelfand, and published in PNAS in August 2026.
Source: Hong Kong University of Science and Technology / Columbia Business School / Stanford University. DOI: 10.1073/pnas.2602128123
