A standard piece of career advice holds that women earn less partly because they negotiate less, and that the remedy is to negotiate more.
The research underlying this is more nuanced than the advice, and the nuance matters for anyone deciding how to act on it.
The initiation finding
Early influential work reported substantial differences in how frequently men and women initiated negotiation over pay.
Subsequent research has complicated this. Meta-analyses have found the difference to be smaller than early work suggested, and heavily dependent on context.
Notably, the difference largely disappears when it's made explicit that pay is negotiable. Where ambiguity exists about whether negotiation is appropriate, differences appear; where the norm is stated, they diminish.
That's an important finding because it points at a fixable feature of the process rather than at a fixed characteristic of candidates.
The backlash finding
The complication that the standard advice tends to omit.
Experimental work has found that women who initiate negotiation can face social costs — evaluators rating them less favourably and expressing less willingness to work with them — that men initiating identical negotiations do not face to the same degree.
The proposed explanation involves prescriptive stereotypes: expectations about how people should behave, with assertive self-advocacy violating expectations for some groups more than others.
Effect sizes vary and the finding has been examined in multiple studies. The broad pattern has reasonable support.
Which produces a genuine bind. Not negotiating costs money. Negotiating may cost something else, and the cost isn't distributed evenly.
What appears to reduce the backlash
Research has examined framings that reduce the social cost.
Communal framing. Negotiating in terms that reference relationships and organisational interest rather than purely individual gain. Studies have found this reduces backlash while preserving negotiation outcomes.
Something like referencing what you've been told about how the organisation values contribution, rather than stating what you want.
Negotiating on behalf of others. Research has found that people negotiate more assertively and face less backlash when advocating for someone else. Framing a request in terms of a team or a role rather than personally can access some of this.
External referencing. Grounding the request in market data rather than personal assessment, which shifts the frame from self-promotion to information.
These are strategies for navigating a biased evaluation, and it's worth being clear that's what they are.
The organisational fix
The version that addresses the cause rather than the symptom.
Remove individual negotiation from pay setting. Structured pay bands with defined progression eliminate the variation that individual negotiation produces. Organisations that have done this report reduced gaps.
This is unpopular with people who negotiate well, which is a reasonable indicator of who the current system serves.
State that pay is negotiable, explicitly, to everyone. Given that ambiguity drives the initiation difference, removing ambiguity is cheap and effective.
Publish ranges. Required in some jurisdictions now. Reduces the information asymmetry that makes negotiation a test of confidence rather than of information.
Ban salary history questions. Prevents past gaps from propagating.
Train evaluators on the backlash effect. Awareness of a specific bias has better evidence than general unconscious bias training, which has produced disappointing results in evaluations.
Practical guidance anyway
Since most people are negotiating within existing systems rather than redesigning them.
Get information first. Salary data for the role, sector and location. Negotiating without information is negotiating against yourself.
Negotiate more than salary. Start date, title, review timing, flexible working, professional development budget, holiday. These are frequently easier to grant and can be worth substantial amounts.
Ask a question rather than making a demand. Asking about flexibility on the figure invites a response and is less confrontational than stating a number, though stating a number first has its own advantages through anchoring.
Have a specific figure ready. Vagueness invites a low offer.
Get it in writing. Verbal agreements about future reviews reliably evaporate.
Be willing to accept. Negotiation frequently ends with the original offer, and accepting graciously is not a failure.
The honest summary
Advice to negotiate more is not wrong. It's incomplete, in a way that matters.
It's advice to take a risk, where the risk is unevenly distributed, given to the people bearing more of it, without acknowledging that.
Knowing that changes how you approach it — as a calculated move with a technique attached, rather than as correcting a personal deficiency you were told you had.
The offer you should not accept immediately
One practical note that applies to everybody and that women in particular report finding difficult.
An immediate acceptance forfeits any negotiation, and the standard advice to take time before responding exists for a reason. A day to consider is entirely normal, is never held against anybody, and creates the space in which a counter can be made.
What generates the pressure to accept immediately is usually anxiety that the offer will be withdrawn, which essentially never happens. An employer who has decided to hire somebody does not un-decide because they asked for a day.