A substantial advice genre holds that women are held back by a confidence deficit, and that the remedy is individual behaviour change — speak up more, apply anyway, stop apologising, adopt certain postures.
Some of this is useful. A good deal of it rests on findings that are weaker than the confident retelling suggests, and the framing locates the problem in a way that's worth examining.
The frequently cited claims
The application statistic. The widely repeated claim that men apply for roles when they meet a minority of the criteria while women wait until they meet nearly all of them.
The origin of this figure has been traced to an internal report, and the underlying survey work suggested that the most common reason women gave for not applying was a belief that they wouldn't be hired because they didn't meet the stated requirements — which is a belief about the process rather than about their own ability.
That's a meaningfully different finding. If women are correctly perceiving that stated requirements are applied more strictly to them, the problem is the process, and advising them to apply anyway addresses the wrong thing.
Power posing. Widely promoted, and subsequent replication attempts have failed to reproduce the hormonal effects originally reported. One of the original authors has publicly expressed reservations about the findings. Effects on subjective feelings of power have replicated more consistently than the physiological claims.
The confidence gap generally. Research on self-assessment does find average differences in how confidently men and women rate their own abilities in some domains. What's less clear is that the difference is the binding constraint on outcomes.
The calibration point
A finding that complicates the standard advice considerably.
If men on average overestimate their abilities and women assess theirs more accurately, then "closing the confidence gap" means asking women to be less accurate.
That's an odd prescription. And there's a reasonable argument that accurate self-assessment is a desirable trait that organisations should reward rather than an obstacle to be trained out.
Which points at a different intervention: assessing candidates on demonstrable ability rather than on displayed confidence, since confidence is a poor proxy for competence and rewarding it selects for something other than capability.
The double bind
The more substantive problem with the advice, and it's well documented.
Research on evaluation has repeatedly found that assertive behaviour is evaluated differently depending on who displays it. Behaviour read as decisive in one person can be read as abrasive in another.
Studies examining negotiation have found that women initiating negotiation face social costs that men initiating identical negotiations do not, in some experimental settings.
Which means advice to be more assertive is advice to take a risk that carries a different cost profile depending on who follows it. That's not a reason not to do it, and it does mean the advice is incomplete without acknowledging the trade-off.
Where the framing goes wrong
The general problem: locating a structural issue in individual psychology.
If evaluation processes are biased, women behaving differently addresses the symptom. If the criteria for advancement reward a pattern of availability incompatible with caregiving, more confidence doesn't help.
There's also a cost to the framing. It implies that women who haven't advanced have failed to work on themselves sufficiently, which adds a personal deficiency to a structural obstacle.
And it's commercially convenient. Individual improvement can be sold as courses, coaching and books. Redesigning evaluation processes cannot.
What has better evidence
Interventions targeting process rather than individuals.
Structured interviews. Same questions, same order, scored against defined criteria. Substantially better predictive validity than unstructured interviews and less susceptible to bias.
Work sample tasks. Assessing what somebody can do rather than how they present.
Blind assessment where feasible. The frequently cited example of orchestral auditions behind screens is contested in its details and the principle has support elsewhere.
Clear promotion criteria. Ambiguous criteria leave room for evaluators to apply different standards, frequently without awareness.
Reviewing who gets stretch assignments. Advancement frequently depends on visible high-stakes work, and the allocation of that work is a decision that can be examined.
What individuals can still do
Not nothing, and framed as navigating a system rather than fixing yourself.
Keeping a record of what you've actually delivered, with figures. Useful in review conversations and useful against your own tendency to discount your contribution.
Understanding the criteria for advancement in your specific organisation, explicitly, by asking.
Building relationships with people who influence decisions, since sponsorship — someone advocating for you when you're not present — has more effect on advancement than mentorship does.
And applying for things you're not certain about, while knowing that this is playing a system rather than correcting a personal flaw.
That distinction matters. Adjusting your behaviour to navigate a biased process is a reasonable strategy. Believing that the bias was in you all along is not.