The projection to nearly $800,000 is worth taking apart, because it is a model rather than an observation. A flat annual advantage of $6,450 carried across four decades sums to about $258,000, so reaching the headline figure requires the initial gap to widen by roughly a factor of three over a career [2]. That is what percentage raises applied to a higher starting base will do, and it is a defensible assumption about how salaries actually move. Nobody followed these graduates for forty years [7]. The account also does not report what base salary the negotiated increment sat on top of [14], which means it cannot be expressed as a share of pay, and a share of pay is the form you would need to compare it with anything.
The five designs do different jobs, and the smallest-stakes one does the most useful job. In the lab study, every participant was told the bonus ran between 50 cents and $3, and every participant was offered 50 cents [8]. Information about the range was handed over for free, and the entire upside of asking was $2.50 [3]. Higher-income participants still asked more often. That weakens the most comfortable explanation on offer, which is that lower-SES workers stay quiet because they do not know what the number could be. Asked directly, respondents gave fear of backlash and a reduced sense of power [9], and the graduates' own phrasing was about not wanting to damage a new relationship [5].
The fifth study is the one the reframing rests on, and it is a vignette study. HR professionals read four written scenarios and produced ratings [10], not hiring decisions with a payroll consequence attached. The thing this does not tell you is whether any real lower-class negotiator was passed over. It does tell you that the fear reported in the survey has a counterpart in how evaluators describe the same behaviour. Phys.org's account does not say how class was conveyed to the raters [15], which matters, because in a live pipeline that cue has to come from something on the application itself.
The Australian data buys one replication outside the US, in a market Gelfand describes as culturally similar, and she says replication across diverse cultures is still missing [12]. Read strictly, this is a finding about the US and one country that resembles it, which is also where individual offer negotiation is most common.
Gelfand describes negotiation as learned rather than innate [13], which is an argument for teaching it. The awkward part sits in the HR ratings: teaching the ask, without changing how the ask is scored, increases the frequency of a behaviour that evaluators already discount more steeply for exactly the people being taught. The lever an employer controls outright is the offer process, and that lever is on the side of the table the coaching model never touches.
