A pair of speakers, two years old, is up for sale online. When no music is playing, they give off an annoying static noise, and the listing does not say so.
A buyer’s message ends with a pointed question: “The speakers don’t have any problems, do they?” In one of the experiments built around that exact question, 85 percent of sellers mentioned the noise, and fewer buyers than that expected their seller to mention a problem.
Shira Garber and three colleagues staged a string of these secondhand sales online across five main studies and a pilot, added a second pilot that surveyed people about what drives negotiators, and reported the results in an open-access paper in Personality and Social Psychology Bulletin this July. Buyers, they found, often underestimated how honest sellers would be. Sellers also reported more concern with being honest, fair and kind, and less concern with the money, than buyers credited them with.
How often did sellers come clean?
In each of the five main studies, some participants played sellers of a secondhand device with a defect they had been told about: an iPod whose battery drained fast, an electric scooter with a faulty battery, the speakers with the static. The rest played buyers, who could earn bonuses for accurately predicting how honest sellers would be.
In the first four studies, a seller counted as honest only if a reply named the defect. Staying quiet counted as dishonest, and so did a dodge like a battery that “isn’t brand new.”
The first pilot set the pattern. Of 76 online workers playing sellers, 63.2 percent picked a truthful opening message over one describing the iPod as being in excellent condition. Of 85 observers given the same information, 43.5 percent predicted the truthful one.
In Study 1, the 176 sellers wrote their own replies to “What can you tell me about it?” and “Anything else?” Of those sellers, 51.7 percent disclosed the battery problem, while the 270 buyers guessed, on average, that 43.2 percent would. With the scooter and a direct question in Study 2, three quarters of sellers disclosed the fault.
The rate swung with the setup. One cause, a direct question, got its own test; other swings the researchers could not fully explain. Until the last study, the exchanges were also largely scripted, with sellers answering pre-written buyer messages rather than a live counterpart.
The way the question is asked
Study 3, with 483 participants, gave half the sellers the pointed “The speakers don’t have any problems, do they?” and half the open “What can you tell me about them?” Buyers here were not told about the static, which is closer to how a real listing works.
The direct question made sellers more likely to mention the noise. Buyers’ predictions did not move with it.
The gap between expected and actual honesty was statistically significant under the direct question and not under the open one, so even inside this paper it did not show up everywhere. Of several possible reasons Study 1 still found a gap with open questions, the authors single out one: sellers there had two chances to come clean, and concealing twice may weigh more on a seller’s conscience than concealing once.
A few minutes of small talk
Study 5 was the only one that paired people live. It matched 422 participants into 211 buyer-seller pairs on an online platform, and some pairs first spent three minutes chatting about something they had enjoyed recently or how they unwind after a busy day.
Across the study, buyers rated their seller’s likely honesty at 6.56 on a ten-point scale, while coders scored the sellers’ actual replies at 7.51. Among pairs who chatted first, that difference was no longer statistically significant. Sellers’ scores did not differ significantly between the two conditions. The chat raised buyers’ expectations of honesty instead.
Study 4 tried a different nudge on 358 participants. Some buyers were asked about the seller’s moral concerns before predicting, others afterward. Thinking about the seller’s conscience first raised buyers’ expectations, and the gap was nearly halved.
It did not close. The authors suggest one reason: 85 percent of sellers in that study disclosed, a rate higher than in the matching condition of Study 3 for reasons they say are not readily apparent.
Where the suspicion comes from
Study 2 measured three motives, adding the pull of money to the two that Study 1 had tracked. The 214 management undergraduates who took part rated, on a seven-point scale, how much the seller had been weighing being decent, feeling guilty, and getting the best price. Sellers put their concern with being a decent person at 4.99, and buyers guessed 3.78. Buyers put the seller’s interest in money at 6.03; sellers reported 5.06.
The sellers still cared about the money. They cared about it less, and about their own decency more, than the buyers pictured.
The paper’s explanation leans on moral typecasting, the idea that people tend to see others as either doers or feelers, rarely both. A seller holds the object and the information, so, the authors argue, a buyer casts them as a doer, and the competitive setting makes that doer look like someone chasing an advantage rather than guarding a clean conscience. In the abstract’s words: “While sellers tended to be honest because of moral identity concerns, buyers expected them to be driven primarily by self-interest.”
Two duller explanations get a hearing too, and the authors say their own mechanism is not the only one. Poor forecasting in general, they point out, gives no reason to expect guesses to run low rather than high. Plain caution, a buyer bracing for the worst to avoid a loss, gets mixed treatment: the Study 4 discussion offers it as a possible added force, while the general discussion says it cannot explain the results, because in the first pilot observers with nothing at stake still underestimated the sellers.
Their mediation analyses deserve a lighter touch than the abstract’s sentence suggests. Buyers and sellers answered parallel versions of the motive questions rather than the same ones, and the authors say the results should be “cautiously interpreted” for that reason. Sellers rated their motives after choosing whether to disclose, which leaves room for a flattering story told afterward, though the authors argue buyers’ parallel ratings may reduce that concern. And in Study 1, the measure of anticipated guilt and related feelings showed no significant difference between buyers and sellers.
Small stakes, narrow scope
The stakes stayed small. Most bonuses were paid by lottery, and a seller in Study 1 had a five percent chance of earning at most $7. The authors grant that low stakes may keep self-interest from fully switching on, while citing earlier work that finds self-interest at work even when stakes are small.
The honesty studied was one kind: telling a buyer about a defect the seller knew. Misrepresenting a bottom price or inventing urgency, both named by the authors, were outside the design. They describe the findings as “based primarily on Western participant samples,” and the sales were one-off exchanges between strangers, chosen to keep reputation largely out of play. Ongoing work relationships and repeat customers sit outside what was measured.
Suspicion is sometimes the sensible response, and the paper says so. A buyer facing a salesperson expected to use persuasion tactics, or shopping where one side knows far more than the other, as with used cars, has good reason for care. The paper does not argue for trusting everyone. It asks for work on when doubting a counterpart’s honesty hurts a negotiator and when it protects one.
Every finding in this piece comes from the published text and tables of one paper, read by an editorial desk that reports on research and does not train negotiators; the supplementary files the paper points to could not be opened. The studies compared predictions with replies across hundreds of people, which says something about buyers and sellers like the ones studied, taken as groups, and cannot tell anyone whether the particular seller in their inbox is being straight with them. For a purchase that matters, sensible safeguards include an independent inspection, a written description of the item’s condition, and the buyer protection some platforms offer.
This paper did not measure what the gap costs. The authors’ illustration is a used car that looks too good to be true, and a buyer who distrusts the seller, walks away, and settles for something worse. Examining what the honesty gap does to the outcomes of negotiations is what they call an important next step; they also call for testing the idea where far more money is at stake.
