RESEARCH_FILE

The Science of Friend Boundaries

The belief that a limit will end a friendship is one of the most-studied prediction errors in social psychology.

SEE THE PRACTICE

Turn a thought this research explains into one clear move.

THE THOUGHT

“Pointing it out would be rude”

YOUR RECORDED RESPONSE

“I don’t need advice on this.”

ONE PRIVATE MOVE

Open your notes app or a blank card and write one boundary sentence you could use with a friend. Keep it private, reread it once, then close it without sending anything.

Assertiveness training entered clinical psychology in 1949 and accumulated decades of supportive trials before being quietly forgotten. Then compliance research found the mirror image: askers underestimate by as much as 50% how likely you are to say yes — because they can't feel how socially expensive refusing is. Money between friends, one-sided venting, and the wording of a no have all since gotten their own data. Here's what the research actually shows about drawing a line without losing the friend.

up to 50%is how much askers underestimated the odds that others would say yes to a direct request in Flynn and Lake's experiments — meaning your friends likely find it far harder to refuse you than you think, and vice versa 1949the year assertiveness training entered clinical psychology — a treatment with decades of supportive trials that a 2018 review had to call 'forgotten' because research and training had quietly dropped it 6 studiesshowed that friends who lend money feel entitled to oversee how it's spent — and stay angry about hedonic purchases even after the loan is fully repaid 7 experimentsdemonstrated that people expect far more spare time in the future than they actually get — which is why you say yes to next month's favor you would refuse this week, then feel trapped when it arrives

How the science changed

1949
Andrew Salter's Conditioned Reflex Therapy introduces training in direct emotional expression — the root of assertiveness training. Refusing, disagreeing and asking are framed as skills that can be practiced, not personality traits. ↗
1973
Spencer Rathus publishes the 30-item Rathus Assertiveness Schedule in Behavior Therapy, giving the field a validated measure — and helping make assertiveness one of the most-studied clinical topics of the 1970s. ↗
2008
Flynn and Lake (now Bohns) show in JPSP that help seekers underestimate by as much as 50% how likely others are to say yes to a direct request — because askers fixate on the costs of helping and miss how socially costly saying no feels to the person asked. ↗
2012
Patrick and Hagtvedt show in the Journal of Consumer Research that the wording of a refusal matters: framing it as 'I don't' (a standing personal policy) rather than 'I can't' (a negotiable excuse) makes the no more persistent and empowering. ↗
2016
Bohns reviews over a decade of studies and consolidates the underestimation-of-compliance effect: people agree to requests in large part because refusing feels awkward and costly in the moment — and askers systematically fail to see that pressure. ↗
2018
Speed, Goldstein and Goldfried call assertiveness training a 'forgotten evidence-based treatment': despite decades of trials showing it reduces distress and tends to improve — not damage — relationships, it had nearly vanished from research and training. ↗
2025
Boundary questions between friends get their own literatures: Angulo, Goldstein and Norton show across six studies that lenders feel entitled to oversight of what friends do with loaned money — and stay angry about 'fun' purchases even after full repayment — while nursing research documents the cumulative toll of one-sided 'trauma dumping' on listeners. ↗

What people believe vs. what the data shows

The belief“A real friend never says no.”

The dataThe fear behind this rule is a measured prediction error: in five experiments with over 2,000 participants, people who declined invitations substantially overestimated how angry the other person would be and how much the relationship would suffer — actual inviters reacted far less negatively than decliners predicted. ↗

The belief“Setting a boundary means being rude or aggressive.”

The dataAssertiveness is defined in the clinical literature as distinct from aggression — stating your own needs without attacking the other person's. Decades of trials reviewed in 2018 found assertiveness training reduces distress and tends to improve interpersonal functioning, not damage it. ↗

The belief“If my asking were really a problem, they'd just tell me no.”

The dataA decade of compliance studies shows the opposite: people say yes in large part because refusing feels awkward and socially costly in the moment, and askers systematically underestimate that pressure. Silence or a yes is not proof your request was easy for them. ↗

The belief“Lending a friend money is just what friends do — no strings attached.”

The dataAcross six studies, lenders felt the money was still partly theirs and believed they deserved oversight of how it was spent — and stayed angry at borrowers who made 'fun' purchases even after being fully repaid. An unstated money boundary is a string; naming expectations up front is what removes it. ↗

The belief“Being available for unlimited venting is what real closeness means.”

The dataResearch on 'trauma dumping' describes a cumulative toll on the listener when heavy disclosure is one-sided, unsolicited and unbounded. Mutual support has limits on both sides; a listening boundary protects the friendship's capacity to keep supporting at all. ↗

TEST_YOURSELF · How well do you know this science?

01 In Flynn and Lake's (2008) studies, by roughly how much did people underestimate others' willingness to say yes to a direct request?
Across experimental and field studies, help seekers underestimated by as much as 50% the likelihood that others would agree — because they focused on the costs of helping while missing the social cost of saying no. The same blind spot means your friends underestimate how hard their asks are for you. source ↗
02 According to Patrick and Hagtvedt (2012), which refusal wording proved more effective at making a no stick?
Framing refusal as 'I don't' invokes a stable personal policy and felt empowering, making the refusal more persistent, while 'I can't' signals a negotiable external constraint that invites pushback. The wording of a boundary is part of the boundary. source ↗
03 Why did Speed, Goldstein and Goldfried (2018) call assertiveness training a 'forgotten' treatment?
Their review documents that assertiveness training — dating to Salter (1949) and heavily studied through the 1970s and 80s — reduces distress and improves interpersonal functioning, yet had largely disappeared from the research literature and clinical training despite that evidence base. source ↗
04 In Angulo, Goldstein and Norton's studies of lending between friends, what kept lenders angry even after they were fully repaid?
Across six studies, lenders were angrier when borrowers made hedonic rather than utilitarian purchases with loaned money — and the anger persisted after full repayment. The mechanism is 'deserved oversight': lenders feel the money is still theirs and that they deserve a say in how it's used. source ↗
05 According to Zauberman and Lynch (2005), why do people agree to future commitments they would refuse if asked for this week?
Across seven experiments, perceived 'slack' — spare resource — was expected to be much greater in the future for time than for money, and this illusion of future free time drove people to take on delayed commitments. It's a core engine of friend overcommitment: the yes is cheap today and expensive when the date arrives. source ↗