Lists of this kind tend to name qualities people already approve of. Warmth, attention, humour, remembering names. Nothing wrong with any of it, and nothing in it that anyone is avoiding.

Two lines of research point at a different category: things that raise how much people like or respect you, which most of us suppress precisely because we expect the opposite. The interesting part is not that they work. It is the size of the gap between how they feel from the inside and how they land from the outside.

The same act, rated twice

Anna Bruk and colleagues at the University of Mannheim published a set of seven studies in the Journal of Personality and Social Psychology in 2018, described in the British Psychological Society’s research digest. They gave the effect a name that has stuck: the beautiful mess.

The design puts the same event in front of two sets of eyes. In some conditions participants imagined themselves doing something exposing, such as confessing romantic feelings or admitting a mistake at work. In others they imagined watching someone else do exactly that.

The ratings diverged consistently. The act that looked like courage in another person looked like weakness when it was your own.

One study moved this out of pure imagination. Participants either expected to improvise a song in front of a jury, or sat on the jury watching someone else do it. The performers judged the exposure of performing more harshly than the people actually watching did.

Why the asymmetry happens

The explanation the researchers offer is about how near the thing is. When you consider your own vulnerability you think concretely: the specific words, the exact face you will make, what happens in the four seconds afterwards. When you consider someone else’s you think in broader terms, and the broader description is the flattering one. Honesty. Courage. Being real.

Same behaviour, two levels of resolution, two verdicts.

The limitation the authors acknowledge is that most of the scenarios were hypothetical, which is a real constraint. Imagining you would admire a colleague’s admission is not the same as admiring it on a Tuesday afternoon when it inconveniences you.

Asking for help, and the three conditions attached to it

The second finding is more specific and comes with better instructions. Alison Wood Brooks, Francesca Gino and Maurice Schweitzer published a set of studies in Management Science in 2015, running eight studies with roughly 1,355 participants.

Their pilot established the belief. Asked whether seeking advice would make them look competent, people expected it to cost them, rating the advice-seeker at 4.50 against 5.69 for someone who did not ask.

The main studies asked the other party. People who had actually been asked for advice rated the person who asked as more competent, not less.

Then the paper does something that most findings of this shape never do, which is to specify when it stops being true.

It only held for difficult tasks. On difficult ones, advice-seekers were rated higher, 3.57 against 3.28. On easy tasks the difference vanished entirely. Asking someone how to do something straightforward gains you nothing, which is roughly what everyone suspected.

It only held when the advice was sought from that person directly. Participants who learned someone had asked a third party showed no such boost, 4.08 against 4.83 for a personal request. The effect appears to run partly on being flattered by the asking.

And it reversed when the request landed outside the adviser’s competence. Asking within someone’s area of strength raised ratings to 5.32. Asking about something they were weak at lowered ratings below asking nothing at all, 4.53 against 4.97.

That last one is the practical sentence in the paper. Asking the wrong person is worse than not asking.

What the two findings share

Both describe a quality that is mispriced from the inside, and both explain why the mispricing survives. You never get to see how your own openness looked to the room, and you rarely learn that the colleague you asked went away feeling useful rather than superior. The feedback that would correct the estimate does not arrive.

What follows from that is narrower than a list of qualities. It is closer to a correction factor: the specific behaviours you are avoiding on the grounds that they will make you look weak are the ones where your estimate has a known bias, and the bias runs in one direction.

It is not an argument for volunteering every insecurity, and the advice research in particular sets its own limits. Difficult problem, the right person, their actual area of expertise. Three conditions, all checkable before you speak.

Where both stop

Neither study measured liking over time, which is what the original question is really about. They measured impressions formed in specific situations, mostly through rating scales, often in scenarios participants imagined rather than lived.

The samples are the usual mix of online participants and students in Western countries, and the advice studies were run in a work context, where competence is the currency being judged. How the same behaviour reads at a dinner table is not established by any of it.

There is also a difference between being rated well on a scale and being someone people seek out for years, and no study of this kind reaches the second thing.

What survives is the asymmetry itself, which is unusually well documented for something this counterintuitive. The behaviours that feel like they will cost you standing are, in a specific and bounded set of circumstances, the ones the other person is quietly upgrading you for.