Deciding for someone else produces a different answer from deciding for yourself

Two answers to one question

Ask someone what a friend in a particular situation should do and you will usually get a clear answer within a minute. Put the same person in that situation and the clarity evaporates, replaced by a much longer list of considerations, most of which concern how difficult the obvious course would be to carry out. Both answers come from the same person, on the same day, about the same facts.

It is tempting to treat one of these as the correct one, and most popular advice treats the distant view as wiser. That is half right at best. The two views differ in what they can see, and each is missing something the other has.

Distance changes which considerations are in view

When a decision belongs to someone else, it presents as a structure: the main objective, the principal obstacle, the eventual outcome. When it belongs to you, it presents as a sequence of concrete difficulties — the conversation you would have to have on Thursday, the money that would be tight in March, the person who would be annoyed. The abstract version is cleaner because the friction has been filtered out by distance, not because it does not exist.

That filtering explains most of the well-known asymmetries. Advice tends to weight the primary criterion heavily and to discount implementation. Personal decisions tend to weight feasibility and short-term discomfort heavily and to lose sight of what the decision was for. Neither is a bias in the sense of a defect; each is an accurate report of the information that survived the distance.

How well established is this

The core observation — that people give more decisive, more principle-driven advice than they act on — appears across many demonstrations and is not seriously disputed. The broad theoretical framework usually invoked to explain it, which links psychological distance to how abstractly things are represented, is a different matter. Parts of that literature have had replication difficulties, and some of the more elaborate predictions derived from it have not held up consistently.

So the sensible position is to take the phenomenon seriously and the theory lightly. You do not need the framework to notice that your advice to others is bolder than your own behaviour, and you should not use the framework to conclude that the bold version is automatically right.

The advice trick, and what it discards

The standard technique is to ask what you would tell a friend in your position. It genuinely helps, and it helps most when the difficulty is that discomfort has crowded out the objective — when you can state the sensible course perfectly well but keep finding reasons to postpone it.

What it discards is private information, and sometimes that information is decisive. You know things about your own tolerance, your obligations, your history with the people involved and your capacity to sustain a hard course that no adviser can see. A friend telling you to leave the job does not know what the last two departures cost you. The trick surfaces the objective; it does not import knowledge, and treating its answer as final throws away the part only you hold.

Deciding on behalf of someone who cannot

A harder case is deciding for a person whose preferences you are meant to represent. Research on this — much of it concerning people asked to predict the wishes of a family member — finds that accuracy is modest, that predictors substitute their own preferences to a noticeable degree, and that confidence in the prediction is not a good guide to whether it is right.

The practical implication is that stated wishes recorded in advance are worth far more than a well-intentioned reconstruction afterwards, because the reconstruction is contaminated in a direction nobody can detect from the inside. Where a decision of this kind has to be made without any record, the honest description is that it is a judgement about what seems best rather than a recovery of what the person wanted.

Using both views deliberately

The useful arrangement is to take the two views in sequence rather than choosing between them. State the decision as you would to a stranger, with the objective named and the implementation left out, and record what that version recommends. Then bring the private information back and ask specifically what it changes — not whether it makes the thing hard, which it will, but whether it changes what the right answer is.

That order matters. Starting from the detail means the objective never gets stated cleanly, and starting from the abstraction and stopping there means you produce advice rather than a decision. The gap between the two answers is not an error to be eliminated. It is a reasonably good measure of how much of your hesitation is information and how much is friction.