โ† Blog
CareerSeptember 18, 2026By Johnson

Attachment Styles at Work, and Why Only One of Them Gets Noticed

The working model that decides how silence from a partner feels also decides how silence from a manager feels. What each attachment pattern looks like at a desk, why the avoidant end reads as professionalism, and where the research stops.

The message goes out at 10:04. By half past eleven there is no reply, and it has been reread four times, once out loud. Somewhere around noon a working theory forms about what the silence means, and by the end of the afternoon that theory has acquired supporting evidence from a meeting last Tuesday.

Nobody calls this attachment. At work it gets called being conscientious, or caring too much, or in the more expensive version, being difficult to manage.

59%

of adults in a nationally representative American sample measured as secure

25%

measured as avoidant

11%

measured as anxious

70%

of the variance in a team's engagement scores sits with the manager

Why attachment shows up at work at all

Attachment research began with infants and then with romantic couples, and for a long time the working assumption was that this was a domain of life with its own rules. The thing being measured turned out to be more general than that. It is an expectation about what other people do with a need, formed early, applied broadly, and it activates wherever someone's regard matters and their signals are unclear.

An office produces unclear signals continuously. A short reply. A meeting that moves. A name left off an invitation. None of these events carries its own interpretation, and the interpretation supplied is the one already installed.

Richards and Schat put this to a proper organisational test in 2011 and found what the theory predicts. Both insecure dimensions were related to turnover intentions and to less of the discretionary helping behaviour that organisations depend on, and the mechanism differed by dimension. High avoidance predicted not seeking support when work got difficult. High anxiety predicted the dysfunctional interaction patterns that colleagues eventually notice. The entry on attachment theory covers the model underneath; what follows is the version with a payroll number.

What each pattern looks like at a desk

PatternAt work it looks likeWhat it gets calledWhat it actually costs
Low on bothAsks the clarifying question, says when something is late, takes feedback as informationEasy to work withVery little, which is why it is boring to write about
High anxietyOver-delivers on vague requests instead of asking; reads brevity as verdict; volunteers for the unwanted task as insuranceKeen, or needy, depending on who is describing itHours, mostly invisible, spent on interpretation
High avoidanceSolves it alone; declines help; goes quiet under load; leaves without a conversationIndependent, low maintenance, a safe pair of handsThe problem nobody heard about until it was large
High on bothWants the manager's attention and finds it unbearable when it arrives; asks for feedback and disputes itInconsistentThe one most likely to be read as a character judgement

The second row is the one that has a name in workplaces already. It is legible, it generates visible behaviour, and it gets managed, sometimes badly. Most of what has been written about attachment at work is about this row, including the anxious version on this site.

The row nobody flags

The third row is the interesting one, because the workplace has no complaint mechanism for it.

Avoidance at work does not produce a symptom. It produces an employee who handles things, who does not escalate, who rarely asks for anything and never appears in anyone's list of difficult conversations. Every one of those is a description of a good hire. The organisational literature keeps finding that the same disposition also predicts not seeking support when the work exceeds what one person can do, and disengaging as difficulty accumulates rather than raising it.

So the pattern that costs an organisation the most is the one its feedback systems are structurally unable to see. A person high on anxiety generates signals continuously, which is unpleasant and also means the problem is known. A person high on avoidance generates none until the resignation, at which point everyone reports being surprised.

There is a further asymmetry worth stating plainly. Workplaces reward the avoidant presentation and penalise the anxious one, and this has nothing to do with which is more functional. Self-containment happens to look like competence, and visible need happens to look like instability, and neither resemblance is earned. The avoidant pattern is not better adapted to work than the anxious one. It is better camouflaged.

Things that are read as work habits

What it is called
What it is doing
Being thorough
Removing every possible reason for someone to be displeased
Preferring to work alone
Avoiding the moment where help has to be requested
Not wanting to bother anyone
Keeping a request small enough that a refusal would not mean much
Reading the room
Monitoring one specific person for a change in weather
Being direct about scope
A limit that is easier to state at work than at home, and worth noticing for that reason

The last line is not a criticism. Work is frequently where insecure patterns are least visible, because the structures carry the load: a scheduled one-to-one supplies the reassurance nobody has to ask for, a written brief removes the ambiguity, and a defined scope draws the boundary that would be a whole conversation at home. That is a real finding about environments rather than people, and it is the practical version of the whole topic.

Why the manager and not the team

Gallup's analysis across millions of employees puts around seventy percent of the variance in a team's engagement at the manager rather than the organisation. The number is usually quoted as an argument about leadership training. It also explains why attachment shows up in this particular relationship and not evenly across the office.

One person controls the resources, the interpretation of performance and the availability of reassurance, and their signals are ambiguous by the nature of the job. That is the structural definition of an attachment figure, whatever anyone thinks about the word. Colleagues at the same level rarely produce the same effect, which is why the same person can be entirely relaxed with a team and rewrite a three-line message to their manager four times.

Where this stops being true

Two honest limits, because the field is younger than the romantic literature and the effect sizes are modest.

The first is that a genuinely erratic manager produces checking, over-delivery and case-building in almost anyone. If the pattern appeared in this job and in no previous one, the reading is about the situation. Attachment is measured across relationships for exactly this reason, and a score taken during a bad quarter under a bad manager is describing the quarter.

The second is that most of the workplace evidence is self-reported, cross-sectional and collected in Western organisations, and it predicts behaviours adjacent to performance rather than performance itself. That is enough to explain a pattern and nowhere near enough to sort people. Any instrument marketed for hiring on this basis is making a claim the research does not support, and the entry on attachment theory is blunter about the limits than most summaries are.

Nothing here is a reason to talk about childhood in a performance review. The useful version is narrower. There is a thing that happens between sending a message and getting a reply, and it is the same thing at ten at night with a partner as it is at 10:04 with a manager, which means it can be measured once.

Sources

Frequently asked questions

Can your boss be an attachment figure?

Structurally, yes, and that is the uncomfortable part. An attachment figure is anyone whose availability matters and whose signals are ambiguous, and a manager controls both the resources and the read on whether things are going well. This does not mean the bond resembles a romantic one in depth or in kind. It means the same expectations about whether closeness is safe get applied to it, which is why a delayed reply from a manager can produce an evening that makes no sense on paper.

Is it unprofessional to talk about attachment style at work?

Using the vocabulary out loud with colleagues usually goes badly, because it imports a clinical register into a relationship that has no consent for one. Using it privately to understand a pattern is different and costs nothing. The version that travels well is behavioural: naming that ambiguity is expensive for you, or that you tend to solve things alone, without the framework attached.

Do attachment styles predict job performance?

Weakly, and anyone selling a hiring instrument built on it is overselling. The organisational findings sit on behaviours that are adjacent to performance, such as whether support gets sought, how emotions are regulated at work and whether someone is thinking about leaving. Performance itself has better predictors, and attachment is not one of them.

Why do I feel fine at work and anxious in my relationship?

Usually because the two relationships differ in how much is at stake and how ambiguous the signals are. Work often has structures that reduce ambiguity, such as scheduled check-ins, written feedback and a defined scope, and those structures do a lot of quiet regulating. A relationship with none of that equivalent scaffolding gives the same system far more to work with.

How do I manage someone who needs a lot of reassurance?

Make the schedule do the work rather than your goodwill. A short, predictable, recurring check-in costs less in total than the unplanned reassurance it replaces, and the predictability is the active ingredient rather than the warmth. What makes it worse is irregular praise, because irregular positive signals are the exact condition that keeps a vigilant system scanning.

Does remote work change any of this?

It removes most of the low-grade evidence that people used to get for free. A corridor, a shared lunch and a face in a meeting all supplied continuous proof that nothing was wrong, and a channel of text messages does not. The reasonable response is not more meetings but more explicit signalling, because ambiguity is the input the whole system runs on.

See this pattern in your own numbers

โœฆ

This article, about your situation

Luna has read the same research โ€” and, if you let her, your results.

โœฆ Talk it through

More notes on people