Your attachment style followed you to work
The relational patterns nobody trained you to see in your team
A few years ago I was coaching an engineering manager, call her Priya, through a conflict with her director. Priya was talented, consistent, well-liked by her team. But every time her director went quiet for a few days, no Slack messages, no check-in, nothing overtly wrong, Priya spiraled. She’d start reviewing her recent work for mistakes. She’d draft and redraft messages trying to get a response. She’d ask peers, carefully, whether they’d heard anything about upcoming changes. By the time her director resurfaced with a routine update, Priya had spent three days in a low-grade crisis that nobody knew about.
I asked her what the silence meant to her. She said: “When someone goes quiet, it means something’s wrong and they’re not telling me.”
That’s not a work problem. That’s an attachment pattern. And it was shaping Priya’s experience of her job, her decision-making, and her relationship with her manager in ways that neither of them could see.
The model, briefly
Attachment theory started with how infants bond with caregivers. The psychologist John Bowlby observed in the 1950s that the quality and consistency of early caregiving creates an internal working model, a template for how relationships function, that persists into adulthood. Mary Ainsworth’s research categorized these patterns into styles, and decades of subsequent work has mapped them onto adult relationships.
The four adult attachment styles are well-documented: secure, anxious-preoccupied, dismissive-avoidant, and fearful-avoidant. The popular conception of these going around captures the basics well enough. What it doesn’t capture is how these patterns show up in the specific dynamics of work: in how people respond to feedback, how they handle ambiguity from leadership, how they behave in conflict, what they need from managers, and where they break under organizational stress.
Research on attachment in the workplace has accelerated in the past decade. A 2024 meta-analysis synthesizing 109 independent samples found that attachment style predicts job performance, burnout, job satisfaction, organizational commitment, and the quality of the leader-member exchange, and that these effects hold even after controlling for the Big Five personality traits. Attachment isn’t personality. It’s the relational operating system that runs underneath personality, and it activates most strongly under threat: reorganizations, ambiguous feedback, conflict with a manager, uncertainty about standing.
This matters for technical leaders because the workplace is a relational system whether you designed it to be one or not. Every direct report has a working model for how authority figures behave, what silence means, whether it’s safe to ask for help, and what happens when you make a mistake in public. Those models aren’t rational assessments of your management style. They’re templates that formed decades before they met you.
How each style shows up in the room
These aren’t types to sort people into. They’re patterns to recognize so you can respond to what’s actually happening instead of what’s visible.
Secure attachment at work looks like what you’d expect: an engineer who can take critical feedback without catastrophizing, ask for help without interpreting it as weakness, tolerate ambiguity from leadership without assuming the worst, and recover from conflict without holding grudges. Securely attached people use their manager as what the research calls a “secure base”: a relationship that provides enough stability to take risks. They’re the ones who’ll push back on a technical decision in a design review and still be fine with you at lunch. Roughly 50-60% of the population trends secure, though workplace stress can temporarily activate insecure patterns in anyone.
Anxious-preoccupied attachment at work is what I saw with Priya. High sensitivity to relational signals. A tendency to interpret ambiguity as rejection or disapproval. Overinvestment in the manager’s mood and availability as a barometer of their own standing. These are often your most dedicated people, deeply loyal and willing to go above and beyond, because the effort is partly driven by a need to secure the relationship. But the cost is significant: they carry a higher cognitive load of relational monitoring that drains resources from the actual work.
An anxiously attached engineer doesn’t just receive feedback. They scan the feedback for what it implies about the relationship. “My manager gave me three specific criticisms in this review” becomes “my manager is dissatisfied with me” becomes “I’m at risk.” The content of the feedback gets processed through a relational filter that amplifies the threat signal. I’ve seen engineers with this pattern spend more time managing their anxiety about a performance review than they spent on the technical work being reviewed.
The organizational footprint: these people are disproportionately affected by managerial inconsistency, delayed feedback, and any ambiguity about their standing. If your culture runs on vibes rather than explicit evaluation, as I described in my hive mind piece, anxiously attached people are the first to suffer. They’re reading the vibes constantly, and the vibes are never clear enough to resolve their underlying question: am I safe here?
Dismissive-avoidant attachment at work looks like extreme self-sufficiency. The engineer who never asks for help, never escalates, never mentions when they’re stuck. They resist pairing, decline mentorship (giving or receiving), and keep managers at arm’s length. Their work is often excellent because they’ve built their identity around not needing anyone. Independence is the armor.
The cost surfaces in two places. First, they don’t signal when they’re in trouble. A dismissive engineer who’s three days into a dead-end approach won’t say so, because asking for help would require acknowledging dependence. By the time the problem is visible, it’s a week behind schedule. Second, they resist the relational infrastructure that teams depend on: retrospectives, one-on-ones, peer feedback. They’ll comply, but they won’t engage. One engineering director told me about a staff engineer who was technically brilliant and relationally invisible. “He ships excellent code and I have no idea what he’s thinking. I’ve never once gotten an honest answer about how he’s doing.”
Research on avoidant attachment and burnout shows that dismissive individuals are at higher risk of burnout because they don’t access the social support systems that buffer against it. They burn alone and leave without warning.
Fearful-avoidant attachment at work is the hardest to read because it oscillates. These people want closeness with their manager and team but are frightened by it. They may be highly engaged one week and withdrawn the next. They approach a promotion conversation with eagerness and then sabotage themselves before the decision is made. They seek feedback and then react defensively when they get it.
This is the pattern most likely to be misread as “difficult” or “inconsistent.” What’s happening underneath is a conflict between the desire for connection and the expectation of harm. The working model says: getting close to authority figures is both necessary and dangerous. The behavior that results looks chaotic from the outside but follows an internal logic that makes sense once you see the template.
I coached a tech lead with this pattern who would build genuine rapport with each new manager, then around month three, start testing the relationship with increasingly provocative behavior: missing deadlines, pushing boundaries in meetings, making sarcastic comments about leadership decisions. Three managers in a row interpreted this as attitude. What was actually happening was a stress test: she was probing to see if the relationship could survive conflict, because her working model predicted it couldn’t.
Why this matters for how you lead
The point is not to diagnose your reports. You’re not a therapist (well, you probably are not a therapist). The point is to recognize that the same managerial behavior produces radically different experiences depending on the attachment template of the person receiving it.
When you go quiet for a week because you’re heads-down on a strategy document, your securely attached engineer barely notices. Your anxiously attached engineer has already drafted three messages they didn’t send and is scanning Slack for signals about whether they’ve fallen out of favor. Your dismissive engineer is relieved by the space. Your fearful-avoidant engineer is oscillating between “they trust me to work independently” and “they’ve stopped caring about me.”
Same silence. Four different internal experiences. And you have no visibility into any of them unless you’ve built the kind of relationship that makes it safe for people to tell you.
This is why research on attachment at work consistently points to the leader-member exchange as the critical variable. The quality of the manager-report relationship doesn’t just affect satisfaction and performance. It functions as the primary attachment relationship in the workplace, the one that either activates or calms the insecure patterns people carry. Managers who are consistent, predictable, explicit about expectations, and available when needed aren’t just being “good managers.” They’re functioning as a secure base that allows people to take the risks that produce their best work.
The control surface
You can’t change someone’s attachment style. It took decades to form and it’s not your job to repair it. But you can change the conditions that activate insecure patterns and the conditions that reinforce secure ones.
Be predictable, especially about availability. Inconsistency is the trigger for anxious attachment and the confirmation of fearful-avoidant expectations. If you’re going heads-down for a few days, say so explicitly. “I’m in strategy mode this week. I’ll be less responsive but nothing has changed between us.” That sentence costs you five seconds and saves your anxiously attached reports days of rumination.
Make evaluation explicit rather than ambient. I keep returning to this across several pieces because it matters structurally. If your people don’t know where they stand, the anxiously attached will assume the worst, the dismissive will disengage from caring, and the fearful-avoidant will oscillate between both. Clear, regular, written feedback isn’t bureaucracy. It’s a stabilizing signal that reduces the relational ambiguity your insecurely attached reports are spending cognitive resources trying to resolve.
Don’t confuse self-sufficiency with health. The engineer who never needs anything is not your easiest report. They’re your most opaque one. Check in on your dismissive people with the same regularity you check in on your anxious ones. They won’t ask for what they need. Build the relationship slowly, consistently, and without forcing intimacy. Over time, predictable availability can create enough safety for avoidant patterns to soften.
Read oscillation as signal, not noise. When someone’s behavior is inconsistent, approach-withdraw-approach, the instinct is to address the behavior. The more useful move is to address the underlying safety question: “I’ve noticed you seem to pull back after we have a direct conversation. I want you to know that disagreement doesn’t change how I see your work or our relationship.” You’re not doing therapy. You’re providing data that the person’s working model doesn’t have.
Signals
You’ll know the relational environment is improving when your anxiously attached people stop over-functioning on Slack at 11pm looking for reassurance signals. When your dismissive engineers start mentioning problems before they’ve become crises. When your fearful-avoidant reports can disagree with you on Tuesday and still show up fully engaged on Wednesday.
Attachment patterns don’t disappear in healthy work environments. They quiet down. The anxious person still scans for threat, but they find less of it. The avoidant person still values independence, but they’re willing to ask for a thought partner on a hard problem. The fearful-avoidant still tests the relationship, but the tests get smaller because the relationship keeps passing them.
You didn’t create these patterns. You inherited them the day each person joined your team. But the relational environment you build determines whether those patterns run hot or cool, and that determines whether your people spend their cognitive resources on the work or on managing an anxiety you can’t see. Every engineer on your team has a working model for what happens when things go wrong. Your job isn’t to fix the model. It’s to build a reality that gives the model less to worry about.