TL;DR: If your toddler is calm, cooperative, and delightful for the teacher, the babysitter, even your own spouse — and then unravels the second she sees you — you are not doing something wrong. Attachment research shows kids save their hardest feelings for the person their nervous system trusts the most. BloomPath’s Ethan Moore breaks down what’s actually happening and what to do with it.

Last Wednesday, Luna’s preschool teacher stopped me at pickup to tell me what a wonderful day my daughter had. Calm. Helpful. Shared the blue crayon without being asked twice. I remember standing there feeling genuinely proud, like I’d done something right.

Four minutes later, in the parking lot, she screamed at me because I opened her car door from the wrong side.

Not cried. Screamed. Full-body, red-faced, “I HATE THIS DOOR” screaming, over a car door she has never once complained about before.

I used to think this meant I was the problem. My wife Mei can pick her up from the exact same preschool, on the exact same day, and get a completely different kid — calm, chatty, no meltdown in sight. For a while I genuinely wondered if I was doing something wrong as a dad. Then Mei sat me down with an article she’d read about attachment research, and it rearranged how I saw every single pickup after that.

Why does my toddler only misbehave with me?

She isn’t misbehaving with you. She’s decompressing because of you. Attachment researcher John Bowlby spent decades studying how children use a trusted caregiver as what he called a “secure base” — a home port they can return to after venturing out into a world that demands constant self-management. Mary Ainsworth’s famous “Strange Situation” experiments in the 1970s found something that still holds up: securely attached toddlers were the ones who got visibly upset when their caregiver left the room, and who came apart with relief — not composure — the moment that caregiver came back.

Here’s the part that reframed it for me. A four-year-old spending six hours at preschool is running an enormous amount of effort just to hold it together: sitting still, sharing, waiting her turn, reading a room full of unfamiliar social rules. That takes real cognitive and emotional labor, even if it looks effortless from the outside. Think of it like a laptop running five browser tabs, a video call, and a backup sync all at once — it’s fine, until the second you close the lid and it finally gets to cool down. You are the lid closing. The meltdown isn’t happening despite how safe she feels with you. It’s happening because of it.

Is it normal for kids to act differently with different caregivers?

Completely. I’ve watched this exact split play out three different ways in our own house: she’s a different kid with her preschool teacher than she is with her grandmother, and a different kid again with Mei versus me. None of that means she loves one of us more or trusts the teacher less. It means she’s calibrating risk. With a teacher, the relationship is newer and the emotional stakes of falling apart feel higher — there’s a social cost to losing it in front of thirteen other four-year-olds. With a parent, especially the parent she’s most attached to, there’s no social cost at all. The relationship can absorb it.

I’ll admit the version of this that stung the most was realizing it’s not always me. Some weeks it’s Mei who gets the meltdowns and I get the easy pickup. It rotates depending on who’s had more one-on-one time that week, who’s been away on a work trip, who put her to bed the last three nights. It’s not a scoreboard. It’s a nervous system deciding, in real time, who’s safest to fall apart in front of.

What is secure attachment theory, actually?

Bowlby’s core idea, later expanded by researchers like Mary Ainsworth, is that children are biologically wired to form one or a few primary attachments — not because it’s emotionally convenient, but because historically it kept them alive. A child who trusts that a caregiver will respond to distress develops what’s called an “internal working model” of relationships: the belief that people can be relied on when things fall apart. That belief doesn’t get built during the easy moments. It gets built during the hard ones — the tantrum in the parking lot, the 2 a.m. wake-up, the meltdown over the wrong-side car door — when the caregiver shows up anyway.

That reframes the whole equation for me as an engineer-brained dad who likes clean inputs and outputs. I used to treat every meltdown like a bug to patch. Now I think of it more like: this is the system doing exactly what it was designed to do, and my job is to be a stable server, not to prevent the request from ever coming in.

What do you actually do in the moment?

None of this makes the parking lot screaming fun. Here’s what’s actually helped us over the past several years, not as a magic fix, but as the difference between a five-minute meltdown and a forty-minute one.

Don’t take the bait to fix it immediately. My instinct is always to problem-solve — open the other door, offer a snack, negotiate. That usually makes it worse in the first sixty seconds, because the feeling hasn’t finished moving through her yet. Waiting quietly for a beat before jumping to solutions gives the wave room to crest.

Say what you see, not what you think. “You’re really mad about the door” lands very differently than “It’s just a door, it’s fine.” One tells her science is happening; the other tells her she’s wrong to feel it.

Keep your own volume down. This one took me the longest to learn. Matching her escalation with mine (even just a sharper “hey, that’s enough”) tends to add fuel. Staying boring and steady — even when I don’t feel steady — helps her system find the calm one in the room.

Expect the rotation. If it’s always the same parent catching every meltdown, that person will burn out fast. Mei and I now openly talk about who’s “on” for hard pickups that week, partly so whoever’s catching it doesn’t quietly start feeling like the family scapegoat.

Resist comparing yourself to the “easy” report. The teacher’s version of your kid and your version of your kid are both real. They’re just measuring different things — one measures compliance, the other measures trust.

What if the meltdowns feel bigger than this explanation?

Most of what I’m describing is developmentally normal and, frankly, a compliment your toddler doesn’t know she’s paying you. That said, secure-base release and something that needs more support can look similar from the outside. If the intensity is escalating over months instead of settling, if there’s ongoing physical aggression specifically toward one parent, or if it’s showing up alongside sleep, speech, or social concerns, it’s worth a conversation with your pediatrician rather than guessing at home. Trust your gut here — you know your kid’s baseline better than any article can.

BloomPath uses illustrated AI characters (Mei and Ethan) to protect our daughter’s privacy. The content is real; the avatars are illustrated. Learn more →

If you’re trying to figure out whether a meltdown pattern is typical for your kid’s age or something to flag, BloomPath tracks developmental stages alongside daily patterns so you’re working from data instead of guesswork at 5 p.m. in a parking lot.

This article is part of our Positive Parenting: The Complete Guide.

Products We Recommend