Published by Unseen Progress, an independent publisher of caregiver research. Last reviewed 2026-05-10. Part of the stepfamily research overview.
Short answer. A stepchild who is warm one-on-one and cold the moment their biological parent enters the room is almost certainly in a loyalty bind — not rejecting the stepparent. The pattern is one of the most consistently documented signatures in the stepfamily literature (Papernow, 2013; Ganong & Coleman, 2017). The child has decided, often unconsciously, that visible warmth toward the stepparent in front of the bio parent constitutes a betrayal — and is regulating the visibility of the relationship accordingly. The research-backed response is the opposite of what most stepparents try.
A loyalty bind is the conflict a child feels when accepting one person seems to require rejecting another. In stepfamilies it is the default state, not an exception (Papernow, 2013). The child's loyalty to the biological parent — the one whose home was disrupted, whose relationship ended, whose grief the child has often absorbed — is unconditional. Warmth toward the stepparent feels, to the child's emotional system, like evidence that the bio parent has been replaceable. The instinct to suppress visible warmth in the bio parent's presence is a protection move, not a judgment.
Ganong and Coleman (2017) synthesise dozens of studies on this exact dynamic: when children are observed across contexts, their warmth toward a stepparent is highest in low-loyalty-load situations (alone, away from home, on neutral activities) and lowest in high-loyalty-load situations (in the presence of the biological parent, during transitions between households, around birthdays, holidays, and milestones). The behavioural difference is not character; it is loyalty load.
Hetherington and Kelly (2002) document the same pattern in their longitudinal data and add an important finding: the contextual flip is more pronounced in children who have a closer relationship with their non-custodial biological parent, not less. In other words, the warmer the child is with the bio parent, the colder they tend to appear to the stepparent in that bio parent's presence — because the loyalty stakes are higher.
The most common descriptions in stepparent communities map to the research with unusual precision:
"When it's just us she's chatty and asks me about my day. The second her dad walks in the room she goes quiet and won't look at me."
"He'll laugh with me in the car and then act like I don't exist at the dinner table."
"Holidays are the worst — she'll have been fine for weeks and then the day her mom is involved she's hostile to me again."
These are not three different children with three different problems. They are three descriptions of the same loyalty-bind signature, viewed from inside the experience.
Stepparents who don't have the research frame typically try one of three responses, and all three make the pattern worse.
"You were nice to me yesterday. Why are you acting like this now?" The child cannot answer this question honestly because they don't have conscious access to the loyalty bind driving the behaviour. Confrontation forces the child to choose a public position — and the position they will choose, every time, is the one that protects the bio parent. The confrontation locks the rejection in.
A stepparent who tries to extract a hug, a kind word, or eye contact in front of the biological parent is asking the child to perform an act of betrayal in front of the person they are trying to protect. The child's resistance escalates predictably. Papernow (2013) identifies this as one of the most reliable ways to harden a loyalty bind into open hostility.
The opposite move — concluding "they don't actually like me, so why try" — also fails. The one-on-one warmth was real. Withdrawing punishes the child for a context-dependent behaviour they didn't consciously choose, and erodes the only part of the relationship that was working.
Across Papernow (2013), Ganong & Coleman (2017), and Deal (2014), the consensus response has three parts.
The relationship is what happens at low loyalty load — the car ride, the one-on-one walk, the moment alone in the kitchen. Those interactions are the actual data. The dinner table with the bio parent present is the noisiest possible measurement of the relationship, not the truest. Treating the cold dinner as the verdict and the warm car as the exception inverts the signal.
The child does not need to be told they are in a loyalty bind. They need the stepparent to behave in ways that lower the loyalty stakes:
Papernow's research shows that the contextual flip softens as the child accumulates evidence, over years, that warmth toward the stepparent has not cost them anything in the bio parent relationship. The flip rarely disappears completely until adolescence is over and the child is no longer protecting the bio parent's emotional position. Year-by-year, however, the cold-when-bio-parent-present interactions get less cold and shorter, and the warm-alone interactions extend their reach.
---
Unseen Progress publishes long-form caregiver research and builds research-backed daily trackers for the families covered. See the full stepfamily research overview for the complete framework.