This is the research companion to The Complicated Kid. Read the main post first.
The main post made a claim it did not stop to prove: that the kid no one can figure out is usually a kid whose working theory is wrong. The behavior is real. The story about its cause is not. That is a hard thing to assert without evidence, because the obvious objection is that you are just guessing, swapping one hunch for another.
So here is the evidence. These five papers do not study “complicated kids” as a category, because no such category exists in the literature. What they study is the machinery underneath the confusion. How often two conditions travel together. How often a child hides the very thing you are trying to see. How a single diagnosis can carry a second, impairing driver that the label never mentions. And whether reading a child along dimensions -- threat, deprivation, adaptation -- outperforms reading them by their box. Read together, they say the same thing the main post said, in numbers: the surface is not the system. Ask what else could be going on.
The hidden driver inside a familiar diagnosis
Shaw, P., Stringaris, A., Nigg, J., & Leibenluft, E. (2014). Emotion dysregulation in attention deficit hyperactivity disorder. American Journal of Psychiatry, 171(3), 276-293.
10.1176/appi.ajp.2013.13070966
What they found: Emotion dysregulation -- not inattention, not hyperactivity, but the trouble managing feelings -- shows up in roughly 25 to 45 percent of children with ADHD and 30 to 70 percent of adults. A large UK population study found mood lability in 38 percent of children with ADHD, about ten times the population base rate. Across 15 study contrasts, aggression was elevated with a pooled effect of about 1.92. Across 19 contrasts, facial-emotion recognition was moderately impaired (about 0.65), and across 21 contrasts, a preference for immediate over delayed reward came in around 0.6. The combination of ADHD plus emotion dysregulation was more impairing than ADHD alone across peer, family, academic, and occupational life, even after controlling for other diagnoses. Imaging implicated a striato-amygdalo-medial-prefrontal network sitting at the interface of bottom-up reactivity and top-down control.
Why this matters for you: A diagnosis is a headline, not the full article. A child can carry the ADHD label and have most of their daily suffering driven by something the label barely names -- the emotional part. If you treat the attention and the meltdowns keep wrecking everything, the working theory was incomplete, not wrong. This is the main post’s point in clinical form: the visible behavior is real, but its cause may be one diagnostic layer down. Ask which engine is actually doing the damage before you assume the named one is the whole story.
What it doesn’t answer: These are clinical ADHD samples, so the prevalence and impairment figures belong to a diagnosed group, not the general population. Definitions of emotion dysregulation vary widely across studies and the meta-analytic heterogeneity is high, so the pooled effects are rough indicators rather than precise estimates. The imaging is mostly small, cross-sectional samples. And the review cannot tell you whether emotion dysregulation is a core feature of ADHD, a distinct entity, or a correlated-but-separable dimension -- it lays out all three models and declines to choose.
When two conditions arrive together
Young, S., Hollingdale, J., Absoud, M., Bolton, P., Branney, P., Colley, W., ... Woodhouse, E. (2020). Guidance for identification and treatment of individuals with attention deficit/hyperactivity disorder and autism spectrum disorder based upon expert consensus. BMC Medicine, 18, 146.
10.1186/s12916-020-01585-y
What they found: A UK multidisciplinary expert panel built unified guidance for the case where ADHD and autism co-occur -- a dual diagnosis that was not even permitted until DSM-5. They cite a meta-analytic rate of about 21 percent comorbidity of autism within ADHD samples, and treat co-occurrence as additive in complexity, not a simple sum of two conditions. The core recommendations: assess with multiple sources rather than any single instrument, and count symptoms parsimoniously, because the same behavior can be double-charged to either condition. Girls are systematically under-identified, partly because rating scales are normed on mostly male samples and partly because girls show fewer disruptive behaviors. Symptoms get masked by home accommodations, by compensatory and camouflaging strategies, and by anxiety. A child diagnosed with autism first should be watched over time for ADHD emerging later, since the two conditions are typically identified at different ages.
Why this matters for you: Comorbidity is not the exception; for neurodevelopmental conditions it is closer to the rule. If your child has one diagnosis and something still does not add up, a second condition hiding underneath is a live possibility, not a long shot. This is exactly the “passed from service to service” child the main post describes -- the one each system treats for its own specialty and no one sees whole. The practical move the panel pushes is the FCD move: gather information from every setting, and resist letting one label do all the explaining.
What it doesn’t answer: This is expert consensus, not a systematic review or a trial -- it carries the weight of agreement, not of pooled data. The authors are explicit that the school-intervention evidence base is thin and that much of the classroom guidance is practitioner-derived. It is UK-centric in its service framing, and it discloses an unrestricted educational donation from a pharmaceutical company to the partnership that convened it. The medication guidance in particular rests on limited data specific to the comorbid group.
The child who hides the thing you are trying to see
Mandy, W. (2019). Social camouflaging in autism: Is it time to lose the mask? Autism, 23(8), 1879-1881.
10.1177/1362361319878559
What they found: Camouflaging is the effortful work autistic people do to pass as non-autistic -- suppressing a stim, studying a peer and copying her gestures, building a working persona. In one online sample, about 70 percent of autistic adults reported consistent camouflaging. Across qualitative studies, people describe it as obligation rather than choice, driven mostly by fear of bullying and rejection. The costs are steep: camouflaging is linked to delayed or missed diagnosis, especially in girls and women, to anxiety and depression, and in one study to elevated suicidality. The interpretive frame is ecological: the problem sits at the interface between the person and an environment that punishes recognizably autistic behavior, not inside the person. School refusal, on this reading, is a fit problem first and an individual problem second. The right intervention target is the environment, not redoubled effort to make the child act typical.
Why this matters for you: A child who looks like they are coping may be paying a hidden tax to look that way. Smooth surface behavior can be evidence of cost, not success -- and the children who mask best are often the ones who get seen latest and helped least. This is the main post’s masking point with data behind it: the signal is being suppressed, so the adults read “fine” when the child is depleting. Build at least one environment -- home, a chosen friendship, an affirming space -- where your child does not have to perform a self. The exhaustion you cannot see is real.
What it doesn’t answer: This is a short editorial summarizing a young literature, not a systematic review. The underlying studies are mostly cross-sectional and self-report, weighted toward autistic adult women, so the developmental picture for young children is largely inferred. Causal direction is unsettled -- an unmeasured environmental third variable, such as bullying itself, could drive both the masking and the mental-health harm. The 70 percent figure comes from an online convenience sample and is not population-representative.
Reading the child by traits, not by category
McLaughlin, K. A., Sheridan, M. A., Humphreys, K. L., Belsky, J., & Ellis, B. J. (2021). The value of dimensional models of early experience: Thinking clearly about concepts and categories. Perspectives on Psychological Science, 16(6), 1463-1472.
10.1177/1745691621992346
What they found: The authors defend a dimensional reading of early adversity against the charge that it just sorts children into new boxes. Their argument turns on distinguishing exposure (the probability a category applies, like “community violence”) from experience (what a child actually lives, like threat of harm or absence of expected input). Dimensional models measure several continuous dimensions at once -- threat, deprivation, unpredictability, harshness -- rather than assigning a child to a single category. They marshal evidence that threat and deprivation predict distinguishable outcomes across amygdala reactivity, aversive learning, cognitive control, pubertal timing, and biological aging, and that harshness and unpredictability map onto different life-history outcomes. Crucially, the dimensions are expected to co-occur; the models are built to handle overlap with latent-class, network, and bifactor methods. A central mechanism, experience-driven plasticity, can operate without the child’s conscious appraisal -- absence of language input shapes the brain much as absence of visual input does.
Why this matters for you: This is the intellectual engine under the main post’s “dimensions over labels.” Two children can share a category -- “adversity,” “trauma,” even a diagnosis -- and need opposite things, because what they actually experienced differs. The functional question is not which box the child is in but which dimensions are loaded: Is this threat, or deprivation, or unpredictability? That question survives co-occurrence, which the category approach does not. It is the same upstream discipline the main post asks of you, given a research grammar.
What it doesn’t answer: This is a commentary with no new data; it assembles existing evidence selectively, and several magnitude claims lean on the authors’ own prior reviews. The authors concede that operationalizing deprivation and unpredictability in children’s real environments is still a measurement frontier, with tools only now emerging. Predictions for higher-order cognition are less well characterized than for sensory and language circuits, and transfer to global and culturally diverse populations is not directly tested -- unpredictability in particular may need cultural calibration.
When the deficit is an adaptation
Ellis, B. J., Abrams, L. S., Masten, A. S., Sternberg, R. J., Tottenham, N., & Frankenhuis, W. E. (2022). Hidden talents in harsh environments. Development and Psychopathology, 34(1), 95-113.
10.1017/S0954579420000887
What they found: Childhood adversity does not only subtract; it can also build stress-adapted skills that fit harsh, unpredictable environments. The authors synthesize evidence that maltreated and previously institutionalized children show enhanced detection and memory for anger and faster threat detection; that children from lower-socioeconomic-status backgrounds show greater attunement to other people, better empathic accuracy, and stronger collaborative problem-solving; and that some young adults from unpredictable homes show enhanced flexibility in switching tasks and updating information, especially when primed with stress and uncertainty. They frame these as adaptive intelligence -- real skill, measurable against standards of speed and accuracy -- and argue that conventional tests, built for quiet controlled settings, systematically under-detect them. The same behaviors that read as deficits in a still classroom (vigilance, fast responding, attention shifting) can be assets in fluid environments.
Why this matters for you: The main post asks you to treat “weird” or “difficult” behavior as a signal worth decoding rather than a defect to suppress. This is the strongest version of that move: some of what looks broken is a system that adapted well to conditions you may not see. Before you set out to correct a behavior, ask what environment it was built for and what it might be good for. Reframing a child’s apparent deficit as a possible adaptation is not wishful thinking -- it is a documented pattern, and it changes what you build on.
What it doesn’t answer: This is a theoretical and integrative review; no intervention is tested, and the authors are repeatedly clear that the hidden-talents evidence base is still emerging and needs replication with preregistered, well-powered, socioeconomically diverse samples. The hypothesized link from stress-adapted skills to success in mainstream school and work is drawn as a dashed line -- not yet demonstrated. Tradeoffs are real: development under stress carries costs in cognitive control and health. And the framework is aimed at children who came of age in harsh settings, not at building resilience in early childhood.
Coming Up
Next time: The Confidence Trap. We celebrate confidence in kids as if it were the goal, but confidence built on never struggling is brittle, and confidence demanded too early can curdle into performance. We look at what actually makes a child feel capable -- and why the route there runs through difficulty, not around it.
Every Functional Child Development post comes with a Dive Deeper companion. Subscribe for the full picture.