Child development

The STAT and How Clinicians Use It

Save

Two things make the STAT unusual among autism screening tools. It is administered to the child rather than filled in by a parent, and it was designed for the second stage of screening — the point where a questionnaire has already raised a question and someone needs a closer, structured look before booking a full evaluation that may be a year away.

Last updated: July 2026

Talk to a clinician

Gale can help you find a clinician in your state and request a visit.

Find care →

What is the STAT?

The STAT is the Screening Tool for Autism in Two-Year-Olds, an interactive, play-based Level 2 autism screener for children between 24 and 36 months. Rather than asking a caregiver to report on behavior from memory, it puts a trained clinician on the floor with the child and structures the interaction so that play, communication, and imitation skills can be observed directly 1.

Its validation was published in two parts. The first study used signal detection analysis in developmentally matched two-year-olds to derive a cutoff, then confirmed the instrument's sensitivity, specificity, and predictive values both in the sample where the cutoff was derived and in an independent validation sample. The second established interrater agreement, test-retest reliability, and the agreement between a child's STAT risk category and their classification on the ADOS-G, a standard diagnostic observation schedule 1.

That second study carries more practical weight than it looks like it should. An instrument that two different clinicians score the same way, and that lines up with a full diagnostic observation, is producing a result about the child rather than about who happened to administer it.

What does a STAT session involve?

Structured play, running across three skill areas: play, communication, and imitation 1. A clinician arranges activities that give a two-year-old natural reasons to do particular things — to request something, to direct an adult's attention toward something interesting, to copy an action, to use an object in a pretend way — and records how the child responds rather than whether the child can be talked into performing on command.

Imitation is the area caregivers least expect to see examined. Copying what another person does with their hands or with an object asks several things at once: noticing that the other person is doing something, treating it as worth attending to, and reproducing it. It is a social act before it is a motor one.

Because everything is administered to the child, the STAT does not lean on a caregiver's recall of what happened eighteen months ago. That is its clearest advantage over a questionnaire, and it is also its cost: it consumes clinician time, requires training, and needs a room with the right materials. Those costs are the reason it sits at the second stage instead of being offered to everyone.

Why this page does not print the STAT's cutoff score

Because the published record that can be verified does not print it. The paper that validated the STAT confirms a cutoff was derived through signal detection analysis and then tested in an independent sample, but the numeric value of that cutoff, the number of items, and the direction of the scale are not stated in the record available here 1. The honest move is to say so rather than to import a number from somewhere shakier.

This is not housekeeping. A cutoff detached from its instrument, its age band, and its administration manual is worse than no number at all, because it looks like information. It is close to the mechanism that makes online autism quizzes misleading: a threshold lifted out of a real study and applied to items that are not the real items, scored by someone with no training in what the responses mean.

There is also nothing a caregiver could do with the figure. The STAT is administered by a trained clinician with specific materials, and its scoring rules live in its manual. Knowing the threshold would not let anyone run the screen at home, and a home approximation of a validated instrument is not the instrument.

What Level 2 means

It describes where an instrument sits in a two-stage screening model, and nothing about how severe anything is. Level 1 screening is universal — a brief instrument offered to every child at set ages, whether or not anyone has raised a concern. Level 2 screening is for the children Level 1 flagged: longer, more specific, and usually administered by a clinician rather than completed by a caregiver. The STAT is a Level 2 tool 1.

It is not the only one. The RITA-T rapid interactive screening test works similar ground: nine clinician-administered interactive activities probing joint attention, social awareness, reaction to emotions, awareness of human agency, and fundamental cognitive skills, administered and scored in roughly ten minutes for children 18 to 36 months 2. Level 2 autism screening is a small family of instruments doing broadly the same job through slightly different activities and age bands.

What this second layer buys is precision at the referral step. A universal questionnaire is deliberately tuned to miss as few children as possible, which means it forwards a substantial number who are not autistic. An observation-based second look is meant to narrow that group before it arrives at a comprehensive diagnostic evaluation.

What comes before the STAT

A Level 1 screen, and the schedule that governs one. The American Academy of Pediatrics model is developmental surveillance at every well-child visit, general developmental screening at 9, 18 and 30 months, and autism-specific screening at 18 and 24 months 3. That autism screening schedule is the layer where most children are first flagged, and it runs on caregiver questionnaires rather than on observation.

The dominant instrument at that layer is the M-CHAT-R/F. Its validation study followed 16,071 toddlers, showed improved detection of autism compared with the original M-CHAT alongside a reduced follow-up burden, and found that children identified through it were diagnosed roughly two years earlier than the national median age of diagnosis 4. Two years, at that age, is most of the life the child has lived.

The m-chat-r/f and the STAT are not competitors, and a family may meet both within a single season. One is a wide net cast over an entire practice; the other is a closer look at what the net brought up. Neither produces a diagnosis, and neither is designed to.

What a STAT risk category does and does not mean

It sorts a child toward or away from a diagnostic evaluation, and that is the entirety of its job. Autism identification runs as a two-step process — developmental screening, then a comprehensive diagnostic evaluation — and autism can be detected by 18 months or younger, with a diagnosis by an experienced professional considered reliable by age two 5. The STAT belongs wholly to the first step.

The direction that gets discussed less is the reassuring one. A screen that places a child in a lower-risk category has not examined that child's development the way an evaluation would. Screening false negatives exist in every instrument ever validated, including the good ones, and the accurate reading of a reassuring screen is 'nothing found today' rather than 'nothing there'.

That is why passed screen still worried is a legitimate sentence to say out loud, and why it lands better with specific examples attached. A screen is a snapshot taken by a stranger, in an unfamiliar room, across a short interval, on a day the child may have been tired. A caregiver's account covers every setting the child lives in and every month they have lived.

The STAT's telehealth descendants

The interactive-screening idea has outlived the requirement that everyone be in the same room. TELE-ASD-PEDS was built on the same premise — structured play activities that make social communication visible — but the caregiver administers the activities at home while a clinician observes remotely and scores what they see 6. The clinician coaches; the caregiver does the playing.

That redesign is not simply the same session delivered over video. Handing the activities to a caregiver changes who is running them, how consistently they run, and what can be seen through a camera aimed at a moving two-year-old. An instrument built for telehealth is designed around those constraints rather than pretending they are absent, which is why it is a separate tool with its own name.

For a family a long way from a developmental clinic, that matters more than the methodological detail. Whether autism can be diagnosed over telehealth has a longer answer than one page can give, but the tools attempting it rest on the same in-person interactive-screening idea the STAT represents: watch the child do something, rather than ask someone to remember it.

Common questions

Children between 24 and 36 months. The name refers to two-year-olds because that is the band it was validated in, and an instrument's accuracy does not automatically extend past the ages it was tested on. For a younger toddler, a clinician would reach for a different instrument rather than administering this one early.

No. It is administered by a clinician trained in the instrument, using specific materials and scoring rules from its manual. The skill being applied is judging how a two-year-old responds during structured play against what is typical at that age, which is built through many prior observations rather than conveyed by a written description.

The validation paper confirms a cutoff was derived through signal detection analysis and tested in an independent sample, but the numeric value is not printed in the record available here, and neither is the item count. Rather than repeat a number that cannot be verified, this page leaves it out. A clinician administering the instrument works from its manual.

The M-CHAT-R/F is a caregiver questionnaire offered to every toddler at routine visits, so it works as a first pass across a whole practice. The STAT is administered directly to the child by a clinician who watches and scores structured play. They occupy consecutive stages of the same process rather than substituting for one another.

Both are clinician-administered interactive screens sitting at the second stage. They differ in age band and in activities: the RITA-T uses nine short activities for children 18 to 36 months, scored in about ten minutes, while the STAT was validated for 24 to 36 months across play, communication, and imitation. A clinician picks one based on the child's age and the setting.

No. It means the next step is a comprehensive diagnostic evaluation. Screening instruments are tuned to over-refer rather than to miss children, so some children flagged at this stage are found to have a language delay, a global developmental delay, or nothing that persists. The evaluation answers the question the screen only raised.

Related

Say it back

How would you explain this to someone you love?

Two or three sentences, just as you’d say it. Gale reflects back what you focused on — a mirror, not a quiz.

Talk to a clinician

Gale can help you find a clinician in your state and request a visit.

Find care →

Things that are looked at without waiting for a screening result

  • —Words, gestures, or social interest a toddler clearly had six months ago and no longer has — a backward step is handled differently from a delay and is not watched and waited on
  • —A child who does not turn to their name called from across a room and has had no hearing test since the newborn screen, since hearing loss produces a very similar picture
  • —Head banging, biting, or other self-injury that leaves marks, or that has become the child's usual response to frustration
  • —Any seizure-like episode in a child being evaluated for developmental concerns — staring spells with unresponsiveness, or stiffening and jerking

A seizure lasting more than five minutes, or a first seizure of any length, is a 911 call rather than a message left for the pediatrician.

This article describes what the STAT is and how clinicians use it. It is not the instrument, cannot be used to screen a child, and does not replace an evaluation by a qualified clinician.

References

  1. 1.Stone WL, Coonrod EE, Turner LM, et al. (2004). Psychometric Properties of the STAT for Early Autism Screening. Journal of Autism and Developmental Disorders 2004;34(6):691-701. doi:10.1007/s10803-004-5289-8 ✓That the STAT is an interactive, play-based Level 2 autism screener for children 24-36 months assessing play, communication, and imitation; that its cutoff was derived by signal detection analysis and confirmed in an independent validation sample; that interrater agreement, test-retest reliability, and agreement between STAT risk category and ADOS-G classification were established; and that the numeric cutoff, item count, and score direction are not stated in the verified record.
  2. 2.Choueiri R, Wagner S. (2015). A New Interactive Screening Test for Autism Spectrum Disorders in Toddlers. The Journal of Pediatrics, 167(2):460-466. doi:10.1016/j.jpeds.2015.05.029That the RITA-T comprises nine clinician-administered interactive activities probing joint attention, social awareness, reaction to emotions, awareness of human agency, and fundamental cognitive skills, administered and scored in roughly ten minutes in children 18-36 months — used here as the comparison Level 2 interactive screener.
  3. 3.American Academy of Pediatrics (2024). Developmental Surveillance and Screening. American Academy of Pediatrics — Patient Care. link ✓The AAP-recommended schedule — developmental surveillance at every well-child visit, general developmental screening at 9, 18 and 30 months, and autism-specific screening at 18 and 24 months — used here to describe the Level 1 layer that precedes a STAT.
  4. 4.Robins DL, Casagrande K, Barton M, Chen CA, Dumont-Mathieu T, Fein D (2014). Validation of the Modified Checklist for Autism in Toddlers, Revised With Follow-up (M-CHAT-R/F). Pediatrics. doi:10.1542/peds.2013-1813 ✓That the M-CHAT-R/F was validated in 16,071 toddlers with improved detection of autism versus the original M-CHAT and reduced follow-up burden, and that screened children were diagnosed roughly two years earlier than the national median — used here to describe the first-stage instrument that typically precedes a STAT.
  5. 5.Centers for Disease Control and Prevention (2024). Screening and Diagnosis of Autism Spectrum Disorder. CDC — Autism Spectrum Disorder (ASD). linkThat autism identification is a two-step process of developmental screening followed by comprehensive diagnostic evaluation, that autism can be detected by 18 months or younger, and that a diagnosis by an experienced professional can be considered reliable by age two.
  6. 6.Vanderbilt Kennedy Center, TRIAD (2024). TELE-ASD-PEDS (TAP). Vanderbilt Kennedy Center — TRIAD. link ✓That TELE-ASD-PEDS consists of caregiver-administered play activities observed remotely by a clinician to support autism evaluation in toddlers via telehealth — used here to describe how the interactive-screening approach was adapted for remote assessment.

6 sources, numbered by first appearance. General health information, not medical advice. AI-assisted editorial content — citations link their sources. Editorial policy

← Back to the health library