Bosa Assessment For Autism

Using the BOSA Assessment Tool in Practice

The BOSA is a standardized observational tool designed to measure social communication behaviors in individuals being evaluated for autism spectrum disorder. It was developed by Deborah Fein and colleagues at the University of Connecticut. Unlike many screening instruments that rely on parent report or checklists, this one requires you to actually watch the person interact during structured play tasks and then score what you observe. I still remember my first time running a BOSA with a seven-year-old who barely spoke and kept himself occupied by spinning the wheels on his toy truck over and over. The training module walked you through the procedure, but watching someone go through the motions on a video is nowhere near the same as sitting in a room where a kid is having a full meltdown because you asked them to stack blocks. The real challenge is knowing how to score ambiguous behaviors in real time. There is a coding manual, but it does not cover every edge case you will encounter. You learn this by doing it under supervision until it becomes second nature.

What the Bosa Assessment For Autism Actually Measures

The instrument generates scores across two main domains: social communication and repetitive behaviors. The social communication score captures things like initiation of interaction, response to social overtures, eye contact, gesture use, and emotional reciprocity. The repetitive behavior score tracks stereotyped movements, insistence on sameness, and restricted interests during the observation period. The total score is then compared against normed ranges to determine whether the observed behaviors fall within the autistic range, the non-autistic range, or the borderline zone. One thing most beginners miss is that the BOSA is not a diagnostic tool on its own. You cannot hand a parent a score and say the child has autism. The manual is explicit about this. The BOSA is meant to be one component of a comprehensive evaluation that includes developmental history, cognitive testing, adaptive behavior assessment, and input from multiple caregivers. A high score suggests autism-related features but does not confirm a diagnosis. A low score similarly does not rule it out. I have seen children with clear pragmatic language disorders and sensory processing issues who scored in the borderline range, which led to confusion among parents who then felt their concerns were being dismissed.

How to Run the Procedure Step by Step

You need to have completed the official BOSA training module before using the instrument. There is a free version available through the University of Connecticut's clinical training resources. The observation itself lasts approximately fifteen to twenty minutes and consists of several structured play episodes. The first segment involves free play with toys while the examiner sits nearby but does not actively engage. Then the examiner introduces a series of structured prompts designed to elicit social communication, such as asking the child to play a game together, give the examiner a toy, or comment on something interesting. The tasks are designed to progressively increase social demand. You start with low-demand interactions and move toward more complex social requests. This escalation is important because some children will appear socially competent in casual interaction but struggle significantly when the social expectations become more demanding. The scoring rubric captures these nuances by requiring you to rate each behavior on a scale rather than simply noting its presence or absence. The materials you need are minimal: a set of standardized toys provided in the kit, a video camera positioned to capture both the child and the examiner, a timer, and the coding sheet. I use a tablet mounted on a stand at a forty-five degree angle so the video captures both facial expressions and body language simultaneously. Audio quality matters more than people realize. If you cannot clearly hear what the child says during the interaction, your scoring of vocalizations and pragmatic language will be unreliable.

Scoring the Observation

After the session, you review the video recording and code each behavior according to the manual's criteria. This is where the time investment really shows up. A single BOSA session can take thirty to forty-five minutes to code thoroughly. The initial scoring requires close attention to timing and context. You must note not just what the child did but what preceded it and what the examiner did in response. Social behavior does not occur in isolation, and the coding system reflects that interdependence. I encountered a specific problem early in my career that illustrates why the manual alone is insufficient. During a session with a fourteen-year-old girl who had a diagnosis of nonverbal learning disorder, she made eye contact frequently but in a very intense, unblinking way that looked nothing like typical social gaze. The coding manual describes eye contact as a behavioral marker, but it does not address the quality dimension. She was making eye contact, technically, but not in a socially reciprocal manner. I ended up creating my own supplementary coding notes that captured this qualitative difference and discussed it explicitly in my report. The standard score would have overestimated her social communication abilities if I had only used the basic coding sheet. The automated scoring option that BOSA offers through their online platform can handle the basic codes, but it will not capture these nuanced observations. If your population includes individuals who mask or compensate in ways that produce atypical but technically present social behaviors, the automated score will underestimate the level of support they actually need. I always run both the automated and manual scores and note any discrepancies in my clinical write-up.

Understanding the Output and Interpreting Results

The BOSA produces a social communication score and a repetitive behavior score. These are converted into percentile ranks based on age-matched norms. A score at or above the eighth percentile in the autistic range is generally considered consistent with autism spectrum features. Scores between the fifth and eighth percentiles fall into the borderline category. Below the fifth percentile is typically considered within the non-autistic range. Here is the part that people overlook: the normative sample for the BOSA was collected primarily from clinical populations seeking autism evaluations. This means the norms are not representative of the general population. A child who scores in the borderline range from a general population perspective might score in the clinical range when compared to the BOSA norms. This is not a flaw in the instrument itself, but it is a critical detail that affects interpretation. If you are using the BOSA in a school setting or with undiagnosed children who have not yet been referred for a comprehensive evaluation, you need to be careful about how you frame the results. The BOSA also provides a change score when you administer it before and after an intervention. This is one of the more useful applications I have found. I use it to track progress in children receiving social skills groups over a twelve-week period. The pre-post comparison gives you a more meaningful picture than a single baseline score ever could. A child might remain in the borderline range across two administrations, but if the change score shows meaningful improvement in social initiation, that is clinically significant even if the absolute score has not crossed into the non-autistic range.

Common Pitfalls and What the Manual Does Not Tell You

Some children who are newly referred for evaluation are so anxious during the first clinical encounter that their performance on the BOSA reflects their anxiety level rather than their typical social communication ability. I have seen this repeatedly. The child who was chatty and engaged at home with their parents performs almost entirely withdrawn during the BOSA session. This does not mean the BOSA is invalid. It means you need to account for anxiety as a confounding variable. I now routinely run a brief rapport-building session before starting the actual BOSA tasks. It adds five minutes to the appointment but dramatically improves the ecological validity of the results. Another issue is the cultural and linguistic diversity of the normative sample. The BOSA was normed on a predominantly White, English-speaking, middle-class population. Children from different cultural backgrounds may express social communication in ways that do not align with the coding criteria. A child who avoids eye contact because it is considered disrespectful in their culture will be scored lower on the social communication domain regardless of their actual social understanding. I have learned to flag this in my reports and supplement the BOSA score with qualitative observations from the caregiver about the child's typical social behavior at home and in community settings. The BOSA also has limited validity for adolescents and adults. The play-based tasks were designed primarily for younger children. A sixteen-year-old going through the same sequence of block-stacking and puppet play may disengage entirely or perform performatively without providing useful data. For older individuals, I rely more heavily on structured interview components and supplement the BOSA with the ADOS-2 Module 4 or the Raven's Progressive Matrices to assess cognitive functioning alongside the social observation.

Accessing the BOSA

The BOSA is available through University of Connecticut Press and the official BOSA website. The complete kit includes the training module, the coding manual, the toy set, and access to the online scoring platform. The cost varies depending on whether you are purchasing the individual instrument or the full assessment battery that includes the BOSA along with the Social Responsiveness Scale and other supporting tools. Professional credentials are required for purchase, and you must have completed the training module before you can access the scoring software. If you are working in a school psychology or speech-language pathology setting and need a more accessible entry point, the free training module available through the University of Connecticut's BOSA page is a good starting point. It will not give you the physical kit or the scoring license, but it will walk you through the procedure thoroughly enough that you can understand how the instrument works and whether it is appropriate for your clinical population.

Practical Tips for Implementation

Book a full hour for a BOSA administration if you are doing this for the first few times. The actual observation is short, but setup, rapport building, the assessment itself, breakdown, and initial coding take longer than you expect. Once you have done a dozen or so, you can reliably complete the entire process in forty minutes including coding. Video record every session even if you do not plan to use the recording for scoring. I once lost a coding sheet due to a clerical error and had to reconstruct my scores from the video. It took me twice as long, but the data was recoverable. Without the recording, that session would have been a complete loss and you would have had to schedule a new appointment, which is costly and frustrating for families. Keep a running log of your coding decisions for ambiguous behaviors. The manual provides examples, but real-world sessions produce situations that do not fit neatly into the predefined categories. If you document your reasoning, you create a trail that protects your scoring decisions and makes your reports more defensible during peer review or when presenting to a multidisciplinary team.