MyTherapyWizard Blog

Tag: Draw a person

  • Built on Evidence. Proven in Practice. Shoutout to Learning Charms’ Team

    Built on Evidence. Proven in Practice. Shoutout to Learning Charms’ Team

    How 3.8 Years of Systematic Clinical Measurement Demonstrates That Occupational Therapy Works

    Stephanie Seymore Wick, MSOT, OT/L | Founder and Clinical Architect, O.T. Wizard | Learning Charms, Inc., Charlotte, North Carolina

    The Problem With Checklists

    For years, occupational therapists working in early childhood settings were collecting data that told them almost nothing. Checklist-style evaluations produced a snapshot: present or absent, yes or no. They could not tell you whether a child improved. They could not tell you which counties had greater concentrations of developmental need. They could not tell you whether your team’s intervention was moving the needle or whether children were simply getting older.

    That was the reality facing Learning Charms in 2022. We were screening and evaluating large numbers of children across Head Start programs, NC Pre-K classrooms, and community settings throughout North Carolina, and we had nothing meaningful to show for it in terms of trend data, geographic insight, or outcome evidence.

    So I built something.

    From Nothing to 8,509 Screenings and Evaluations

    The FUNdamental Foundations (FF) screener was designed and developed by a managing pediatric occupational therapist with 25+years of clinical experience. It was built as a structured, multi-domain developmental tool designed from the outset to generate analyzable data. It was not designed for publication. It was designed to answer clinical questions: What does this population look like? Where are the gaps? Is what we are doing making a difference?

    Thirty clinicians on the Learning Charms team tested and used each version in the field, providing the real-world feedback that drove every refinement. They made the transition from paper-and-pencil evaluations to digital data entry on a tablet or laptop, mid-session, with children in front of them. That is not a small ask. The early weeks required support with the technology. There were growing pains. The team did it anyway, and they did it without much, if any complaint.

    The FF tool went through two versions, each refined based on team feedback. Version 6 ran from June 2022 through July 2023. Version 7, with improvements including date of birth capture and an updated item structure, ran from August 2023 through May 2025. In late 2025, the practice transitioned to O.T. Wizard, a fully rebuilt clinical intelligence platform designed and built by the same therapist.  O.T. Wizard was built with Rasch psychometric architecture, 15 guided evaluations, and integrated outcome tracking across 12 domains.

    The table below summarizes what 3.8 years of that effort produced.

    Table 1. Clinical Data Collected Across the Full Evidence Ecosystem (2022-2026)

    PlatformPeriodRecordsEvaluationsScreeningsE1-E2 Pairs
    FUNdamental Foundations V6Jun 2022 – Jul 20232,9281,5111,417324
    FUNdamental Foundations V7Aug 2023 – May 20254,9622,3682,594477
    O.T. WizardSep 2025 – Mar 2026619619—97
    TOTAL3.8 years8,5094,4984,011898

    Note. E1-E2 pairs = children with two complete evaluations allowing pre-to-post comparison. FF V6 pairs are V6-only matches. FF V7 pairs include V7-only and cross-version (V6 E1 to V7 E2) matches. OTW pairs matched by Student_ID. pp = percentage points.

    In total: 8,509 individual assessment records. 4,498 full evaluations. 4,011 developmental screenings. 898 pre-to-post evaluation pairs. Over 245,000 item-level data points. Collected by a single clinical team, through routine practice, over less than four years.

    What the Data Shows: Gains That Exceed Maturation

    The central question in any clinical outcome dataset without a randomized control group is this: how do you know the gains are from intervention and not just from children getting older?

    We address this directly.

    Using cross-sectional developmental data from our own E1 (Initial Evaluation) dataset, we calculated the expected rate of developmental growth per month for each skill area based on age alone. This gives us a maturation baseline specific to this population. We then compared that expected gain to the gains actually observed in children who received OT services between E1 and E2 (Re-evaluation), over a mean interval of 5.4 months.

    The results are consistent across all three measured domains and across both independent datasets.

    Table 2. Observed Gains vs. Expected Maturation Over Mean 5.4-Month Interval (FF n=801 pairs, OTW n=94 pairs)

    ItemFF Observed GainExpected (Maturation)RatioOTW Observed Gain
    Draw a Person (0-4 scale)+1.24 pts+0.39 pts3.2x+1.06 pts
    Functional Pencil Grasp+25.3 pp+9.7 pp2.6x+27.7 pp
    Finger Touching (54-mo milestone)+20.3 pp+11.7 pp1.7xn/a
    Cohen’s d (DAP)0.94——0.84

    Note. Expected gain calculated from cross-sectional linear regression of E1 scores on age in months using the full FF evaluated dataset. pp = percentage points. Cohen’s d: 0.2 = small, 0.5 = medium, 0.8 = large effect. OTW finger touching item not directly comparable due to different item structure.

    Draw a Person improved at 3.2 times the expected developmental rate in FF and 2.6 times in OTW. Functional pencil grasp improved at 2.6 times expected in FF and 3.1 times in OTW. These are not marginal differences from what maturation alone would predict. They are two to three times larger. And they replicate across an entirely independent dataset collected with a different tool, by the same team, with different children.

    Why This Is Not Just Children Getting Older

    If the gains above were driven primarily by maturation, we would expect children at all starting points to show similar improvement. A child who enters at score 0 would gain roughly as much as a child who enters at score 3, because age-related development does not care where you start.

    That is not what we see. The table below shows Draw a Person gains stratified by E1(Initial Evaluation)  score, combining FF and O.T. Wizard data. The pattern is unambiguous.

    Table 3. Draw a Person Gain by E1 Score: FF (n=801) + OTW (n=93) Combined

    E1 ScorenE2 MeanMean Gain% Improved% Same% Declined
    0 (no parts)365+291.74+1.7474%26%0%
    1 (approximations)203+162.37+1.3781%13%6%
    2 (head, no body)165+312.66+0.6654%37%9%
    3 (recognizable)60+112.96-0.0429%46%25%
    4 (6+ body parts)8+63.36-0.360%57%43%

    Note. n column shows FF count + OTW count at each E1 score level. E1 score 0 = no recognizable approximations. Score 4 = recognizable person with 6 or more body parts. Gains decline systematically as E1 score increases, reflecting ceiling effects at higher starting points rather than absence of progress.

    Children who started at score 0 improved by an average of 1.74 points, with 74% showing measurable gains. Children who started at score 3 or 4 were near the ceiling of the scale and showed flat or slightly negative scores at E2, exactly as ceiling effects predict.

    This score-dependent gain gradient is the signature of a real treatment effect. Maturation produces relatively uniform gains regardless of starting point. Intervention produces the largest gains in children with the most room to grow. That is what we observe, and it replicates point-for-point across both the FF and OTW datasets independently.

    The grasp and finger touching data tell the same story from a different angle.

    Table 4. Skill Transition Rates: What Happened Between E1 and E2

    ItemStatus at E1nOutcome at E2
    Pencil GraspNon-functional377 (FF) + 65 (OTW)60% converted to functional
    Pencil GraspFunctional417 (FF) + 29 (OTW)94% maintained functional
    Finger TouchingFail407 (FF)58% passed at E2
    Finger TouchingPass394 (FF)82% maintained pass

    60% of children with non-functional pencil grasp at E1 had functional grasp by E2. 94% of children with functional grasp at E1 maintained it. Skills were not fluctuating randomly. They were moving in one direction and holding. That is not maturation. That is intervention.

    Two Tools, Three Years Apart, Same Answer

    The FF screener and O.T. Wizard are different instruments. FF was a clinician-developed Google Form with embedded scoring anchors and standardized stimulus materials. O.T. Wizard is a fully architected clinical platform undergoing Rasch psychometric validation, with 596 data variables per evaluation and item-level calibration. The two tools share several core items, including Draw a Person, pencil grasp classification, and finger touching. They do not share overlapping children. The DAP scale is directly comparable across both tools at the 0 to 4 range, with identical scoring anchors at each level. O.T. Wizard extended the ceiling by adding two higher-level descriptors, bringing the OTW scale to 6 points total. For this analysis, OTW DAP scores were capped at 4 to ensure a valid cross-tool comparison.

    Yet when we calculate the cross-sectional developmental growth rate for Draw a Person from FF E1 data, we get 0.072 points per month. From OTW E1 data, we get 0.082 points per month. Two tools, thousands of children, the same underlying developmental trajectory captured within 0.01 points of each other per month.

    When two independent measurement systems produce convergent developmental slopes and convergent gain ratios, that is not a coincidence. That is construct validity. Each dataset serves as an independent replication of the other’s findings, and both point to the same conclusion.

    Danielle, an OTR/L out of NC, administers an evaluation with a 3 year old using OT Wizard

    What This Means for OT Practice and Clinical Infrastructure

    Pediatric occupational therapists have long known that their interventions make a difference. The challenge has been demonstrating it systematically, at scale, in a form that partners, funders, schools, and insurance providers find credible.

    The Learning Charms team built that demonstration over 3.8 years, starting from scratch, with no research funding, no university partnership, and no IRB. They built it by replacing meaningless checklists with structured clinical measurement, by training a team of 30 clinicians to collect data consistently, and by iterating their tools until the data was worth analyzing.

    O.T. Wizard is the current iteration of that infrastructure. It is not a platform claiming efficacy. It is a platform whose evidence base already exists, built by the same team that built the platform, using the same children, in the same communities, over the same years. The data in this paper is not a promise of what O.T. Wizard will eventually show. It is a record of what systematic clinical measurement has already demonstrated.

    OT works. The data, replicated across tools and years and nearly 900 pairs of children, shows it.

    Disclosure

    Stephanie Seymore Wick is the founder and clinical architect of O.T. Wizard and owner of Learning Charms, Inc. All data was collected through routine clinical practice and contracted screening partnerships. No external funding was received. The FUNdamental Foundations screener was a clinician-developed field tool and has not undergone formal psychometric validation. O.T. Wizard is currently undergoing Rasch analysis validation. All findings should be interpreted as practice-based clinical evidence rather than results from a randomized controlled trial.

    About O.T. Wizard

    O.T. Wizard is a clinical intelligence system for pediatric occupational therapy professionals. The platform supports evaluation, documentation, goal planning, and scheduling across 12 domains including fine motor skills, visual-motor integration, praxis, visual perception, executive functioning, activities of daily living, and participation. O.T. Wizard is undergoing Rasch analysis validation to establish psychometrically sound, norm-referenced scoring with living norms that update as the clinical database expands. Learn more at otwizard.com.

    About Learning Charms

    Learning Charms is a pediatric occupational therapy group that employs roughly 25 OTP’s in the Charlotte , NC and surrounding counties. Learning Charms is now focused mainly on preschool aged children in their school environment.

  • Understanding the Draw-A-Person Task: A Data-Based OT Perspective

    Understanding the Draw-A-Person Task: A Data-Based OT Perspective

    What the Draw-A-Person Task in O.T. Wizard Is Actually Showing Us

    A data-based look at the Draw A Person Task in preschool OT evaluations inside O.T. Wizard.

    Important context: The data presented here were drawn from preschool children who did not pass an occupational therapy screening and were subsequently evaluated. This sample is not a normative population and should not be interpreted as representative of typically developing preschoolers.

    The Draw-A-Person task is a familiar tools in pediatric occupational therapy. It is widely used, information rich, and often referenced in evaluation reports. At the same time, many therapists find it challenging to interpret, especially when scores are low.

    Rather than debating the value of Draw-A-Person conceptually, this article looks at how the task behaves in real evaluation data when it is administered consistently across a preschool sample. You may have heard of “The Good Enough Draw A Person” drawing assessment. The O.T. Wizard uses a similar but more simple version.

    This analysis is descriptive only. No Rasch or item response modeling has been applied yet.

    Sample overview

    • Number of evaluations included: 404
    • Approximate number of students: 404
    • Evaluations with Draw-A-Person data: 401
    • Total item-level data points across the evaluation system: 19,062
    • Referred clinical sample, not normative population

    Draw-A-Person was part of the standard evaluation battery and was administered when the child tolerated the task. Missing responses were excluded rather than scored as zero.

    Draw-A-Person scoring rubric

    Draw-A-Person was scored using the following criteria:

    • Score 0 (0.0): No approximations
    • Score 1 (0.2): Approximations emerge
    • Score 2 (0.4): Head and parts present but no body
    • Score 3 (0.6): Recognizable person with body and at least four body parts
    • Score 4 (0.8): Recognizable person with six or more body parts
    • Score 5 (1.0): Recognizable person with twelve or more body parts

    It is important to note that even a score of 1 reflects emerging representational drawing rather than an absence of skill. The score was converted to a normalized score where a raw score of 5 = normalized score of 1.0 (most credit).

    Average Draw-A-Person performance

    Across the combined preschool sample:

    • Draw-A-Person scores were analyzed using normalized values derived from the scoring rubric. The average normalized score across the preschool sample was 0.38, corresponding to an average rubric level of approximately 1.9. Clinically, this places the average child between “approximations emerge” and “head and parts present but no body.”
    • While individual scores varied, this average suggests that most preschoolers in the sample demonstrated emerging representational drawing skills rather than fully recognizable figures

    Distribution of Draw-A-Person scores

    When responses are mapped directly onto the scoring rubric, the distribution looks like this:

    • Score 0, no approximations: 290 children, approximately 72 percent
    • Score 1, approximations emerge: 99 children, approximately 25 percent
    • Score 2, head and parts without a body: 7 children, approximately 2 percent
    • Score 3, recognizable person with body and at least four parts: 4 children, approximately 1 percent
    • Score 4, recognizable person with six or more parts: 1 child, less than 1 percent
    • Score 5, recognizable person with twelve or more parts: 0 children

    This is a heavily floor-weighted distribution.

    What this distribution tells us

    In this preschool sample, most children did not produce a recognizable person. Nearly three quarters of children showed no recognizable approximations, and only a very small percentage produced a clearly recognizable person with a body.

    Draw-A-Person age anchors suggest that by approximately 36 months, children often demonstrate a head with emerging parts, by 48 months a recognizable person with a body and multiple body parts, and by 64 months increasingly detailed human figures. In contrast, the average performance in this referred preschool sample falls between “approximations emerge” and “head and parts present but no body.” This indicates that, as a sample group, these children are demonstrating representational drawing skills that are less mature than would be expected based on age anchors, which is consistent with a population that did not pass an occupational therapy screening rather than a normative sample.

    This helps explain why Draw-A-Person often feels like a difficult task for young children and why it frequently stands out in evaluation reports.

    Why Draw-A-Person behaves differently than many other tasks

    Compared to many visual perceptual, gross motor, or participation-based tasks, Draw-A-Person requires multiple skills to work together at once. These include visual motor integration, visual perception, motor planning, body awareness, fine motor control, and representational thinking.

    Because of this high level of integration, Draw-A-Person tends to function as a high-threshold task. It separates children who are beginning to integrate these skills from those who are not yet developmentally ready to do so.

    The data supports what many therapists experience clinically. Draw-A-Person is informative, but it should not be interpreted in isolation.

    Interpreting Draw-A-Person results responsibly

    Based on this data, several points are important for clinical interpretation:

    • Low Draw-A-Person scores are common in preschoolers
    • Progress from score 0 to score 1 is clinically meaningful
    • Scores of 3 or higher represent a small minority of children
    • Draw-A-Person should be interpreted alongside visual perceptual, fine motor, praxis, and participation data

    Whether the task is developmentally ambitious, exhibits a floor effect, or is a candidate for item misfit will be evaluated during future Rasch analysis. At this stage, the data supports careful, contextual interpretation rather than over-weighting the score.

    Why this matters for practice

    Most therapists have an intuitive sense that Draw-A-Person is hard for young children. Very few have seen how strongly that intuition is reflected in actual data.

    Seeing the full distribution helps recalibrate expectations and supports clearer communication with parents, teachers, and teams. It also reinforces the importance of viewing Draw-A-Person as one piece of a much larger evaluation picture.

    As this dataset grows and formal psychometric analysis is completed, these descriptive patterns will be tested and refined. For now, they provide a grounded, data-anchored explanation for why Draw-A-Person feels the way it does in real preschool OT evaluations.