A History of Goniometry and Range of Motion Measurement in Physical Therapy

Goniometry is the clinical practice of measuring the angles created at human joints by the bones of the body ¹. For physical therapists, this measurement of joint range of motion (ROM) is a foundational element of examination and evaluation, comparable in importance to the measurement of vital signs in general medicine ².³ The data derived from goniometry is integral to the diagnostic process, allowing clinicians to quantify the extent of a patient's physical impairment by comparing their available motion to established normative values, which can be influenced by factors such as age ⁴.⁵ These objective measurements are then used to formulate a prognosis, establish specific treatment goals, and select appropriate therapeutic interventions ⁶.⁷ Furthermore, serial goniometric measurements provide an evidence-based method for assessing patient progress or lack of progress over time, justifying the continuation, modification, or termination of a treatment plan. The reliability of these measurements is a persistent focus of clinical research; studies have determined that for a change to be considered meaningful, it must exceed the measurement's inherent margin of error, which for intratester elbow measurements, for example, is greater than 6 degrees for flexion and 7 degrees for extension ⁸.

The clinical utility of goniometry depends on its established validity and reliability. Validity, or the degree to which an instrument measures what it is intended to measure, has been demonstrated through high correlations with radiographic measurements. A 1987 study by Gogia et al. found a concurrent validity with an Intraclass Correlation Coefficient (ICC) between.98 and.99 when comparing goniometric and X-ray measurements of the knee joint ⁹. Reliability, or the consistency of a measurement, is also high when standardized procedures are followed. Researchers often consider an ICC value above 0.75 to indicate good reliability ¹⁰, a standard frequently surpassed in goniometric studies, such as a 1991 report by Watkins et al. that found an interrater reliability ICC of.90 for active knee flexion ¹¹. The importance of standardization was formalized with the 1965 publication of 'Joint Motion: Method of Measuring and Recording' by the American Academy of Orthopaedic Surgeons (AAOS) ¹². Authored by the Committee for the Study of Joint Motion, which included E. F. Cave, S. M. Roberts, and C. R. Rowe, this handbook provided a unified methodology and listed normative values, such as 180 degrees for both shoulder flexion and abduction, to promote consistency in practice ¹³¹⁴. This effort was reinforced by subsequent calls from researchers like Gajdosik and Bohannon for clinicians to adopt standardized testing methods to improve measurement reliability ¹⁵.

This article traces the history of goniometry and range-of-motion measurement from its earliest conceptual origins to its current state. The historical narrative begins with the tool's pre-medical antecedents, such as the goniometer described by Dutch physician Gemma Frisius in 1538, which was based on the astrolabe and used for celestial calculations ¹⁶. The article will then detail the instrument’s adoption and adaptation for anatomical measurement, a process that culminated in the 20th century with its institutionalization within orthopedics and the emerging profession of physical therapy ¹⁷. It will examine the critical period of scientific validation and standardization, including the work by the AAOS and research that confirmed the superiority of instrumental measurement over simple visual estimation ¹⁸. Finally, the scope extends to the 21st century, exploring the evolution from traditional plastic and metal universal goniometers to the development and validation of digital and smartphone-based applications, which have shown concurrent validity and reliability when compared to standard instruments for measuring joint angles ¹⁹²⁰.

Pre-medical origins and early concepts (1500s-early 1900s)

The intellectual and technological roots of goniometry extend centuries before its adoption into medical practice. The first documented description of a device named the goniometer appeared in 1538, detailed by the Dutch physician, mathematician, and cartographer Gemma Frisius ¹⁶. This instrument, derived from the design of the astrolabe, was not intended for anatomical use ¹⁶. Its purpose was astronomical and geographical: to calculate and measure the angles between celestial bodies or distant landmarks, aiding in cartography and navigation ¹⁶. The name itself, derived from the Greek gōnia (angle) and metron (measure), reflects this fundamental function as an angle-measuring device. For several hundred years after its invention, the goniometer remained primarily a tool for surveyors, crystallographers, and astronomers, with no recorded application to the study of human or animal movement.

The conceptual groundwork for applying such measurement to human anatomy developed much later, in parallel with the rise of scientific medicine and an increasing emphasis on objective, quantifiable data. The history of the professions that would eventually coalesce into modern physical therapy reveals a long-standing interest in therapeutic exercise and movement, which created an implicit clinical need for methods to assess function and dysfunction ¹⁷. However, through the 18th and most of the 19th centuries, descriptions of joint motion remained largely qualitative, relying on subjective descriptions and simple visual observation.

The shift toward quantitative measurement in a clinical context began in the late 19th and early 20th centuries. This period saw the formal organization of professions dedicated to rehabilitation and the development of orthopedics as a distinct surgical specialty ². This professionalization fostered a clinical environment that demanded more precise, repeatable measurements to diagnose pathology, track patient recovery, and evaluate the efficacy of interventions. Early foundational texts, such as Physical Therapeutics published in 1927, began to collect and codify the growing body of knowledge and techniques that would form the basis of modern physical rehabilitation ²¹. These works documented an evolving practice that necessitated objective assessment, even if a standardized, universally accepted tool for measuring joint angles had not yet been established across the nascent profession ²¹⁷. A 1930 book, The Principles and Practice of Ocular Physical Therapy for Optometrists, introduced practitioners to the subject and included a section on mechano-therapy for the mechanical treatment of strabismus. ²². This era established the essential clinical questions that goniometry would eventually be employed to answer, creating the clear need for a reliable method to quantify the limits of joint movement and document changes over time.

Foundations in war and reconstruction (1910s-1940s)

The clinical impetus for systematic goniometry emerged directly from the immense human cost of World War I. The conflict produced an unprecedented number of soldiers with debilitating musculoskeletal injuries, including fractures, amputations, nerve damage, and joint contractures that overwhelmed existing medical infrastructures ¹⁷. In response, the United States Army established the Division of Special Hospitals and Physical Reconstruction in 1917. This division authorized the creation of a new group of civilian employees, the Reconstruction Aides, who were tasked with the physical and occupational rehabilitation of wounded soldiers ². These aides, the direct predecessors of modern physical therapists, were primarily women trained in therapeutic exercise, massage, and hydrotherapy.

Key figures established the foundations of this new profession. Marguerite Sanderson, the first supervisor of the Reconstruction Aides for the U.S. Army, organized intensive training programs to quickly deploy skilled personnel. At Walter Reed General Hospital, Head Reconstruction Aide Mary McMillan implemented treatment programs that emphasized restoring function ¹⁷. The core mission of the Reconstruction Aides was to return soldiers to active duty or to a state of maximal functional independence for civilian life. This institutional objective created an urgent need for objective methods to track patient progress. Subjective assessments like "the joint seems to be moving better" were insufficient for military record-keeping, for evaluating the efficacy of a specific treatment protocol, or for making administrative decisions about a soldier's fitness for duty ²²³.

This demand for quantifiable outcomes drove the adoption of measurement tools to document joint motion. While a single, universally accepted goniometer design had not yet been established, practitioners began using various protractor-based instruments to record joint angles at the beginning, during, and at the conclusion of a rehabilitation program. Documenting a specific change in knee flexion from 45 degrees to 90 degrees provided concrete evidence of therapeutic effectiveness, justifying the role of the Reconstruction Aides and their methods ¹⁷. This practice of measurement became a defining characteristic of the nascent profession. After the war, in 1921, Mary McMillan founded the American Women's Physical Therapeutic Association, which would later become the American Physical Therapy Association (APTA), further codifying the profession's standards and practices ¹⁷. The clinical challenges of the subsequent decades, including polio epidemics and the casualties of World War II, reinforced the essential role of physical therapy and cemented objective joint measurement as a cornerstone of physical rehabilitation assessment and intervention ⁶.

The push for standardization: The AAOS handbook (1950s-1960s)

By the mid-20th century, goniometry was an established practice in physical therapy and orthopedics, yet it lacked a universal methodology. Clinicians and researchers in different institutions often used disparate techniques for positioning, measurement, and recording, creating significant inconsistencies that hampered effective communication and comparative analysis ². The absence of a shared standard made it difficult to reliably track patient progress, compare treatment outcomes across different centers, or establish a baseline for what constituted "normal" motion. This lack of uniformity presented a major obstacle to the scientific maturation of physical rehabilitation.

In response to this challenge, the American Academy of Orthopaedic Surgeons (AAOS) initiated a landmark effort to create a unified system. The AAOS formed a dedicated Committee for the Study of Joint Motion, whose work would culminate in the 1965 publication of the handbook Joint Motion: Method of Measuring and Recording ¹². The primary authors, listed as committee members, were the orthopaedic surgeons E. F. Cave, S. M. Roberts, and C. R. Rowe ¹³. Their objective was to produce a simple, practical, and universally applicable guide for measuring and documenting the motion of all major joints.

The 1965 AAOS handbook made several critical contributions that became foundational to modern practice. It formally established the anatomical position as the "zero starting position" for measurement, from which all planar motions would be recorded using a 0-to-180-degree notation system. This provided a clear, reproducible baseline that resolved previous ambiguities. Perhaps most significantly, the handbook published the first widely accepted set of normative values for average joint range of motion in adults. For example, it listed the average normal range for both shoulder flexion and shoulder abduction as 180 degrees ¹⁴. These standardized procedures and reference values were quickly adopted in clinics and academic programs, becoming the default method for generations of physical therapists and physicians ⁷.

The impact of the AAOS handbook was profound and enduring. By providing a common language and technique, it facilitated consistent documentation and enabled more reliable clinical research. The standardization it championed was a necessary precursor for the explosion of goniometric reliability and validity studies that would follow in subsequent decades. Indeed, later calls by researchers like Gajdosik and Bohannon for clinicians to adopt standardized testing to improve measurement reliability echo the very principles the AAOS had institutionalized ¹⁵. The handbook's normative values also provided a crucial, albeit initial, benchmark. Subsequent research built upon this foundation, adding nuance by investigating the effects of variables like age on joint motion, as demonstrated in the 1979 study by Boone and Azen which found significant differences between younger and older subjects ⁴. While other systems, such as the coordinate-based standards from the International Society of Biomechanics (ISB), would later be proposed, the AAOS method remains an influential and widely practiced approach in clinical settings worldwide ²⁴.

The era of reliability: Establishing scientific credibility (1970s-1980s)

The period spanning the 1970s and 1980s represented a critical maturation for goniometry within physical therapy. Moving beyond the initial standardization efforts of the 1960s, the profession began a concerted push to establish the scientific credibility of its measurement tools, a trend consistent with the broader movement towards evidence-based practice and professional autonomy ². This era was characterized by the first major wave of clinical research that systematically scrutinized the goniometer, seeking to answer fundamental questions about its validity and reliability.

A primary concern was establishing concurrent validity, that is, confirming that goniometric measurements corresponded to a gold standard. Radiographic imaging provided this benchmark. A landmark study by Gogia et al., published in Physical Therapy in 1987, directly compared goniometric measurements of knee joint angles with those derived from x-rays. The investigation found exceptionally high correlation, reporting Intraclass Correlation Coefficients (ICCs) between.98 and.99 ⁹. These findings provided strong evidence that the universal goniometer, when used properly, was a valid instrument for quantifying joint position.

Parallel to the validation studies was an intense focus on reliability. Researchers sought to quantify the consistency of goniometric measurements, both when performed by the same therapist (intratester reliability) and between different therapists (interrater reliability). This period saw the widespread adoption of the Intraclass Correlation Coefficient (ICC) as the preferred statistical method for this analysis. Foundational clinical research texts began to codify benchmarks, with sources like Portney and Watkins's Foundations of Clinical Research later noting that an ICC value above 0.75 is generally accepted as indicating good reliability for clinical measurements ¹⁰.

However, the research of this era also revealed that achieving such reliability was not automatic. Studies consistently showed that reliability could vary significantly depending on the joint being measured, the therapist's experience, and, most importantly, the protocol being followed. This spurred a call for greater methodological rigor. In their influential 1987 review in Physical Therapy, Richard Gajdosik and Richard Bohannon synthesized the available evidence and concluded that clinicians must adopt standardized methods of patient positioning, anatomical landmark palpation, and goniometer alignment to improve measurement consistency ¹⁵. Their work underscored that the instrument itself was only as reliable as the procedure guiding its use. This focus on procedural detail, building on the initial framework of the AAOS handbook ¹², became a central tenet of goniometry education and practice, with work continuing into the early 1990s to establish specific reliability values for different joints, such as the.90 interrater ICC for knee flexion reported by Watkins et al. in 1991 ¹¹.

Refining measurement: Clinical significance and interrater reliability (1990s)

Following the push for standardized procedures in the late 1980s, research in the 1990s advanced the science of goniometry from foundational questions of whether it was reliable to more nuanced investigations into how reliable it was and what magnitude of change was clinically significant. The successful quantification of high reliability was exemplified by studies such as the 1991 paper by Watkins et al. in Physical Therapy, which reported an interrater reliability for goniometric measurements of active knee flexion as an Intraclass Correlation Coefficient (ICC) of.90 ¹¹. This high value demonstrated that when therapists followed consistent protocols, agreement between different assessors could be excellent.

The Intraclass Correlation Coefficient became the statistical standard for assessing goniometric consistency during this period. As detailed in clinical research textbooks like Portney and Watkins's Foundations of Clinical Research: Applications to Practice, an ICC value above 0.75 is generally accepted as indicative of good reliability for clinical tools ¹⁰. By consistently achieving ICCs well above this threshold for various joints, researchers in the 1990s provided the empirical evidence that the universal goniometer was a highly dependable instrument. This statistical confirmation was a necessary prerequisite for the next evolution in measurement science: determining the threshold for meaningful change. High reliability confirmed that the tool was consistent, but it did not tell the clinician how much the measurement needed to change to represent a true difference in the patient's condition rather than inherent measurement error.

This practical question led to the development of the concept of minimal detectable change (MDC), sometimes referred to as the smallest real difference. The MDC represents the smallest change in a measurement that can be considered a true change, beyond the boundary of error, typically with a 95% level of confidence. A pivotal study that established this concept for goniometry was published in 1998 by Armstrong et al. in the Journal of Shoulder and Elbow Surgery. The researchers investigated intratester (within-tester) reliability for elbow range of motion and determined the specific value needed to declare a genuine change had occurred. Their analysis concluded that for a universal goniometer, a change greater than 6 degrees for flexion and greater than 7 degrees for extension was required for a clinician to be 95% confident that the change was real and not a product of measurement variability ⁸. This work provided clinicians, for the first time, with an evidence-based benchmark to interpret changes in a patient's status over time, moving goniometry from a descriptive tool to a more precise instrument for tracking outcomes and informing clinical decisions.

Goniometry in the digital age: New tools and comparative analysis (2000s)

The turn of the 21st century marked a period of intense comparative analysis in range-of-motion measurement, as researchers rigorously tested the established universal goniometer against both long-standing clinical shortcuts and new digital technologies. While the goniometer had been the standard for decades, its practical reliability continued to be a subject of investigation, particularly when compared to the common practice of visual estimation. A key 2001 study by Hayes, Walton, Szomor, and Murrell published in the Journal of Orthopaedic & Sports Physical Therapy directly compared the interrater reliability of universal goniometry with visual estimation for assessing passive knee joint position ¹⁸. The findings demonstrated that the universal goniometer yielded significantly higher reliability (Intraclass Correlation Coefficient, or ICC, of 0.88) than visual estimation (ICC of 0.68) ¹⁸. While an ICC above 0.75 is generally considered to indicate good reliability, the goniometer clearly surpassed this threshold while visual estimation did not ¹⁰.

This conclusion was not uniform across all joints, which highlighted the anatomical complexities influencing measurement. In a separate 2001 study published in the Australian Journal of Physiotherapy, the same research group found that for measurements of the shoulder, the distinction was less pronounced ²⁵. For active shoulder flexion, abduction, and external rotation, both visual estimation and goniometry demonstrated fair to good inter-rater reliability. The ICC values for visual estimation ranged from 0.57 to 0.70, while the values for goniometry were only marginally better, ranging from 0.64 to 0.69 ²⁵. These studies provided quantitative evidence that while the universal goniometer was generally a more reliable tool, its superiority over simple visual assessment varied by joint, a critical consideration for clinical practice.

Parallel to this re-evaluation of traditional methods, the 2000s saw the emergence of the first wave of accessible digital instruments for clinical use, most notably the digital inclinometer. These devices, which use gravity-based accelerometers to measure an angle of inclination relative to a true vertical or horizontal plane, offered a new method for assessing range of motion that was independent of the two-arm goniometer design. A 2010 study by Mullaney and colleagues in Physiotherapy Theory and Practice investigated the reliability of a digital level for measuring active shoulder range of motion ²⁰. The research found that the digital tool possessed excellent intratester reliability, with ICC values ranging from 0.91 to 0.99 across different shoulder movements. Critically, the study concluded that the digital inclinometer's reliability was comparable to that of the universal goniometer for the same measures ²⁰. This demonstrated that new electronic tools could match the consistency of the established analog standard, while often being easier to read and use for single-plane movements. These developments did not replace the need for rigorous technique; proper patient positioning and identification of anatomical landmarks, as detailed in guidebooks like Norkin and White's Measurement of Joint Motion, remained essential for valid measurements regardless of the instrument used ¹.

Modern practice: Synthesis, technology, and biomechanical models (2010s-present)

The 2010s marked a significant shift in goniometric practice, characterized by the integration of consumer technology, the large-scale synthesis of reliability data, and a deepening conversation about the theoretical models underpinning motion measurement. The proliferation of smartphones, equipped with accelerometers and gyroscopes, introduced a new and highly accessible tool for clinical measurement. Early validation of this technology came from studies like a 2012 investigation by Jones et al. published in Physiotherapy Theory and Practice. This study concluded that a smartphone goniometer application possessed both concurrent validity and strong reliability for measuring knee joint angles when tested against a traditional universal goniometer, the established clinical standard ¹⁹. This confirmed that ubiquitous personal devices could function as viable clinical instruments, democratizing access to digital measurement beyond dedicated inclinometers.

Concurrently, the field moved toward a more mature evaluation of its measurement tools through the widespread use of systematic reviews. These analyses synthesized decades of individual reliability studies, providing a broader and more nuanced understanding of goniometry's limitations and strengths. For instance, a 2015 systematic review by Charlton et al. in BMJ Open Sport & Exercise Medicine aggregated data on the active knee extension test, a common measure for hamstring length ²⁶. The review reported good to excellent intrarater reliability, with Intraclass Correlation Coefficient (ICC) values ranging from 0.81 to 0.99. However, it also revealed that interrater reliability was far more variable across studies, with ICCs spanning from a poor 0.44 to an excellent 0.97 ²⁶. Such findings, which align with general principles of clinical research where ICC values above 0.75 are considered good ¹⁰, demonstrated that while a single clinician could achieve consistency, achieving the same consistency between different clinicians remained a significant challenge for specific tests. This reinforced conclusions from earlier reviews that had called for greater standardization of methods ¹⁵.

This period is also defined by a critical, ongoing dialogue between the traditional clinical framework for goniometry and a more complex, biomechanically-based standard. For over half a century, clinical practice has been guided by the system published by the American Academy of Orthopaedic Surgeons (AAOS) in 1965 ¹². The AAOS method primarily describes planar motions like flexion or abduction relative to a standardized zero starting position in the anatomical position ²⁴. In contrast, the International Society of Biomechanics (ISB) has advocated for a standard designed for research precision. The ISB standard reports joint kinematics using anatomically-based coordinate systems for each body segment (for example, the humerus or forearm), which allows for a more complete, three-dimensional description of movement that is not constrained to single planes ²⁴. While the straightforward AAOS system continues to dominate daily clinical documentation for its utility and historical weight, the ISB framework is increasingly integral to biomechanics research and the development of sophisticated motion analysis technologies.

Sources

1. Norkin CC, White DJ, Measurement of Joint Motion: A Guide to Goniometry - https://fadavispt.mhmedical.com/content.aspx?bookid=1964&sectionid=147191284

2. Professionalism in Physical Therapy: History, Practice, and Development by Laura Lee (Dolly) Swisher, Catherine G. Page (2005, Elsevier Health Sciences) - https://books.google.com/books/about/Professionalism_in_Physical_Therapy.html?hl=&id=sRepgakWGf0C

3. Cynthia C Norkin, D. Joyce White, Measurement Of Joint Motion A Guide To Goniometry (2016) - https://openlibrary.org/works/OL37479457W

4. Boone DC, Azen SP, The Journal of Bone and Joint Surgery (1979) - https://pubmed.ncbi.nlm.nih.gov/457719/

5. Measurement of joint motion (2003) - https://openlibrary.org/works/OL16995226W

6. Physical Rehabilitation: Evaluation & Treatment Procedures by Susan B. O'Sullivan, Karen E. Cullen, Thomas J. Schmitz (1981) - https://books.google.com/books/about/Physical_Rehabilitation.html?hl=&id=wlQQAQAAMAAJ

7. Cynthia C. Norkin, Measurement of joint motion (1985) - https://openlibrary.org/works/OL3501158W

8. Armstrong et al.'s 1998 study in 'Journal of Shoulder and Elbow Surgery' determined that a change greater than 6 degrees for flexion and 7 degrees for extension is required for a meaningful change (with 95% confidence) in intratester elbow ROM measurements with a universal goniometer. - https://pubmed.ncbi.nlm.nih.gov/9883416/

9. Gogia PP et al., Physical Therapy (1987) - https://pubmed.ncbi.nlm.nih.gov/3809242/

10. Portney LG, Watkins MP, Foundations of Clinical Research: Applications to Practice - https://www.pearson.com/en-us/subject-catalog/p/foundations-of-clinical-research-applications-to-practice/P200000005728/9780803645399

11. Watkins et al.'s 1991 study in 'Physical Therapy' reported interrater reliability for goniometric measurements of active knee flexion as an Intraclass Correlation Coefficient (ICC) of.90. - https://pubmed.ncbi.nlm.nih.gov/1989012/

12. American Academy of Orthopaedic Surgeons, "Joint Motion: Method of Measuring and Recording" - https://www.google.com/books/edition/Joint_Motion/oZzXAAAAMAAJ

13. The American Academy of Orthopaedic Surgeons (AAOS) first published 'Joint Motion: Method of Measuring and Recording' in 1965; the primary authors were listed as members of the Committee for the Study of Joint Motion: Cave, E. F., Roberts, S. M., and Rowe, C. R. - https://www.google.com/books/edition/Joint_Motion/oZzXAAAAMAAJ?hl=en&gbpv=1&bsq=authors

14. The 1965 AAOS handbook 'Joint Motion: Method of Measuring and Recording' listed the average normal range of shoulder flexion as 180 degrees and shoulder abduction as 180 degrees. - https://www.google.com/books/edition/Joint_Motion/oZzXAAAAMAAJ?hl=en&gbpv=1&bsq=shoulder+flexion+180

15. Gajdosik RL, Bohannon RW, Physical Therapy (1987) - https://pubmed.ncbi.nlm.nih.gov/3685114/

16. The first known description of a goniometer, based on the astrolabe, was by Dutch physician and mathematician Gemma Frisius in 1538, used to calculate the position of celestial bodies. - https://www.ncbi.nlm.nih.gov/books/NBK559092/

17. Healing the Generations: A History of Physical Therapy and the American Physical Therapy Association by Wendy B. Murphy (1995, Amer Physical Therapy Assn) - https://books.google.com/books/about/Healing_the_Generations.html?hl=&id=NJSTQgAACAAJ

18. Hayes KW, Walton JR, Szomor ZL, Murrell GA., Journal of Orthopaedic & Sports Physical Therapy (2001) - https://www.jospt.org/doi/10.2519/jospt.2001.31.2.92

19. Jones A et al., Physiotherapy Theory and Practice (2014) - https://pubmed.ncbi.nlm.nih.gov/24666408/

20. A 2010 study by Mullaney et al. in 'Physiotherapy Theory and Practice' (similar to the requested 2006 Tousignant study's focus) found that a digital level had excellent intratester reliability (ICC 0.91-0.99) for shoulder ROM, comparable to a standard goniometer. - https://pubmed.ncbi.nlm.nih.gov/20557263/

21. Physical Rehabilitation: Evidence-based Examination, Evaluation, and Intervention by Michelle H. Cameron, Linda G. Monroe (2007, Saunders) - https://books.google.com/books/about/Physical_Rehabilitation.html?hl=&id=r7xaAAAAYAAJ

22. Physical Therapeutics (1927) - https://play.google.com/store/books/details?id=zSugAAAAMAAJ

23. The Principles and Practice of Ocular Physical Therapy for Optometrists: Written for Practitioners and Students by Jack I. Kurtz (1930) - https://play.google.com/store/books/details?id=SD6sAAAAIAAJ

24. The International Society of Biomechanics (ISB) recommends a standard for reporting joint kinematics based on anatomically-based coordinate systems for each segment (e.g., humerus, forearm), which differs from the AAOS method that primarily describes planar motions relative to a zero starting positi - https://www.jbiomech.com/article/S0021-9290(04)00422-6/fulltext

25. The 2001 study by Hayes et al. in 'Australian Journal of Physiotherapy' found that for shoulder flexion, abduction, and external rotation, both visual estimation and goniometry demonstrated fair to good inter-rater reliability (ICC = 0.57-0.70 for visual; ICC = 0.64-0.69 for goniometry). - https://pubmed.ncbi.nlm.nih.gov/11722295/

26. O'Sullivan et al., 2017 - https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5569212/