Date of Award

5-2026

Document Type

Scholarly Project

Degree Name

Doctor of Physical Therapy (DPT)

Department

Physical Therapy

First Advisor

Kevin O'Brien

Keywords

Titleist Performance Institute, TPI, TPI Movement Screen, inter-rater, reliability, Gwet’s AC1, Krippendorff’s Alpha, functional movement analysis, golf swing biomechanics, movement screen, golf, physical therapy, golf fitness

Abstract

Background: The Titleist Performance Institute (TPI) Movement Screen is a golf-specific physical assessment widely used by professionals to identify physical limitations that may influence swing mechanics and injury risk. Despite its popularity, there is limited peer-reviewed evidence regarding the screen’s inter-rater reliability. Without established reliability, clinical and performance decisions based on TPI scores may be questionable. Prior research on similar tools, such as the Functional Movement Screen (FMS) and Selective Functional Movement Assessment (SFMA), shows that reliability varies based on evaluator experience, scoring objectivity, and standardization—all relevant concerns for the TPI Movement Screen.

Purpose: To determine the inter-rater reliability of the TPI Movement Screen among raters with varying levels of clinical experience and certification.

Methods: Thirty-two healthy adults aged 18–60 with diverse golf and fitness backgrounds were assessed using the 16-item TPI Movement Screen. Participants with injuries or conditions affecting participation were excluded. Eight independent raters (seven DPT students and one licensed physical therapist with TPI Level 1 certification) evaluated each participant either live or via video. All raters received standardized instruction from the TPI-certified rater. Scoring was completed using the TPI Pro application. Each rater independently scored all 16 tests per participant, yielding 6,144 data points. Inter-rater reliability was evaluated using Percent Agreement, Krippendorff’s Alpha, and Gwet’s AC1.

Results: Overall Percent Agreement was high (87.9%). Krippendorff’s Alpha yielded a coefficient of 0.696, suggesting moderate to substantial agreement, while Gwet’s AC1 was 0.877, indicating excellent reliability. Discrepancies between the two metrics were greatest in tests with skewed score distributions—conditions under which Krippendorff’s Alpha may underestimate agreement (known as the “Alpha Paradox”). Tests such as Bridging, Single-Leg Balance, and Toe Touch showed near-perfect reliability. Tests like Pelvic Tilt and Overhead Squat had lower agreement, likely due to more scoring subjectivity. No significant reliability differences were observed between the certified rater and the student raters.

Conclusion: The TPI Movement Screen demonstrates moderate to excellent inter-rater reliability, even among novice, non-certified raters. Gwet’s AC1 provided more stable reliability estimates in skewed data. These findings support the TPI screen’s use in golf performance and rehabilitation, with further study needed on test-retest and intra-rater reliability. Across all raters and tests, Percent Agreement was high (87.9%). Krippendorff’s Alpha yielded a coefficient of 0.696, suggesting moderate to substantial agreement. In contrast, Gwet’s AC1 produced a value of 0.877, indicating excellent inter-rater reliability. Discrepancies between the two metrics were most pronounced in tests where one score category dominated the distribution—a scenario in which Krippendorff’s Alpha is known to produce lower and potentially misleading estimates, a phenomenon sometimes referred to as the “Alpha Paradox.” Tests such as Bridging, Single-Leg Balance, and Toe Touch showed near-perfect reliability, whereas tests like Pelvic Tilt and Overhead Squat demonstrated lower agreement, likely due to greater scoring subjectivity. No significant differences in reliability were observed between the TPI-certified rater and the DPT students, suggesting that even novice raters can achieve high levels of agreement with basic training.

Download

Share

COinS