What is ICC (Absolute Agreement, Not Adjusted)?

ICC (not adjusted) / Absolute Agreement refers to one specific type of ICC model that assesses the absolute agreement between raters or measurements without adjusting for systematic differences between raters. This model variation evaluates how close the actual values of the measurements or ratings are to one another, considering both the consistency of ratings and the actual level of agreement.

• Absolute Agreement examines whether different raters or measurement methods give the same value (or close to the same value). For example, if two raters evaluate a patient’s blood pressure, the absolute agreement would assess if both raters report very similar blood pressure values.
• Not Adjusted means no correction is made for systematic differences between raters or instruments. This variation contrasts other forms of ICC that might adjust for variability between raters (e.g., a rater who consistently scores higher or lower).

Practical Applications

1. Medical Diagnostics
   • Imaging Studies: In radiology, multiple radiologists may interpret the same MRI or X-ray images. ICC (Absolute Agreement) assesses if they reach the same conclusion regarding a diagnosis.
   • Clinical Measurements: Nurses or doctors may take blood pressure readings. ICC (Absolute Agreement) would measure whether their readings are closely aligned, ensuring the reliability of the data across healthcare providers.
2. Psychological Testing
   • Behavioral Assessments: When multiple psychologists or therapists rate patient behavior, ICC (Absolute Agreement) can be used to check if they are giving similar scores for the same patient.
   • Cognitive Tests: If different test administrators score subjective tests, such as IQ or personality tests, ICC ensures that the scores are consistent across raters.
3. Sports Science and Kinesiology
   • Physical Performance Measurements: When assessing an athlete’s performance, such as sprint times or heart rate measurements from different instruments, ICC (Absolute Agreement) can verify that the results from multiple devices or observers are reliable.
   • Skill Rating in Sports: Different coaches might evaluate athletes’ skills, and ICC can help confirm that evaluations are consistent.
4. Manufacturing and Quality Control
   • Consistency of Measurements: In quality control, ICC (Absolute Agreement) assesses whether different inspectors or machines produce the same measurements when inspecting the same product.
   • Inter-rater Reliability: It checks if different quality control personnel rate product defects similarly.
5. Education
   • Grading and Assessment: Teachers grading exams or assignments may vary in leniency or strictness. ICC (Absolute Agreement) helps ensure that scores are comparable across different graders, ensuring fairness.

Why Use Absolute Agreement?

The absolute agreement ICC is chosen when the exact or near-exact agreement between measurements or raters is important. This measure is particularly useful when there is no tolerance for systematic differences (e.g., in medical diagnosis or quality control), where minor discrepancies could have serious consequences.

In contrast, consistency models of ICC are used when you are more concerned with the rank or relative standing of subjects rather than exact agreement. Absolute Agreement is thus stricter and more conservative than consistency-based models.

Summary

ICC (Absolute Agreement, Not Adjusted) measures how closely different raters or instruments agree on the same measurements without accounting for systematic differences between raters. It is applied in healthcare, psychology, sports, and manufacturing, where exact agreement is critical for decision-making.

1. Shrout, P. E., & Fleiss, J. L. (1979). Intraclass correlations: uses in assessing rater reliability. Psychological Bulletin, 86(2), 420-428.
2. Koo, T. K., & Li, M. Y. (2016). A guideline of selecting and reporting intraclass correlation coefficients for reliability research. Journal of Chiropractic Medicine, 15(2), 155-163.
3. McGraw, K. O., & Wong, S. P. (1996). Forming inferences about some intraclass correlation coefficients. Psychological Methods, 1(1), 30-46.