Noise refers to the undesirable variability in judgments. It can be broken down into three components: level noise, pattern noise, and occasion noise. Level noise occurs when there are differences in the average level of judgments made by different judges. Pattern noise arises from variability in judges' responses to particular cases, while occasion noise is the random variability within a judge's judgments across different occasions.

One study that illustrates the concept of system noise is the noise audit of criminal sentencing conducted by Bartolomeo et al. In this study, federal judges were asked to evaluate 16 cases and assign sentences. The results showed significant variability in the sentences assigned by different judges for the same case, indicating the presence of system noise.

Another example comes from the field of performance evaluation in organizations. Research has shown that 70-80% of the variance in job performance ratings can be attributed to system noise, with level noise and pattern noise being the main contributors. This variability can lead to inconsistencies in performance evaluations and ultimately impact employee morale and productivity.

In medicine, system noise can have serious consequences. For instance, when two doctors provide different diagnoses for the same patient, at least one of them is wrong. This variability in judgments can lead to incorrect treatments and potentially harm the patient.

Understanding and addressing the components of system noise is crucial for improving the quality of judgments in various fields. Reducing level noise can be achieved by ensuring that judges maintain uniformity in their use of cues, weights assigned to cues, and the use of the scale. Tackling pattern noise, on the other hand, requires addressing the stable individual differences that contribute to variability in judgments. By focusing on these components, organizations and professionals can work towards reducing system noise and improving the accuracy and consistency of their judgments.