The two unknown components of the reliability equation that can only be estimated, not directly measured, are the true score and the error score. Researchers use statistical methods to estimate these values in an effort to understand the reliability of a measurement in producing consistent, error-free results.
When calculating reliability, the two unknown components that we can only estimate are the true score and the error score. While the observed score is known, it is not possible to directly measure the true score or the error score. Instead, researchers estimate these components using statistical methods. The true score refers to the actual level of the attribute or ability being measured if there were no errors in measurement, and the error score represents the difference between the observed score and the true score due to various factors. Factors contributing to measurement error can include the precision of the measurement tool, the skill of the person making the measurement, and variances in the object being measured.
the reliability of a measurement is a reflection of how consistently it can produce the true score free from random error. To estimate reliability, researchers look at the consistency of scores across items, times, or raters and use statistical methods to estimate the true and error scores that cannot be directly observed.