percentCorrect

Percentcorrect is a measure of accuracy that expresses the share of correct responses in a set of attempts as a percentage. It is equivalent to the concept of accuracy or proportion correct in many contexts.

Calculation is straightforward: percentcorrect = (number of correct responses / total number of responses) × 100. When some data are missing, the denominator can be adjusted to reflect only non-missing responses. In software and data analysis, percentCorrect is often stored as a field name representing this value for an individual or a group.

Percentcorrect is widely used across fields such as education, psychology, cognitive science, and machine learning. In education, it summarizes how well a student performed on a test or assignment. In machine learning, classification accuracy equals the percentcorrect on a held-out test set, indicating the proportion of correctly labeled instances.

Limitations should be considered. Percentcorrect does not account for guessing, especially in multiple‑choice tasks, where chance performance can inflate the value. It also does not account for class imbalance or the relative importance of different errors. When these issues are relevant, practitioners may use chance-adjusted metrics like Cohen’s kappa, or consider other evaluation measures such as precision, recall, F1-score, or the confusion matrix.

Example: if a person answers 20 of 25 questions correctly, the percentcorrect is (20/25) × 100 = 80%.

See also: accuracy, confusion matrix, Cohen’s kappa, precision, recall, F1-score.