The relationship between grading leniency and grading reliability
基于五万多个课程数据，发现评分越宽松，评分可靠性越低，尤其在宽松评分课程中更明显。
Grading reliability for a class can be measured as the correlation between the students’ grade point average (GPA) and the grades they received in that class. Grading leniency can be measured as the difference between the average grade for the class and the students’ average GPA. Data from more than fifty-thousand course sections show that lenient grading is associated with lower grading reliability. This association is particularly strong among course sections with lenient grading. Although grading leniency may be a symptom rather than a cause of low-grading reliability, reducing grading leniency may lead to improved grading reliability.