Teacher Assessment vs. Standardized Testing

Teachers’ evaluations of student performance can be as reliable as standardized tests scores and help avoid many of the negative consequences of the “exam culture”.

Many countries around the world make extensive, and sometimes exclusive, use of high-stakes examinations to determine entry to the next stages of education. The rationale behind such systems is that standardized tests are thought to be the best (most fair and reliable) way to predict and compare students’ academic ability.

However, according to the authors of a recent study on the subject “there are downsides of testing that need to be considered, especially if the exams are very high stakes” (all quotes are from the original article) :

• “High-stakes exams could be detrimental to the learning process if teachers ‘teach to the test’ in a manner that undermines the accumulation of knowledge and understanding of subjects.”

 

• “Exams are associated with anxiety and distress, and research has shown that high test anxiety is negatively associated with achievement… High-stake exams also have an effect on teachers’ morale as well as pupils’ wellbeing and mental health.”

For those reasons, the researchers set out to investigate the “added benefit of standardized tests over teacher assessments in predicting educational achievement.” To do so, the team analyzed a large amount of data from the UK National Pupil Database and compared teacher assessments and test scores, over the course of their education (from age 7 to 18), of thousands of students.

One of the main findings of the study is that “teacher assessments and exam scores are highly correlated both contemporanously and over time.” Consequently, standardized test scores actually add very little, in terms of predicting power, to teacher evaluation: “Earlier teacher ratings provided 90% of the combined power of teacher ratings and test scores to predict educational outcomes. After controlling for teacher ratings, test scores explained on average 10% of the variance in GCSE, 5% A-level and 3% university entrance exam scores.” Such percentages are arguably very small, considering that “the deck is stacked against teacher assessments in these analyses because the educational outcomes we considered (GCSE and A-level outcomes) are themselves test scores.”

According to the authors, this “raise(s) questions about the value of the testing culture that characterizes compulsory education in the UK” and elsewhere. They are “not suggesting that students’ progress should not be monitored or arguing against testing in general. For example, it is possible that in an increasingly technological society light-touch frequent testing can aid learning. However, we question the value of high-stakes exams so early and regularly in children’s lives.”

Indeed, teacher evaluation comes with a number of benefits:

• “Teacher ratings assess performance over time, not just a few critical and highly stressful days of testing.”

 

• “Ongoing teacher monitoring of student performance is essential for effective teaching.

• “Teacher ratings themselves are to some extent based on in-class tests, which promote consolidation of learning, especially when done regularly and frequently (unlike a single high stakes examination).

Of course, a system based on teacher assessments comes with its own limitations:

• “Teacher assessments do not always accurately reflect students’ ability because of biases and stereotypes held by teachers”

• The need for an “accountability system for monitoring schools and teachers” is “more challenging without external tests".

According to the authors, the first issue can be addressed by standardizing teacher evaluations. Thus, in the UK, teachers follow strict guidelines to assess students’ performance on clearly identified items. As for the second issue, it could potentially be addressed through "random" testing.

Reference: Rimfeld, Malanchini, Hannigan, Dale, Allen, Hart, and Plomin (2019), “Teacher assessments during compulsory education are as reliable, stable, and heritable as standardized test scores”, The Journal of Child Psychology and Psychiatry.