Connecting dots: Historical perspective
Starting from my opposition to the extreme reliance of tests, I am starting to look into the history of education, rather schooling (focus on US where I am right now). Also wondering the impact of data mining and machine learning as they become more and more common.
The advantage of testing, as many proponents of testing point out is that it enables policy makers to evaluate schools, teachers, curriculum...most importantly student progress. What other method is as scalable, reliable and non-discriminatory. It is a fact that standardised testing has shed light on student achievement between ethnic groups, economically disparate groups, differently abled students etc. But like any other facet of life, too much is not too good.
Schools are already collecting the test data and mining it. This is another great opportunity. Machine learning can allow early intervention, highlight problems, guide learning outside classroom etc. But again, blind over-reliance to punish schools, teachers or worse students based on numbers is very likely. Something policy makers should watch out for. What if a machine predicts one student is more likely to get the required grades than another based only on numerical indicators? Is education, or schooling, moving in this direction?
But to provide an answer to this question, it is very important to understand how schooling fits into the education system historically.
More posts to follow to elaborate these topics.
The advantage of testing, as many proponents of testing point out is that it enables policy makers to evaluate schools, teachers, curriculum...most importantly student progress. What other method is as scalable, reliable and non-discriminatory. It is a fact that standardised testing has shed light on student achievement between ethnic groups, economically disparate groups, differently abled students etc. But like any other facet of life, too much is not too good.
Schools are already collecting the test data and mining it. This is another great opportunity. Machine learning can allow early intervention, highlight problems, guide learning outside classroom etc. But again, blind over-reliance to punish schools, teachers or worse students based on numbers is very likely. Something policy makers should watch out for. What if a machine predicts one student is more likely to get the required grades than another based only on numerical indicators? Is education, or schooling, moving in this direction?
But to provide an answer to this question, it is very important to understand how schooling fits into the education system historically.
More posts to follow to elaborate these topics.
Comments
Post a Comment