Personalized learning is the holy grail of education. Imagine if each learner had an educational plan customized exactly to his/her strengths and weaknesses? Back in 1984, Benjamin Bloom (yes, the same one from Bloom’s Taxonomy) published a research paper that showed the benefits of one-on-one tutoring.
Bloom demonstrated that one-on-one tutoring can have a huge effect on leaner achievement. These results have come to be known as the “tutoring effect” or the “two sigma effect,” because his study showed that personally tutored students had a two standard deviation improvement in performance. To some extent this isn’t surprising. A good tutor, working one-on-one with a learner, can pinpoint weaknesses and customize instruction for that learner.
Remember, this was before the widespread use of computers as learning aids, when almost all instruction was instructor-led, group instruction. So, Bloom asked: Are there methods of group instruction that can achieve this level of benefit? He researched a number of different instructional methods and published this paper:
He found that some methods of group instruction were effective (e.g. mastery learning), but none approached the two sigma effect size.
But what Bloom could not know in 1984 was that we would one day have devices that can provide anytime, anywhere, on-demand, one-on-one instruction. In theory, computers can diagnose learner weaknesses and target remediation, just as a human tutor can:
So what are the results? Thus far the results are mixed. Some studies show an improvement in achievement, others not as much. But I think this is to be expected. This technology is new, and most new technologies initially show mixed results. We will improve our methodologies over time, particularly as we incorporate machine learning algorithms into diagnosing learning gaps and recommending remediation.
Will we ever achieve the two sigma effect of human tutors? Hard to say, but we can certainly improve on one-size-fits-all educational experiences.