Study of personalised learning shows promising evidence – or does it?

The US National Education Policy Center’s Think Twice Think Tank Review Project recently reviewed a RAND study on personalised learning. The RAND study examined the effects of three school-wide personalised learning initiatives on pupil achievement to try to find evidence linking specific learning strategies to achievement outcomes.

RAND defined “personalised learning” (PL) as incorporating five specific characteristics including data-supported pupil goals accessible to teachers and pupils, and personalised learning of each pupil’s choice with in-school support and learning outside school.

The RAND study compared the MAP reading and maths scores of 11,000 pupils in 62 schools who had been using a personalised learning approach for two years to the scores of pupils matched at baseline to serve as a comparison. Researchers found higher achievement scores for the PL group, especially at primary age. In addition, the study showed that personalised learners’ scores increased at a greater rate than the nation’s scores. Overall, researchers deemed PL promising practice.

However, the Think Twice Think Tank Review Project disagrees. In a review of the study researchers felt that the study’s limitations prevented it from demonstrating true evidence of promising practice. First, reviewers noted that only pupil involvement in analysing their own data and goal setting was associated with consistent gains. They pointed out that two of the attributes ascribed to the success of PL – flexible learning environments and student grouping – were also used in schools not using PL. They noted that the largest departure from usual classroom practice (competency-based progression) was not used in the majority of the experimental schools, casting doubt on its pertinence. In addition, reviewers were dubious about the generalisability of the findings because 90% of the study schools were charter schools.

Think Tank reviewers concluded that the study suggests there may indeed be personalised practices associated with test score gains, but that the practices in the three experimental models weren’t drastically different than practices in untreated schools. The study’s limitations cast doubt on the models’ generalisation.

Source: Continued Progress Promising Evidence on Personalized Learning (2015), RAND Corporation, and Review of Continued Progress: Promising Evidence on Personalized Learning (2016), NEPC.

Leave a Reply

Your email address will not be published. Required fields are marked *