In the comments, Christian Bokhove provided a link to the working paper that the report drew upon. A number of things stood out.
For instance, on page 74 there is a box explaining fully how “index of memorisation” was calculated. After reading the “Ten questions…” report, I was concerned about what the alternatives were to the memorisation responses and I was right to be. I think this is an extremely weak construct.
The most stunning thing that I have found, however, is PISA’s definition of good teaching:
“In its Analytical Framework (OECD, 2013), PISA defines the three dimensions of good teaching as: clear, well-structured classroom management; supportive, student-oriented classroom climate; and cognitive activation with challenging content (Klieme et al, 2009; Baumert et al, 2010; Lipowsky et al, 2009; Kunter et al 2008).” My emphasis.
It may now seem less surprising that PISA haven’t highlight the following relationship in their own data, even though they have pointed out weaker correlations in the same data set: