In Puentedura's presentation (2006), he asks "what type of things improve performance". This was to try to obtain a 2 standard deviation (SD) improvement for EVERY student, following his presentation of the result of US students in the OECD PISA Report. These tests emphasised performance rather than what courses the students take in each country. Conclusion - Commonwealth countries (and Japan, Korea, China, Finland) do better than average, most of Europe and the USA are in the average category.
He quotes Bloom's work from 1984 into research on how to accomplish improvement - conclusion that one-to-one tutoring by experienced tutors produce a 2 standard deviation gain in learning. But other aspects give improvement too (warning - not additive):
1.00 Feedback-corrective (Mastery Learning)
1.00 Cues and explanations
1.00 Student classroom participation
1.00 Student time on task
1.00 Improved reading/study skill
0.80 Cooperative learning
0.80 Homework (graded)
0.60 Classroom morale
0.60 Initial cognitive prerequisites
0.50 Home environment intervention
0.40 Peer and cross-age remedial tutoring
(Bloom 1984, Walberg 1984)
All these are interesting and I will follow up this work, but, but, Puentedura maintains that Computer use will produce AT LEAST 0.4 SDs and up to 2.0 SD. This work was done in Mathematics.
You have to do things well to achieve these results, thinking in terms of the SAMR and TPCK models.
He highlights four technological avenues to transformation (here transformation being used specifically to produce these types of radical improvements in learning):
Visualisation and Simulation