From A Theory-Based Meta-Analysis of Research on Instruction by RJ Marzano:

The next two techniques displayed in Table 7.2 employed the information processing function of idea representation.  Techniques that provided students with metacognitive strategies for using visual memory had an effect size of 1.04, indicating a percentile gain of 35 points.  Presumably, these strategies help students represent information they are reading in nonlinguistic form…  From these findings, one might infer that idea representation is a key aspect of the reading process…

[R]eading studies [that] addressed techniques that attempt to enhance the idea representation information processing function during reading using pictorial aids…were considered as a group in themselves (as opposed to grouping them with the idea representation techniques in Table 7.2) because they did not employ the metacognitive system.  Rather, they were considered manipulations of the environment designed to stimulate idea presentation in students.  These studies had an average effect size of .46 (n=16; SD=.20), indicating a percentile gain of 28 points.  Table 7.4 displays the differential effects of techniques within this category

I may be misreading this, but what it seems to say is that interventions that use pictures as cues presented to students while reading underperform interventions that teach students to make up their own pictorial representations by a lot (0.46 is a fairly average effect size seen in numerous interventions, whereas 1.0 and above is a fairly rare effect size).

The key seems to be that information organized for the students into pictorial form allows them to disengage from metacognitive strategies, whereas organizing their ideas into pictorial form themselves engages metacognitive processing. These findings are in line with Hattie’s findings on audio-visual aids and the like. 

But I’ve only just skimmed this at this point, and the focus is not higher education here, so one must tread carefully. I’ll come back to this tomorrow I guess. 

This all relates to a crazy idea for the Quantitative Literacy course next semester…