Holliday, W. G. (1973). Critical analysis of pictorial research related to science education. Science Education, 57(2), 201-214.

Conclusion. Analysis of pictures and their instructional utility has produced a few interesting results and many more potentially-fruitful areas for further experimentation. For example, most media textbooks do not seem to distinguish between the mechanisms for verbal and pictorial concept formation. The usefulness of pictures in science instructional materials is an important, complex issue. Like so many other instructional variables, there is no simple answer to many of the questions that have been raised regarding their role. For example, color, a multi-dimensional variable, should be examined for specific kinds of cognitive and affective grounds. However, there exists conflicting evidence of support for many of the hypotheses posed by these arguments. This area of instructional design is in its infancy. The potential benefits inherent in visuals to science education intuitively are abundant. Varying information contexts need to be analyzed systematically in terms of the stimuli presented to the perceiver, the responses required and the hypothesized intervening mechanisms occurring within the learner. This approach to the problem is more likely to produce a theoretical base from which relatively more accurate predictions will be possible.