Ten Common but Dubious Reasons to Use Multimedia Learning

Paul A. Kirschner & Mirjam Neelen

Dick Clark and Dave Feldon (D&D for short) wrote a chapter in the second revised edition of the Cambridge Handbook of Multimedia Learning in which they raise questions with regards to ten commonly vented reasons that people (educators, administrators, policy makers, etc.) give when implementing multimedia in education[1]. They discussed five of these reasons in their chapter in the first edition of the book (2005) on the basis of strong empirical research. In the second edition of the book they do the same for five new positively perceived learning benefits of multimedia (in addition to the original five that were in the first edition of the book). The first five were that multimedia instruction should be used because it:

  1. Leads to more learning. D&D show that there is no convincing nor plausible proof that one medium or a combination of various media, including multimedia, lead to better learning. If there is proof of better learning then this can be explained by “non-media” factors such as instructional theory (e.g., the method / pedagogy), intelligence, social economic status, the ability level of the learner, the quality of the teacher, etcetera.
  2. Is more motivating. It’s possible that multimedia instruction seems more attractive for learners and that they’ll prefer this type of learning approach if available, however this interest doesn’t equal motivation for learning nor does it lead to better learning achievement (also see our blog on effects of motivation on learner achievement).
  3. Uses pedagogical agents that support learning. The idea is that such agents make learning more personal, though what they actually do is provide more common just-in-time scaffolding (e.g., hints, procedures, feedback, instruction). Designing, developing, and implementing such agents requires more time and is thus more expensive. They’re often also, according to D&D, quite silly and learners get bored with them fairly quickly (remember “Clippy”, Microsoft’s paperclip agent?).


Good Old Clippy 🙂

It’s interesting to note that Ruth Clark and Richard Mayer (2016) approach this topic a bit differently. On one hand they state that preliminary results of studies that are trying to identify if agents improve learning suggest that it “might be worthwhile to consider the role of animated pedagogical agents as aids to learning” (p 193). However, they go on to explain that a cartoon character isn’t necessarily better than a real character and that a lifelike image is not always an essential component in an effective agent. What it basically seems to come down to is the importance of a conversational style (not be confused with a “locker room” tone, which is too informal and would distract from learning), which suggests that a conversational style voice over might be enough (assuming that explanation through speech is the most appropriate), which then in turn suggests that an agent might not be required at all‽

  1. Adapts to various learning styles and thus optimises learning for more learners. First, get a piece of chalk or a whiteboard marker, go to the board, and write on the board 100 times: “Learning styles don’t exist. Learning styles don’t exist. Learning styles don’t exist. Learning Styles…. D&D explain that multimedia make it possible to map the differences between learners with regard to what they learn and perhaps also to offer various versions of instructional material. However, there’s no such thing as interaction between multimedia and non-existing learning styles, except for mathemathantic interactions (activities that kill learning – see Clark, 1987 or our blog here).
  2. Facilitates learner-directed, constructivistic, and discovery learning approaches. The question shouldn’t be if multimedia learning facilitates these types of learning, it should be if constructivistic and discovery learning facilitate learning in the first place (also see our blog on inquiry-based learning). See also the following graphs from PISA 2015 where it is clear that on the one hand (see the first figure) teacher directed instruction has a strong positive effect on achievement while on the other hand (see the second figure) enquiry learning has a medium to strong negative effect on achievement:

Pisa 1
Pisa 2


D&D show that guided instruction is more effective (also see Kirschner, Sweller, and Clark, 2006) and that the only types of learners who are good at discovery or inquiry-based learning are experts (also see the next principle).

The next (new) five focus on the expectation that multimedia instruction is good for learning because it provides:

  1. Autonomy and control over the sequence of instruction. Although many multimedia environments allow for this, there are two challenges associated with learner control. First, learners usually don’t have the required knowledge to effectively and efficiently determine their own learning sequence. When allowed this type of control, they usually achieve less. Second, experts would be really good at controlling their own learning sequence but experts usually don’t require instruction (expertise reversal effect).
  2. Higher order thinking skills. There is limited evidence that multimedia instruction does support these, however, even if it does we’re running into the same challenges as with the previous principle. It’s more likely that the instructional approach is supporting the increase in higher order thinking skills and not the multimedia. To cite Dick Clark (1983) on media in education: “Media are mere vehicles that deliver instruction but do not influence student achievement any more than the truck that delivers our groceries cause changes in our nutrition.” In other words, multimedia instruction is only the vehicle, it’s the content (the instructional approach) that does or doesn’t make the difference.
  3. Incidental learning of enriching information. Incidental learning takes place when you learn more than planned for in the instruction. For example, by watching Donald Trump speak you might pick up on some nonsensical government policies and additionally you might “learn” how to berate and belittle an opponent (this could also be a good reason to NOT watch Trump speak, by the way). Multimedia research shows that if you emphasise something in instruction, this will be learned (intentional learning) and all remaining points are at a disadvantage. In other words, the risk of emphasising something is always that only what you emphasise is remembered (i.e., what you intend the learner to learn or the viewer to see) and not the rest (i.e., that which is incidental to what you have emphasised). This goes for all types of instruction and not just for multimedia instruction. Think here of learning prompts, adjunct questions, stated learning aims, etcetera).
  4. Interactivity. This refers to, for example accessing content through hyperlinks, which actually doesn’t add to learning but rather distracts from it. Gaby Solomon called this the Butterfly Defect (fluttering like a butterfly from one piece of seemingly interesting information to another without actually learning). Furthermore, it seems that sometimes proponents of multimedia instruction forget that the instructor is most likely the most interactive medium and an intelligent one as well.
  5. Authentic learning environments and activities. The idea is that the advantage lies in increasing motivation (see the second principle) and/or facilitating transfer. However, according to D&D there is hardly any evidence that this is actually the case.

Just to make it crystal clear: D&D are not against multimedia. They’re only saying that 1) multimedia is not a panacea, 2) multimedia in itself achieve nothing, 3) using multimedia the wrong way does more harm than good, and 4) the selected instructional methodology is and remains the most critical determining factor. And this fourth one is and remains the learning professional’s – such as the teacher’s – expertise!


Clark, R. E. (1983). Reconsidering the research on learning from media. Review of Educational Research. 53, 445-459

Clark, R. C., & Mayer, R. E., (2016). E-learning and the Science of Instruction (4th Ed.). San Franciscio, CA: Pfiefer.

Clark, R. E. (1989) When Teaching Kills Learning: Research on Mathemathantics. In H. Mandl, E. De Corte, N. Bennett, & H. F. Friedrich (Eds.) Learning and instruction. European Research in an International Context. Volume II. Oxford: Pergamon.When Teaching Kills Learning: Types of Mathemathantic Effects. Retrieved from: https://www.researchgate.net/publication/234744652_When_Teaching_Kills_Learning_Types_of_Mathemathantic_Effects

Clark, R. E. & Feldon, D. F. (2014). Ten common but questionable principles of multimedia learning. In R. E. Mayer (Ed.), The Cambridge Handbook of Multimedia Learning (2nd ed.) (pp. 151-173). Cambridge, MA: Cambridge University Press

Kirschner , P. A. , Sweller , J. , & Clark , R. E. ( 2006 ). Why minimal guidance during instruction does not work: An analysis of the failure of constructivist, discovery, problem-based, experiential and inquiry-based teaching. Educational Psychologist, 41, 75–86

[1] This chapter is a progression/extension of a chapter in the first edition, in which D&D discussed five principles. This chapter is accessible through: http://www.davidlewisphd.com/courses/EDD8121/readings/2005-ClarkFeldon.pdf