I am curious about the poor student achievement outcomes associated with student control over learning.

Student control over learning has multiple meanings.  In one meaning, we want students to know what to do when they do not know what to do; we want them to become their own teachers.  In another meaning, we have to teach them to acquire these skills, and in turn this requires a deeper understanding of the content to which they need to know and understand to have appropriate control over learning. In another meaning, it means we give students control over their next learning steps – and here is where the notion fails.  For most of us, when we reach the edge of what we know and can do, we need expertise to optimally help us work out the next steps. There are too many blind alleys in next steps and it is not efficient and, in some cases, can turn students off the challenge of learning, if we make them take too many blind alleys. While many teachers use the notion ‘we want students to have control over their learning’ they usually do this by offering (few) choices-- what do you think about doing this or that? Would you rather lose your left or right arm – your choice.  And this smaller selection of options avoids too many blond alleys but enables students to make judgement about next steps and thus learn some of the skills of being their own teacher. Karich, Burns, & Maki (2014) investigated student control within educational technology.  Their meta-analysis showed that ‘including learner control within educational technology was almost zero (g= 0.05), and were also near zero when examining most characteristics of control and classroom contextual factors.’  ‘There does not seem to be an advantage to give the learning control over any particular instructional component’ (p 404).

Carolan, Hutchins, & Wickens (2014) also found negligible effects.  They concluded on average, training designs with more learner freedom have not translated to more effective transfer performance than training with less learner freedom.  They reasoned that most learning tasks affect the level of working memory required for processing. As complexity increases, this intrinsic load can increase, requiring management through appropriate instructional strategies. The instructional environment and methods can impose additional cognitive load during training, for example, the control decisions and search requirements of active learners, that is extraneous to the task being learned (e.g., Kirschner et al., 2006; Tuovinen & Sweller, 1999). This extraneous load limits the resources available for active cognitive processing and schema construction that are germane to meaningful learning (Kalyuga, 2011; Sweller, 2010).

Extraneous cognitive load can affect executive functions such as attention to self-monitoring, and self-assessment for LC decisions while also engaging in the learning task (Kostons, van Gog, & Paas, 2010).  There was a slightly higher effect on process than knowledge tasks. In Visible Learning (2009), I noted: The effects of student choice and control over learning is somewhat higher on motivation outcomes (d = 0.30) than on subsequent student learning (d = 0.04, (Niemiec, Sikorski, & Walberg, 1996; Patall, Cooper, & Robinson, 2008). Indeed, the more instructionally irrelevant choices had higher outcomes (e.g., color of pen to use, what music to listen to when learning). Such irrelevant choices are less effortful, not particularly important or having major consequences on the learning, and too many choices may be over whelming” (John Hattie, personal communication, June 20, 2018).