Why prior knowledge could change what AR does to a learner
Cognitive load is the mental effort a task demands. If a learner already knows the basics, an AR overlay may mostly confirm what they expect; if they do not, the same overlay can add things to process on top of the experiment itself. A thematic review of extended reality in education, including AR, notes that learning objectives and learner characteristics can influence cognitive burden, and that burden depends on how the technology is matched to the task [2]. That is a general statement about design fit, not a test of prior knowledge in a lab.
The engineering-education study gives a related clue. Students using 3D AR with hints learned more than those using AR alone or conventional instruction, with a large effect for AR plus hints and a moderate effect for AR alone [1]. Hints are not the same as prior knowledge, but they act like a substitute for it: they supply the guidance a well-prepared learner might not need. The study did not measure learner prerequisites, so it cannot show that weaker students benefited more.
The lab study measured readiness, not cognitive load
The one study here set in an AR-supported laboratory was a hydropower activity using rule-based scaffolding, which means the AR gave students structured prompts tied to steps in the exercise [5]. It measured students' digital readiness, their perceptions of AR quality, and their engagement. It did not measure cognitive load or learning outcomes, and the authors state they cannot conclude that cognitive load was reduced [5].
That gap matters for the question. Digital readiness is a learner prerequisite of a sort, but the study does not report whether it changed engagement differently for more and less prepared students, nor whether it interacted with cognitive load. So it supports only a narrow point: AR lab activities can be studied with learner characteristics in view, but this one does not establish moderation.
Support tools can lower load, but that is not the same as moderation
A controlled experiment with high school students found that giving learners a rubric for self-assessing scientific abstracts improved judgment accuracy and reduced cognitive load during self-assessment [3]. This shows that a structured support can make a demanding task feel lighter. It does not involve AR or laboratory work, and it did not test whether the benefit was larger for students with weaker prior skills.
A broader analysis of extended reality in education cautions that some design factors do not improve learning outcomes directly, and it frames cognitive load theory as central to when learning is effective [4]. Read together with the AR-with-hints result [1], the pattern is consistent with support reducing load and improving outcomes, but none of these studies isolates learner prerequisites as the thing that changes the size of that effect.
What you can reasonably conclude, and what would settle it
On the current evidence, the honest answer is that the moderating role of learner prerequisites in AR lab work is untested. The strongest adjacent finding is that scaffolding in AR was associated with better learning and that cognitive load related negatively to performance [1], and a review suggests learner characteristics belong in the design equation [2]. But the lab-based AR study explicitly did not measure cognitive load [5], and the rubric study, while showing reduced load, was not AR and not a lab [3].
To answer the question properly, a study would need to measure prior knowledge or skill before an AR lab task, measure cognitive load during it, and test whether the load–learning relationship differs across prerequisite levels. The existing work points toward that design rather than delivering its result.
Sources used in this answer
Enhancing Cognitive Load, Collaboration Tendency, and Enjoyment Through 3D Augmented Reality in Engineering Education
In engineering education, 3D AR with hints produced the largest learning gains, cognitive load negatively predicted performance, and enjoyment and collaboration tendency positively predicted it, but learner prerequisites were not examined and the setting was not laboratory work.
Presence, cognitive load, and learning in extended reality: A thematic analytical review
A thematic review of extended reality in education indicates that learning objectives and learner characteristics can influence cognitive burden depending on how technology is matched to task requirements.
Rubrics enhance accuracy and reduce cognitive load in self-assessment
High school students given a rubric for self-assessing scientific abstracts showed more accurate self-judgments and lower cognitive load during self-assessment than students without one.
Extended Reality in Education: Analysis of Benefits, Risks and Design Challenges
An analysis of extended reality in education notes that some design factors do not improve learning outcomes directly and highlights cognitive load theory as central to effective learning.
Embedding Rule‐Based Augmented Reality Scaffolding in Experiential Learning Cycles: An SEM Study of Student Engagement in Lab‐Based STEM
An AR-supported hydropower laboratory activity with rule-based scaffolding measured digital readiness, AR quality perceptions, and engagement, but did not measure cognitive load or learning outcomes and could not conclude that cognitive load was reduced.