There has been a steady increase in the use of video-based instruction in secondary and tertiary education over the past two decades, and a rapid acceleration in implementation during the COVID-19 pandemic. As the world moves towards a new normal, there is a need to focus on developing high quality video-based resources as research indicates not all video-based instruction is equally effective at supporting learning. Previous studies have found video-based instruction can be improved by minimising extraneous cognitive load, and by fostering a social partnership between the instructor and the learner (social agency). This study was guided by a pragmatist philosophical paradigm and employed a convergent QUAN + QUAL mixed methods research design to investigate how instructor visibility and familiarity influences cognitive load, social agency and learning gain in a high school science context.

The first phase of the study, the instructor visibility study, involved 116, Year 9 and 10 students viewing a video lesson on lightning formation where the instructor was visible (transparent whiteboard), or not visible (digital drawing board). The second phase of the study, the instructor familiarity study, involved 147, Year 9 and 10 students viewing a digital drawing board video lesson on lightning formation where the instructor was either familiar (participant's science teacher) or unfamiliar (science teacher from another school). Learning gain was measured using a pretest-posttest instrument, and cognitive load and social agency were measured using closed-ended Likert-type scale instruments, open response questions, and focus group interviews.

Critical analysis of the instructor visibility study supported five findings: (1) There was no statistically significant difference in learning gain between the instructor visibility conditions (p = .65); (2) Extraneous cognitive load was statistically significantly lower when the instructor was visible (p = .02) and the effect size was small (r = .21); (3) Germane cognitive load was statistically significantly higher when the instructor was visible (p < .05) and the effect size was small (r = .19); (4) Participants rated the instructor's teaching style statistically significantly more engaging when the instructor was visible (p = .04) and the effect size was small (r=.19); and (5) participants reported instructor visibility provided attentional cues through pointing gestures, comprehension cues through iconic gestures and facial expressions and social cues engaged and motivated the learner. Critical analysis of the instructor familiarity study supported four findings: (1) There was no statistically significant difference in learning gain between instructor familiarity conditions (p = .83); (2) There was no statistically significant difference in cognitive load between the instructor familiarity conditions (intrinsic p = .08; extraneous p = .44; germane p = .23); (3) Participants rated measures of social agency statistically significantly higher when the instructor was familiar (instructors' teaching style more engaging (p = .02); instructor working with me (p < .02)); and (4) Participants reported instructor familiarity provided attentional cues due to familiarity with the instructor's voice and teaching style and established credibility, and provided social cues that engaged and motivating the learner.

This study adds to the extant body of research on instructor visibility in videos in an authentic learning setting by demonstrating extraneous cognitive load was reduced when the instructor was visible in the transparent whiteboard video. In addition, this is one of the first empirical studies investigating instructor familiarity in video-based instruction. The thesis outlines implications for practice and directions for future research on instructor visibility and familiarity in video-based instruction.