Abstract:
Educational videos are a central component in formal and informal learning contexts. When optimising the design of such educational videos, the question arises whether an instructor presenting the content should be visible next to the learning content. This often involves so-called talking heads, where the head of the instructor speaking into the camera is visible in one corner of the educational video. Theories such as the social agency theory and the cognitive theory of multimedia learning (e.g., Mayer, 2021) suggest both potential advantages (social presence, deeper processing) and disadvantages (distraction, cognitive load) of a visible instructor. For the talking head format in particular, it is questionable whether the potential disadvantages outweigh the advantages because talking heads attract visual attention but offer fewer advantages than an instructor’s full body representation (e.g., due to the lack of gestures). Research concerning the talking head format is still limited, especially with regard to measures beyond learning outcomes.
The aim of this thesis was to expand the findings on the talking head format and at the same time to investigate whether certain characteristics of the learning material (e.g., the type of content alongside the talking head) represent boundary conditions for the use of talking heads. Four experiments with a total of N = 488 participants served as the empirical basis for the dissertation, which investigated the effects of talking heads in educational videos with narrated slides. The investigations included a variety of dependent variables, for instance different learning outcomes, eye movements, cognitive load, social presence, and satisfaction ratings. Three experiments were conducted online, and one experiment took place in the laboratory using eye tracking. In all experiments, the presence of a talking head in the educational videos was varied (either as a between- or within-subjects factor), with Experiments 1a, 1b, and 2 additionally varying the slide type (graphic- vs. text-based slides) and Experiments 1b and 2 additionally varying the presentation type of the content (sequential step-by-step presentation or static all at once presentation). The results largely showed no effects of the talking head on learning outcomes (Experiments 1a, 1b, and 3), although in one investigation (Experiment 2) there was a small detrimental talking head effect on learning outcomes. With regard to learners’ ratings, the findings were heterogeneous, with Experiment 2 showing strong positive effects on different ratings (e.g., satisfaction), whereas such effects were mostly absent in the remaining experiments (Experiments 1a, 1b, and 3). However, at the level of process data measured with learners’ eye movements, there was a strong distraction from the learning content due to the talking head (Experiment 3). To further substantiate the findings, internal meta-analyses across the four experiments were conducted on the key dependent variables captured in all experiments. Analyses revealed that the talking head had no overall effect on
learning outcomes across all experiments, but enhanced learners’ ratings of satisfaction and decreased their self-reported extraneous cognitive load. Taken together, considering the findings of the internal meta-analyses, the advantages outweigh the disadvantages, given that a talking head can improve individuals’ learning experience.