Abstract:

Introduction:
Instructional videos are a popular means for teaching new concepts and delivering information for learning software tasks (van der Meij & van der Meij, 2013). The growth of instructional videos has been encouraged by the convenience of producing videos rapidly.
Cognitive load theory (CLT) is a theoretical framework linking instructional design principles with concepts of long‐term and working memory (Sweller et al., 1998). One of the interventions from this framework is the expertise reversal effect (Kalyuga, 2007), which postulates that a learner's existing cognitive resources can influence the efficacy of instructional techniques. CLT proposes that there are three components that influence the cognitive difficulty of a task. Intrinsic load (ICL) is the material’s essential difficulty and is increased by element interactivity. Extraneous load (ECL) is caused by inessential pieces of information that interfere with the learning process. Germane cognitive load (GCL) is the result of the constructive method of handling information, in a way that contributes to learning. This study aspires to fill this gap in the literature by examining the effects of learner expertise on the three types of cognitive load through instructional videos.

Methodology and Participants:
The participants consisted of 31 students (male:18, female:13) of a physics department from a Greek university. The students answered a pretest with ten multiple choice items that measures students’ expertise in python programming language. Regarding the results of the pretest, we allocated students into three conditions: low expertise (n= 10), medium expertise (n=14) and high expertise (n= 7).

Materials:
Six long instructional videos (40 min) were developed and covered main aspects on python programming language. A Demonstration-Based Training (DBT) approach was followed to create the videos (Brar & van der Meij, 2017).
The Cognitive Load Questionnaire (Klepsch et al. 2017) was a subjective scale that concentrates on ICL, ECL and GCL. Answers are given on a 7-point Likert scale which range from absolutely false (1) to absolutely true (7). Reliability analyses for the three measures were ICL α = 0.80, ECL α = 0.83, and GCL α=0.80.

Results and Discussion:
A one-way ANOVA revealed a significant effect of the learner expertise on ICL, F(2,30) = 3.55, p = .02. Comparisons between conditions showed the high expertise group indicated significantly lower ICL than both medium expertise and low expertise groups. This finding aligns with CLT (Sweller, 2020), as it postulates that novices need to invest more mental effort to construct their mental representations in long-term memory.
A one-way ANOVA revealed a significant effect of the learner expertise on ECL, F(2,30) = 4.65, p = .018. Tukey’s HSD post hoc test showed the low expertise group indicated higher ECL than both medium expertise and high expertise groups. A possible explanation may lie in the instructional design principles that were used. A holistic adoption of DBT in conjunction with CTML principles may influence differently ECL (Mayer et al., 2020).
A one-way ANOVA revealed a significant effect of the learner expertise on GCL, F(2,30) = 3.97, p = .03. Tukey’s HSD post hoc test showed the low expertise group indicated higher GCL than both medium expertise and high expertise groups. This finding is in line with CLT (Sweller, 2020) as experts find the process easier for them.

Keywords:

Multimedia learning, instructional videos, cognitive load.