4. Method
4.1. Participants and design
The experiment was conducted with 127 participants: 64 Ger-man college students and 63 participants in continuing education (81 females, 46 males; mean age = 30.69; SD = 12.73).1 They re-ceived either a personalized or formal version of a computer-based program explaining how human visual perception works. The partic-ipants were randomly assigned to one of four experimental condi-tions. Table 2 shows the 2 (formal or personalized) 2 (college students or continuing education participants) between-subjects factorial design.
The required number of participants was calculated using the computer software G*Power (Cunningham & McCrum-Gardner,
2007). Based on the literature review on personalization effects in multimedia learning, a large effect size was estimated for trans-fer. The a priori power analysis (a = .05, f = .40, F-test, one-sided) revealed a sample size of at least N = 112 (1 b = .95). The exper-iments in this study were conducted with a total sample of 127 participants.
Stiller and Jedlicka (2010) found that prior knowledge may influence personalization effects. Hence, before the learning phase, the participants of the presented study estimated their prior knowledge on a five-point rating scale (1 ‘very low’ to 5 ‘very high’). To control the distribution of the prior knowledge across the four groups, a Kruskal–Wallis test was employed (v2 = 6.534, p = .088). The test revealed an equal distribution of prior knowl-edge across the groups.
The continuing education participants attended Microsoft Office courses. Since these were beginner courses, it can be assumed that they have a rather limited knowledge regarding computer use. In contrast, it is assumed that college students are used to working with computers. For example, several university lectures are car-ried out as online or blended-learning courses.