Context and present study:
Students are increasingly relying on videos from platforms for educational or informative purposes. However, these videos may be misleading, and students can have difficulties in evaluating the quality of information and sources. Particularly, it is challenging for them to assess the competence and trustworthiness of sources accurately. Developing students' skills for sourcing—i.e., identifying and representing —a source's characteristics to predict, interpret, and evaluate its content and relevance - in videos is thus a major educational challenge.
Several studies have focused on text evaluation, but few have addressed video formats. Our study aims to evaluate the impact of a teaching activity on the development of video sourcing among middle- and high-school students.

Méthod:
This study involved a total of 193 students from 6 to 12 grades. The mean age was 13.7 years (SD = 2.3) and 118 students identified themselves as girls, 67 as boys, and 8 unspecified in gender. For their recruitment, 16 teachers implemented the intervention with at least one of their classes.
The study consisted of five phases: a pretest, three teaching activities and a posttest. Each teaching activity addressed a different source evaluation criterion: credibility, expertise, and trustworthiness. Each session followed the very same structure : two videos introducing one of the criteria, three videos related to the teacher’s topic, varying in source credibility (reliable, unreliable, or questionable). Finally, students selected the video they thought was the most credible.
Students’ skills to evaluate sources in videos were measured using comparable pre- and post-tests administered before and after teaching activities. We measured students' assessment of source credibility for four videos: two with experts and two with non-expert sources. Firstly, students assessed the competence and trustworthiness of each source on a 9-point Likert scale.
We performed the first repeated measures ANOVA with the evaluation expertise score as a dependent variable, source (expert vs. non-expert) and time of assessment (pretest vs. posttest) as repeated measures factors, then school level as a between-subject factor.

Results:
The teaching activities had differential effects on the evaluation expertise scores on school level. The ANOVA indicated a significant interaction effect between source, time of assessment, and school level variables (F(1, 191) = 4.297, p = .004 ; η²p =.022). For middle-school students, post hoc tests revealed no significant difference between pre- and post-test scores for lay people or experts. However, high school students evaluated laypeople source more favourably in pre-intervention than in post-intervention (pbonferroni = .001, Mdifference = 0.831, SE = .205) while evaluation for expert sources remained consistent (pbonferroni = 1.000, Mdifference = 0.140, SE = .205).

Discussion:
Our study showed that the teaching activities influenced students' source evaluation skills. These results suggest that they fostered a more critical evaluation of videos that differed according to the students' grade level. Teaching activities seem to have benefited high-school students. The latter adopted a more discerning, criterion-based judgement. They demonstrated a more sophisticated use of the evaluation criteria, with the expertise of the source emerging as a key evaluation criterion.

Keywords: Teaching activities, middle school students, high school students, source evaluation, videos.