We present the results of an 2x2 factorial design experiment conducted with 182 undergraduate students comparing 4 treatments to teach the seasons: an explanatory text and video, and a refutation text and video. Both videos have the same length, present similar geocentric and allocentric perspectives and use the explanatory and refutation texts as voice-over. Pretest results were similar for the 4 treatment groups. Results of a mixed-design ANOVA show a robust learning effect between the pre- and post-test with a clear advantage for refutation, and a more modest but significant effect of the video format. These results suggest that refutation plays a more decisive role than the presentation medium (i.e., video vs. text) in fostering conceptual learning, but video format helps as well.