I conducted two second-language word learning studies. Participants were exposed to some training trials and then attempted to chose the right word for a series of videos. The videos were made by us and carefully pre-tested and selected to represent the word meanings at stake. As such, the video stimuli were not a random selection of anything -- unlike the sample of actual words or sentences or sounds, etc., in many psycholinguistic experiments. In fact, in the second study, the word meanings themselves were invented by us, so any other representations of them are purely hypothetical. Do I need to consider the videos a random factor in analysis?