3.3 Survey Procedure and Measures

Participants were randomly assigned to view one of the four videos. Roughly half saw an organic spokesperson and half a synthetic spokesperson (286 organic and 275 synthetic). Participants were not informed that the video may be AI-generated, to avoid biasing impressions and in order to test H3. IRB approval was granted with this understanding, due to the minimal realistic harm such a deception would produce. Instead, participants were asked to give their feedback under the guise that they were viewing a possible pitch video for a business expansion opportunity, without being told that such venture was fictitious. Following the viewing, participants answered a series of series of survey questions designed to assess various dimensions of the video’s impact, including persuasive effectiveness and brand impression. Answer options were given on a five-point Likert scale in ascending value from Strongly Disagree to Strongly Agree.
Video effectiveness (H1) was measured by the composite average of three questions. While argumentation is complex, three foundational characteristics were drawn from the literature to provide a composite view of how well the video pitch effectively could move the viewer. After viewing the Plant Haul expansion pitch video, participants rated how strongly they agreed that the pitch was understandable (Clark, 2008; Hoffjann, 2022) and memorable (Duarte, 2012). Participants then directly whether or not they found the proposal convincing.
Brand impression (H2) was likewise measured here as a composite score of three characteristics. Three measures salient in the brand image literature were likewise measured in this study in order to produce a composite brand image (H2) rating. These include how competent (Aaker et al., 2010), trustworthy (Chaudhuri & Holbrook, 2001), and likeable (Nguyen et al., 2013) the brand is perceived to be by respondents. After viewing the video, participants rated how strongly they agreed that the fictitious business, Plant Haul, was competent, trustworthy, and likable.
Participants were then offered the chance to provide open-ended responses on the strengths and weaknesses of the pitch.
After rating the video, participants were asked to note whether the video they viewed was an organic human spokesperson video or an AI-generated avatar video (H3). The survey also collected demographic data, providing an opportunity for a more nuanced analysis of the impact across different viewer segments.