Publication: Multimodal critical discourse analysis of advertising persuasion : assumptions and eye tracking evidence
Abstract
Critical Discourse Analysis (CDA) is often hailed for its efficacy of explicating social processes through close examination of discursive practices. However, the causal relationship between discursive practices and social-ideological changes is not always clear. This is certainly the case in the current Multimodal Critical Discourse Analysis (MCDA) of advertising persuasion. Current research, while focusing on the description and interpretation of social semiotic resources used in advertising, often extrapolates advertising persuasion and its social-ideological repercussions from problematic analyst-centered assumptions on viewer perception of ads. The assumptions indicate tensions between analyst-centered descriptive frameworks widely used in MCDA and its explanatory purposes.<br><br>As the first step to ease the tensions, this study is aimed to test some of the assumptions with an eye-tracking experiment. Specifically, this study began with a thorough review of interdisciplinary research into advertising persuasion and formulated three research questions emerging from the review: (1) What do critical viewers see in a print advertisement? (2) What do average viewers see in the same advertisement? (3) Are there differences between the experts’ engagement with the advertisements and that of the average viewers? If so, what are the differences in terms of the common scanpath shared by each type of viewer and the visual perception influenced by the scanpath? To answer the research questions, a match design between-group eye-tracking experiment including two instruction conditions that model critical and ordinary viewers was conducted. The experiment involved 44 Chinese university students, EyeLink© 1000 plus (an eye-tracker), and 10 authentic Chinese print advertisements. In each trial, one ad was randomly presented to a participant who viewed the ad as instructed while the eye-tracker recorded his or her eye movements. Immediately after this, the participant was instructed to verbalize the viewing experience while the reporting was audio-recorded.<br><br>Data collected were analyzed with both quantitative and qualitative analytical components, including the analysis of variances, the scanpath trend analysis, and the AOI-based verbal triangulation. Results from the data analysis revealed differences between the two groups of viewers. To begin with, statistical analysis results indicated that the critical viewers spent significantly more time on ad viewing than the ordinary viewers did. In addition, results from the scanpath trend analysis revealed that the common scanpaths of the critical viewers were consistently longer and covered more details in the ads than those of ordinary viewers. Furthermore, verbal reporting of the critical viewers involved a higher degree of semantic specifications than that of ordinary viewers. Overall, the findings invalidated two MCDA assumptions while supported claims in some social semiotic postulations.<br><br>This study contributes to the emerging multimodal reception research by testing some assumptions of social semiotics. The study also demonstrates the feasibility and potentials of using a data clustering algorithm to discover the common scanpath from eye movement metrics, which is significant for current efforts to expand genre theory to multimodality and develop multimodal analysis frameworks. On a more practical level, the findings can also be used to guide analysts’ descriptions of advertisements in MCDA studies.