Understanding Influences and Biases in Decision Making: Three Meta-Analyses on Information Processing
Abstract
In this dissertation, I conduct three meta-analyses exploring how consumers process information in the presence of marketing appeals meant to influence their decision-making. In Essay 1, I undertake a meta-analysis of 26 empirical studies yielding 443 sales elasticities to examine how important variables—relating to specific features of the reviews, the websites on which reviews appear, and the nature of the products being reviewed—impact retail sales. Results suggest that online product reviews have a significantly greater influence on sales elasticities when they are delivered by a critic or expert and appear on a third-party, non-seller website. Further, while both review valence and review volume exert an influence on sales elasticities, observations based on review valence have significantly higher sales elasticities than those based on review volume. Interestingly, sales elasticities were statistically invariant to a host of design-related variables, such as publication outlet, sample geography, whether the data was contemporaneous or longitudinal, and if the sample included unreviewed products.
Essay 2 investigates the persuasive impact of goal framing—where the goal of an action or behavior is framed either in terms of a positive (gain) or negative (loss) consequence—on behaviors and evaluations. I conduct a meta-analysis of 178 papers—including 231 studies, containing 806 goal framing effects and 41,502 observations—that examine goal framing effects. A hierarchical linear model (HLM) is used to account for the nested structure of the meta-analytic dataset. Estimation results confirm that overall—across a variety of samples, contexts, and issues—goal framing effects are significant. Furthermore, we find evidence of a negativity bias, wherein loss frames, on average, exert a stronger influence on subjects than do gain frames. Gain frames have greater persuasive impact when argument strength is low and the issue being framed is not health-related, while loss frames are found to be more effective when the recommended behavior is one subjects are likely to approach with vigilance rather than eagerness.
In Essay 3 I meta-analyze 388 effects across 117 studies from 61 separate manuscripts exploring proportional reasoning—a tendency among potential donors to disregard the magnitude of quantitative outcomes when deciding whether to support a charitable cause—in decisions to victims in need of help. I document the presence of a significant proportional reasoning effect, wherein potential donors are more likely to render aid when a larger proportion (smaller number) of victims will benefit from their help. Using construal level theory as a theoretical framework, I show that the proportional reasoning of potential donors (1) intensifies when an appeal provides concrete details about a larger proportion of victims affected by an issue that is physically distant, and (2) diminishes when an appeal provides abstract information about a smaller proportion of victims affected by issues that are temporally and socially close.
Together these essays substantiate the prevalence of biases and heuristics in decision-making, and shed light on how practitioners can craft marketing appeals more effectively to maximize their persuasive impact and affect consumer behavior.