How do we Optimize Message Matching Interventions? Identifying Matching Thresholds, and Simultaneously Matching to Multiple Characteristics

Abstract

Matching the content of persuasive messages to the characteristics (e.g., motives, personality) of people receiving them is an effective technique to improve persuasion. However, little is known about how to optimize matching beyond simply using the technique. We propose that matching interventions can be strengthened by matching messages to multiple characteristics at a time, and we introduce the concept of matching thresholds to improve the way interventions assign messages. Matching thresholds are defined as the points along characteristics where people change from being most responsive to one message type to another. We provide statistical and methodological tools to estimate matching thresholds, and evaluate these tools in two simulation studies. We then report an online experiment (N=568) where we find an advantage for simultaneously matching messages to two characteristics (promotion focus and interdependent self-construal) and provide estimates of the matching thresholds to guide the assignment of gain/loss frames, and independence/interdependence appeals.Keywords: message matching; message tailoring; persuasion; matching thresholds; independent and interdependent self-construal; promotion and prevention focus OPTIMIZING MESSAGE MATCHING 3 How do we Optimize Message Matching Interventions? Identifying Matching Thresholds, and Simultaneously Matching to Multiple CharacteristicsResearch on persuasion has a long-standing focus on identifying and optimizing strategies to improve persuasive communication, and one of the most well-documented strategies for increasing persuasion is message matching: altering the content of a message to match the characteristics, needs, concerns, and/or preferences of the individuals to whom they are delivered (Carpenter, 2012;Lavine & Snyder, 2000; Maio & Olsen, 2000;Noar, Benac, & Harris, 2007).For example, messages emphasizing the career outcomes of volunteerism are more persuasive for individuals highly motivated by career concerns, whereas messages emphasizing interpersonal outcomes are more persuasive for those highly motivated by social concerns (Clary, Snyder, Ridge, Miene, & Haugen, 1994;Clary et al., 1998;Joyal-Desmarais & Snyder, 2016). Message matching ensures people receive only the most personally relevant information, and has been used to influence many types of behaviors, including physical health (Gallagher & Updegraff, 2012;Noar et al., 2007), mental health (Lueck, 2018), consumer (Snyder & DeBono 1985), prosocial (Clary et al., 1994;1998), pro-environmental (Tangari & Smith, 2012, and political behaviors (Bertolotti & Catellani, 2014;Voelkel & Feinberg, 2018). However, beyond the general idea that matched messages are more persuasive, principles for how to design matched messages to maximize their effectiveness have yet to be specified (Updegraff & Rothman, 2013).We address this limitation by extending the message matching literature in several ways.First, we demonstrate that matching a message to multiple personality characteristics at once can provid...

Appealing to motivation to change attitudes, intentions, and behavior: A systematic review and meta-analysis of 702 experimental tests of the effects of motivational message matching on persuasion.

“…The electronic search was developed in consultation with an information specialist; it used a large set of terms that describe the message matching phenomenon across different traditions of research (e.g., including variants of “message matching,” “functional matching,” “attitude functions,” “framing,” “tailored communication,” “targeting,” “congruency,” “personalization,” “message fit,” “individualization”), along with terms tied to specific forms of message matching (e.g., “gain-frame,” “loss-frame,” “cultural appeal,” “value-expressive congruence”). Before conducting our review, we evaluated our search terms using a set of 60 empirical publications on message matching and found the search to identify 82% of these publications (see Joyal-Desmarais, 2020, for details); given the scope of this area of research and the lack of standardized terminology across studies, this coverage rate was considered indicative of a good sensitivity–specificity trade-off. The backward citation search made use of 81 key sources reviewing message matching effects (e.g., narrative reviews, systematic reviews, meta-analyses, chapters), and the forward citation search used the same 81 sources along with 33 influential and/or foundational reports of empirical studies on message matching.…”

“…Good reliability was established as a percent agreement of at least 80%, r of at least .80, and an ICC of at least .80 (Belur et al, 2021; Neuendorf, 2002; Syed & Nelson, 2015). An in-depth report of our interrater reliability analyses is provided in Joyal-Desmarais (2020), which includes analyses by coder and for each variable. Average percent agreement was 95.3% across variables, and the r s and ICCs for continuous variables were always above .80 (average: r = .98, ICC = .97).…”

“…Coders were also provided with a coding dictionary that supplied definitions for each element contained in the codebook. The codebook and dictionary are provided in Joyal-Desmarais (2020) and on our project page.…”

“…Perhaps because research centered on motivational matching, message framing, message tailoring, and context matching have each used distinct terminology and different approaches to designing message-based interventions (Joyal-Desmarais, 2020), reviews of empirical research on message matching have tended to follow these demarcations closely. For example, reviews by Krebs et al (2010), Noar et al (2007), and Sohl and Moyer (2007) are explicitly concerned with message tailoring; reviews by Carpenter (2012), Lavine and Snyder (2000), and Shavitt (1990) focus on “functional” (i.e., motivational) matching; and reviews by Gallagher and Updegraff (2012), O’Keefe and Jensen (2006), and Xu and Huang (2020) are focused on message framing.…”

“…Most motivational matching research adheres to a similar notion of what constitutes a matched message : Messages that are aligned with the characteristic(s) to which they are matched. However, there is markedly less clarity on what constitutes a message that is not matched (Joyal-Desmarais, 2020; Rothman et al, 2020). For instance, some studies compare matched messages to messages that are systematically mismatched (i.e., incongruent with one’s values; e.g., Lavine & Snyder, 1996), whereas other studies compare matched messages to generic messages (that use the same form regardless of who receives them; e.g., Alexander et al, 2010).…”