Promoting Healthy Eating Practices through Persuasion Processes
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
BASIC AND APPLIED SOCIAL PSYCHOLOGY
2021, VOL. 43, NO. 4, 239–266
https://doi.org/10.1080/01973533.2021.1929987
Promoting Healthy Eating Practices through Persuasion Processes
Blanca Requeroa , David Santosb , Ana Cancelaa ~olc
, Pablo Brin , and Richard E. Pettyd
a
Universidad Villanueva; bIE School of Human Sciences and Technology; cUniversidad Autonoma de Madrid; dOhio State University
ABSTRACT
The successful impact of healthy eating campaigns often depends on the extent to which
messages are effective in changing attitudes and behaviors over time. The present work
proposes that healthy eating campaigns can be designed taking into consideration elabor-
ation and validation processes so that the degree of attitude change is maximally influenced
and is consequential. The first set of studies described in this review demonstrates the
importance of considering elaboration in determining initial attitudes toward healthy foods
as well as the subsequent attitude strength consequences (e.g., stability, prediction of
behavior, spreading). The second set of studies focuses on the role of perceived validity of
one’s thoughts in the domain of eating as a potential mediator of the persuasion process.
These studies include campaigns promoting positive attitudes toward healthy eating (e.g.,
eating of vegetables and fruits), and interventions oriented to decreasing the intake of
unhealthy food (e.g., taxing junk food). We also discuss the role of modality of information
presentation (e.g., verbal and visual information vs. direct physical experience) in those stud-
ies. Finally, the review offers a tutorial with concrete recommendations that researchers,
practitioners and public policy makers can follow in order to predict both short and long-
term attitude and behavior changes.
The success of healthy eating campaigns depends Taking into account these mechanisms and the most
largely on the fact that the campaigns are effective in relevant research on healthy eating consumption and
modifying the attitudes (e.g., toward the consumption food preference, we also include a number of practical
of fruits and vegetables) and preferences of the audi- recommendations for interventions and applications
ence as expected, and that these changes are conse- based on these two processes of change.
quential and lasting (McDermott et al., 2015;
McEachan et al., 2016; Riebl et al., 2015). Not all
Outline of the review
changes in attitudes are consequential (Krosnick &
Petty, 1995). Thus, in the present review, we explain The review is structured as follows. First, it briefly
and provide evidence for the role that two psycho- introduces the Elaboration Likelihood Model (ELM)
logical mechanisms play in producing consequential of attitude change (Petty & Bri~ nol, 2012; Petty &
attitude change in the domain of healthy eating con- Cacioppo, 1986), as a conceptual framework to under-
sumption: elaboration of the information presented stand how attitudes and preferences are formed and
and validation of people’s thoughts to that modified. This initial section covers two key variables
information. that affect the amount of thinking as illustrations –
The process of elaboration emphasizes the distinc- attitudinal ambivalence and inconsistencies between
tion between mechanisms of attitude change that actual and desired attitudes–, and also addresses the
require a relatively large versus small amount of consequences of elaboration (e.g., for attitudes to
thinking, and the content of the audience’s cognitions guide behavior in the long-term; Petty & Bri~ nol,
to a health campaign as determinants of that attitude 2020). Next, this review highlights that the evaluations
change (Petty & Cacioppo, 1986). The process of val- people have of the thoughts that result from elabor-
idation distinguishes between primary and secondary ation are also relevant in determining attitude change.
cognition, and focuses on the evaluation that people Special attention is paid to the perceived origin and
make of the cognitive responses they have generated destination of the thoughts, and the role that emo-
regarding the communication (Bri~ nol & Petty, 2009). tions play in influencing thought evaluation. Finally,
CONTACT Blanca Requero brequero@villanueva.edu Universidad Villanueva, Calle de la Costa Brava, 6, Madrid, 28034, Spain.
ß 2021 Taylor & Francis Group, LLC240 B. REQUERO ET AL.
the review offers practical recommendations and resources allocated to preventing and eradicating these
raises some key questions that practitioners and public causes such as obesity (and its health-related prob-
policy makers can ask themselves when crafting a lems) are spent on designing health campaigns to pro-
health campaign designed to be persuasive. mote well-being by creating favorable attitudes and
behaviors toward eating a healthy diet (Mayne et al.,
Scope of the review 2015). However, the data about the success of healthy
eating interventions have shown mixed results (Bonell
This review is unique in focusing on the core psycho-
et al, 2015; Rekhy & McConchie, 2014). The results of
logical mechanisms that can lead to attitude and
previous studies of the health campaigns and interven-
behavior change. Previous reviews in the domain of
persuasive campaigns promoting healthy food prefer- tions are diverse, producing positive effects, null
ence have focused on other relevant topics such as effects, and even adverse effects. But, there is some
social categories (Tarrant & Butler, 2011), self-efficacy room for optimism.
(Brug, 2008; Conner & Armitage, 2002), motivational First, it is notable that some health campaigns have
factors (Ajzen & Fishbein, 2005; Bandura, 2001), or produced the intended results. Importantly, even
goal setting and goal pursuit (Schwarzer et al., 2011). when some of the effect sizes have been small, these
Indeed, making healthy food choices is a complex small effects can accumulate over time to be conse-
phenomenon shaped by multiple factors (Mata et al., quential and meaningful (Abelson, 1985; Loyka et al.,
2018; Spiteri-Cornish, 2016; Wilson et al., 2016). In 2020). For instance, Talvia et al. (2011) designed a
order to contribute to this domain, the present review longitudinal intervention in which parents received
offers the science of persuasion as a foundation so nutritional education about a specific dietary topic
that practitioners and public policy makers can under- (saturated and unsaturated fat, visible and invisible
stand and improve the efficacy of their persuasive fat, and the role of fruits and vegetables and whole-
attempts before any efforts have been undertaken. grain products in healthy eating) along with counsel-
Health behavior change represents a complicated pro- ing dealing with the child’s and family’s diet. After
cess with many challenges and understanding the core the intervention, parents’ attitudes toward general
principles of the psychology of attitude change repre- health maintenance improved (dmother ¼ 0.39 and
sents a promising starting point. dfather¼0.53, small and medium effect respectively;
Finally, it should be noted that the research papers Cohen, 1988) and these attitudes were associated with
that comprise this review were selected based on a decrease in saturated fat intake (d ¼ 0.32, small
whether (1) the attitude object belonged to the healthy effect; for other examples of the importance of food
eating domain, (2) a persuasive attempt or treatment attitudes of parents, see Lwin et al., 2017; Pettigrew et
was attempted, (3) the process underlying attitude al., 2016; Romanos-Nanclares et al., 2018).
change was related to one of the two main mecha- Beyond parents, schools have also become the
nisms of change highlighted here, elaboration or val- scene of interventions aimed at improving the atti-
idation, and (4) whether there was any consequence tudes and healthy food intake of students (Dudley et
associated with those psychological processes in terms al., 2015; Kulik et al., 2019). For example, Francis et
of attitude change or strength (e.g., impact on behav- al. (2010) examined the impact of a nutritional educa-
ior; Petty & Krosnick, 1995). The aim was not to be tion program on attitudes related to healthy dietary
exhaustive in our coverage of all the studies pertinent habits. Results showed improvements in students’ atti-
to attitude change in the domain of healthy eating tudes toward a better diet (d ¼ 0.20, small effect), and
consumption, but rather to offer an illustrative selec- also improvements in healthy dietary choices
tion of studies that show the importance of consider- (d ¼ 0.21, small effect; for additional examples of edu-
ing two of the fundamental processes of attitude cational programs in schools, see Cespedes et al.,
change – elaboration and validation. 2013; Koo et al., 2019; Mikkelsen et al., 2014; Yoder et
al., 2014).
Another way to expose people to persuasive mes-
Promoting positive attitudes toward sages can be through cooking classes or workshops
healthy food (Brown & Hermann 2005; Condrasky & Hegler, 2010;
Unhealthy eating habits are associated with the main Larson et al. 2006; see also, Hersch et al., 2014 for a
causes of diseases and death in many societies (Micha review). For example, after attending a 10-week pro-
et al., 2017; Pearson-Stuttard et al., 2017; World gram with practical and educational sessions, partici-
Health Organization [WHO], 2009). Most of the pants showed an increase in vegetable preferenceBASIC AND APPLIED SOCIAL PSYCHOLOGY 241
(d ¼ 0.35, small effect), and more favorable food and of non-desirable healthy food evaluations and behav-
healthy cooking attitudes (d ¼ 0.29, small effect; iors (Lorenc & Oliver, 2013; Salmon et al., 2013).
Cunningham-Sabo & Lohse, 2013; see also, Burton et Several factors can explain these ironic reverse effects
al., 2017; Wolfson et al., 2020). As another illustration, of health campaigns such as the social pressure to
Winham et al. (2014) tested an interesting program look thin (Myers & Biocca, 1992) or the psychological
integrating theater elements into interactive cooking reactance motivating people to maintain a sense of
workshops. After several weeks, these researchers personal freedom (e.g., Kim & So, 2018; Lewis et al.,
observed that participating in these experiences 2010; Puhl et al., 2013).
prompted more positive attitudes toward healthy eat- Taking into consideration these contradictory
ing (d ¼ 0.41, medium effect). In fact, another way to results and the lack of satisfactory explanations, it can
work with persuasive proposals is through role-play- be challenging to predict when health campaigns will
ing or interpretation games (see Cheadle et al., 2012; produce the intended effects, no effects at all, or
Joronen et al., 2008; Perry et al. 2002). Bush et al. boomerang effects. Even when health campaigns are
(2018) examined the impact of a school program that effective in producing positive results immediately, it
included a 60-min live theater performance addressing still seems complicated to predict if those changes will
nutrition content areas. Results showed that the the- be lasting over time, which is key to achieving an
ater performance improved children’s healthy eating adequate level of well-being. As explained further
attitudes (d ¼ 0.62, medium effect). shortly, the study of the fundamental psychological
As another example, an Intuitive Eating program mechanisms of attitude change allows researchers to
encouraged individuals to focus on internal body sig- predict future changes in people’s behavior and
nals (i.e., hunger, satiety, and appetite) as a guide for
improve the effectiveness of health campaigns (Lowe
attitudes relevant to eating (Warren et al., 2017). This
et al., 2015, Petty et al., 2009; Rees et al., 2018;
intervention focused on the idea that awareness of
Salovey & Wegener, 2003).
internal hunger and satiety cues serve as a guide about
when and how much to eat. Education programs
based on intuitive eating have demonstrated positive Summary
changes in eating attitudes (d ¼ 0.63, medium effect;
Because attitudes are one of the most important
Healy et al., 2015; see also, Cole & Horacek, 2010;
(though not only) determinants of behavior, exploring
Humphrey et al., 2015). For example, patients in an
the mechanisms of attitude change should be useful
eating disorder center who participated in an intuitive
for those interested in bringing about healthy lifestyle
eating program showed healthier attitudes toward
habits. We focus on two critical psychological mecha-
food and eating (d ¼ 0.91, large effect) and also
applied these attitudes to make healthier choices in nisms that have been responsible for attitude change
their eating behavior (d ¼ 0.31, small effect; Richards in the domain of healthy eating – elaboration and val-
et al., 2017). These interventions often involve a idation. These two processes have also been critical
mindfulness component in which people are encour- for understanding how attitudes change or resist
aged to mentally separate from their own thoughts change over time in other domains (Petty & Bri~ nol,
(Bernstein et al., 2015; Luttrell et al., 2014). In the 2020, Teeny et al., 2017).
current review, we will describe studies using this Another relevant feature of this review is the con-
meta-cognitive approach to encourage healthy eating sideration of implicit measures of attitude change.
by getting people to separate from their nega- Some recent work has included assessments of atti-
tive thoughts. tudes with measures that tap into automatic rather
Although there are many examples of successful than deliberative evaluations (Petty et al., 2009;
healthy eating campaigns, there are also examples of Wittenbrink & Schwarz, 2007; for reviews). An assess-
interventions that have not been successful. That is, ment of a person’s automatic evaluative response in
some work has found that certain intervention strat- the domain of healthy eating consumption can be
egies are not effective, leading to no changes in rele- important because deliberative self-reports can be con-
vant outcomes (e.g., Gill et al., 2005; Noar, 2006; taminated by social desirability concerns, and auto-
Snyder & Hamilton, 2002). Finally, unexpected nega- matic evaluations have been shown to determine
tive effects of some health campaigns have been judgments and behaviors engaged in spontaneously
observed. That is, some health interventions have pro- (e.g., Dimofte, 2010; Gawronski & Payne, 2011;
duced a boomerang effect and increased the number Sheeran, Bosch, et al., 2016; Trendel & Werle, 2016).242 B. REQUERO ET AL.
Although measures of automatic evaluation often 2014) model, the dual attitudes approach (e.g., Wilson
produce the same results that explicit self-reports do et al., 2000), and the meta-cognitive model (MCM) of
(e.g., both revealing that people like fast food), they attitude structure (Petty & Bri~nol, 2006; Petty et al.,
can also show different responses. Previous work has 2007). We return to the MCM later in the review
shown that explicit and implicit measures of attitudes when we discuss some determinants of elaboration.
are capable of impacting behavior individually (e.g., Before getting to that, we discuss the role of elabor-
Greenwald et al., 2015) and in combination (e.g., ation in producing health-related persuasion and
Johnson et al., 2017). Thus, in the present review we some important consequences of high elaboration atti-
show how these measures can be modified using dif- tude change.
ferent persuasion procedures. Furthermore, we review
previous research showing how the basic mechanisms
Elaboration of healthy-eating messages:
of persuasion are critical for predicting both short and
promoting attitudes toward vegetable
long-term changes in deliberative and auto-
consumption
matic attitudes.
As an illustration of the role of elaboration in promot-
ing healthy food attitudes, Cancela et al. (2016) con-
Elaboration
ducted a study where participants read a health
Elaboration, how much thinking people engage in message favorable to vegetable consumption com-
regarding a persuasive treatment, is a fundamental prised of compelling or specious arguments. Varying
element in the ELM (Petty & Cacioppo, 1986). This the quality of the arguments in a message is a com-
theory proposes that judgments can be influenced by mon technique in the persuasion literature to examine
mechanisms that require different degrees of informa- whether some variable affects the degree of message
tion processing, with some persuasion processes elaboration (Petty et al., 1976). In particular, if some
requiring more cognitive effort than others. The proc- variable increases thinking about a message, then this
esses involved in changing attitudes, however, and the variable should increase the extent to which people’s
consequences that occur differ depending on the attitudes following the message are contingent on the
degree of thinking in which the person is engaged quality of the message. If thinking is low, however,
(Petty & Bri~ nol, 2020). Specifically, because high elab- then attitudes should not be affected much by the
oration persuasion processes produce more conse- cogency of the message (see Carpenter, 2015).
quential attitudes, we focus on high elaboration rather Variables external to the message are most likely to
than cue-based change in this review (see Table 1). affect the amount of elaboration when it is not already
We also highlight that the ELM explains that high constrained to be very high or very low (e.g., the mes-
elaboration can produce changes to both deliberative sage is on a topic of moderate rather than very high
as well as automatic attitudes. or low importance). In the Cancela et al. (2016) study,
The ELM is just one example of dual-process theo- following exposure to the strong or weak message,
ries that deal with both deliberative and automatic participants reported their evaluations of vegetable
processes. There are other relevant frameworks, but consumption. Importantly, participants also completed
we focus on the ELM because most of the persuasion a measure of perceived elaboration consisting of one
studies addressing elaboration processes have been question regarding the amount of attention they paid
guided by that theory. Other models like the to the message.
Associative Propositional Evaluation model (APE; The results showed that the degree of processing of
Gawronski & Bodenhausen, 2006) posit that individu- the message influenced attitude change toward vege-
als react positively or negatively to attitude objects as table consumption. Under relatively high reported
a function of the feelings or the propositions that elaboration, more argument quality discrimination
come to mind when confronted with the attitude was obtained. However, at relatively low elaboration,
object or issue. These responses are then evaluated for attitude change was similar for both messages. Viewed
their validity by examining whether the responses are differently, increased processing enhanced the influ-
compatible with the person’s knowledge. There are ence of strong arguments (making attitudes toward
still other relevant models that deal with how attitudes vegetables more positive) but reduced the influence of
are stored and structured in memory, such as the weak arguments (making attitudes less positive).
MODE (Motivation and Opportunity as DEterminants Beyond the intake of vegetables, elaboration processes
of attitude to behavior processes, Fazio & Olson, have been found to be relevant in changing attitudesTable 1. Elaboration of persuasive messages about healthy and unhealthy eating.
Independent
Study Health Intervention Variable/Predictor Dependent Variable Results Effect Size Practical Implication Questions (from Tutorial)
Cancela et al. (2016) Promotion of 1. Argument quality: 1. Attitudes toward Greater elaboration led to 2-way interaction for Attitudes toward vegetables What is the amount of
eating vegetables. Message composed of eating vegetables. enhanced argument attitudes: are more likely to guide thinking about
strong or 2. Behavioral intentions: quality effect on attitudes. d ¼ 0.40 food choices when the proposal?
weak arguments. Behavioral intentions Elaboration enhanced Main effect on behavioral elaboration is manipulated
2. Extent of elaboration: regarding vegetable attitude-behavioral intentions: to be high rather
High vs. low consumption. intention correspondence. d ¼ 0.40 than low.
manipulated through a
matching induction.
Cancela et al. (2021) Promotion of 1. Argument quality: 1. Attitudes toward Greater elaboration led to 2-way interaction for Not only short-term Is the attitude perceived
eating vegetables. Message composed of eating vegetables. enhanced argument attitudes: persuasion but also to be stable and
strong or 2. Attitude strength: quality effect on attitudes. d ¼ 1.3 expected long-term resistant to change?
weak arguments. Measures of the Elaboration enhanced Main effect on perceived consequences depend on
2. Perceived extent of perceived stability attitude strength. attitude strength: the extent of elaboration.
elaboration. and resistance d ¼ 1.7
to change.
Horcajo et al. (2010) Promotion of 1. Thought direction: List 1. Implicit measures Elaboration led to changes in Main effect on automatic Merely thinking about eating Is the attitude change
(study 3 & 4). eating vegetables. positive or negative (IAT) assessing automatic attitudes. attitude toward eating vegetables (without a generalized to other
aspects of automatic attitudes Greater automatic vegetables: message) can lead to related attitudes?
eating vegetables. toward self–vegetable associations d ¼ 0.58 attitude change and
2. Implicit measures (IAT) eating vegetables. after listing positive 2-way interaction for attitude strength as
assessing self-esteem. 2. Implicit measures aspects of eating automatic self-vegetable assessed with
(IAT) assessing the vegetables and when association: implicit measures
automatic link implicit self-esteem d ¼ 0.48
between vegetables was positive.
and the self.
Horcajo et al. (2010) Promotion of 1. False feedback to Implicit measures (IAT) More favorable implicit 2-way interaction for Increasing the perceived
(study 4). eating vegetables. increase or decrease assessing automatic attitudes toward eating automatic attitude: linkage between self and
the perceived linkage attitudes toward vegetables when the false d ¼ 0.72 vegetables can lead to
between the self eating vegetables. feedback increased self- more favorable attitudes
and vegetables vegetables linkage and toward eating vegetables.
2. Implicit measures (IAT) when implicit self-esteem
assessing the was positive.
self-esteem.
Bri~
nol et al. (2004) Promotion of 1. Argument quality: Attitudes toward A greater extent of Elaboration can increase In which emotional state
eating vegetables Message composed of eating vegetables. unpleasantness led to when people feel bad. is the individual before
strong or enhanced processing
weak arguments. argument quality. information?
2. Extent of
unpleasantness
associated with
ambivalence.
Clark et al. (2008) Promotion of taxing 1. Measures of the Attitudes toward taxing When the message was 3-way interaction for Ambivalence is an antecedent Does the person
foods with high levels previous attitudes food proposal. initially relatively attitudes: of elaboration: experience evaluative
of saturated fat. toward taxing agreeable, greater d ¼ 0.46 ambivalence increases conflict toward the
food proposal. ambivalence led to information processing of attitudinal object?
2. Measures of explicit enhanced argument messages that are
ambivalence toward quality effects consistent with the
taxing food proposal. on attitudes. dominant
3. Argument quality: evaluative reaction.
Message composed of
strong or
weak arguments
Bri~
nol, Petty, and Promotion of 1. Self-esteem discrepancy Attitudes toward When the message was Explicit-implicit discrepancies
BASIC AND APPLIED SOCIAL PSYCHOLOGY
Wheeler (2006) eating vegetables. calculated through the eating vegetables. framed as self-relevant, are an antecedent of
(study 4). difference between the greater explicit-implicit elaboration increasing
automatic and self-esteem discrepancy motivation to process
(Continued)
243244 B. REQUERO ET AL.
toward other health-related outcomes and services,
Questions (from Tutorial)
Would the individual like
including ecological food consumption (Royne et al.,
different attitudes?
2011), nutrition counseling (Kerssens & van Yperen,
1996, Ko et al., 2014; Wilson, 2007), interventions to
to have
prevent eating disorders (Unikel-Santoncini et al.,
2019; Withers & Wertheim, 2004), and even attitudes
relevant to the object for
toward energy bars and other food-relevant products
(Cancela et al., 2021).
guiding behavior above
elaboration but also in
important not only for
Practical Implication
information that is
discrepancy exists.
Desired attitudes are
actual attitudes.
Elaboration and consequences of change:
and beyond
which the
consequential attitudes toward vegetable
consumption
Although many studies have demonstrated the bene-
fits of enhancing elaboration of strong communica-
d ¼ 0.56; and by actual
intentions predicted by
Main effect on behavioral
tions for producing attitude change (Petty &
desired attitudes:
Effect Size
Wegener, 1998), elaboration is important not only
because it determines the extent of attitude change,
attitudes:
d ¼ 0.44
but also because persuasion under high degrees of
thinking is usually more consequential in the long-
term (Bri~nol & Petty, 2006; Petty et al., 2009; Petty &
led to enhanced argument
restaurant over and above
participants’ likelihood of
Bri~nol, 2020; Petty & Krosnick, 1995). For example,
Desired attitudes predicted
eating in a fast food
recent research by Cancela et al. (2021) revealed that
that predicted by
actual attitudes.
Results
attitudes toward vegetables were more likely to guide
quality effects
on attitudes.
food choices when elaboration was manipulated to be
high rather than low.
Beyond actual elaboration, the mere perception of
thinking is also important in moderating the conse-
quences of attitudes. For example, consider the study
regarding eating in a
Dependent Variable
fast food restaurant.
by Cancela et al. (2016) described earlier. In that
Behavioral intentions
research, participants not only reported their level of
elaboration and their attitudes toward eating more
vegetables, but also reported the extent to which they
considered their attitudes to be consequential. In
accord with the elaboration-strength principle, results
attitudes toward a fast
attitudes toward a fast
Message composed of
showed that participants perceived that their attitudes
Measures of desired
self-relevant or not.
Message framed as
Measures of actual
Argument quality:
Variable/Predictor
weak arguments.
esteem measure.
deliberative self-
food restaurant.
food restaurant.
were stronger (more stable and resistant to change)
Independent
when they believed they had engaged in a high degree
strong or
of processing of the ad. Importantly, prior research
had shown that when people think their attitudes are
based on thinking, they in fact become stronger (e.g.,
2.
3.
1.
2.
having a greater impact on behavior even when con-
Health Intervention
trolling for objective levels of thinking; Barden &
Fast food restaurants.
Petty, 2008; see also Rucker et al., 2014). Similarly, a
recent study about a proposal to tax junk food showed
that the greater the perceived elaboration, the more
attitudes guided behavioral intentions related to the
Table 1. Continued.
proposal (Requero et al., 2020). The results of these
DeMarree et al. (2017)
studies may help provide understanding of the lack of
effectiveness of some persuasive campaigns and inter-
(study 1).
ventions. For example, if attitudes toward the inter-
Study
vention message were based on responses to simpleBASIC AND APPLIED SOCIAL PSYCHOLOGY 245
cues rather than elaborative processing, the attitudes positivity. That is, when high implicit self-esteem indi-
would not be expected to result in behavior change. viduals (self-good) came to like vegetables more, the
‘vegetables-self’ automatic link was strengthened. In
contrast, for low implicit self-esteem participants (self-
Elaboration and generalization of change:
bad), the ‘vegetables-self’ link tended to strengthen
spreading from vegetable consumption to a
when they thought about negative features of vege-
new identity
table consumption. Another study in this line of
Beyond stability, resistance to change, and prediction research (Horcajo et al., 2010, experiment 4) showed
of intentions and behavior, another consequence of that false feedback enhancing (vs. decreasing) the per-
elaboration related to attitude strength is that attitude ceived linkage between the self and vegetables created
change that occurs under high elaboration conditions more positive automatic evaluations about vegetables,
for the targeted attitude can lead to change in a but only for those with relatively high implicit
related attitude by a process of spreading activation self-esteem.
between associated cognitions (Glaser et al., 2015; As noted, our proposal is that attitude change leads
Petty et al., 2012). It is highly feasible that attitude to more spreading consequences when the attitudes
change mechanisms that involve high elaboration are changed through high elaboration mechanisms. As
about an object produce an activation of the concepts another illustration, Moreno et al. (2021) examined to
(e.g., beliefs, values) with which it is closely related what extent changing attitudes toward a healthy (or
(McGuire, 1981). Thus, if attitudes toward vegetable unhealthy) diet through high elaboration processes
consumption change through a high elaboration pro- would increase (or reduce) prejudiced attitudes toward
cess, this change could generalize and transfer to other a related attitude issue – overweight people.
associated healthy consumption attitudes (e.g., toward Participants were first asked to generate positive
fast food) creating a consistent and holistic relation- thoughts either about a healthy or an unhealthy diet
ship with a healthy diet (Bui & Fazio, 2016). (see also Rudolph & Hilbert, 2017). After the thought-
In research relevant to this idea of attitudinal generation task, participants responded to a series of
spreading as a consequence of high elaboration, items regarding their attitudes toward the assigned
Horcajo, Bri~ nol, and Petty (2010, experiment 3) diet as the focal attitude measure. Elaboration was
showed that when participants thought about eating measured in this study merely by counting the num-
vegetables, associated changes on related automatic ber of thoughts listed by each participant. In addition
measurements were observed. Extensive thinking was to manipulating attitudes toward diets and measuring
explicitly encouraged in this study to ensure that con- elaboration, participants were also asked to rate a
ditions fostered high elaboration processes. number of social groups as part of an ostensibly unre-
Specifically, participants were instructed to generate lated study. The key group of interest embedded in
positive or negative thoughts about increasing the this list was people who were overweight. Thus, atti-
consumption of vegetables in their diet. This was tudes toward obese people were the distal (indirect)
designed to link vegetables more generally to either attitudes of this study.
good or bad. After generating their thoughts, partici- The results showed that the manipulation of atti-
pants completed implicit association tests (IATs) that tudes toward diets was successful. That is, participants
assessed the automatic link between vegetables and asked to generate positive thoughts toward healthy
the self (i.e., a measure of how well vegetables were diets reported higher liking for eating healthier than
associated with the self) as well as between the self those assigned to generate positive thoughts toward
and the concepts of good and bad (i.e., a measure of unhealthy diets. Most importantly, those focal atti-
implicit self-esteem; Greenwald et al., 2002). tudes (toward diets) were more related to distal atti-
In accord with balanced identity theory (Greenwald tudes (toward overweight people) among participants
et al., 2002; see also, Gawronski et al., 2007; Langer et who showed higher levels of thinking. Specifically,
al., 2009; Walther & Trasselli, 2003), when partici- high thinking participants who wrote positive
pants were engaged in high elaboration about the thoughts about healthy eating produced significantly
positive aspects of vegetable consumption, they more negative attitudes toward obese people than low
reported stronger associations between the self and thinking individuals. Also, high thinking individuals
vegetables compared to those who thought about the induced to like unhealthy foods produced more favor-
negative aspects. However, this impact only occurred able attitudes toward obese people than low thinking
when participants strongly associated the self with participants. These findings were replicated in other246 B. REQUERO ET AL.
studies in which elaboration was measured differently believes cigarettes are bad). That is, with implicit
(e.g., by assessing reading time) and when elaboration ambivalence an individual also has an attitude object
was manipulated rather than measured. linked to both positivity and negativity in memory
To summarize, this set of studies revealed that atti- (the same as explicit ambivalence), but one of these
tudes unrelated to healthy eating can be generalized to reactions is endorsed (the explicit attitude) but the
other important domains (e.g., prejudiced attitudes other is tagged as invalid (the implicit attitude). This
toward obese people). As noted, this indirect change individual does not recognize being ambivalent
is facilitated, at least in part, by high elaboration proc- because the individual does not consider both reac-
esses with respect to the focal attitudes. Importantly, tions to be valid. An individual’s evaluative reaction
there was some spreading also in the low thinking to an attitude object might be seen as invalid for a
conditions, but it was not as pronounced as observed variety of reasons including that (a) the individual
under high thinking. Furthermore, there are other fac- believes the reaction is a mere cultural association
tors that can contribute to spreading of attitude (e.g., from the media) and does not represent what he
change beyond elaboration, such as processes relying or she truly believes (e.g., “I have a negative reaction
on mere association, attitude accessibility, and moral towards vegetables because they are portrayed as not
conviction (e.g., Blankenship et al., 2015; Brannon et tasty on TV, but I know that is not true”) and (b) the
al., 2019; Cvencek et al., 2021; Glaser et al., 2015; reaction represents a prior attitude (e.g., “I used to
Leippe & Eisenstadt, 1994; Maio et al., 2009; Walther, like smoking cigarettes, but now I no longer do”;
2002). We focused on elaboration as a mechanism of Petty et al., 2006; see also Wilson et al., 2000). Next,
change because it has received relatively less attention we describe two examples illustrating how explicit and
regarding its role in producing indirect change. In the
implicit ambivalence are consequential by affecting
next section, we turn to another important determin-
elaboration in the domain of healthy eating
ant of elaboration: attitudinal ambivalence. This factor
consumption.
is especially relevant in the health domain given the
contradictory reactions that a person can have toward
Explicit ambivalence: taxing junk food
healthy and unhealthy foods (Norris et al., 2019;
Previous work has shown that explicit discrepancies
Yan, 2015).
are seen as negative experiences (e.g., produce feelings
of conflict; Priester & Petty, 1996), and therefore indi-
Elaboration as a mechanism to cope with viduals try to handle the discrepancy in some way.
ambivalence For example, by paying attention to information rele-
There are many factors that can motivate (e.g., per- vant to the object of discrepancy people often hope to
sonal relevance) or enable (e.g., message repetition) find a solution to alleviate the general unpleasantness
people to elaborate on a message when it is not that comes from psychological conflicts (e.g., Abelson,
already constrained by other factors to be high or low et al., 1968; Festinger, 1957; Nordgren et al., 2006).
(Petty & Cacioppo, 1986). However, to illustrate the Taking into account further information, they expect
importance of elaboration in persuasion, we focus on to know more details of the positive or negative
a variable – ambivalence – that is particularly relevant aspects so that the discrepancy and the subjective dis-
in a health context. One of the more interesting and comfort decrease (Durso et al., in press; H€anze, 2001;
useful aspects of the MCM (Petty & Bri~ nol, 2006), Jonas et al., 1997). Some studies have revealed that
mentioned earlier, is that it points to two different people reporting relatively high (vs. low) levels of felt
kinds of ambivalence that a person can experience ambivalence toward healthy food consumption usually
about a health topic. One type, called explicit ambiva- pay more attention to relevant details to disinhibited
lence, occurs when people have an attitude object eating such as food size and calories (Cornil et al.,
linked in memory to both positivity and negativity 2014; Goldstein et al., 2014).
and they further believe that both of these reactions In one study illustrating ambivalence affecting elab-
are valid (e.g., “Hamburgers taste good, but they are oration (Bri~nol et al., 2004), individuals feeling high
also unhealthy”). A second type, called implicit evaluative conflict showed greater processing of a per-
ambivalence, occurs where there are discrepancies suasive communication related to vegetable consump-
between people’s automatic and deliberative attitudes tion compared to those feeling low evaluative conflict.
(e.g., some automatic positivity comes to mind when The unpleasantness associated with psychological con-
a person sees cigarettes, but the person consciously flict led to an increase in elaboration as indicated byBASIC AND APPLIED SOCIAL PSYCHOLOGY 247
enhanced argument quality discrimination in attitudes the merits contained in the arguments. However,
toward vegetable consumption. when the advertising message was framed as not
In another illustrative example, Clark et al. (2008) important to the self, discrepancy did not lead to dis-
first asked participants to report their attitudes and crimination based on argument quality. That means
the degree of explicit ambivalence toward a proposal that discrepancies may not always increase informa-
of taxing foods with high levels of saturated fat. Then, tion processing – only when the object for which the
individuals received an advertising message comprised discrepancy exists is relevant for the self. Moreover,
either of compelling or specious arguments about the the direction of the created discrepancy index (i.e.,
benefits of the tax proposal. Next, participants was explicit self-esteem greater or less than implicit),
reported their attitudes about the proposal. When did not influence the outcome. Just as inconsistencies
people were relatively ambivalent before they read the would lead to more processing of an advertising mes-
ad, they showed a greater degree of information proc- sage about vegetable consumption when the ad is said
essing than when relatively unambivalent. Therefore, to be self-relevant, if the implicit-explicit discrepancy
ambivalence enhanced persuasion for the strong ad was about vegetables per se (e.g., positive explicit atti-
but decreased it for the weak ad. Importantly, the tude but negative implicit), this discrepancy would
effects of ambivalence on amount of processing were generally lead to more processing of ads about vegeta-
present especially for participants who had a relatively bles because such messages would be directly relevant
agreeable previous opinion. Pro-attitudinal messages to the attitudinal discrepancy.
were perceived as potentially more relevant to reduce Previous work has shown that explicit attitudinal
ambivalence since it would bolster their existing atti- ambivalence is related to low certainty in one’s atti-
tude. In other words, participants in this paradigm tude (Bargh et al., 1992; Jonas et al., 1997). As a
were more interested in processing an ad that was
result, ambivalent attitudes can reduce action readi-
consistent with their dominant evaluative reaction
ness (van Harreveld et al., 2009) and encourage
rather than a conflicting one since the former message
greater message elaboration to enhance certainty
might more easily resolve the ambivalence. When
about the object (Hodson et al., 2001). Similarly,
ambivalence increases the extent of processing, it is
implicit ambivalence is related to implicit uncertainty
likely to result in attitudes that are consequential.
about the attitude object (Petty & Bri~ nol, 2009) as
well as a general feeling of discomfort that is not dir-
Implicit ambivalence: unrecognized conflict toward
ectly attributed to the attitude object (Rydell & Durso,
vegetable consumption
Although there are a number of studies that have 2012). An attempt to reduce this discomfort is likely
examined the consequences of ambivalence, some behind the increased information processing observed
studies have shown the potential presence of and con- from implicit ambivalence as it is from explicit
sequences of implicit ambivalence. For example, ambivalence (Johnson et al., 2017).
Bri~nol, Petty and Wheeler (2006, experiment 4) found
that increases in the extent of discrepancy between Elaboration and other forms of ambivalence:
explicitly and implicitly measured attitudes was associ- Implications for fast food consumption
ated with enhanced processing of a communication
about healthy eating consumption. In this research, Another source of evaluative conflict has recently
participants’ self-evaluations were first assessed with been identified: the attitudes people actually have
an automatic measure (IAT, Greenwald & Farnham, toward different objects (e.g., cake, genetically modi-
2000) as well as a self-report measure (Rosenberg, fied foods, individuals with obesity, etc.) can be differ-
1965). Then the difference between both indicators ent from the attitudes they would like to have
was estimated and an index of implicit-explicit self- (DeMarree et al., 2014). These divergent attitudes can
esteem discrepancy was calculated. Then, participants influence the behavior and thoughts of an individual
read either a strong or a weak advertising message in different ways. For example, the actual attitude
about vegetable consumption framed as important for might encourage the person to eat fast food whereas
the self or not. the desired attitude might encourage the person not
The results showed that when the advertising mes- to eat it. These competing evaluative tendencies, like
sage was framed as important to the self, the greater that resulting from implicit-explicit attitude discrepan-
the discrepancy between a person’s explicit and impli- cies, can produce feelings of conflict about the attitude
cit self-esteem, the more the participant discriminated object.248 B. REQUERO ET AL.
Prior research has shown that inconsistencies processes stemming from variables ranging from the
between desired and actual evaluations lead to subject- ease with which thoughts are produced to the per-
ive ambivalence with all of the various consequences ceived origin of the thoughts are also relevant to influ-
previously mentioned. This type of ambivalence has encing health attitudes (see Table 2).
been shown for many health topics such as taxing
junk food and engaging in exercise (DeMarree et al.,
Validation through ease: promoting
2014). In some cases, when actual attitudes are not
healthy-eating
congruent with desired attitudes, the latter can motiv-
ate behavioral repertoires designed to achieve behav- One of the most studied variables affecting validation
iors related to the desired attitudes independently of processes is the ease with which one’s thoughts are
actual attitudes. For instance, in one study (DeMarree generated (Schwarz et al., 1991). Specifically, the easier
et al., 2017, study 1), participants completed a meas- it is to generate a thought, the more confidence peo-
ure about their attitudes toward the McDonald’s res- ple have in it (Tormala et al., 2002). This effect occurs
taurant chain (both desired and actual) and then especially under high elaboration conditions when
reported their behavioral intentions (e.g., “Over the people are not only motivated and able to generate
next month, how likely is it that you will eat at least thoughts but also to evaluate them. In one study rele-
one meal at McDonald’s?”). Results showed that the vant to healthy food campaigns, students from a
behavioral intentions were predicted by the desired course on preventing eating disorders listed the bene-
attitudes toward this restaurant above and beyond the fits or detrimental aspects about a healthy diet
actual attitudes. As is the case with any form of evalu- (Requero et al., 2015). Following the thought listing
ative conflict, one could expect that the greater the task, all participants completed the attitude measure
discrepancy between people’s actual and desired atti- regarding the diet and reported the ease with which
tudes the greater the interest in receiving more infor- they generated their thoughts. Results showed that for
mation pertinent to the topics for which the people who rated it relatively easy to generate
discrepancy exists (DeMarree et al., 2014, 2017). thoughts, those writing positive aspects of the diet
showed more favorable attitudes toward that diet than
did those writing negative aspects. However, for indi-
Validation
viduals who perceived it was relatively difficult to gen-
Having described the importance of elaboration proc- erate thoughts, those who wrote positive aspects
esses in persuasion, the second persuasion process tended to be less favorable toward the diet than did
that is a key component of the ELM has been more those who wrote detrimental aspects.
recently described and involves thought validation, a This study suggested that the meta-cognitive
meta-cognitive process that refers to thinking about experience of ease can be an important determinant
one’s own thinking. Importantly, according to the of attitudinal outcomes in the domain of food-related
ELM, any persuasion variable (e.g., ambivalence) can health persuasion. That is, individuals showed more
not only affect the extent of elaboration (i.e., how favorable evaluations regarding the diet when they
much individuals think), the valence of that thinking were asked to think in a positive direction and less
(i.e., whether individuals are largely favorable or favorable attitudes when they were required to think
unfavorable in their thoughts), but can also influence in a negative direction at high but not at low levels of
how individuals perceive or feel about their thoughts. perceived ease. Interestingly, individuals who were
When people consider their thoughts as valid or feel relatively low in perceived ease of thought generation
good about them, they use them when evaluating. showed a non-significant opposite trend. Indeed, as
When people have doubt about or do not like their noted earlier, when people have strong doubts associ-
cognitive responses, they do not rely on them to form ated with the validity of their thoughts in one direc-
their evaluations. When people are highly doubtful of tion they might infer that the opposite direction
their thoughts, they can even make judgments oppos- might be more likely to be true. It is important to
ite to the valence of the thoughts generated. This note that fluency is operating through a validation
mechanism of self-validation is based on secondary or process in this study because it has an opposite effect
meta-cognition because it involves a reflection on pri- on persuasion depending on whether the validated
mary cognition, and it takes relatively high thinking thoughts are positive or negative. Beyond ease, recent
conditions to operate (Bri~nol & Petty, 2009). In the research has shown that other experiences such as the
next sections, we illustrate how thought validation mere feeling of readiness and preparation can increaseTable 2. Validation of thoughts polarizes attitudes toward healthy and unhealthy eating.
Independent Questions
Study Health Intervention Variable/Predictor Dependent Variable Results Effect Size Practical Implication (from Tutorial)
Requero et al. (2015) Eating Mediterranean diet. 1. Thought direction: Attitudes toward eating a Greater experience of ease 2-way interaction for Meta-cognitive experiences Do the individual’s thoughts
List positive or negative Mediterranean diet. in generating thoughts attitudes: of ease polarize attitudes about the health issue
aspects of eating a produces greater impact d ¼ 0.42 by validating thoughts. come to mind easily?
Mediterranean diet. of those thoughts in
2. Thought validity: forming attitudes: ease
Measure of the ease of increased persuasion for
thought generation. positive thoughts but
decreased persuasion for
negative thoughts.
Gasc
o et al. (2018) Eating Mediterranean diet. 1. Thought direction: Attitudes toward eating a People relied on their 2-way interaction for People consider the origin of Is the origin of the thought
Eating fast food. List positive or negative Mediterranean diet or thoughts more in forming attitudes toward eating a their thoughts as to perceived as a valid or
aspects of eating a fast food. attitudes when the Mediterranean diet or fast whether they are valid or reliable source?
Mediterranean diet or thoughts were internally food: invalid. Origin affects
fast food. rather than d ¼ 0.50 thought usage and
2. Thought validity: externally originated. polarizes attitudes.
Thought origin with
internal (high validity)
vs. external (low)
perceived validity.
Eating fast food. 1. Thought direction: 1. Index of attitudes People relied on their 2-way interaction for index of People consider the origin of
List positive or and behavioral thoughts more in forming attitudes and behavioral their thoughts as to
negative aspects of intentions regarding attitudes when the intentions: whether they are valid or
eating fast food. eating fast food. thoughts were internally d ¼ 0.36 invalid. Origin affects
2. Thought validity: 2. Measures of rather than externally Thought validity as a thought usage and
Thought origin with thought validity. originated mediator: Indirect effect polarizes attitudes and
internal (high validity) This effect was mediated by 95% CI ¼ [0.01, 0.10] behavioral intentions.
vs. external (low) thought-validation.
perceived validity.
Bri~
nol et al. (2013) Eating Mediterranean diet. 1. Thought direction: Index of attitudes and When thoughts were placed 2-way interaction for index of People evaluate the What does the person do with
List positive or negative behavioral intentions in one’s pocket attitudes and behavioral destination of their those thoughts after
aspects of eating regarding eating a participants used their intentions: thoughts as whether it generating them?
Mediterranean diet Mediterranean diet. thoughts more in forming d ¼ 1.08 implies validity or
2. Destination of attitudes and behavioral invalidity. Invalidity
thoughts: Placing in intentions than when reduces thought usage
one’s pocket (high placed in the trash. and attenuates attitudes
validity) vs. the trash and behavioral intentions.
(low validity).
Kim et al. (in press) Eating fast food. 1. Thought direction: 1. Behavioral intentions When the meaning of the 2-way interaction for Meaning of the destination Is the meaning of the key
List positive or negative regarding eating destination implied behavioral intentions: that is associated with variable (ease, origin,
aspects of eating fast food. validity, participants used d ¼ 0.48 validity affects thought destination) associated
fast food. 2. Measures of their thoughts more in Thought confidence as a confidence and this effect with validity?
2. Meaning of the thought confidence. forming behavioral mediator: Indirect effect polarizes behavioral
destination of thoughts: intentions. 95% CI ¼ [0.07, 0.52] intentions.
Moving to save (high This effect was mediated by
validity) vs. moving to thought confidence.
delete (low validity).
BASIC AND APPLIED SOCIAL PSYCHOLOGY
249250 B. REQUERO ET AL.
confidence in thoughts affecting food-relevant atti- the benefits or detrimental aspects of eating a fast
tudes even when those thoughts are unrelated to the food diet. After that, individuals were assigned to the
domain in which the person was prepared (Carroll et thought origin manipulation. Then, individuals com-
al., 2020). pleted measures of attitudes and behavioral intentions
regarding fast food. Importantly, perceived thought
validity was also assessed as a potential mediating pro-
Validation depending on perceived origin of
cess. Results showed that individuals in the internal
thoughts: changing attitudes toward the
origin condition used their thoughts to guide evalu-
Mediterranean diet, and fast food
ation to a greater extent than those in the external
Research on thought validation has found that there origin condition, and this result was mediated by the
are many individual factors (e.g., extent of happi- perceived validity of their thoughts.
ness) and situational variables (e.g., being in a confi-
dent posture) that are associated with perceived
validity in addition to ease that can interact with Validation as a function of subjective destination
thought valence to impact judgments (see Bri~ nol & of thoughts: healthy food
Petty, 2009, for a review). For example, the per-
In addition to considering the origin of their thoughts,
ceived origin of one’s thoughts can impact attitude
people can also consider their destination. For
change by affecting the validity of people’s thoughts.
example, in one experiment, Bri~ nol et al. (2013), as
The origin of thoughts denotes the perceived source
part of a course on dietary habits, had participants
of a particular primary cognition (e.g., “Where did
write the benefits or the detrimental aspects of a par-
this thought come from?” and “Did I think of this
ticular diet on a sheet of paper. After that, they were
myself, or did I hear somebody else say it?”). The
asked to either throw that paper in the garbage (an
perceived origin of thoughts is relevant because if it
invalid destination), move it to inside their pockets (a
is related to validity, then it can impact thought use
destination associated with high validity), or just leave
and the extent of attitude change.
it on a table (control condition). It was shown that
In one relevant study (Gasc o et al., 2018), adoles-
placing one’s thoughts in a high validity location led
cents who participated in a workshop on preventing
to greater use of the thoughts in forming evaluative
eating disorders were asked to list either the benefits
judgments and behavioral intentions regarding the
or detrimental aspects of different diets – a fast food
diet than placing the thoughts in a low validity loca-
diet or the Mediterranean diet. After that, individuals
tion. In fact, individuals in the invalid location condi-
were assigned to the thought origin induction. In this
tion not only used their thoughts less but they even
manipulation, individuals were required to choose an
used them in a reverse way. In sum, perceptions about
origin for their thoughts among possibilities that listed
the origin and destination of thoughts can have an
exclusively external options (i.e., from one’s peer
impact on thought usage thereby affecting attitudes
group or the media) or internal options (e.g., from
and behavioral intentions toward healthy and
one’s personality or experience). The options pre-
unhealthy foods.
sented to the participants required them to consider
and select only external or internal sources of their
thoughts. Results showed that the thought valence
Validation is meaning dependent: choosing
manipulation had a larger impact on judgments when
between healthy and unhealthy options
the participants were assigned to the internal origin
condition compared to the external origin condition. Experiments have shown that sometimes the very
Furthermore, this result was obtained regardless of the same response with respect to one’s thoughts can be
type of diet they had to think about. Thus, this experi- interpreted differently, and the meaning is critical for
ment suggested that when people perceive that their the outcome. For example, Kim et al. (in press) asked
thoughts are internally originated, they consider those individuals to type out the benefits or the detrimental
thought to be more valid by default compared to aspects of McDonald’s food using a computer key-
when they perceive that their thoughts are exter- board. Then, they had to save or delete their thoughts.
nally originated. That is, in the high validity condition, individuals
A final experiment in this line of research tested were told to save their thought list on the computer
the psychological mechanism of the effect using a whereas in the invalidity condition they were told to
mediational approach. Participants first thought about delete the thought list from the computer before theYou can also read
