CHOCOLATE AND POLITICS: A CROSSNATIONAL, SURVEY-BASED
EXPERIMENT ON RECRUITMENT TO A BOYCOTT CAMPAIGN*

Shelley Boulianne and Nicole Houle†

People rarely participate in political activities. Many scholars explain this lack of participation
in terms of “nobody asked” (Schlozman and Brady 2022; Verba, Schlozman, and Brady 1995).
This theory motivates hundreds of studies (and millions of dollars in campaign spending)
related to door-to-door canvassing in elections (see Green, McGrath, and Aronow 2013). There
is also a good deal of research on the role of “being asked” in relation to participation in marches
and demonstrations (see Walgrave and Wouters 2014). Yet, we know little about the role of
“being asked” to participate in everyday political activities, such as boycotting. Jennifer Earl,
Lauren Copeland, and Bruce Bimber (2017) argue that boycotts and participation in marches
and demonstrations are protest tactics used by social movement organizations, and thus, it
makes sense to connect these scholarships. Within the protest scholarship are widespread
debates about whether organizations or interpersonal networks are more important in the
mobilization process (Earl, Copeland, and Bimber 2017).
This article uses a survey-based experiment to examine the effects of being asked to
participate in a boycott. Experiments are advantageous in establishing causality (Mutz 2011)
and moving scholarship beyond retrospective self-reports about being asked, which can have
problems related to recall. Furthermore, this study examines the important question of whether
interpersonal prompts are more effective than organizational prompts in the willingness to
engage in boycotts, which is a key theoretical question related to political mobilization. This
study is distinctive in considering whether crossnational differences exist concerning the
effectiveness of interpersonal versus organizational prompts to participate in politics, offering
insight into global patterns of political mobilization. Finally, we examine the nuances concerning the effectiveness of recruitment attempts. We consider a series of moderators of the
relationship between being asked by a friend to participate and willingness to participate in a
boycott: views about the targeted product, the perceived effectiveness of boycotts, or prior
* The authors would like to thank Karolina Koc-Michalska and Lauren Copeland for the advice on the topic of this
survey experiment.
† Shelley Boulianne is the R. Klein Chair in Communication (full professor) at Mt. Royal University, Canada. She has
held professor positions at University of Southampton, MacEwan University, and Université Catholique de Lille (France).
She earned her PhD in sociology from the University of Wisconsin-Madison. She conducts research on media use and
public opinion, as well as civic and political engagement, using meta-analysis. Nicole Houle (BA, Honours) worked as a
research assistant at MacEwan University. Dr. Boulianne supervised her honour’s thesis about gender differences in
political participation during the pandemic. Direct correspondence to SBoulianne@mtroyal.ca.
© 2025 Mobilization: An International Quarterly 30(1): 73-88
DOI 10.17813/1086-671X-30-1-73

Downloaded from http://meridian.allenpress.com/mobilization/article-pdf/30/1/73/3501511/i1938-1514-30-1-73.pdf by Simon Fraser University user on 25 March 2025

Organizations spend millions of dollars to encourage citizens to participate in their campaigns;
however, organizations' mobilization effectiveness has been under question. This report uses a
survey-based experiment (n = 6,290) to examine the extent to which a friend’s versus an
organization’s endorsement affects people’s willingness to boycott chocolate because of the use
of child labor. The survey data were gathered in autumn 2019 in the United States, United
Kingdom, France, and Canada. We find that organizational endorsements are ineffective in
influencing a subject’s willingness to participate in a boycott. Instead, prompts from friends
increase the willingness to participate. Views about chocolate moderate the effectiveness of a
friend’s endorsement of the boycott. The findings provide insight into the roles of organizations
and interpersonal ties in mobilizing citizens to engage in political activities.

74

Mobilization

knowledge of this real-life boycott campaign. We also find that a friend’s endorsement of the
boycott campaign is more effective in recruitment for US and English-speaking Canadian
subjects than other subjects, offering new insights about international differences in the
effectiveness of political mobilization tactics.
ORGANIZATIONAL PROMPTS TO PARTICIPATE
Downloaded from http://meridian.allenpress.com/mobilization/article-pdf/30/1/73/3501511/i1938-1514-30-1-73.pdf by Simon Fraser University user on 25 March 2025

Lisa Neilson and Pamela Paxton’s (2010: 19) seminal piece documents the correlation between
organizational memberships and political consumerism, explaining that organizations “are a
potential source of information about social and environmental issues as well as a source of
motivation to take action.” A recent meta-analysis of research on political consumerism finds
that being a member of an organization was a significant and positive predictor of political
consumerism in nineteen studies (Copeland and Boulianne 2022), albeit the research is
exclusively based on cross-sectional surveys that cannot establish causality.
To better understand political consumerism, we build on theories and research on protest
participation and election campaigns. Alan Schussman and Sarah Soule (2005: 1083) find that
being a member of an organization increases the likelihood of being asked to participate and
that “being asked to protest is the strongest predictor of participating in protest.” When both
organizational membership and being asked are included in a model, being asked remains
significant, whereas organizational memberships are no longer statistically significant (De Vydt
and Ketelaars 2020; Schussman and Soule 2005). Jeroen Van Laer (2017) surveyed a sample
of organizational members of four environmental groups in Belgium. In predicting the willingness to participate in a climate protest, he finds both informal (interpersonal ties) and formal
(organizations) links matter. However, for actually showing up to the event, informal links are
significant, while formal links are not. Eva Anduiza, Camilo Cristancho, and José M. Sabucedo
(2014) studied protestors in Spain and found that organizations are less effective in recruitment
to protest than personal networks. Dana Fisher and Marije Boekkooi’s (2010) study of protestors in the United States suggests personal networks are slightly more important in hearing
about demonstrations than organizations, although the most important channel for hearing
about protest events is the Internet. The role of organizations merits a re-evaluation with a crossnational (and more representative) sample.
While the effectiveness of organizations’ recruitment initiatives is debated within the
protest scholarship, door-to-door canvassing is deemed one of the most effective methods of
increasing electoral turnout. Donald Green, Mary McGrath, and Peter Aronow (2013) conducted a meta-analysis of these initiatives in the United States, offering conclusive evidence of
their effectiveness. Yosef Bhatti et al. (2016) conducted a meta-analysis of research on the
effectiveness of door-to-door canvassing in Europe. They find these organizational initiatives
are effective, but less so than in the United States due to higher turnout in Europe, cultural
differences, and differing political systems (e.g., first-past-the-post versus proportional
representation). Again, the findings suggest crossnational differences in the effectiveness of
recruitment campaigns.
Outside of election campaigns, political participation may not rely heavily on organizational recruitment. Instead, digital media enable communication through online networks,
replacing the need for organizations to coordinate protest events (Anduiza, Cristancho, and
Sabucedo 2014; Bennett 2013; Earl, Copeland, and Bimber 2017). In a sample of American
respondents, of those who participated in political consumerism, people largely reported
participating in self-directed boycotts (72 percent) and buycotts (77 percent) instead of
organizationally directed campaigns (Earl, Copeland, and Bimber 2017). The authors suggest
this could be due to the shift toward self-directed political activism, especially when engaging
in political consumerism in the United States. Newer studies suggest that the United States is
distinctive in terms of political consumerism and the role of groups and organizations. Shelley
Boulianne, Lauren Copeland, and Karolina Koc-Michalska (2024) find that group ties are more

Chocolate and Politics

75

INTERPERSONAL PROMPTS TO PARTICIPATE
Using a sample from Sweden, Carolin Zorell and Thomas Denk (2021) find that self-reported
“influence from others” is a significant predictor of participating in any type of political consumption activity. Yet, these self-reports are far from conclusive. Drawing on the large body of
research on protest, we borrow some of the theoretical claims explaining this correlation.
Stefaan Walgrave and Ruud Wouters (2014) offer an extensive review of the literature about
being asked to participate in protest and subsequent participation, building on Klandermans and
Oegema’s (1987) model of protest participation. In particular, the decision to participate is
based on (1) agreement with the movement goals, (2) perceptions that participation will be
effective, and (3) network influences, such as being asked to participate (Klandermans and
Oegema 1987). In a reent iteration of the model, Klandermans and van Stekelenburg (2022:
174) argue that “social networks are indispensable in the processes of mobilization. Individual
grievances and feelings are transformed into group-based grievances and feelings within social
networks.”
Contemporary research affirms the importance of being asked. For example, Michiel de
Vydt and Pauline Ketelaars (2020: 13) find “being asked is by far the strongest variable in the
model” that predicted intentions to participate in the Grand Parade in Brussels. While they
initially offered distinctions about who asked (partner, family, friends, acquaintances, colleagues, other students, or members of an organization), they do not analyze these distinctions
and instead assess whether anyone asked the respondent to participate. They report that 69
percent of their sample was asked to participate in the protest. In contrast, María Inclán and
Paul Almeida’s (2017) work finds only twenty-eight percent of their sample of protestors in
Mexico City were asked to participate in a demonstration (56 percent were not asked, and
sixteen percent could not remember). The probability of being asked and the effectiveness of
this recruitment attempt may depend on one’s culture.
In relation to protest participation, studies show that strong ties to the person asking one to
participate are more effective at recruitment than weak ties to the person (McAdam and Paulsen
1993; Snow, Zurcher, and Ekland-Olson 1980). In other words, friends would be more effective
at recruitment than acquaintances, albeit research is split (see Walgrave and Wouters 2014).
Concerning protest, Walgrave and Wouters (2014) argue that being asked to participate is
important, not only for one’s own mobilization but also because the recipient of the invitation
is more likely to extend the invitation to others. Being asked has a direct effect on an individual’s participation and an indirect effect within one’s personal network as the person spreads
the invitation to participate amongst their friends, further expanding participation within the
social network. Asking others is important in creating a known other who is also participating
in the event. Studies show that people attend marches and demonstrations with others (Doherty
and Schraeder 2018; Fisher and Boekkooi 2010). Our primary research question is about the
effectiveness of these prompts, and our secondary question is about the robustness of these
findings across national contexts:

Downloaded from http://meridian.allenpress.com/mobilization/article-pdf/30/1/73/3501511/i1938-1514-30-1-73.pdf by Simon Fraser University user on 25 March 2025

strongly predictive of political consumerism in France and the United Kingdom than in the
United States. As such, the effectiveness of organizational prompts in political consumerism
may be country specific.
In a systematic review of more than 300 studies on political consumerism, Dawn Yi Lin
Chow, Ga-Eun (Grace) Oh, and Amitabh Anand (2022) document the overrepresentation of
North America in the existing scholarship and suggest further research consider external factors
affecting political consumerism, such as macroeconomic, social, cultural, historical, religious
and environmental factors. We offer a modest contribution to this research gap with some
insights using a four-country sample with two language groups represented.

76

Mobilization

RQ1: Are prompts from friends more effective than prompts from organizations in increasing
the willingness to boycott?
RQ2: Are there crossnational differences in the effectiveness of these prompts in increasing the
willingness to boycott?
MODERATED EFFECTS ON THE WILLINGNESS TO BOYCOTT
Downloaded from http://meridian.allenpress.com/mobilization/article-pdf/30/1/73/3501511/i1938-1514-30-1-73.pdf by Simon Fraser University user on 25 March 2025

The perceived effectiveness of an activity is a major factor when deciding to incur the costs
(time, resources) associated with participation (Klandermans and Oegema 1987). Stolle,
Hooghe, and Micheletti’s (2005) classic work on political consumerism documents a positive
correlation between engaging in this political activity and viewing it as effective. This finding
is replicated in contemporary work (Kelm and Dohle 2018). We consider whether the perceived
efficacy of boycotting influences the willingness to boycott and the effectiveness of prompts to
participate.
Several recent experimental studies have been conducted on the willingness to boycott.
One study examines the willingness to boycott a tea shop depending on whether the shop
supports or does not support same-sex marriage legalization (Wang, Chang, and Chen 2021).
The key moderator was one’s views about same-sex marriage legalization. If the respondents
support the legalization and the tea shop does not, the respondents (n = 177) are, on average,
more willing to boycott the tea shop. In another study included in this report (Wang, Chang,
and Chen 2021), they altered the information about whether friends and relatives liked/disliked
the brand being boycotted to determine how one’s social network influences the willingness to
boycott. They find that the effectiveness of their manipulation depends on social comparison
information. People who do not care about comparing themselves to others will follow their
own ideological beliefs when deciding whether to boycott rather than follow the direction of
their friends and relatives. In sum, the experiments suggest the effectiveness of boycott pleas
will depend on one’s views about the target of the boycott, the reason for the boycott (alignment
with one’s views), and social networks. In other words, the mobilization model replicates two
of three elements of Klandermans Dirk Oegema’s (1987) model, excluding the efficacy of participation.
Marc Jungblut and Marius Johnen (2021) argue that social network effects are different for
boycotting than buycotting campaigns because boycotting is more collective in nature and has
a negativity bias (punishing companies). In their experiments, they suggest differential effects
when the dependent variable is boycotting versus buycotting intentions. They find people’s
reactions to corporations taking a political stand will depend on their views about the product
targeted by the boycott/buycott campaign. The results of their experiments support these
moderated effects. As such, views about the products should be considered when understanding
reactions to calls to boycott, adding yet another element to understanding mobilization.
Finally, Cheng Hong and Cong Li (2021) conducted an experiment related to Ben and
Jerry’s (an ice cream company) position on gun control and people’s willingness to boycott or
buycott this company. The study is distinctive in using a factual situation and a real brand,
which enables them to account for brand loyalty in determining the effectiveness of their
stimuli. Hong and Li repeat the approaches of other experiments with respect to considering the
effects of one’s own views on gun control (support/do not support) as a predictor of willingness
to boycott or buycott. Continuing on the theme of social networks, they consider whether or not
participants thought most Americans favored (or not) gun control as a moderator of the effect
of their manipulation.
Considering this line of research, we also study prior views about the product, prior
awareness of a real-life boycott campaign, and perceptions about the efficacy of boycotting as
moderators of the effects of recruitment attempts on willingness to boycott. Our third research
question is:

77

Chocolate and Politics

RQ3: To what extent does the effectiveness of prompts depend on other factors (efficacy of
boycotting, views about the product, prior awareness of the campaign)?
METHODS

Downloaded from http://meridian.allenpress.com/mobilization/article-pdf/30/1/73/3501511/i1938-1514-30-1-73.pdf by Simon Fraser University user on 25 March 2025

This survey-based experiment was administered to an online panel in the United States, United
Kingdom, France, and Canada from September to November 2019. The experiment received
human-subjects ethics approval prior to data collection (File No. 101662). The survey was
funded by a grant from the Social Sciences and Humanities Research Council of Canada (Grant
No. 435-2019-04-94). Given the funding source, the choice of countries is based on Canada
being the focal point. The US was chosen as Canada’s only neighbor, with the UK and France
being founding nations, leaving institutional and cultural legacies. We use a similar case design
(Seawright and Gerring 2008) in that we choose countries with similar scores for the macroindicators that have been often considered in this field of research, such as income inequality
and democracy score (see Boulianne, Copeland, and Koc-Michalska 2024; Chow, Oh, and
Anand 2022). Freedom House (2021) scores are based on political rights and civil liberties.
These scores are consistently high for the four countries: the US (83), the UK (93), France (90),
and Canada (98). In terms of income inequality (World Bank 2021), the Gini coefficients for
these countries are as follows: the U.S. (41.4 based on 2018 data), the UK (35.1 based on 2017
data), France (32.4 based on 2018 data), and Canada (33.3 based on 2017 data). Considering
the global variation in these indicators, these countries are considered similar. However, it is
worth noting that the US and Canada are at opposite ends with these values. The US has the
lowest democracy scores, Canada has the highest; the US has the highest income inequality,
and Canada (and France) have the lowest income inequality.
Kantar was hired to administer the survey to their online panel. Respondents are recruited
into the online panel through mobile apps and online spaces. Kantar (2024) uses “a mix of
curated partners, media, social channels and loyalty schemes . . . which leads to respondents
from dif-ferent cultures, backgrounds, education levels, ages, genders and sexualities.” Their
online panel has more than 100 million respondents who have agreed to complete surveys.
Kantar has been used in other published crossnational surveys (e.g., Kaun, Larsson, and Masso
2024) as well as in survey-based experiments in political communication (e.g., Haugsgjerd,
Karlsen, and Steen-Johnsen 2023). Both studies have been published in top journals in the field
of sociology and political science.
Kantar uses incentives and rewards for participants, but these incentives and rewards are
not attached to this specific project. Per ethics requirements, respondents sign a project-specific
consent form affirming that participation is voluntary. Prior to providing the survey data file to
a client, Kantar (2024) uses “identity validation, machine AI learning, Honesty Detector, digital
fingerprinting technology and in-survey quality controls.” For this specific project, they removed speedsters and poor-quality respondents prior to sending the data file to the researchers.
Quotas were used to ensure representation of the population in terms of age, education, and
sex. Table 1 in the online appendix (see Boulianne and Houle 2024 in the reference section for the
link) presents a side-by-side comparison of sample and population statistics. According to the 2016
Canadian census, twenty-three percent of Canadians speak French as their first official language
(Canadian Heritage 2019). As such, we included a quota to ensure a similar portion of our
sample (374 of 1539 Canadians) completed the survey in French. The data and replication files
are available online at the Figshare site (see Boulianne 2024 in the references for the link).
Table 1 outlines the different experimental conditions. Table 2 offers a robustness check,
demonstrating that the random assignment process produced three groups/samples with similar
characteristics for the three conditions. The random assignment process helps rule out
spuriousness in the relationship (Mutz 2011) between recruitment prompts and willingness to
participate in a boycott.

78

Mobilization

Table 1. English and French Versions of Different Conditions

Your best friend asked you to boycott
chocolate, because of the industry’s use of
child labor.

Dans quelle mesure êtes-vous susceptible de
participer à un boycott du chocolat en raison du
recours au travail des enfants par la filière du cacao ?

How likely are you to participate in a boycott
of chocolate because of the cocoa industry’s
use of child labor?

Condition 2: Organizational Endorsement
L’organisation Ethical Consumer vous a demandé
de boycotter le chocolat, en raison de l’utilisation du
travail des enfants par l’industrie chocolatière.

Ethical Consumer organization asked you to
boycott chocolate, because of the industry’s
use of child labor.

Dans quelle mesure êtes-vous susceptible de
participer à un boycott du chocolat en raison du
recours au travail des enfants par la filière du cacao ?

How likely are you to participate in a boycott
of chocolate because of the cocoa industry’s
use of child labor?

Condition 3: No Endorsement
Dans quelle mesure êtes-vous susceptible de
participer à un boycott du chocolat en raison du
recours au travail des enfants par la filière du cacao?

How likely are you to participate in a boycott
of chocolate because of the cocoa industry’s
use of child labor?

Scale
Pas du tout probable
Pas vraiment probable
Assez probable
Très probable
Extrêmement probable

Not at all likely
Not very likely
Somewhat likely
Very likely
Extremely likely

Table 2. Random Assignment Check
No
Endorsement
n = 2098

Friend
Endorsement
n = 2097

Organizational
Endorsement
n = 2096

Like chocolate
(Responses: 1 to 5)

3.82
(SD = 1.09)

3.79
(SD = 1.11)

3.79
(SD = 1.12)

.446

Efficacy of boycotts
(Responses: 1 to 5)

2.20
(SD = 1.16)

2.23
(SD = 1.18)

2.25
(SD = 1.16)

.393

Awareness of boycott of cocoa
industry for the use of child labor
(Responses: 0,1)

15%

16%

14%

.337

Females (Responses: males, females)

51%

51%

52%

.666

Age in years
(Responses: 18 to 91)

48.42
(SD = 17.30)

48.38
(SD = 17.32)

48.54
(SD = 17.34)

.950

Education (Responses: high school or
less, some college, bachelor’s degree,
more than bachelor’s degree)

1.92
(SD = 1.05)

1.99
(SD = 1.06)

1.93
(SD = 1.05)

.088

p-value

Downloaded from http://meridian.allenpress.com/mobilization/article-pdf/30/1/73/3501511/i1938-1514-30-1-73.pdf by Simon Fraser University user on 25 March 2025

Condition 1: Friend Endorsement
Votre meilleur ami vous a demandé de boycotter le
chocolat en raison de l’utilisation du travail des
enfants par l’industrie chocolatière.

Chocolate and Politics

79

Downloaded from http://meridian.allenpress.com/mobilization/article-pdf/30/1/73/3501511/i1938-1514-30-1-73.pdf by Simon Fraser University user on 25 March 2025

Experiments are widely criticized for being artificial, particularly those in lab settings. Most
survey-based experiments employ fictitious vignettes to examine attitude differences (Mutz
2011). We designed a realistic intervention involving a real organization and political issue. For
these reasons, this project has a stronger claim to validity than other types of survey-based
experiments using fictitious scenarios. Diana Mutz (2011: 87) argues that “the single most important factor in implementing effective manipulations is probably the length of the experimental
treatment . . . keep manipulations short and to the point.” Our manipulation is a single sentence.
We chose to examine chocolate given the longstanding call to boycott the cocoa industry’s
use of child labor for cocoa production in West Africa (Balch 2021; Onishi 2001; Owens 2023;
Whoriskey and Siegel 2019). The call to participate in a boycott or buycott may arrive from
various groups and organizations, including labor unions, environmental groups, animal rights
groups, religious groups, political parties, or consumer organizations. We chose to focus on
Ethical Consumer as an organization because this organization has been involved in and advocated for boycott campaigns against the cocoa industry. Furthermore, it is a media company
(printed magazine with a supporting digital and social media presence) that promotes sustainable
consumption. Their website includes a guide to the ethical consumption of chocolate, which
involves avoiding products from companies with unethical practices or environmentally
damaging production lines (Wexler 2018). As such, the experimental manipulation is plausible;
the organization exists and does advocate for boycotts of certain chocolate brands. This
organization is based in the United Kingdom but targets an international and English-speaking
audience.
We split the sample into three conditions: being asked to participate by the Ethical
Consumer organization, being asked to participate by a friend, and not being asked to participate
(see table 1). Our survey question is, “How likely are you to participate in a boycott of chocolate
because of the cocoa industry’s use of child labor?” We offered the following response
categories: not at all likely (1), not very likely (2), somewhat likely (3), very likely (4), and
extremely likely (5).
We focus on child labor, a simplistic issue compared to other rationales for political
consumerism. For example, Chih-Chien Wang, Shu-Chen Chang, and Pei-Yeng Chen (2021)
consider same-sex marriage legalization in Taiwan, which is an issue with two sides. Jungblut
and Johnen (2021) consider gun control (also Hong and Li 2021) and the border wall in the
United States, which are also issues with two competing sides. In contrast, our issue concerns
child labor, which is far less contentious. We focus on boycotting instead of buycotting,
following Jungblut and Johnen’s claim that social networks matter more for boycott campaigns
than buycott campaigns.
Moving scholarship forward, we sought to examine whether the effectiveness of organizations and interpersonal prompts depends on various factors (RQ3). First, we examine whether
prior awareness of the campaign impacts the effectiveness of these prompts. We asked
respondents whether they had heard of the campaign to boycott the cocoa industry for the use
of child labor. This question is important because prior knowledge of the campaign could
impact the effectiveness of the various manipulations. Second, we consider whether the
effectiveness of the prompts depends on views about the target—chocolate. We stated, “We
would like to know how much you like chocolate.” The response options were 1 (not at all), 2
(a little), 3 (a moderate amount), 4 (a lot), and 5 (a great deal). Third, we account for the perceived efficacy of boycotting. We asked, “For each activity, could you tell me whether or not
you think this activity is effective for influencing political leaders? d) Joining in boycotts.”
Respondents were given a response scale as follows: 1 (not at all), 2 (a little), 3 (a moderate
amount), 4 (a lot), and 5 (a great deal). As described above, we examine whether perceptions
about the efficacy of this activity influence willingness to engage in boycotts, and whether these
perceptions moderate the effects of the prompts.
Based on reviewer feedback, we also tested an alternative measure of efficacy, which
focuses on influencing businesses. We asked, “In general, how much of an effect do you believe
your buying decisions have on changing the behavior of businesses?” The response options are

80

Mobilization

the same as the measures listed above. This measure does not mention boycotting per se, but a
reviewer asked that the results be presented for readers to consider. As such, all of the analyses
were repeated using the alternative measure of efficacy (Boulianne and Houle 2024). This
alternative measure of efficacy (which does not mention boycotting) produces the same results
as the measure that asks explicitly about the efficacy of boycotting.
RESULTS

Table 3. Willingness to Boycott Chocolate
Friend endorsement
2.76
(SD = 1.20)
2.66
(SD = 1.21)
2.75
(SD = 1.22)

Organizational
endorsement
2.67
(SD = 1.23)
2.60
(SD = 1.30)
2.61
(SD = 1.14)

2.80
(SD = 1.20)

2.83
(SD = 1.26)

.301

All countries
n = 6290

No endorsement
2.62
(SD = 1.20)

p-value

United States
n = 1700
United Kingdom
n = 1542
France
n = 1509

2.43
(SD = 1.18)
2.60
(SD = 1.21)
2.91
(SD = 1.16)

Canada (English)
n = 1165

2.59
(SD = 1.23)

2.91
(SD = 1.15)

2.66
(SD = 1.15)

<.001

Canada (French)
n = 374

2.57
(SD = 1.08)

2.71
(SD = 1.19)

2.62
(SD = 1.25)

.623

.001
.004
.074

Table 4 summarizes the results from an ordinary least squares (OLS) multivariate
regression on the willingness to participate in a chocolate boycott. The table displays the effects
of the treatments (friend endorsement and organizational endorsement) compared to noendorsement (reference category). The coefficients indicate the change in the averages when
compared to the no-endorsement condition. As model 1 indicates, compared to no-endorsement,

Downloaded from http://meridian.allenpress.com/mobilization/article-pdf/30/1/73/3501511/i1938-1514-30-1-73.pdf by Simon Fraser University user on 25 March 2025

Table 3 displays the average scores on the willingness to participate in a boycott of chocolate.
Pooling results across the various samples, we find those prompted with a friend’s endorsement
of the campaign are more willing to participate in the boycott (M = 2.76) compared to those
who did not receive an endorsement (M = 2.62) or received an endorsement from an organization (M = 2.67). We conducted an analysis of variance (ANOVA) to compare the three
averages and report on the statistical significance of the F-test. The difference between the three
averages is statistically significant (p = .001). However, the largest difference is clearly in
relation to being asked by a friend (2.76 vs. 2.62/2.67), which increased the willingness to
boycott chocolate (see the regression analysis for further clarity about this point).
We also find crossnational differences in the effectiveness of the endorsements (RQ2).
Again, we conducted an ANOVA within each country. The results show significant effects of
endorsements for the American sample (p = .004) and English-speaking Canadian sample (p <
.001), but not the sample from France (p = .301) or French-speaking Canadian sample (p =
.623). The ANOVA F-test considers differences in the three averages: the control group, the
organizational-endorsement group, and the friend-endorsement group. Using this analytical
approach, the United Kingdom sample does not show significant effects (p = .074); however,
looking specifically at the friend-endorsement group, we find that the experiment increases the
willingness to participate (see regression analysis). The control group and the organizational
endorsement condition do not differ in the United Kingdom sample. In sum, the experiment’s
effects are strongest in English-speaking Canada and the United States.

81

Chocolate and Politics

Table 4. Multivariate Analysis of Experimental Effects on Willingness to Boycott Chocolate

b
Model 1: Experiment
(Reference: no endorsement)
Friend endorsement
Organization
endorsement
R-square,
sample size

0.139
0.046

Model 2: Experiment + Moderators
(Reference: United States)
United Kingdom
France
Canada (English)
Canada (French)
Like chocolate
Awareness of boycott
Efficacy of boycotts
R-square, sample size

Model 1
SE

p-value

0.037
0.037
.002

<.001
.220
6291

b

Model 2
SE

p-value

0.150
0.043

0.035
0.035

<.001
.216

0.126
0.271
0.146
0.119
0.036
0.418
0.336

0.040
0.040
0.043
0.065
0.013
0.041
0.013
.149

.002
<.001
.001
.066
.005
<.001
<.001
6092

Table 5 includes a country-specific analysis to help understand the crossnational differences in the effectiveness of recruitment attempts. In a multivariate analysis by country, we do
not find that the organizational prompt significantly increases willingness to participate in the
boycott (except in the United States; b = 0.185, p = .006). As observed in other results, the
friend endorsement condition is most effective in influencing English-speaking Canadians (b =
0.342, p < .001) and respondents from the United States (b = 0.231, p = .001). In these models,
the effect of the interpersonal prompt is statistically significant (b = 0.158, p = .026) in the
United Kingdom (in contrast to the ANOVA analysis in table 3). Finally, these country-specific
results show France and French-speaking Canada are distinctive groups. In France, the
endorsement from a friend decreases willingness to participate in the boycott (b = -0.097, p =
.182), which is in sharp contrast to other countries where a friend’s endorsement of the
campaign increases willingness to participate. However, the difference does not reach statistical
significance. We noted this identical pattern in table 3 (ANOVA analysis). In the case of
French-speaking Canadians, a friend’s endorsement does not significantly affect willingness to
boycott (b = 0.186, p = .189).

Downloaded from http://meridian.allenpress.com/mobilization/article-pdf/30/1/73/3501511/i1938-1514-30-1-73.pdf by Simon Fraser University user on 25 March 2025

subjects receiving the friend endorsement treatment express a greater willingness to participate
in a chocolate boycott (b = 0.139, p < .001). In contrast, the organizational endorsement condition does not have a significant effect compared to receiving no endorsement (b = 0.046, p =
.220). The findings replicate the results in table 3 (bivariate analysis).
The same pattern of treatment effects remains after controlling for country and moderators
(table 4, model 2). Adding country and moderators significantly increases the explained variance of the model from .002 to .149. Compared to the United States (reference group), subjects
in other countries are more willing to boycott chocolate. French speakers in Canada are the
exception, as they do not significantly differ from American respondents' willingness to boycott
chocolate (b = 0.119, p = .066).
As for the moderators, the three variables are significant predictors of willingness to boycott
chocolate after controlling for the experimental conditions and country. Liking chocolate, being
aware of the campaign, and the perceived efficacy of boycotting increase people’s willingness to
participate in a boycott of chocolate.

82

Mobilization

Table 5. Multivariate Analysis of Experimental Effects on Willingness to Boycott Chocolate by Country
b
0.231
0.185
0.065
0.474
0.396

b
0.158
-0.003
-0.010
0.429
0.350
b
-0.097
-0.087
0.038
0.445
0.271
b
0.342
0.077
0.057
0.391
0.316
b
0.186
0.010
-0.009
0.077
0.353

SE
0.067
0.067
0.023
0.087
0.024
.199
SE
0.071
0.070
0.026
0.087
0.026
.143
SE
0.072
0.072
0.029
0.077
0.024
.112
SE
0.082
0.080
0.029
0.092
0.031
.138
SE
0.141
0.147
0.058
0.184
0.051
.131

p-value
.001
.006
.006
<.001
<.001
1646
p-value
.026
.970
.698
<.001
<.001
1479
p-value
.182
.229
.196
<.001
<.001
1480
p-value
<.001
.337
.048
<.001
<.001
1124
p-value
.189
.944
.878
.674
<.001
363

To identify the extent to which the endorsements depend on other factors, we introduced a
series of interaction terms composed of friend-endorsement treatment with, respectively, liking
chocolate, the perceived efficacy of boycotting, prior knowledge about the campaign, and
country (table 6). As model 1 (table 6) shows, the interaction between the friend-endorsement
and liking chocolate is significant and negative (b = -0.059, p = .030), meaning that, for the
people who love chocolate, a friend’s endorsement of boycotting chocolate is slightly less
effective. Prior awareness of the campaign (model 2, b = -0.163, p = .056) and perceived effectiveness of boycotting (model 3, b = 0.024, p = .346) do not significantly alter the effects of a
friend’s endorsement on willingness to boycott (the main effects are significant as noted above
in relation to table 4). Finally, as observed in tables 3 and 5, the friend-condition is more
persuasive for English-speaking Canadians (model 4, b = 0.211, p = .008) than for other subjects
included in the experiment.

Downloaded from http://meridian.allenpress.com/mobilization/article-pdf/30/1/73/3501511/i1938-1514-30-1-73.pdf by Simon Fraser University user on 25 March 2025

United States
Friend endorsement
Organization endorsement
Like chocolate
Awareness of boycott
Efficacy of boycotts
R-square, sample size
United Kingdom
Friend endorsement
Organization endorsement
Like chocolate
Awareness of boycott
Efficacy of boycotts
R-square, sample size
France
Friend endorsement
Organization endorsement
Like chocolate
Awareness of boycott
Efficacy of boycotts
R-square, sample size
Canada (English)
Friend endorsement
Organization endorsement
Like chocolate
Awareness of boycott
Efficacy of boycotts
R-square, sample size
Canada (French)
Friend endorsement
Organization endorsement
Like chocolate
Awareness of boycott
Efficacy of boycotts
R-square, sample size

83

Chocolate and Politics

Table 6. Differing Effectiveness of Friend Endorsement
b
0.376

Friend
endorsement
Organization 0.044
endorsement

Model 1
SE
p
0.110 .001

b
0.175

Model 2
SE
p
0.037 <.001

b
0.096

Model 3
Model 4
SE
p
b
SE
p
0.067 .151 0.113 0.038 .003

0.035

.208

0.044

0.035

.213

0.044

0.035

.212 0.043

0.035

.217

0.125

0.040

.002

0.126

0.040

.002 0.126

0.040

.002

0.126

0.040

.002

France

0.270

0.040

<.001 0.271

0.040 <.001

0.271

0.040

<.001 0.271

0.040 <.001

Canada
(English)
Canada
(French)
Like
chocolate
Boycott
awareness
Efficacy of
boycotts

0.145

0.043

.001

0.144

0.043

.001

0.146

0.043

.001 0.080

0.050

.111

0.121

0.065

.061

0.120

0.065

.065

0.120

0.065

.063 0.121

0.065

.062

0.056

0.016

<.001 0.036

0.013

.006

0.036

0.013

.005 0.036

0.013

.005

0.417

0.041

<.001 0.474

0.051 <.001

0.417

0.041

<.001 0.420

0.041 <.001

0.336

0.013

<.001 0.337

0.013 <.001

0.328

0.015

<.001 0.336

0.013 <.001

0.024

0.026

0.346

Interaction Effects
Friend ×
-0.059 0.027
Like chocolate
Friend × Boycott
awareness

.030
-0.163 0.085

0.056

Friend ×
Efficacy
Friend × Canada
(English)
R-square, sample
size

0.211
.150

6093

.150

6093

.149

6093

0.079

0.008

.150

6093

DISCUSSION
Neilson and Paxton (2010) examine the role of social networks in political consumerism,
explaining the importance of social ties for increasing awareness and motivation to participate
in efforts towards social change. We offer insights into the role of social networks in political
behavior using a survey-based experiment to examine whether friends or organizations are more
effective in increasing the willingness to participate in political consumerism. Overall, people
who receive a prompt from a friend to boycott are more willing to participate than those who
receive a prompt from an organization (RQ1), albeit the effect sizes are small.
Organizations effectively influence subjects in the United States (table 5) but not in other
countries. In contrast, correlational analysis suggests following social groups (on social media)
predicts political consumerism in the United States, United Kingdom, and France (Boulianne,
Copeland, and Koc-Michalska 2024). This study also suggested groups may be less effective
in mobilizing citizens to participate in political consumerism in the United States compared to
the United Kingdom and France (versus the current study suggesting the opposite). The crossnational variations in the effectiveness of organizations merit additional attention. Earl, Copeland, and Bimber (2017: 142) report that only eight percent of their sample boycotted because of
organizationally directed campaigns; they suggest this is because today’s political consumer is
entrepreneurial, may distrust the government and organizations, and can use digital media to
make their own self-directed consumer choices. While organizational prompts may be effective
in the United States (see table 5), citizens rarely report engaging in boycotts led by an organization.

Downloaded from http://meridian.allenpress.com/mobilization/article-pdf/30/1/73/3501511/i1938-1514-30-1-73.pdf by Simon Fraser University user on 25 March 2025

United
Kingdom

84

Mobilization

Downloaded from http://meridian.allenpress.com/mobilization/article-pdf/30/1/73/3501511/i1938-1514-30-1-73.pdf by Simon Fraser University user on 25 March 2025

As noted, Boulianne, Copeland, and Koc-Michalska (2024) use a three-country sample
(United States, United Kingdom, France) to explore the role of social groups in political consumerism. The present experiment adds a fourth country (Canada), which helps to untangle
crossnational variations related to the effectiveness of interpersonal versus organizational
prompts to participate. The English-speaking Canadian sample reacts similarly to the American
sample, meaning interpersonal prompts are more effective than organizational prompts.
While the United Kingdom sample is also English speaking, this sample did not respond
to the organizational prompt. In particular, the averages are identical for the no-endorsement
control condition and for the organizational endorsement. This finding is surprising, considering
the organization in this experiment (Ethical Consumer) is based in the United Kingdom. Further
research should investigate whether this organization has low credibility in the United Kingdom, which would explain the null effects in that country.
In the United Kingdom, United States, and English-speaking Canada, interpersonal prompts
effectively increase willingness to participate (table 5, RQ2). France and French-speaking Canada
offer some unusual findings. In all countries, a friend’s endorsement of the campaign increases
willingness to participate compared to the control (no endorsement) condition. However, for
France, being exposed to the prompt from a friend decreases, on average, the willingness to
participate. This set of findings implies that social network effects may operate differently in these
four countries and across language groups in the different countries.
We consider several explanations to explain the distinct effects in France and Frenchspeaking Canada. First, one could argue that respondents in France feel quite passionate about
French chocolate, making them impervious to attempts to manipulate the consumption of
chocolate. However, this explanation is not plausible given the high averages for willingness to
boycott (table 3, M = 2.91) for respondents from France compared to other countries. In the
absence of recruitment attempts, respondents from France are more willing, on average, to boycott chocolate. A second possible explanation could be an error in the translation, which also
seems unlikely because the French-speaking groups react differently to the experiment. Being
prompted by a friend to boycott decreases willingness to participate in a boycott among respondents in France yet slightly increases willingness to participate in a boycott among
Canadians respondents who are French speaking (albeit neither effect reaches statistical
significance). French-speaking respondents in Canada demonstrate a similar pattern to other
countries, but perhaps due to the smaller sample size (n = 374), the effect is not statistically
significant. A third and most plausible explanation relates to a combination of culture (aside
from language) and country differences. Perhaps people in French-speaking cultures are more
resistant to social influence. As noted, Wang, Chang, and Chen (2021) found that the effectiveness of their manipulation depended on social comparison information; perhaps Frenchspeaking cultures care less about following the direction of their friends and relatives. In France,
this resistance to social influence may decrease willingness to participate, whereas, in Frenchspeaking Canada, it merely reduces the effect size of social influence.
Existing scholarship suggests that income inequality and democracy scores are positive
predictors of participation in political consumerism (Boulianne, Copeland, and Koc-Michalska
2024; Chow, Oh, and Anand 2022). Based on these macrolevel indicators, we would expect the
US respondents to report lower levels of political consumerism than other groups of respondents. Aside from French-speaking Canadians, the other groups express greater willingness to
boycott than US respondents. However, in terms of our experiment, these contextual factors do
not explain the pattern of effects. While the U.S. and Canada differ in income inequality and
democracy scores, they respond similarly to the experiment. The effects are similar in the U.S.
and English-speaking Canada (RQ2). Instead, the key distinctions seem to relate to language,
as noted in the prior paragraph.
In a study of social influence on political behavior, Bruce Bimber and Homero Gil de
Zúñiga (2022) measured whether the respondents were aware of others’ political activities (for
related experiments, see Bond et al. 2017 and Haenschen 2016). These studies fall under the
domain of social network effects on political behavior. They note that the literature in this field

Chocolate and Politics

85

Downloaded from http://meridian.allenpress.com/mobilization/article-pdf/30/1/73/3501511/i1938-1514-30-1-73.pdf by Simon Fraser University user on 25 March 2025

is based “on experimental studies of single behaviors in one country” (Bimber and Gil de
Zúñiga 2022: 135; e.g., Bond et al. 2017). Using a crossnational survey, they find that “social
influence on political behavior has a high degree of context independence, despite falling somewhat short of being thoroughly universal” (136). Our study includes one country from their list
of countries demonstrating large effects of social influence on political consumerism (United
Kingdom) and one from their list of countries showing null effects (United States). In other
words, in their nineteen countries, the United Kingdom exhibited a large social influence on
political consumerism, and the United States was among the countries with the lowest effects
related to political consumerism. Neither Canada nor France is included in their survey. We
find both the United States and the United Kingdom exhibit social network effects (i.e., interpersonal prompts increase participation). In contrast to self-reports, this experiment finds strong
differences in the role of social networks on political behavior across samples.
Studies of interpersonal influence on political consumerism have rarely been conducted.
Zorell and Denk (2021) find people are more likely to report being influenced by others to
participate in political consumerism in their sample of Swedish citizens. However, the finding
is correlational and relies on self-reports about the motivation to participate. As noted in relation
to Earl, Copeland, and Bimber (2017) versus the current study versus Bimber and Gil de Zúñiga
(2022), self-reports offer different findings than experiments. Yet, the role of “being asked” to
participate in political participation is well-supported in studies of protest (Anduiza, Cristancho,
and Sabucedo 2014; De Vydt and Ketelaars 2020; Schussman and Soule 2005; Walgrave and
Wouters 2014) and elections (Bhatti et al. 2016; Green, McGrath, and Aronow 2013). We
contribute towards understanding the conditions under which being asked to participate has a
greater impact on citizens’ participation.
These findings have implications for mobilization practices of social and political campaigns.
In particular, friends and interpersonal social networks are important for mobilization. Social
media help create and sustain these interpersonal networks, and, as such, can be critical tools
for mobilization. People can use social media to mobilize their friends to participate in political
consumerism, as illustrated in a cross-sectional survey of Canadians (Boulianne 2022). Also,
social media communication can span borders, and thus, campaigns need to consider what type
of messaging might influence people in different parts of the world. In particular, there may be
more effective methods to mobilize participants in France than friends and interpersonal ties.
Campaigns need to consider alternative messaging to persuade respondents in France to join a
boycott campaign. While our study is limited to only four countries and two language groups,
we suggest that further research consider more languages and countries to understand mobilization strategies for global campaigns.
Finally, the small effect sizes are based on a single sentence presented to respondents.
While we find a significant effect, it is likely temporary. Instead, mobilization campaigns
should consider findings from advertising research. A meta-analysis of advertising research
suggests that respondents are persuaded by at least eight exposures (Schmidt and Eisend 2015).
In the context of this experiment, we would need to expose respondents to the prompts at least
eight times to effectively change boycotting in the long term.
We examine how opinions about the target of activism (chocolate) interact with a recruitment attempt (RQ3). Those who like chocolate are less affected by a friend’s prompt to
boycott chocolate (see table 6). We might extend this finding to other forms of activism.
Requests to participate in protests directed at corporations or governments will depend on how
the protest is framed. For example, if a protest or election campaign is framed in terms of
environmental issues, a person will be more responsive to the recruitment attempt if they care
about the environment. The recruitment attempt will be less effective if they do not care about
the environment. As mentioned, previous research identifies pre-existing attitudes as a factor
that can affect people’s willingness to engage in political consumerism (Hong and Li 2021;
Jungblut and Johnen 2021; Wang, Chang, and Chen 2021). This implies that, even with a
friend’s encouragement, people will participate if their views align with the cause, as noted by
Klandermans and Oegema (1987).

86

Mobilization

REFERENCES
Anduiza, Eva, Camilo Cristancho, and José M. Sabucedo. 2014. “Mobilization Through Online Social
Networks: The Political Protest of the Indignados in Spain.” Information, Communication & Society
17(6): 750-764.
Balch, Oliver. 2021. “Mars, Nestlé and Hershey to face child slavery lawsuit in the U.S.” The Guardian,
February 12, 2021. https://www.theguardian.com/global-development/2021/feb/12/mars-nestleand-hershey-to-face-landmark-child-slavery-lawsuit-in-us.
Bennett, W. Lance. 2013. The Logic of Connective Action: Digital Media and the Personalization of Contentious Politics. Cambridge University Press.
Bhatti, Yosef, Jens Olav Dahlgaard, Jonas Hedegaard Hansen, and Kasper M. Hansen. 2016. “Is Door-toDoor Canvassing Effective in Europe? Evidence from a Meta-Study Across Six European
Countries.” British Journal of Political Science 49(1): 279-290.
Bimber, Bruce, and Homero Gil de Zúñiga. 2022. “Social Influence and Political Participation Around
the World.” European Political Science Review 14(2): 135-154.
Bond, Robert M., Jaime E. Settle, Christopher J. Fariss, Jason J. Jones, and James H. Fowler. 2017. “Social
endorsement cues and political participation.” Political Communication 34(2): 261-281.
Boulianne, Shelley, Lauren Copeland, and Karolina Koc-Michalska. 2024. “Digital Media and Political
Consumerism in the United States, Britain, and France.” New Media & Society 26(4): 2110–2130.
Boulianne, Shelley, Nicole Houle. 2024. “Chocolate and politics: A cross-national, survey-based
experiment on recruitment to a boycott campaign.” Online Appendix.
https://figshare.com/articles/online_resource/Boulianne_Houle_Chocolate_Politics_Appendix_pdf/
27852300?file=50617566.
Boulianne, Shelley. 2024. “Survey-based experiment on recruitment to a boycott campaign about chocoFighalate.” (a dataset). Figshare. Available at https://doi.org/10.6084/m9.figshare.25555893.v1
Boulianne, Shelley. 2022. “Socially mediated political consumerism.” Information, Communication &
Society 25(5): 609-617.
Canadian Heritage. 2019. Some Facts on the Canadian Francophonie. Government of Canada.
https://www.canada.ca/en/canadian-heritage/services/official-languagesbilingualism/publications/facts-canadian-francophonie.html.
Chow, Dawn Yi Lin, Ga-Eun (Grace) Oh, and Amitabh Anand 2022. “Exploring the patterns in political
consumption: A review and identification of future research agenda.” International Journal of
Consumer Studies 46(6): 2128-2152.
Copeland, Lauren, and Shelley Boulianne. 2022. “Political Consumerism: A Meta-Analysis.” International Political Science Review 43(1): 3-18.
De Vydt, Michiel, and Pauline Ketelaars. 2020. “Linking Consensus to Action: Does Frame Alignment
Amongst Sympathizers Lead to Protest Participation?” Social Movement Studies 20(4): 439-458.

Downloaded from http://meridian.allenpress.com/mobilization/article-pdf/30/1/73/3501511/i1938-1514-30-1-73.pdf by Simon Fraser University user on 25 March 2025

Being aware of the boycott campaign and one’s perception of the efficacy of boycotts have
a significant impact on people’s willingness to boycott. Still, these variables do not impact the
effectiveness of the experimental manipulation (RQ3). In other words, being aware of the
campaign increases the likelihood of participating but does not influence the effectiveness of
recruitment attempts. People are more willing to boycott if they think boycotting is an effective
way to influence the government (Kelm and Dohle 2018; Stolle, Hooghe, and Micheletti 2005),
but views about the efficacy of this activity do not alter the effects of recruitment attempts.
In sum, this experiment contributes to scholarship about the role of social networks in
political behavior, engaging debates about whether organizations still matter. In particular, the
effectiveness of organizational recruitment efforts pale compared to interpersonal influences.
However, organizational prompts cannot be completely dismissed; their effectiveness may
depend on social context (country). In particular, further crossnational research will help
examine how organizations matter and for what types of political activities. Our results suggest
interpersonal networks will matter more for recruitment than organizational efforts, particularly
related to noninstitutionalized forms of political participation. The findings from this experiment on political consumerism align with the results from self-reported data coming from
surveys of protestors (Anduiza, Cristancho, and Sabucedo 2014; Van Laer 2017).

Chocolate and Politics

87

Downloaded from http://meridian.allenpress.com/mobilization/article-pdf/30/1/73/3501511/i1938-1514-30-1-73.pdf by Simon Fraser University user on 25 March 2025

Doherty, David, and Peter J. Schraeder. 2018. “Social Signals and Participation in the Tunisian Revolution.” The Journal of Politics 80(2): 675-691.
Earl, Jennifer, Lauren Copeland, and Bruce Bimber. 2017. “Routing Around Organizations: Self-Directed
Political Consumption.” Mobilization: An International Quarterly 22(2): 131-153.
Fisher, Dana R., and Marije Boekkooi. 2010. “Mobilizing Friends and Strangers: Understanding the Role of the
Internet in the Step-It-Up Day of Action.” Information, Communication, & Society 13(2): 193-208.
Freedom House. 2021. Freedom in the World, 2021. Available at: https://freedomhouse.org/report/freedomworld.
Green, Donald P., Mary C. McGrath, and Peter M. Aronow. 2013. “Field Experiments and the Study of
Voter Turnout.” Journal of Elections, Public Opinion and Parties 23(1): 27-48.
Haenschen, Katherine. 2016. “Social pressure on social media: Using Facebook status updates to increase
voter turnout.” Journal of Communication 66(4): 542-563.
Haugsgjerd, Atle, Rune Karlsen, and Kari Steen-Johnsen. 2023. “Uninformed or Misinformed in the
Digital News Environment? How Social Media News Use Affects Two Dimensions of Political
Knowledge.” Political Communication 40(6): 700–718.
Hong, Cheng, and Cong Li. 2021. “Will Consumers Silence Themselves when Brands Speak Up about
Sociopolitical Issues? Applying the Spiral of Silence Theory to Consumer Boycott and Buycott
Behaviors.” Journal of Nonprofit & Public Sector Marketing 33(2): 193-211.
Inclán, María, and Paul D. Almeida. 2017. “Ritual Demonstrations Versus Reactive Protests: Participation
Across Mobilizing Contexts in Mexico City.” Latin American Politics and Society 59(4): 47-74.
Jungblut, Marc, and Marius Johnen. 2021. “When Brands (Don’t) Take My Stance: The Ambiguous
Effectiveness of Political Brand Communication.” Communication Research 49(8): 1092-1117.
Kantar, 2024. How to Get a Representative Sample for Your Online Surveys. https:www.kantar.com/
inspiration/research-services/how-to-get-a-representative-audience-for-your-online-surveys-pf.
Kaun, Anne, Anders O. Larsson, and Anu Masso. 2024. “Automating Public Administration: Citizens’
Attitudes towards Automated Decision Making across Estonia, Sweden, and Germany.” Information,
Communication & Society 27(2): 314-332.
Kelm, Ole, and Marco Dohle. 2018. “Information, Communication and Political Consumerism: How
(Online) Information and (Online) Communication Influence Boycotts and Buycotts.” New Media &
Society 20(4): 1523-1542.
Klandermans, Bert, and Dirk Oegema. 1987. “Potentials, Networks, Motivations, and Barriers: Steps
Towards Participation in Social Movements.” American Sociological Review 52(4): 519-531.
Klandermans, Bert, and Jacquelien van Stekelenburg. 2022. “Social Psychology and Political Participation.” Pp. 167-182 in The Oxford Handbook of Political Participation, edited by Marco Giugni
and Maria Grasso. New York: Oxford University Press.
McAdam, Doug, and Ronnelle Paulsen. 1993. “Specifying the Relationship Between Social Ties and
Activism.” American Sociological Review 51(3): 464-481.
Mutz, Diana. 2011. Population-based survey experiments. Princeton University Press.
Neilson, Lisa. A., and Pamela Paxton. 2010. “Social Capital and Political Consumerism: A Multilevel
Analysis.” Social Problems 57(1): 5-24.
Onishi, Norimitsu. 2001. “The bondage of poverty that produces chocolate.” New York Times. July 29,
2001. https://www.nytimes.com/2001/07/29/world/the-bondage-of-poverty-that-produceschocolate.html.
Owens, Jasmine. 2023. Ethical Chocolate. Ethical Consumer. https://www.ethicalconsumer.org/fooddrink/shopping-guide/chocolate.
Schlozman, Kay Lehman, and Henry E. Brady. 2022. “Political Science and Political Participation.” Pp.
25-44 in The Oxford Handbook of Political Participation, edited by Marco Giugni and Maria Grasso.
New York: Oxford University Press.
Schussman, Alan, and Sarah A. Soule. 2005. “Process and Protest: Accounting for Individual Protest
Participation.” Social Forces 84(2): 1083-1108.
Schmidt, Susanne, and Martin Eisend. 2015. “Advertising repetition: A meta-analysis on effective
frequency in advertising.” Journal of Advertising 44(4): 415-428.
Seawright, Jason, and John Gerring. 2008. “Case selection techniques in case study research: A menu of
qualitative and quantitative options.” Political Research Quarterly 61(2): 294-308.
Snow, David, Louis A. Zurcher Jr., and Sheldon Ekland-Olson. 1980. “Social Movements: A Microstructural Approach to Differential Recruitment.” American Sociological Review 45(5): 787-801.
Stolle, Dietlind, Marc Hooghe, and Michele Micheletti. 2005. “Politics in the Supermarket: Political Consumerism as a Form of Political Participation.” International Political Science Review 26(3): 245-269.

88

Mobilization

Downloaded from http://meridian.allenpress.com/mobilization/article-pdf/30/1/73/3501511/i1938-1514-30-1-73.pdf by Simon Fraser University user on 25 March 2025

Van Laer, Jeroen. 2017. “The Mobilization Dropout Race: Interpersonal Networks and Motivations
Predicting Differential Recruitment in a National Climate Change Demonstration.” Mobilization: An
International Quarterly 22(3): 311-329.
Verba, Sidney, Kay Lehman Schlozman, and Henry Brady. 1995. Voice and equality: Civic Voluntarism
in American politics. Cambridge, MA: Harvard University Press.
Walgrave, Stefaan, and Ruud Wouters. 2014. “The Missing Link in the Diffusion of Protest: Asking
Others.” American Journal of Sociology 119(6): 1670-1709.
Wang, Chih-Chien, Shu-Chen Chang, and Pei-Yeng Chen. 2021. “The Brand Sustainability Obstacle:
Viewpoint Incompatibility and Consumer Boycott.” Sustainability 13(9): 1-23.
Wexler, Jennifer. 2018. Ethical Chocolate. Ethical Consumer. https://www.ethicalconsumer.org/fooddrink/shopping-guide/chocolate
Whoriskey, Peter, and Rachel Siegel. 2019. “Cocoa’s Child Laborers.” Washington Post, June 5, 2019.
https://www.washingtonpost.com/graphics/2019/business/hershey-nestle-mars-chocolate-childlabor-west-africa/.
World Bank. 2021. Gini index. https://data.worldbank.org/indicator/SI.POV.GINI
Zorell, Carolin V., and Thomas Denk. 2021. “Political Consumerism and Interpersonal Discussion
Patterns.” Scandinavian Political Studies 44(4): 392-415.