The Exodus Hosts:
Framing Effects of the Venezuelan Refugee Crisis in Latin American Citizens
Master’s Thesis
Handed in by Kimberly Yánez (12091472) on May 29th, 2020 Supervisor: dhr. dr. R. Rachid Azrout
University of Amsterdam
Faculty of Social and Behavioural Sciences
Department of Communication Science. Graduate School of Communication Erasmus Mundus Master’s “Journalism, Media & Globalisation”
Abstract
This study is set to extend the knowledge about framing effects in non-Western societies where immigrants and the host society have less cultural differences. It explores the framing effects of the Venezuelan refugee crisis in Latin American citizens by measuring their attitudes towards mass immigration. I tested the exposure to the victim and threat frames in a survey experiment (n = 372) to determine if they have negative or positive effects on
anti-immigration attitudes. Interaction models with three moderator variables (socioeconomic status, number of immigrants and identification) were carried out to account for existing conditions that can affect the strength of news content. Results show expected outcomes from the main predictors and no significant interactions were found to be weakening their strength. Findings suggest that the theory of frame valence used in North American and European studies can also be determinants of immigration attitudes in Latin America. They also evidence the challenges of facing with an unprecedented flow of refugees for the host countries of Latin America, their governments and the media.
Introduction
The Venezuelan refugee crisis across Latin America is reaching Syrian proportions since it became noticeable in 2014 (Organization of American States [OAS], 2019a).
Accordingly, governments of the host countries created special immigration policies to grant refugee and political asylum (The UN Refugee Agency [UNHCR], 2018). For their citizens such a mass welcoming can be either perceived as beneficial for the economy, a natural response to a humanitarian turmoil, or the reason for the rising of anti-Venezuelan sentiments (Bennouna, 2019). Studying news frames on immigration is key to understand what shapes citizens' attitudes towards it (Bosilkov & Drakaki, 2018).
This being said, Bleich, Bloemraad and de Graauw (2015) argue the importance of understanding the way host societies deal with mass immigration based on how this phenomenon is presented in media. That influence has societal consequences that shape public opinion and lead to consensus in policy changes (Lahav & Courtemanche, 2012). Hence, it is relevant to counter potential threat perceptions, reducing uncertainty (Esses, Medianu & Lawson, 2013), and to address issues of assimilation between the two groups more pragmatically (Croucher, 2013).
Moreover, compared to extensive research about the refugee crisis tackled in Europe (e.g. Figenschou, Beyer & Thorbjørnsrud, 2015) and Canada (e.g., Lawlor & Tolley, 2017) due to conflicts in the Middle East and Africa, studies about the crisis in Latin America and its representation in media are rather scarce. In their findings, immigration framing in Western media presents newcomers as victims, threats to the security or the economy or as agents for multiculturalism (e.g. Ramasubramanian & Miles, 2018; Esses et al., 2013).
Although they are helpful to understand media perceptions of this particular topic, they too pose a challenge for when an immigration phenomenon occurs in a non-Western society with a highly reported situation yet almost unexplored field of study.
Apart from organizations’ reports and briefings (e.g. Fundación Ideas Para la Paz [FIP], 2019), the little academic knowledge on the societal consequences, how media frame the issue and its effects in people’s attitudes towards Venezuelan newcomers, exhibit the relevance to conducting this study. More specifically, most research about immigration and framing focuses on cultural differences between immigrants/refugees and their host societies (e.g, Lecheler, Bos & Vliegenthart, 2015), but in Latin America, many ties bind Venezuelans with the rest of the region, such as language, religion, and history. Thus, these findings are a first attempt to shed light on how cultural similarities and the media influence host receptivity (Croucher, 2013).
What does not change is that news framing is said to explain the journalistic conventions for presenting certain topics; a process that leaves an impression on news audiences about these topics (de Vreese, 2005). Accordingly, frames influence people’s attitudes towards immigration (e.g. Lecheler & Vliegenthart, 2015). Exploring that process’s dynamic could enable a better understanding of the media’s influential role in how policies are perceived. Therefore, I pose the following research question: To what extent does the
framing of the Venezuelan refugee crisis in Latin American media have an effect on the attitudes of citizens towards mass immigration?
Theoretical Framework
Framing effects on immigration attitudes
According to de Vreese (2005) frames in journalism “are parts of political arguments, journalistic norms, and social movements’ discourse. They are alternative ways of dening issues, endogenous to the political and social world” (p. 53). It is different from
agenda-setting which increases the prominence of an issue, while framing is about how the audience should think about it with the use of subtle wording and syntaxis (Scheufele, 2000). In essence, culture motivates journalists to make sense of stereotypes, values, and shared beliefs (Van Gorp, 2005) using social and syntactic elements for a process understood as framing. As such, there are individual and societal consequences to how news is presented (Scheufele, 2000). They convey audience interpretations of them and are known as framing effects (de Vreese, 2004). In politics, such effects reflect how elites influence citizens’ opinions (Druckman & Nelson, 2003).
On that note, studies about the framing of immigrants have been taken into
consideration due to the mass immigration coming from countries in a war in the Middle East and in Africa (e.g. Ramasubramanian & Miles, 2018). The attention devoted to the study of media coverage of migrants and minorities helped to identified frames suited to these issues like their economic impact, ethnicity, rights, services, security and the deservingness of immigrants to arrive in a new country (Lawlor & Tolley, 2017). Studies suggest that in many instances, immigration news is negative and tends to dehumanize the newcomers;
transforming uncertainty into a crisis for a host society that feels threatened (Esses, Medianu & Lawson, 2013).
Nonetheless, refugees are different from regular immigrants (O’Rourke & Sinnott, 2006). Thus, it is common to victimizing them for being forced to leave their home countries; a humanized focus that appears alongside the threat frame in content analysis research (Van Gorp, 2005; Benson, 2013). Because these findings merit further interpretation, scholars analyse other aspects that have an influence on people’s opinions, like emotions and predispositions that characterize the country who receives them (e.g. Lecheler, et al., 2015; Bosilkov & Drakaki, 2018). These latter conditions could moderate the effects of the frames on immigration attitudes.
Framing and perceptions of the Venezuelan refugee crisis
In order to examine the extent to which the framing of the Venezuelan refugees has an effect on the attitudes of people from Latin America, I built on the work of previous research to choose two issue-specific frames. These are the victim and the threat frames of Benson (2013) found in a French-American comparison on immigration news. On the one hand, the victim frame addresses “immigrants are victims of unjust government’s policies” and “suffer from poverty, lack of access to health care, dangers related to border crossing, etc.” (p. 8). On the other hand, the discursive indicators of threat frames reflect the idea that immigrants are committing crimes, taking jobs, and -if entering illegally- breaking the law from the
beginning. A dichotomy also identified previously by Van Gorp (2005) who found framing of asylum seekers and illegal immigrants in Belgian press presented them as victims or
intruders, who indicated empathy is linked to victimization and stricter border policies are linked to threats.
This is also true for Oxfam (2019) that reported there is ambivalence of immigration attitudes in citizens from Perú, Ecuador, and Colombia:
More than 80% of people in Ecuador, Colombia and Peru say that they understand the circumstances that are forcing millions to migrate and empathize with the situation of the people coming from Venezuela, although around 70% clearly state that they would like stricter border policies to be enforced. (p. 3)
Those discourses are similar to the findings in the exploratory analysis of Fundación Ideas para la Paz (FIP) (2018) concluding that Colombians’ perceptions of Venezuelans immigrants are shaped by the media and the political elites whom the media interview. According to the FIP, the most recurrent frames were the negative perception one
“caracterizada por señalar que la migración venezolana generaba aumentos en la actividad delictiva, especialmente en delitos como homicidio, hurto y contrabando” [characterized for pointing out that Venezuelan migration generated increases in criminal activity, especially in crimes such as homicide, theft and smuggling] (FIP, 2018, p. 10); and the Venezuelan as a
victim where “la población venezolana comienza a considerarse como tal (víctima) están
relacionados con el tráfico de personas, la explotación laboral, la discriminación y la
estigmatización” [the Venezuelan population begins to be considered as such (as victim) are related to human trafficking, labor exploitation, discrimination and stigmatization.] (FIP, 2018, p. 10).
In psychology, frames carry cognitive information that influences retrospective judgments since encountering a positive or negative frame has pronounced dominance over opinions (Isaac & Poor, 2016). In other words, frames have valence. For de Vreese and Boomgaarden (2003) when exposed to a framed content with an inherently negative valence contributes to negative opinions; the same for positive valence. They explain that “given this valence, news frames can be expected to influence public support for various policy
measures.” (p. 362). Regarding the victim/intruder dichotomy van Gorp (2005) also argues their inherent values: framing refugees as victims have a positive connotation because it implies more tolerant reporting and flexibility; while framing them as intruders sends a negative message, motivating people to ask for border restrictions. Therefore, the victim frame leads to pro-immigration attitudes and the threat frame to anti-immigration ones (Benson, 2013). Within this distinction I expected to nd the following effects:
H1: Exposure to news about the Venezuelan refugee crisis presented in a threat frame
will have a positive effect on anti-immigration attitudes.
H2: Exposure to news about the Venezuelan refugee crisis presented in a victim frame
will have a negative effect on anti-immigration attitudes.
By considering these possible scenarios and the “ambivalence” in Latin American attitudes, testing these hypotheses should respond to RQ1: Which frame has stronger effects
Preconditions towards immigration in the host societies
In integrated threat theory, perceived threat is related to less receptivity and leads to the unwillingness to adapt or to assimilate the presence of the other group, true for both
minorities and dominant groups (Stephan & Stephan, 1996, as cited in Croucher, 2013). This theory can explain the extent to which host receptivity is affected by mass media portrayals and/or other factors influencing the hosts’ attitudes (Croucher, 2013). The climate shown in media and preconditions of the host society has a symbiotic relationship. To illustrate this, overtime, sociodemographic variables like economic crises, level of education and changes in political leaning determined the attitudes of Spanish people in three dimensions: tolerant (open to immigration), reluctant (opposed to it) and ambivalent (tenuous levels of tolerance); but also, the representation of immigrants set the tone of public opinion by stereotyping them or reinforcing the prejudice that immigration is a problem (Cea D’ancona, 2015).
Mirroring the premise of integrated threat theory, recent research shows that in different contexts people have access to political information that is not exclusive to the news and are able to deliberate over a topic (Druckman & Nelson, 2003). Conditions previous to any exposure to frames moderate their effects, like the people who depend on media are
influenced more strongly by the news than those who do not follow the news for guidance, yet the latter develop firm opinions grounded in other aspects anyways (Boomgaarden et al, 2011). For instance, using the victim frame would suggest that journalists promote
pro-immigration attitudes, but it is possible “some readers predisposed to such attitudes will find evidence in such coverage to reinforce their worldviews.” (Benson, 2013). That is why I choose to account for moderator variables. Baron and Kenny (1986) define moderators as
variables that affect the strength of the independent variable over the dependent variable. Hence, I pose the last research question:
RQ2: Which moderators affect the strength of the threat/victim frames over
immigration attitudes?
Socioeconomic status (SES). The level of education in Denmark, France, Great Britain, The Netherlands, and Sweden induces “tolerant attitudes toward immigration later in life among students that would otherwise have left secondary education” (Cavaille &
Marshall, 2019; p. 261). On that same note -and based on trade theory,- O’Rourke and Sinnott (2006) found that “high-skilled are less anti-immigrant than the low-skilled”, but it is a factor which correlates with and it is limited by the level of inequality of the host society like socioeconomic position and sociodemographic characteristics (p. 848). The 24 countries they studied are also European, where inequality is much lower than in Latin America, including those in Eastern Europe (Amarante, Galván & Mancero, 2016; Gaviria, 2007). Influenced by inequality, it is expected that high-skilled people in the poorest countries will prefer more restrictions on immigration, and unskilled people could be more open to
immigration. To contextualize it, mass immigration in countries with fewer job opportunities and high poverty rates increases competition while lowering wages (O’Rourke & Sinnott, 2006).
O’Rourke and Sinnott (2006) also made a distinction between immigrants and refugees and they found people are more positive attitudes towards refugees in general; therefore the skill level is not as relevant as if it were for only immigration. Nonetheless, I argue their stance on trade theory and skill level is still valid for this study about refugees. Even though
refugees are less likely to live in their neighborhoods or to have the opportunity of accessing similar jobs, the refugee crisis has represented a significant brain drain to Venezuela, where highly skilled professionals such as scientists, academics, engineers and doctors left the country after seeing a contraction in the demand and could not earn enough income for an acceptable standard of living (García, 2020). Therefore, in these circumstances, a higher SES in a country with high inequality can derive in stronger positive effects on an-immigration attitudes, which is an opposite premise to previous findings where high-skilled people are more tolerant (e.g. Cavaille & Marshall, 2019).
In that sense, SES establishes an index of education, income and occupation that estimates levels of inequality and it is a benchmark on assimilation of immigration groups in comparison with the native-born ones (Waters & Jimenez, 2005). On that note, I expect the following:
H3: Exposure to the threat frame will have stronger positive effects on anti-immigration
attitudes among Latin American citizens with a higher socioeconomic status compared to citizens with a lower socioeconomic status.
Number of immigrants. Contact and conflict theories, in many ways opposite to each other, could explain how exposure to a larger number of immigrants either has positive or negative effects. The contact hypothesis serves to explain policy preferences towards immigration (Ellison, Shin & Leal., 2011). On the one hand, in Allport's intergroup contact hypothesis prejudice reduction is possible when different groups experience an equal status over a situation and have common goals (Pettigrew, 1998). This has greater effects when the exposure is mediated with positive news stories about outgroup members (Wojcieszak &
Azrout 2016). Accordingly, more contact or exposure to a large number of immigrants with different groups is more likely to generate acceptance than distance (Ellison et al., 2011). Based on that outcome, H4 is formulated below:
H4: Exposure to a victim frame will show stronger negative effects on anti-immigration
attitudes in the Latin American citizens from the countries with the highest numbers of Venezuelan refugees than those whose countries have a lesser flow of them.
On the other hand, realistic conflict theory gives a different angle to the previous argument. It is based on the salience of an issue that creates hostility towards immigrants. Physical or mediated exposure to a larger number of them leads to distrust when it becomes prominent is social debates (Azrout, van Spanje, & de Vreese, 2013). It means that in the absence of them the issue is less important, hence the threat feels far away. Thereupon:
H5: Exposure to a threat frame will have stronger positive effects on anti-immigration
attitudes in the Latin American citizens which countries host the highest number of
Venezuelan refugees and will have a weaker effect in those whose countries receive fewer people.
Identification. Latin America’s history of instability led to intergenerational migration flows (Gaviria, 2007; Pagnotta, 2014). Those living outside their country of origin and the Venezuelan refugees are co-experiencers. In some cases being socially distance
(unacquaintance) with others does not differ in experience from those who are socially close, although the latter scenario amplifies the experience (Boothby, Smith, Clark & Bargh, 2016). In that premise, “it is important to note that social proximity is not the only way
attitudes and emotions play a role in the co-experiencers feeling closer (Boothby, et al., 2016, p. 1438). Accordingly, in interpersonal understanding sympathy works as a way to identify with another person’s experience (Meneses & Larkin, 2017). Therefore, sympathetic understanding is tied up in having the same experience with someone else. I thus expect:
H6: Latin American citizens who are immigrants are less likely to be affected by the
threat frame and will have stronger negative effects on anti-immigration attitudes than those who are not immigrants.
Methodology
Experimental design
An experiment was designed to account for the framing effects of immigration news on the Venezuelan refugee crisis. The method is there to measure the effect of an independent variable on a dependent one, where the former is presented as the experimental stimulus (Babbie, 2010). Nonetheless, scholars who study media effects argue experimental surveys expose subjects to frames and test their effects immediately, which could have strong internal validity, but not enough external validity (Barabas & Jerit, 2010). Therefore, an experiment is a way of news consumption that does not resemble real-life habits but relies on the
randomization of conditions as a trustworthy measure of causal effects.
Despite the criticism, survey experiments are “the most suitable methodological approach to disentangle the complex psychological mechanism that might underlie framing effects” (Lecheler et al., 2015, p. 818). In light of these considerations, I addressed the operationalization of news frames and their effects, where the frames of the Venezuelan
refugee crisis are the independent variable, and the Latin American citizens' attitudes towards immigration are the dependent variable.
Participants and data collection
Recruitment was restricted to a varied group of Latin American citizens -except for Venezuelans- living outside or in their home countries. To reach them I applied a
convenience sampling strategy because of time and economic restrictions. I distributed the survey through social media channels. Some limitations of conducting a study with this type of sampling increase the possibility of bias as the first people who could take the
questionnaire are my friends. Data collection started on April 5th and ended on April 22th, 2020. Out of 563 participants, 191 were excluded for having missing data or not completing the survey (150), for being Venezuelan (26), living in Venezuela (1), or not having a Latin American nationality (14). The valid sample number was 372 (see Appendix E for
demographics).
Procedure
To conduct the experimental design I laid out an online survey experiment (see
Appendix A for the questionnaire). It started with participants giving their consent to take the survey after I complied that their personal information will remain anonymous. Before the manipulation, Latin Americans were asked about their demographics for the formulation of the moderator variables. Next, each respondent was randomly assigned to one of three conditions: [1] a victim frame, [2] a threat frame, or [3] a control condition without frame exposure. The distribution was the following: those exposed to the victim frame were 111 (29.8%), to the threat frame 125 (36.3%), while the control group had 136 (36.6%)
participants. Both the indicators and the framed articles were transcribed in Spanish, but the results are translated into English. I tested if the randomization across conditions was consistent with a series of Chi-squares. The randomization was successful across the
variables of age group (X2 (10, N = 372) = 10.87, p = .092), nationality (X2 (44, N = 372) = 17.43, p = .245) and identification (X2 (2, N = 372) = .486, p = .784). But it failed for gender (X2 (6, N = 372) = 4.004, p = .135), where the distribution of females (n = 249) and not females (n = 123) was not equal among conditions.
After the manipulation, I measured the DV, which means that all subjects indicated their level of agreement to immigration statements based on perceived threat. Other elements included in the experiment are a reading attention check to ensure better validity and a manipulation check. Because of time restrictions, there was not a pilot test, hence the manipulation check was displayed once the participants responded to the questions about immigration attitudes. It was considered that checking the manipulation right after the experimental groups had read the articles could prime them to answer in concordance with that reflection. When the experiment was completed I displayed a debriefing explaining the experimental purpose of it and, in case they did not belong to the control group, the news articles shown were manipulated to fit two different frames.
Stimulus material
The independent variable is the exposure to the frames of the Venezuelan refugee crisis (see Appendix B for material). For this study, articles with the victim and threat dichotomy were created considering the elements highlighted in previous research of Benson (2013) and van Gorp (2005). The victim frame presents immigrants as sufferers of unjust situations in
their home countries and of traveling by foot to leave, while the threat frame tackles the fear that immigrants take jobs, suppress wages, and are more likely to commit crimes than non-immigrants. I used their indicators as guidance for frame-building.
Thereupon, findings in the news from different Latin American media outlets helped me select parts of different articles that can function as realistic elements of manipulation: Negative and positive connotations about Venezuelan refugees. To have a regional approach to the crisis, I focused on what is the general sentiment and policies in Latin America and used some countries as examples. The quotes were displayed to be from unknown sources like members of NGOs, although some of them are originally from politicians. The goal was to avoid political leaning from becoming an unintended effect. And lastly, both articles had feature images. I used the visual criteria from van Gorp (2005) to choose the pictures. Those who read the victim frame saw a man in distress with a crying child in arms waiting at the border; while the participants exposed to the threat frame saw a crowd of single men trying to cross the border and being stopped by the police. The participants in the control group
responded to the manipulation check as well. In that regard, it tested whether respondents took over the frame in their mind: “To what extent do you perceive the Venezuelan refugees as victims?” and “To what extent do you perceive the Venezuelan refugees as intruders?”. The Likert scale consisted of seven points, where 1 is “Not at all” and 7 is “Completely”.
Regardless of being exposed to the victim frame (M = 4.46, SD = 1.678) or the threat frame (M = 4.51, SD = 1.649), both groups rated them very much as victims. But in
comparison to the control group (M = 4.31, SD = 1.575), people in the threat frame rate them higher, thus the manipulation check was not successful due to non-significant results of the
ANOVA, F(2, 369) = 0.547, p = 0.579. A different pattern occurs when rating them as intruders. Participants in the victim frame rate them less as intruders (M = 2.05, SD = 1.461) than those exposed to the threat frame (M = 2.43, SD =1.466), hence this showed
non-significant results as well, F(2, 369) = 2.18, p = .123. Nonetheless, there are not
contrasting differences. On average, participants across conditions were inclined to consider Venezuelan refugees as victims than as a threat. I will analyse this failed manipulation check in the discussion section.
Operationalization and measures
Dependent variable. The indicators to measure the immigration attitudes are based on previous research and set to measure the opinions of the subjects. In line with the research of Lecheler et. al (2015), the indicators were constructed to measure issue-specific opinions on a seven-point Likert-type scale, where the values closer to seven illustrate agreement with the statements and those closer to one measure disagreement with the statements. Therefore, a higher value indicates positive effects on anti-immigration attitudes and lower values show negative effects.
I emulated the measurements of perceived threat in the research of Wojcieszak and Azrout (2016), as perceived threats are predictors of prejudice towards immigrants.
Nonetheless, I excluded the statement about religious practices enriching the way of living as most Latin American religions are rooted in Christianity. I used items like “Venezuelan immigrants abuse my country’s public services because they take more out than they put in” and “Venezuelan immigration is good for my country’s labour market”. Although there are less cultural differences among Latin Americans than those between Europeans and Middle
Easterns, I included one item from the work of Velasco, Verkuyten, Weesie, and Poppe (2008), where the focus was on the cultural integration of Muslims in the Netherlands and adapted it to the context of this study: “My country’s identity is being threatened because there are too many Venezuelans” (see Appendix A for all items).
Only one factor was extracted in the factor analysis, although one item does not load that strongly on the factor. It suggests that the scale measures similar patterns of responses. There was 60.7% of the variance with factor loadings from .163 to .851. Besides, the alpha score showed the measurement scale to reach acceptable reliability, meaning there is internal consistency between the items (Cronbach α = .830). The coefficient of item number two (about Venezuelans being good for the economy) is the less consistent (Cronbach’s alpha if item is deleted, α = .863). It could be problematic as the score might be higher without it, but it is expected because reverse coded items tend to be less correlated; thus I kept it. The descriptives of the anti-immigration attitudes scale are M = 2.93 and SD = 1.45.
Moderators. I included some variables that might alter the effect of the independent variable: socioeconomic status, number of immigrants and identification:
Socioeconomic Status: This variable is an index of education, occupation, and income. I designed the following criteria: The measure for education had 10 options, from “No schooling completed” to “Doctorate”, including the option to write other for better clarification. Income was measured by assessing it subjectively: participants were asked where do they think they are financially on a 7-point Likert scale. One was “very poor” and seven was “very wealthy”. For occupation, I consider the International Standard
ten groups with different skill levels. I also included the option of unemployment with the possibility to write down an unlisted reason.
Then, I coded responses of all constituent variables into dichotomies of low (1.00) versus high (2.00) and took the means. These were labeled according to the reduced scale proposed by Dudeja, Bahuguna, Singh, and Bhatnagar (2015), which condenses all in four points ‘Lower’ (1.00), ‘Lower Middle’ (1.33), ‘Upper Middle’ (1.67), and ‘High’ (2.00). For example, if the participant has a master’s degree (2.00), is unemployed (1.00), and considers to position his income on the third point of the scale (1.00), then, her SES is lower middle (1.33) (see coding criteria on Appendix C). In this sample, 59 (15.9%) of them have a low SES, 115 (30.9%) are in the lower middle SES category, 129 (34.7%) account for upper middle SES, and 69 (18.5%) have a high SES.
Number of immigrants: This variable is based on their nationality. To test the H4, I categorized the responses into “countries with a smaller number of Venezuelans” and “countries with a larger number of Venezuelans” (see Appendix D for criteria). Those in the first category host less than 100,000 Venezuelans, while the second category is for countries that host more than 100,000. This criterion is based on a simple overview of the countries that media and politics focus on; where reporting on the countries with over 100,000 Venezuelans is what determines newsworthiness and also mediates people’s perception of the flow of immigrants in their home country when they are reminded more often (e.g. Toro, 2019; Bermúdez, 2019). I counted on 320 (86%) of people from countries with a larger number of refugees and 52 (14%) participants from countries with a smaller number of them.
Identification. I asked the participants if they live in their home country or if they are immigrants themselves. For this, they answered a yes-or-no question: “Do you live in the country of your nationality?”. The participants who live in their country of nationality are 214 (57.5%) and 158 (42.5%) are immigrants themselves.
Data analysis
I tested two datasets from the same sample conducting. One included people who did not pass the attention check (n = 372) (see Appendix E for detailed test results). The other one only had people who passed the attention check (n = 332) (see Appendix F for these results). I chose to analyse the first dataset because it provided more room for analysis of the main effects and it offered a better distribution of the conditions, reducing the possibility of systematic bias. Without those who did not pass, the distribution is highly uneven since the control group did not have an attention check. Despite the differences in the number of participants, the results of both versions were very similar. Regardless, the outcome with the first version might have more measurement error.
Due to a failed randomization check for gender, the following findings are controlled for gender to increase the validity. I performed an ANOVA for the main effects where the conditions were the IV and anti-immigrant attitudes of the DV. I conducted three two-way ANOVA’s to test moderations on the IV: each time anti-immigration attitudes scale was the DV, the conditions as the first factor, and one of the moderators, the second factor.
Results
To respond RQ1 about which frame has stronger effects on people's attitudes, I tested H1 and H2. The former about the possible correlation between exposure to a victim frame and pro-immigration attitudes, and the latter regarding the correlation between exposure to a threat frame and anti-immigration attitudes. One can observe in Figure 1, in comparison to the control group (M = 2.85, SD = 1.47), on average exposure to the threat frame leads to higher values of anti-immigration attitudes (M = 3.18, SD = 1.45). By contrast, when compared to the control group exposure to the victim frame leads to lower values of anti-immigration attitudes (M =2.76, SD =1.41).
Figure 1
Mean Effects of the Victim and Threat frames across Conditions
Note. Means range on a scale from 1 to 7. A higher number indicates more anti-immigration attitudes.
The Levene’s test suggested accepting the null hypothesis that the variances of groups are more or less equal, F(2, 369) = 0.50, p = .605. Hence, it is appropriate to continue since the assumptions have been met. Univariate tests showed the correlation is marginally
significant , 1 F(2, 368) = 2.80, p = .062. Although it's not strong support for the hypotheses, it
is also not a reason to fully reject the hypothesis . Concerning RQ1, by looking at the 2
absolute differences, the threat frame seems to have a bigger difference from the control group (3.18 - 2.85 = 0.33) than the victim frame (2.76 - 2.85 = - 0.09). In detail, a pairwise comparisons illustrates a significant difference in attitudes between the estimated marginal means for the threat and victim frames (mean difference = .414 , p = .030), a marginally significant difference between the control group and the threat frame (mean difference = -.338 , p = .060), but non-significant differences between the victim frame and the control group (mean difference = -.076 , p = .685) . With that said, IV elicits a marginally significant 3
effect on DV and indicates that framing Venezuelans as threatening has stronger effects on host societies than portrait them as victims.
Turning to answer RQ2, about which moderators affect the strength of the frames over immigration attitudes. As follows, H3 which theorizes higher SES and exposure to the threat frame might have a positive effect on anti-immigration attitudes. The assumption was met according to Levene’s test F(11, 360) = .544, p = .873, given the three groups have more or less equal population variances. The main effects seem to be consistent. All the four groups
1 The support of the hypotheses is not strong when considering conventional levels of significance, but I must reject the null-hypothesis with 93.8% certainty. Such a trend towards significance can lead to a 95% of certainty by increasing the number of participants.
2 The second sample version, which excludes the 40 people did not pass the attention check showed the same differences in means, but had a non-significant result, F(2, 328) = 1.84, p = .160.
3 The estimated marginal means used for the pairwise comparison are slightly different from the descriptives of the anti-immigration attitudes scale, because the former ones are controlled for gender (see tables E22 and E23).
of SES, showed more anti-immigration attitudes when exposed to the threat frame than those exposed to the victim frame.
The interaction effects observed in Figure 2 when compared to the control group, indicate low SES (3.40 - 2.93 = 0.47) and lower middle (3.03 - 2.64 = 0.39) moderated the effects of the threat frame more than upper middle SES (3.30 - 2.95 = 0,35) and high SES (3.04 - 2.93 = 0.11), which is not in line with the hypothesis. By contrast, when exposed to the victim frame high SES (2.92 - 2.93 = -0.01) and upper middle (2.78 - 2.95 = - 0,17) had more positive effects on anti-immigration attitudes than low SES (2.69 - 2.93 = -0,24) and lower middle (2.63 - 2.64 = -0.01). Nevertheless, a lack of difference in differences is reflected in the tests of between-subjects effects: the interaction of SES on those exposed to the threat frame is not statistical significance, F(6, 359) = .164, p = .986. These factors do not influence the strength of the IV and positive effects on anti-immigration attitudes. On that account, I must accept the null hypothesis that there is no moderation and reject H3.
Figure 2
For the variable of the number of immigrants, I examined its moderation from two perspectives. Firstly, according to contact theory, if participants are from countries with a larger number of Venezuelan immigrants and exposed to the victim frame, then there will be lower values of anti-immigration attitudes than those from countries with less of them (H4). Secondly, according to conflict theory, larger exposure to outgroups and a threat frame would result in higher values of anti-immigration attitudes than those with less exposure (H5). To start, the Levene’s test reported significant difference variances, thus the assumption is not met, F(5, 366) = 2.82, p = .016. This is due to the notable difference between the groups. There are 268 more people from countries with a larger number of Venezuelans (n = 320) than the ones exposed to a smaller number (n = 52).
Additionally, in this test the main effects showed to be consistent with the previous ones: across the groups of this variable, the frames have the expected outcome of H1 and H2. Looking at the interaction in Figure 3, the difference between the victim frame and the control condition for participants from countries with a large number of refugees (2.87 - 2.94 = -0.07) and the same frame and the control conditions for individuals from countries with a small number of refugees (2.30 - 2.04 = 0.26), indicated some moderation. In line with H4, with more refugees, the victim frame negatively affects anti-immigration attitudes, but with fewer refugees, the victim frame positively contributes to anti-immigration attitudes.
But the trend is not the same as the difference between the threat frame and the control condition for people from countries with more Venezuelans (3.28 - 2.94 = 0.34) and that same difference among those from countries with fewer Venezuelans (2.60 - 2.04 = 0.56). In both groups, the threat frame leads to more anti-immigration attitudes despite the number of
immigrants, thus there is no sign of moderation. Having said that, there is no significant interaction with the IV that could explain the results, F (2, 365) = .211, p = .810; therefore, I must reject H4 and H5. However, it is important to remark that the tests of between-subjects effect show the number of immigrants the host countries receive affects the strength of the predictor (p = .001).
Figure 3
Estimated Marginal Means of Immigration Attitudes by Interaction of Number of Immigrants and Conditions
The variable of identification was used to test H6 by considering the moderated effect of being an immigrant or not on the group that was exposed to the threat frame. It was expected that those who are immigrants themselves would still show negative effects on anti-immigration attitudes. No significant results of the Levene’s test meant the assumption of variances was met, F(5, 366) = 1.12, p = .348. When looking at the difference between the threat frame and the control group of those who are immigrants (3.08 - 2.91 = 0.17) and
between those who are not (3.25 - 2.79 = 0.46), the interaction is in line with the hypothesis. Identifying with Venezuelans who left their home country like them affects the strength of the threat frame by reducing anti-immigrant sentiments, as seen in Figure 4.
Figure 4
Estimated Marginal Means of Immigration Attitudes by Interaction of Identification and Conditions
Nevertheless, the tests of between-subjects effects suggest there is no statistical significance of this variable affecting the strength of the IV, F(2, 365) = .329, p = .720. I must accept the null hypothesis and reject H6. Overall, the moderator variables were not statistically significant to respond to RQ2 with much accuracy, but there was evidence of some differences of differences with the influence of the number of immigrants and
identification. Just like the other interaction models, the main effects were constant when testing H6.
Discussion and Conclusion
This study tested through an experiment the extent to which the framing of the Venezuelan refugee crisis influences the immigration attitudes of Latin Americans. The findings suggest the inherent valence of the victim and the threat frames leads to the expected directions and show consistent patterns across all tests, while there were non-significant moderations on the IV. To begin with, a failed manipulation check, which reflects some aspects of Latin Americans. The participants in the threat frame thought of Venezuelans as victims and not as intruders when asked. That adds up to the reported ambivalence in their attitudes because they might “feel empathy and a desire to welcome migrants, while also harbouring fears and concerns” (Oxfam, 2019, p. 3). Ambivalence could also be the outcome of cultural similarities between Venezuelans and their hosts since in experiments of framing effects in Europe (e.g. Lecheler et al., 2015) and North America (e.g. Lahav &
Courtemanche, 2012) cultural differences made individuals more susceptible to stimuli.
Nevertheless, a failed manipulation check does not contradict the results of the main effects. When portraying Venezuelans as victims participants showed pro-immigration
attitudes while framing them as threatening led to more anti-immigration sentiment. Although media plays an important role in influencing opinions about policies (De Vreese &
Boomgaarden, 2013), and there is evidence in this study that could explain why that is the case, it is not to a large extent the assumption of the research question can be met. More importantly, this interpretation supports H1 and H2 with a weak statistical significance, hence
the following analysis should be addressed carefully. Moreover, regardless of the interaction models’ results, the framing effects remained constant in all of them as well.
The likelihood of having different effects is low, thus it is argued there is not random chance due to a systematic trend. A pattern of the main effects implies the inherent values of the frames are strong determinants for attitudes despite the pre-existing conditions of the participants. For instance, the threat frame was stronger than the victim frame as “it seems that negative framing does enhance attitude strength more generally.” than positive framing (Bizer, Larsen & Petty, 2011, p. 77). It could have detrimental consequences if news content provides considerations based on intolerance (Lecheler et al., 2015) and the dehumanization of refugees (Esses et al., 2013).
According to Barabas and Jerit (2010), it is difficult to ascertain if those effects are likely to happen in a real-life situation where ounces of such frames come from different sources and contexts in a longer period (de Vreese, 2004; Figenschou et al., 2015). Yet, in the Latin American scenario, where these contrasting frames tend to interact with little exposure to alternative ones (FIP, 2018), not only people can base their opinions and political
deliberation on them, but also contribute to polarization when engaging in interpersonal conversations; and from there becoming a powerful tool for elites to spread extremist views (Druckman & Nelson, 2003). Still, the lack of random chance in both versions of the sample is not a test of significance. As such, a way to reach for significance is increasing the sample size. Also, a cross-sectional study or a pretest could better simulate a real-life situation. Frames have greater effects in motivating opinions with repetition and with a believable context (Chong & Druckman, 2017).
Regarding context, it was necessary to detach the study from some ground rules in previous work. Firstly, I considered the cultural similarities among Venezuelans and host societies, which castaway most items normally used to measure symbolic threat. Secondly, the level of inequality in the region sets different expectations from research in Europe, where highly skilled people show less anti-immigration attitudes (e.g. Cavaille & Marshall, 2019). And thirdly, because Latin America has a long history of political and economic instability leading most of these countries to periods of migration (Pagnotta, 2014), I took into account a variable of identifying oneself as an immigrant or not. The scale to measure DV was valid and reliable, but the moderator variables need improvement.
One of the variables requiring to be improved is socioeconomic status. In general, it is complex to measure levels of SES with limited time and resources (Dudeja et al., 2015). I presume the reason SES is not a moderator as proposed by trade theory is that I grouped participants based on their nationality and not where they live. It might happen that the Latin Americans who live in Denmark (n = 72) or Germany (n = 23) -who represent the majority of immigrant participants- see less inequality despite the fact they are aware of the high
inequality levels back home. Those who live in welfare countries might be causing a
measurement error. Moreover, on average, high and upper middle SES weakened the strength of the victim frame. Going back to trade theory, an article about allowing Venezuelans to start a new life in their country by providing no restrictions for entry would motivate higher SES people to “favour both protection and immigration restrictions” (O’Rourke & Sinnott, 2006, p. 841), since “residents in Latin America are quite pessimistic when assessing their own mobility experiences” (Gaviria, 2007, p. 56). More research is necessary to assess if the
expectation for H3 was contradictory. I then wonder what are the implications for frame building in non-Western countries when determinants like inequality need to be considered?
As a moderator, the number of immigrants was in line with the hypotheses, but its strength on the frames was weak. An interesting observation happened with this variable: even when there was no difference in differences within the threat condition, there is more anti-immigration sentiment in countries with more refugees. Despite integration can best be pursued when immigrants and the host society share the same cultural heritage (Lundy & Darkwah, 2018), other factors are playing a role in the perception of Venezuelans as a threat. This is better explained by integrated threat theory. As an outgroup, immigrants are targets of prejudice and stereotypes; an increasing number of them lead to stricter immigration policies (Azrout et al., 2013; Croucher, 2013), as it has become the case in various Latin American countries that host most of them (Bennouna, 2019). For more trustworthy results, participants from countries with a large number of Venezuelans and those from countries with fewer of them needed to be similar, which was a sampling limitation. Still, it poses a challenge for media and governments in countries with increasing anti-Venezuelan sentiment.
Turning to the variable of identification, being an immigrant weakens the strength of the threat frame. In this case, shared experiences with Venezuelans defined their identity as a group, and collective identity means similar opinions and understandings (Meneses & Larkin, 2017). However, it might cast doubt on the participants' capability to differentiate the
determinants of being a refugee from being an immigrant (and identify themselves as one of them) since attitudes towards each group are different too (O’Rourke & Sinnott, 2006). In that case, it is argued that not all shared experiences are so much about deep emotional
bonding as much as they are about triggering reminiscences of presumed similarities for individual introspection (Meneses & Larkin, 2017). Regardless, when studying people’s attitudes towards immigration and how they influence immigration policies there is little to say about interpersonal commonalities -beyond the role of emotions- between the hosts and the newcomers that could impact such policies too. Further research about integration and assimilation should explore this indicator.
To conclude, this experiment represents a first attempt to analyse framing effects of the Venezuelan refugee crisis. Findings showed, on the one hand, that theory of frame valence extends its applicability to non-Western societies, meanwhile, it poses new implications for frame building when there is high inequality. On the other hand, cultural similarities are a determinant for negative and positive attitudes towards immigration as well as for ambivalent behaviour, or what Cea D’ancona (2015) described as tenuous tolerance. Yet, regardless of the similarities, negative frames are more powerful than positive ones, thence a tool to influence intolerance and debilitate the efforts of integration.
References
Amarante, V., Galván, M. & Mancero, X. (2016). Inequality in Latin America: a global measurement. Economic Commission for Latin America and the Caribbean (ECLAC). Retrieved from: https://bit.ly/37XIumX
Azrout, R., van Spanje, J., & de Vreese, C. (2013). Focusing on Differences? Contextual Conditions and Anti-immigrant attitudes’ Effects on Support for Turkey’s EU Membership. International Journal of Public Opinion Research, 25(4), 480–501. https://doi.org/10.1093/ijpor/edt016
Babbie, E. (2010). The basics of social research (12th ed.). Belmont: Wadsworth Thomson
Learning.
Baron, R., & Kenny, D. (1986). The Moderator-Mediator Variable Distinction in Social Psychological Research: Conceptual, Strategic, and Statistical Considerations. Journal
of Personality and Social Psychology, 51(6), 1173–1182. doi:
https://doi.org/10.1037/0022-3514.51.6.1173
Barabas, J., & Jerit, J. (2010). Are survey experiments externally valid? American Political
Science Review, 104(02), 226-242. doi: https://doi.org/10.1017/S0003055410000092
Bennouna, C. (2019, October 7). Latin America shuts out desperate Venezuelans but Colombia’s border remains open – for now. The Conversation. Retrieved from: https://bit.ly/2OhKI9K
Benson, R. (2013). Shaping Immigration News: A French-American Comparison. Cambridge: Cambridge University Press, 2-20. doi:
Bermúdez, A. (2019, July 29). Cómo impacta la crisis migratoria venezolana en la economía
de otros países de América Latina. [How the Venezuelan immigration crisis affects the
economy of other Latin American countries]. BBC. Retrieved from: https://bbc.in/3cw3699
Bizer, G., Larsen, J., & Petty, R. (2011). Exploring the Valence-Framing Effect: Negative Framing Enhances Attitude Strength. Political Psychology, 32(1), 59–80.
https://doi.org/10.1111/j.1467-9221.2010.00795.x
Bosilkov, I.& Drakaki, D. (2018). Victims or Intruders? Framing the Migrant Crisis in Greece and Macedonia. Journal of Identity and Migration Studies, 12(1), 26-45. Retrieved from: https://bit.ly/36NbXPR
Boomgaarden, H. Van Spanje, J., Vliegenthart, R., & de Vreese, C. (2011). Covering the crisis: Media coverage of the economic crisis and citizens’ economic expectations. Acta
Politica, 46(4), 353–379. doi: https://doi.org/10.1057/ap.2011.18
Boothby, E., Smith, L., Clark, M., & Bargh, J. (2016). Psychological Distance Moderates the Amplification of Shared Experience. Personality and Social Psychology Bulletin,
42(10), 1431–1444. https://doi.org/10.1177/0146167216662869
Bleich, E., Bloemraad, I. and de Graauw, E. (2015). Migrants, Minorities and the Media: Information, Representations and Participation in the Public Sphere. Journal of Ethnic
and Migration Studies, 41(6), 857–873.
https://doi.org/10.1080/1369183X.2014.1002197
Cavaille, C., & Marshall, J. (2019). Education and Anti-Immigration Attitudes: Evidence from Compulsory Schooling Reforms across Western Europe. American Political
Cea D’ancona, M. (2015). The effects of economic crisis in the framing and evolution of Spanish public opinion towards immigration. Migraciones, 37(37), 29–52. doi: https://doi.org/10.14422/mig.i37.y2015.002
Chong, D., & Druckman, J. (2007). Framing Public Opinion in Competitive Democracies.
American Political Science Review, 101(4), 637–655. doi:
https://doi.org/10.1017/S0003055407070554
Croucher, S. (2013). Integrated Threat Theory and Acceptance of Immigrant Assimilation: An Analysis of Muslim Immigration in Western Europe. Communication Monographs,
80(1), 46–62. https://doi.org/10.1080/03637751.2012.739704
De Vreese, C. (2004). The Effects of Frames in Political Television News on Issue Interpretation and Frame Salience. Journalism & Mass Communication Quarterly,
81(1), 36–52. doi: https://doi.org/10.1177/107769900408100104
De Vreese, C. (2005). News Framing: Theory and Typology. Information Design Journal
13(1), 51–62. doi: https://doi.org/10.1075/idjdd.13.1.06vre
De Vreese, C. H., & Boomgaarden, H. (2003). Valenced news frames and public support for the EU: Linking content analysis and experimental data. Communications: The
European Journal of Communication, 3, 261-281. doi:10.1515/comm.2003.024.
Druckman, J. N., & Nelson, K. R. (2003). Framing and deliberation: How citizens’
conversations limit elite influence. American Journal of Political Science, 47, 729-745. doi: https://doi.org/10.1111/1540-5907.00051.
Dudeja, P., Bahuguna, P., Singh, A., & Bhatnagar, N. (2015). Refining a socio-economic
status scale for use in community-based health research in India.(Original Article)(Report). 61(2), 77–83. https://doi.org/10.4103/0022-3859.150442
EFE. (2019, April 30). Venezolanos en El Salvador respaldan a Juan Guaidó y exigen la
salida de Maduro. [Venezuelans in El Salvador support Juan Guaidó and Ask for
Maduro to Leave]. EFE. Retrieved from https://bit.ly/2X7nBT9
Ellison, C., Shin, H., & Leal, D. (2011). The Contact Hypothesis and Attitudes Toward Latinos in the United States. Social Science Quarterly, 92(4), 938–958.
https://doi.org/10.1111/j.1540-6237.2011.00798.x
Esses, V. M., Medianu, S., & Lawson, A. S. (2013). Uncertainty, Threat, and the Role of the Media in Promoting the Dehumanization of Immigrants and Refugees. Journal of
Social Issues, 69(3), 518–536. doi: https://doi.org/10.1111/josi.12027
European Parliament (2018, December). The Venezuelan migrant crisis: A growing
emergency for the region. Retrieved from: https://bit.ly/31jUk9a
Figenschou, T., Beyer, A., & Thorbjørnsrud, K. (2015). The Moral Police. Nordicom Review,
36(1), 65–78. https://doi.org/10.1515/nor-2015-0006
Fundación Ideas Para la Paz (2018, August). Seguridad ciudadana y migración venezolana:
Análisis exploratorio [Citizen Security and Venezuelan Migration: Exploratory
Analysis]. Retrieved from: https://bit.ly/2ubF5TN
Garcia, D. (2020). Brain drain in Venezuela: the scope of the human capital crisis. Human
Resource Development International, 23(2), 188-195, doi:
10.1080/13678868.2019.1708156
Gaviria, A. (2007). Social Mobility and Preferences for Redistribution in Latin America.
Economía 8(1), 55-96. doi:10.1353/eco.2008.0003.
International Labor Office (2019). International Standard Classification of Occupations:
https://www.iomp.org/wp-content/uploads/2019/02/international_standard_classificatio n_of_occupations.pdf
Isaac, M., & Poor, M. (2016). The sleeper framing effect: The influence of frame valence on immediate and retrospective judgments. Journal of Consumer Psychology, 26(1), 53–65. https://doi.org/10.1016/j.jcps.2015.05.002
Lahav, G., & Courtemanche, M. (2012). The Ideological Effects of Framing Threat on Immigration and Civil Liberties. Political Behavior, 34(3), 477–505.
https://doi.org/10.1007/s11109-011-9171-z
Lawlor, A & Tolley, E. (2017). Deciding Who’s Legitimate: News Media Framing of Immigrants and Refugees . International Journal of Communication 11(2017), 967–991. doi: https://go-gale-com.proxy.uba.uva.nl:2443/ps/i.do
Lecheler, S., Bos, L., & Vliegenthart, R. (2015). The Mediating Role of Emotions: News Framing Effects on Opinions About Immigration. Journalism & Mass Communication
Quarterly, 92(4), 812–838. doi: https://doi.org/10.1177/1077699015596338
Lundy, B., & Darkwah, K. (2018). Measuring Community Integration of Lusophone West African Immigrant Populations Through Needs Assessment, Human Security, and Realistic Conflict Theory. Journal of International Migration and Integration, 19(2), 513–526. https://doi.org/10.1007/s12134-018-0551-7
Meneses, R., & Larkin, M. (2017). The Experience of Empathy: Intuitive, Sympathetic, and Intellectual Aspects of Social Understanding. Journal of Humanistic Psychology, 57(1), 3–32. https://doi.org/10.1177/0022167815581552
Organization of American States. (2019a). OAS Working Group to Address the Regional
Crisis Caused by Venezuela’s Migrant and Refugee Flows. Retrieved from:
https://bit.ly/2uerT0l
Organization of American States. (2019b). Informe de la situación de los refugiados y
migrantes venezolanos en Bolivia. [Report of the situation for Venezuelan refugees and
migrants in Bolivia] Retrieved from: https://bit.ly/2X5UiQU
Organization of American States. (2019c). Informe de la situación de los refugiados y
migrantes venezolanos en Brasil. [Report of the situation for Venezuelan refugees and
migrants in Brazil] Retrieved from: https://bit.ly/2X9g9Xt
Organization of American States. (2019d). Informe de la situación de los refugiados y
migrantes venezolanos en Costa Rica. [Report of the situation for Venezuelan refugees
and migrants in Costa Rica] Retrieved from: https://bit.ly/2zTDeFx
O’Rourke, K. H. and Sinnott, R. (2006). The Determinants of Individual Attitudes Towards Immigration. European Journal of Political Economy, 22(4):838–861. doi:
https://doi.org/10.1016/j.ejpoleco.2005.10.005
Osorto, M. (2019, April 22). Masivo ingreso a Honduras de nicaragüenses y venezolanos. [Massive entry to Honduras of Nicaraguans and Venezuelans]. El Heraldo. Retrieved from https://bit.ly/2WKFqIJ.
Oxfam (2019, October). Yes, But Not Here: Perceptions of xenophobia and discrimination
towards Venezuelan migrants in Colombia, Ecuador and Peru. Retrieved from:
Pagnotta, C. (2014). Latin American migration within the continent between 1980 and 2000: a historical perspective. Nuevo Mundo Mundos Nuevos. doi:
https://doi.org/10.4000/nuevomundo.67174
Pettigrew, T. (1998). Intergroup contact theory. Annual Review of Psychology, 49, 65–85. Ramasubramanian, S. & Miles, C. (2018). Framing the Syrian Refugee Crisis: A
Comparative Analysis of Arabic and English News Sources. International Journal of
Communication 12(2018), 4488–4506. https://ijoc.org/index.php/ijoc/article/view/8174
Response For Venezuelans. (2019). 2019 Regional Refugee And Migrant Response Plan For
Refugees And Migrants From Venezuela. Retrieved from
http://reporting.unhcr.org/node/21600
Scheufele, D. A. (2000). Agenda-setting, priming, and framing revisited. Another look at cognitive effects of political communication. Mass Communication & Society, 3, 297–316. https://doi.org/10.1207/S15327825MCS0323_07
Soy502. (2018, March 22). Comunidad venezolana en Guatemala: dudas y preocupación por
las visas. [Venezuelan Community in Guatemala: Doubts and Worries regarding
Visas]. Soy502. Retrieved from https://bit.ly/3e22ubH
Toro, M. (2019, June 7). Ya son más de 4 millones de migrantes y refugiados venezolanos en
el mundo, según ACNUR. [There are already 4 million Venezuelan refugees and
migrants in the world, UNHCR reports]. CNN en Español. Retrieved from https://cnn.it/35UXJOa
Van Gorp, B. (2005). Where is the Frame?: Victims and Intruders in the Belgian Press Coverage of the Asylum Issue. European Journal of Communication, 20(4), 484–507. https://doi.org/10.1177/0267323105058253
Velasco González, K., Verkuyten, M., Weesie, J., & Poppe, E. (2008). Prejudice towards Muslims in The Netherlands: Testing integrated threat theory. British Journal of Social
Psychology, 47(4), 667–685. https://doi.org/10.1348/014466608X284443
Waters, M., & Jimenez, T. (2005). Assessing Immigrant Assimilation: New Empirical and Theoretical Challenges. https://doi.org/10.1146/annurev.soc.29.010202.100026 Wojcieszak, M., & Azrout, R. (2016). I Saw You in the News: Mediated and Direct
Intergroup Contact Improve Outgroup Attitudes. Journal of Communication, 66(6), 1032–1060. https://doi.org/10.1111/jcom.12266.
Appendix A
English version of the survey experiment questionnaire
Start of block: Introduction Welcome!
I am researching my master’s thesis at the Graduate School of Communication, University of Amsterdam. I am interested in exploring Latin Americans' opinions towards the Venezuelan immigration. Please be assured that your responses will be kept completely confidential.
The study should take you around 10 minutes to complete. Your participation in this research is voluntary. You have the right to withdraw at any point while you are taking it without any prejudice.
If you would like to contact us to discuss this research, please email me to kimberly.yanezolivero@student.uva.nl.
By clicking the button below, you acknowledge that your participation in the study is voluntary. You may choose to terminate your participation in the study at any time and for any reason. Please note that this survey will be best displayed on a laptop or desktop computer. Some features may be less compatible for use on a mobile device.
Kimberly Yánez Consent:
○ I consent, begin the survey
○ I do not consent, I do not wish to participate End of block: Introduction
Start of block: I do not consent, I do not wish to participate We are sad to see you leave. Thank you for your time.
End of block: I do not consent, I do not wish to participate
Start of block: I consent, begin the survey
End of block: I consent, begin the survey
Start of block: Randomization check/demographics Age - How old are you?
Gender - Which gender do you identify with?
○ Female ○ Male ○ Other
○ Prefer not to say
End of block: Randomization check/demographics
Start of block: Randomization check/moderators
Education - What is the highest degree or level of school you have completed? If currently enrolled, tell me your highest degree received.
○ No schooling completed ○ Some high school, no diploma
○ High school graduate, diploma or the equivalent ○ Some college credit, no degree
○ Trade/technical/vocational training ○ Associate degree ○ Bachelor’s degree ○ Master’s degree ○ Professional degree ○ Doctorate degree ○ Other:
Income: Compared to other people in your country, are you relatively poor or relatively wealthy?
1 Very poor
2 3 4 5 6 7
Very wealthy Compared to other people
Occupation: Tell me about your current employment status.
I am unemployed because:
○ Manager ○ Professional
○ Technician or associate professionals ○ Administrative support worker ○ Clerical support worker
○ Service or sales worker
○ Skilled agricultural, forestry and fishery worker ○ Craft and related trades worker
○ Plant and machine operator, or assembler
○ Elementary occupation (cleaner, helper, vendor, transport, construction) ○ Commissioned Armed Forces Officer
○ Non-commissioned Armed Forces Officer ○ Armed Forces Occupation, Other Rank
Other occupation:
Number of immigrants: What is your nationality? If many, with which do you most identify?
Identification: Do you live in your country of your nationality? ○ Yes (1)
○ No (2)
End of block: Randomization check/moderators
Start of block: Country where they live that is not their country of nationality. Identification 2: In which country do you live?
End of block: Country where they live that is not their country of nationality.
Start of block: Experimental group 1 - Treat frame Thank you for getting this far!
Now I will present you a news article. Please read it carefully. (See appendix B for stimulus material)
End of block: Experimental group 1 - Treat frame
Start of block: Experimental group 2 - Victim frame Thank you for getting this far!
Now I will present you a news article. Please read it carefully. (See appendix B for stimulus material)
End of block: Experimental group 2 - Victim frame
Start of block: Control group Thank you for getting this far!
Now I want to know your opinion about the Venezuelan immigration in your country. End of block: Control group
Start of block: Dependent variable/Attitudes towards Venezuelan immigration
On a 7-point scale where 1 is “strongly disagree” and 7 is “strongly agree”, to what extent do you agree with the following statements?:
1 2 3 4 5 6 7
Venezuelan immigrants
public services because they take more out than they put in
Venezuelan immigration is good for the
(nationality/denomyn) labour market
○ ○ ○ ○ ○ ○ ○
Venezuelan immigrants are a threat to the security of
(nationality/denomyn) people
○ ○ ○ ○ ○ ○ ○
Venezuelan immigrants are an important cause of crime in my country
○ ○ ○ ○ ○ ○ ○
My country’s identity is being threatened because there are too many Venezuelans
○ ○ ○ ○ ○ ○ ○
End of block: Dependent variable/Attitudes towards Venezuelan immigration
Start of block: Manipulation check
Thank you for getting this far! Now I am interesting in knowing the following:
1 Not at all 2 Very little 3 Little 4 Neith little not much 5 Much 6 Very much 7 Completely To what extent do you perceived the Venezuelan refugees as victims? ○ ○ ○ ○ ○ ○ ○ To what extent do you perceived the Venezuelan refugees as intruders? ○ ○ ○ ○ ○ ○ ○
