The Effects of Othering on Perceptions of TerrorismDownload

Avery Roe

Abstract
The United States has been subject to many acts of mass violence, yet not all of these attacks have been perceived as terrorism. In this paper, I argue that the media’s use of othering or changing the discourse between an in group and out group leads specific groups of people to be more likely to define an attack as terrorism when the out group is Muslim people. Based on an original survey experiment, I find that targeted demographics, political characteristics, and media habits respond differently to othering and the likelihood of defining an attack as terrorism is altered when the attack is tied to someone of the Islamic faith.
Keywords: Othering, Terrorism.
Different Perceptions of Terrorism and Othering
Shootings, bombings and other violent attacks happen frequently in the United States. The broadcast news media, later referred to as the media for simplicity, has developed a pattern of reporting through the high rate of these horrific events. As an example, the word terrorism is used frequently, but only in certain cases. In 2009, Major Nidal Malik Hasan killed 13 people at the Fort Hood military base. In 2013, Aaron Alexis entered the Navy Yard in Washington D.C. and killed 12 people. These attacks were almost identical as they both took place at military bases and targeted American soldiers, both were committed by a single gunman and both resulted in 13 deaths, as Alexis was killed on scene. The main difference is that Major Hasan was a Middle Eastern Muslim while Alexis was African-American.
The coverage of the two instances was radically different. Hasan was portrayed as an Islamic extremist and Alexis was portrayed as a mentally ill lone wolf (Morin, 2016). This difference in reporting is referred to as othering. Othering is defined as changing the discourse between an ingroup and an outgroup (Dervin, 2015). Because Major Hasan was Muslim and Middle Eastern, he was considered to be in the outgroup for most people.

Avery Roe, Peace and Conflict Studies, Butler University, Indianapolis, United States of America.  
E-mail Id.: aroe@butler.edu

The media chose to highlight this aspect of his identity which led to him being seen more as a terrorist in the eyes of the viewers. Alexis was considered in the ingroup because people perceived him as an American, thus he was perceived as a lone wolf.
The PATRIOT Act of 2001 defines international terrorism as “activities that appear to be intended to affect the conduct of government by mass destruction”, and domestic terrorism as “activities that occur primarily within U.S. jurisdiction, that involve criminal acts dangerous to human life, and that appear to be intended to intimidate or coerce a civilian population, to influence government policy…or to affect government conduct” (“H.R. 3162”, 2001). While incredibly important in a court of law, the government’s definition of terrorism is not necessarily how most citizens conceptualize the topic. In fact, it is likely not considered at all when the public or the media label an act of violence as terrorism or not. Deciding what should and should not be considered terrorism is not the goal of this study. The question this study seeks to answer is how the media’s othering influences the public in defining terrorism. Understanding public opinion of terrorism is necessary because it is often the driving factor behind any responses to an attack.
This is an important question because it will help speak to the media’s role in intergroup relations and responses to attacks around the country. It has the ability to address intergroup relations because othering highlights the differences between people which has the ability to further the tension between groups. If the media is successful in portraying Muslims as terrorists, it is reasonable to believe that it could be a contributing factor to tense relations between non-Muslim Americans and the Muslim community. These tense relations have significant ramifications including impacting public policy. Public views towards policy relating to Muslims are indicative of less trust of the Muslim community (Jamal, 2008; Ogan et al., 2014). This study has the ability to show that the media has an impact on the policy views by shaping the public’s view of Muslims.
Previous literature suggests that, in general, people would be more likely to classify an attack as terrorism if the perpetrator is presented as Muslim. By relying on work more focused on intergroup relations and othering, I argue that the effect of othering will be more pronounced among certain demographic characteristics where Muslims are more likely to be viewed as members of an outgroup. Likewise, the effect of othering should be more noticeable among certain media habits since the media is the vehicle of othering.  To test my hypotheses, I conducted an original survey experiment through Amazon Mechanical Turk. Respondents in the treatment group were exposed to a hypothetical attack where the perpetrator was identified as Muslim while respondents in the control group received the same vignette without any demographic information. All respondents were then asked how they defined the attack, with one of the multiple-choice options being terrorism.
In evaluating the results, I found no overall relationship between othering and how respondents defined terrorism. However, I did find support for my hypotheses that both demographic characteristics and media influence people’s views on whether or not an attack is terrorism. Overall, these results suggest that othering can be very effective for targeted populations, particularly when it has the ability to highlight whether a perpetrator is a member of an ingroup or outgroup. This signifies that the media plays a large role in shaping public opinion; including influencing how people define terrorism.
Othering and the Perceptions of Muslims and Terrorism
In the plethora of attacks that have occurred in the United States in recent years, a difference in the media responses has become clear. If a perpetrator is viewed as American, they are typically said to be a lone wolf with a mental illness; if the perpetrator is Muslim and/or Middle Eastern, they are quickly labeled as terrorists. This is known as othering and it is presenting people in the outgroup differently from the ingroup highlighting the differences between the groups (Dervin, 2015; Johnston, 2001). Othering works through the media’s negative treatment of Muslims, which is important due to the influence that the media has been proven to have on public opinion (Barnett, 2003; Gallup, 2013; Morin, 2016; Ogan et al., 2014). In addition to othering, the American public already has uniquely negative views of Muslims that are often associated with violence (Huddy & Feldman, 2011; Jamal, 2008; Kam & Kinder, 2007; Sides & Gross, 2013). The dynamics of othering suggest that certain demographic groups, that see Muslims as an outgroup, will be more affected by the use of othering (Kalkan et al., 2009; Tajfel & Turner, 1979; Wike and Grim, 2010). Throughout all of the prior literature, it has not been seen how othering influences public perceptions on the definition of terrorism and how an individual’s demographic characteristics influence that outcome.
The way in which the media chooses to present a violent incident is important because it is the primary means through which most Americans receive information about the news. In 2013, Gallup News did a poll asking Americans where they got their news. Only 6% of respondents reported their primary news source as not being some form of media (Gallup, 2013). Given the technological advances that have occurred since 2013, there is reason to believe that if this statistic has seen any significant changes that more people are relying on the media as their primary news source. This signifies that the media has a wide reach, and the effects of media consumption affect society as a whole. The media has been shown to increase people’s political knowledge. Research on the 2006 midterm election showed more exposure to the media led people to acquire greater knowledge about the campaigns (Wei & Lo, 2008). Similar results were found when looking at current events in the European Union; further media exposure increased political knowledge (de Vreese&Boomgaarden, 2006). People gain their political knowledge primarily through news media. This is important when discussing current events regarding religion and terrorism because it shows that the media has significant potential to have the influence on people’s perception as to what occurs because their factual knowledge primarily comes from the news.
The way the media presents a story has been proven to impact the audience’s feelings about the subjects of the story. For example, in an experiment, Brooke Barnett evaluated people’s perception of alleged criminals. She found that if manufactured newscasts had either visual or auditory biases, including use of restraints and different clothing types, the audience viewed the person accused as guiltier and more threatening (Barnett, 2003). The choices made in the presentation of the story impacted how people viewed the subjects of the story, giving the role of the media incredible importance. Another study compared the frames the media used about gun control before and after 9/11 and their impact on public opinion in an experimental setting. The frames before 9/11 were not focused on terrorism while those after were focused on terrorism. This study found that the frames used before 9/11 had almost no impact on policy views. However, the frames used after 9/11 proved to be very effective in shaping views around gun control and generating a negative reaction to Muslim people (Schnell & Callaghan, 2005). Together, these studies suggest that how the media reports a story has the ability to impact people’s policy views.
Non-Muslim Americans have been known to perceive Muslims in a different, more antagonistic manner than they do other subgroups in the American population. In 2013, one survey of white, non-Muslim, Americans found that Muslims ranked third from lowest on a feeling thermometer, only above the LGBT community and illegal immigrants. It also found that Muslims were ranked both the most violent and untrustworthy group compared to groups such as Hispanic and Asian Americans, as well as black and white people (Sides & Gross, 2013). This study illustrates that there is a difference between how Muslims and members of other outgroups are thought of in the United States. Muslims are viewed in a uniquely antagonistic light throughout America. They are also most commonly associated with violence when compared to any other outgroup.
In addition to more general negative feelings towards Muslims, the American public tends to have strong policy views against them;specifically, there is general support for taking away their human rights. One study found that people who were exposed to issues relating to Muslims in the media were less likely to believe Muslims should have the same rights as people of other religions (Ogan et al., 2014). This study additionally proved the impacts of media othering on Muslims. In 2003, 49% of the general population would have supported increased surveillance of Muslim and Arab Americans and 41% would uphold detention of suspicious Arabs or Muslims without sufficient evidence (Jamal, 2008). Both of these studies show a unique inclination towards taking away the rights of Arabs and Muslims. The results suggest a lack of trust towards Muslims by the general population which impacts people’s policy views.
Harsh policy views are particularly noteworthy in the aftermath of an attack committed by a Middle Eastern person. For example, the main predictor of support for the American War on Terror was ethnocentrism (Kam & Kinder, 2007). Another psychological study found that those who felt more threatened and insecure after the 9/11 attacks were more likely to support strong foreign and domestic policies (Huddy & Feldman, 2011). Americans had strong policy views relative to Middle Eastern people after the attacks on September 11th. These policy views associate Muslims with violence and bring feelings of threat and insecurity.Public response among Americans is notably less intense after an attack committed by a white American in comparison to a Middle Eastern person. Since 1990, Gallup has conducted a regular poll examining public opinion toward gun control, noting mass shootings occurring during that time frame. The study features attitudes after seven major shootings committed by domestic actors.None of them created a significant, long lasting change in public opinion (Gallup, 2017). This is a stark contrast to American public opinion after attacks such as 9/11. As shown previously, Americans had a strong policy response after the attack, but after shootings committed by domestic actors; the response in policy views was not as strong. The school shooting in Parkland, Florida is a clear exception to the general trend. In recent polls support for gun control has risen significantly and is at the highest it has been since the early 1990s when crime rates were at an all-time high (Gallup, 2018). A potential explanation for this is the survivors of the shooting keeping the media focus on gun control longer than groups have in the past. The overall public views on gun control suggest that the majorityof people trust white Americans more than Muslims because they do not see as much of a need to respond to shootings committed by actors perceived as domestic. If an attack is committed by a Muslim a policy response is perceived as more necessary.
A range of prior literature has found that Americans view Muslims negatively. This often comes with mistrust and an association with violence. Thus, my first hypothesis is:
H1: Americans will be more likely to identify acts of violence committed by Muslims as terrorism compared to acts of violence committed by other groups.
The theory of othering relies on the pervasive ability of the media to present stories differently in a systematic manner. This includes the significant differences in the media’s treatment of Muslims. One study looking for potential causes for the common anti-Muslim feeling found that those who paid more attention to the media were more likely to think of Islam as a violent religion (Ogan et al., 2014). More specifically, in the context of reports on their conflict, both the Sunni and Shia sects of Islam have been consistently portrayed in a negative manner and associated with terrorism in the western media (Douai &Lauricella, 2014). The authors theorized that this portrayal was another arm of the war on terror as it had begun to influence journalists. The western media representation of Sharia Law, or Islamic religious law, has also been studied. Stories about Sharia Law most often feature topics with negative connotations in Western Culture (Hoewe et al., 2014). It is apparent that western media presents Muslims differently than other groups. Muslims are portrayed negatively and are often associated with violence. In doing this, the media separates the Muslim population from the rest of the American public.
This uniqueness in coverage is amplified in the context of an act of mass violence. The media coverage after the 2009 Fort Hood Shooting and the 2013 Navy Yard Shooting has been compared due to the similarities between the shootings. The main difference came in the identities of the shooters. The Fort Hood shooter was both Muslim and Middle Eastern, whereas the Navy Yard shooter was African-American. Within the media, The Fort Hood shooter was mostly associated with terrorism and his religion. The media’s portrayal of The Navy Yard shooter, however, was mostly related to his mental health issues (Morin, 2016). This analysis shows that the common media association between Islam and violence is unique to that religion. The Navy Yard shooter being African-American shows that the othering is not only because Muslims are commonly viewed as not white. In another example, Kimberly Powell looked at media coverage of all the terrorist attacks after 9/11 that were either carried out or stopped on the day they were supposed to be carried out. She found a variety of patterns including; the quick identification of a perpetrator as Muslim or questioning if they were, the portrayals of domestic agents as mentally unstable, and the allusion to a future threat (Powell, 2011). People who are Middle Eastern but not Muslim are still othered in the media as Muslim. This suggests that the othering of religion is both the intention of the media and it is what gives the othering its effect.
The association between all Arab people who commit acts of terrorism with the Muslim faith without the knowledge of the connection is also very harmful to perceptions of the Islamic faith. In the context of an act of mass violence, it is clear that the media treats Muslims differently from those who are not Muslim. Thus, I hypothesize that people’s media habits will impact whether they identify acts of violence committed by Muslims as terrorism. For example, those with positive views towards the media might be more affected by othering because they are more inclined to believe the media; and those who watch more international news might be less affected because of the higher international prevalence of Muslims (Pew Research Center, 2018b; Pew-Templeton Global Religious Futures Project, 2010). More specifically I hypothesize:
H2a: Americans that have warm feelings toward the media will be more likely to identify acts of violence committed by Muslims as terrorism compared to acts of violence committed by other groups.
H2b: Americans that watch international news will be less likely to identify acts of violence committed by Muslims as terrorism compared to acts of violence committed by other groups.
Othering, however, should not have a similar impact across all demographic groups. Due to the fact that Muslims are a minority in America, (Pew Research Center, 2018b) they are an outgroup for many non-Muslim Americans. However, for Muslims and those who are part of larger groups that often include Muslims they are perceived as an ingroup. This should create a difference in reaction to othering because the society’s outgroups are not necessarily the outgroups for all individuals (Johnston, 2001). Often these groups are based on demographic characteristics. For example, a racial minority might be a societal outgroup; but they may not be considered an outgroup among other racial minorities. 

The effect of being an outgroup had the strongest effect on negative views of Muslims because they were seen as a big group of others due to them being considered both religious and racial minorities (Kalkan et al., 2009). In reference to Muslims, the ingroup’s perception of the outgroup is primarily determined by threat perception (Wike and Grim, 2010). The general American population views Muslims as less favorable than they view other groups, primarily due to a perception of threat or them being part of an outgroup. As previously shown, the media others Muslims and is effective in creating feelings of threat and influencing public opinion. This is evidence that the media’s impact does hold in influencing public opinion of Muslims.
Intergroup dynamics suggest that different groups will have different responses to othering in the media even with everyone being exposed to the same negativity. The importance of othering in this instance is that in differing the discourse on something such as an attack in a way that can lead to the ingroup judging themselves to be morally superior the media sets up a major schism in public perception. In the example of the othering of Muslims those who are in that outgroup see themselves as better because it is not people like them that they see associated with terrorism.
This effect is magnified when Social Identity Theory is considered. Social Identity Theory states that people derive a sense of their identity and self-esteem from the groups that they belong to. It explains discrimination against outgroups as the consequence of social competition for status (Tajfel & Turner, 1979). Social Identity Theory and othering interact because othering further disadvantages outgroups by further differentiating between the in- and outgroup. Therefore, my third hypothesis is:
H3: Americans that belong to groups that are more likely to view Muslims as outgroups will be more likely to identify acts of violence committed by Muslims as terrorism compared to acts of violence committed by other groups.
In general, this ingroup/outgroup dynamic would likely hold true for any group but here my application is specific to those groups that are likely to perceive Muslims as an outgroup. These groups will likely include religious and political groups; as well as those with negative views about racial minorities.
First of all, Christians clearly view Muslims as the outgroup. Because both groups are defined by their religion, being in one group means that someone is not in the other group. This effect is magnified for those who take the Bible as the literal word of God. Due to the ideological differences between Christianity and Islam, those who take the Bible as the literal word of God are more likely to see Muslims as an outgroup to themselves.
Conservatives and Republicans are also expected to see Muslims as outgroups. In 2017 only 13% of Muslims said they were either Republican or lean Republican. This is compared to the 66% who said they were either Democrats or leaned Democrat (Pew Research Center, 2018a). Because the meaning of Conservative and Republican are often mixed in the public discourse the same effect likely holds for both groups. This shows that those who are Republican or Conservative do not share political affiliation with many Muslims, so they could be seen as an outgroup. Because a significantly higher percentage of Muslims are Democrats, they are in all likelihood seen as an ingroup in that political context.
For those who have cold feelings towards racial minorities, Muslims are an outgroup. Because the media and public discourse has often fused Arab people and Muslims into one category (Cainkar, 2002), those with cold feelings towards racial minorities are not likely to view Muslims as an ingroup. Because they equate Arabs, whom they do not like, with Muslims, they likely view Muslims vastly different from them, thus defining them as an outgroup.
In summary, I hypothesize that Americans will be more likely to define acts of violence committed by Muslims as terrorism. Due to the effects of othering I also hypothesize that people’s opinions of the news and consumption habits will have an impact as well as their individual in and outgroups. Due to the wide variety of responses to acts of mass violence in the United States, it is clear that some demographic groups are differently affected by the factor that determines response, which I suggest is othering. Because defining an attack as terrorism plays a large role in the responses to an attack, it is important to determine how society comes to see attacks as terrorism.
Methodology
The question this study seeks to answer is whether or not the media’s othering of a perpetrator influences people’s perception of attacks as terrorism. It has been proven that the media is very effective in influencing the public and others people based on religion, including Muslims. Opinion polls have shown that the general American public has negative views towards Muslims, suggesting that the media is affecting public opinion. However, it is still unknown if othering- shaping the discourse by highlighting the religion of the perpetrator- is what is moving public opinion. Based on prior literature, I developed three hypotheses; first, that people will be more likely to view an attack as terrorism if the assailant is associated with Islam, that people’s media habits will have an impact on their views of a hypothetical attack, and that those viewing Muslims as an outgroup will be more effected by the othering. I will be using an original survey experiment to test this because a survey is the best way to holistically understand public opinion. A survey allows me to best assess whether people classify an act of violence as terrorism, not relying on any legal or academic definition which is less relevant for my hypotheses. The experiment will also allow a direct comparison of the results of the twoscenarios which will provide leverage on causation that othering is what leads individuals to identify an act of violence as terrorism or not. 
To test my hypotheses, I utilized an online survey administered through the online survey platform Qualtrics. The survey included questions designed to tap into political attitudes and behaviors. The sample was recruited from Amazon’s Mechanical Turk crowdsourcing program. Mechanical Turk, in spite of being a more recent tool for recruiting survey respondents, is inexpensive and documented to produce reliable data (Buhrmester, Kwang, and Gosling 2011; Mason and Suri 2012; Berinsky, Huber, and Lenz 2010; Levay, Freese, and Druckman 2016). The survey, available for one week during the spring of 2018, recruited respondents by paying them $0.50 upon completion of the survey. The survey had a total population of 751 respondents.
The population studied in this survey consisted of American Citizens over the age of 18 with internet access. This is because the survey instrument was online and open to citizens over the age of 18. The survey respondents being American is relevant because I am looking at othering that occurs in American media. While the survey relied on a convenience sample, the experimental design gave me the required leverage to test my hypothesis comparing differences between the treatment and control groups.
To test my hypotheses about the effects of othering, I designed an original survey experiment(see Figure 1). The split sample questions described a hypothetical bombing. Half of the respondents were randomly assigned to the control group and given the following question: “Please consider this hypothetical scenario: At least 15 people have died and many more are injured after a bomb was detonated at a mall. Authorities say the suspect is in custody. How do you define this attack?”. The other half of the respondents were randomly assigned to the treatment group. They received the same hypothetical scenario but were given additional information regarding the possible religious affiliation of the suspect. The question included the phrase “and it has been suggested that he has ties to local Mosques” immediately after stating that the suspect is in custody.

Figure – 1
Survey Experiment


Treatment Group
(Split Sample) Please consider this hypothetical scenario: At least 15 people have died and many more are injured after a bomb was detonated at a mall. Authorities say the suspect is in custody, and it has been suggested that he has ties to local Mosques. How do you define this attack?
  • Use of a Weapon of Mass Destruction
  • Mass Murder
  • Terrorism
  • Other (Please Specify)

Control Group
(Split Sample) Please consider this hypothetical scenario: At least 15 people have died and many more are injured after a bomb was detonated at a mall. Authorities say the suspect is in custody. How do you define this attack?

  • Use of a Weapon of Mass Destruction
  • Mass Murder
  • Terrorism
  • Other (Please Specify)

By randomly assigning respondents to the treatment or control group and only varying the religious affiliation of the suspect, I am able to attribute any statistically significant differences between groups to the effects of othering. I was careful not to make the hypothetical scenario too similar to a real, high-profile attack. This decision ensured that respondents were reacting to my hypothetical scenario and not drawing upon their recollections of their attitudes and responses to media coverage after specific attacks. I also did not add a location or name to prevent the attack from being too close to any of the respondents if I chose a significant name or city. In addition to all of these factors, keeping the wording of the question the same other than specifying the religion of the attacker allows these questions to measure the effects of othering.
My independent variable is the othering of the perpetrator. The treatment condition in this instance is the mention of the mosque while the control is no mention of the mosque. Therefore, if the treatment and control groups had a statistically significant difference in how they respond to the hypothetical scenario, I can confidently attribute this difference to the mention of the mosque in the hypothetical senario.
The dependent variable is how people define the hypothetical scenario. I will measure it in the responses to both questions. Both the treatment and control groups received the same response options which included; use of a weapon of mass destruction, mass murder, terrorism, and the option to write in a different response. My hypothesis is concerned with whether othering influences perceptions of terrorism so I dichotomized the variable into res ponses of terrorism and collapsed all other responses into a not terrorism category. The responses for both parts of the experiment are the same and in the same order. This will allow me to compare what people believe depending on their knowledge of the perpetrator. 
To statistically evaluate my results, I will use a series of t-tests. The t-tests will take the difference in means of the treatment and control group then determine both the direction and any statistical significance of the difference. I expect that I will find the people who receive the treatment will be more likely to identify the scenario as terrorism compared to the control group. To test the second and third hypotheses, that some demographic groups and those with specific media habits will be more affected than others, I will perform further t-tests comparing the average treatment effect among key religious, political, racial characteristics and media habits. If I find statistically significant differences in the average responses between those in the treatment and control group, I can confidently attribute the impact to the othering of the perpetrator as Muslim.

Survey Results
Prior to conducting the survey, I first hypothesized that people would be more likely to identify an attack as terrorism if the assailant is presented as having connections to the Muslim faith. In addition, I hypothesized that those with certain media habits and certain demographic groups would be more effected by the othering. The survey had 751 total respondents. A total of 49.8% received the control question and 50.2% received the treatment question mentioning the connection to Mosques. In general, there was a balance between the treatment and control groups, as seen in Table 1. This balance extends across the theoretically relevant variables to test my hypotheses. Moreover, there is a large enough sample size in each group. Taken together, this gives me confidence in testing my hypotheses because the balance allows valid conclusions in the comparisons.
Table 1


Participant Demographics in Treatment and Control Groups

Demographic Groups

Control Group
(Total People)

Treatment Group (Total People)

Total People

Overall

49.8% (374)

50.2% (377)

751

Television Main Media Consumption

51.39% (74)

48.61% (70)

144

Television not Main Media Consumption

49.42% (300)

50.58% (307)

607

Positive Feelings Towards Media

49.44% (132)

50.61% (135)

267

Negative Feelings Towards Media

49.9% (241)

50.1% (242)

483

Primarily Consume International News

61.36% (54)

38.64% (34)

88

Primarily Consume Local and
National News

48.27% (320)

51.73% (343)

663

Conservatives

48.22% (122)

51.78% (131)

253

Non-Conservatives

50.71% (250)

49.29% (243)

493

Republicans

45.45% (90)

54.55% (108)

198

Non-Republicans

51.36% (284)

48.64% (269)

553

Christians

51.15% (200)

48.85% (191)

391

Non-Christians

48.33% (174)

51.67% (186)

360

Take the Bible Literally

53.21% (58)

46.79% (51)

109

Do Not Take the Bible Literally

49.14% (315)

50.86% (326)

641

Negative Feelings Towards Racial Minorities

48% (96)

52% (104)

200

Positive Feelings Towards Racial Minorities

50.45% (278)

49.55% (273)

551

 To test my hypotheses, I performed a series of t-tests. To test the first hypothesis, I performed a t-test on the overall population of respondents which is shown in Figure 2 and Table 2. In both the treatment and control group, a majority of respondents defined the attack as terrorism; 61.5% in the treatment group identified the attack as terrorism while 58.3% of the control group did. While this 3.2 % difference is in the direction of my hypothesis, the results are not statistically significant. Thus, I am unable to confirm my first hypothesis that othering leads more Americans to define an attack as terrorism.
Theoretically, there are at least two plausible explanations for these null results. The first is to take the results at face-value. Most respondents, regardless if they were randomly assigned to the treatment or control group identify the attack as terrorism. Perhaps othering has no impact because most Americans would identify the bombing of a public space an act of terrorism. This would be unexpected in light of the variety of literature that states that the American public has negative views towards Muslims and perceptions that they are violent. Yet, there are reasons to suspect that even if most people defined the attack as terrorism, othering should still have an independent effect. In light of the demographic variance in the sample this strikes me as a less plausible explanation.
A more plausible explanation, given my second and third hypotheses, is that the null results mask the asymmetric nature of othering; thanks to the outgroup dynamics of the question an offsetting effect occurred. It is likely that while the treatment condition made some groups more likely to call the attack terrorism it made other groups less likely due to their awareness of the common narrative about Muslims. For example, different media habits could lead to some being more likely and others being less likely to call the attack terrorism. This would lead to an offsetting effect which would create null results when evaluating the entire population of respondents.
Figure – 2

 

Table 2


Likelihood of Calling the Attack Terrorism

Demographic Groups

Control Group

Treatment Group

Difference

Overall

0.583

0.615

0.032

Television Main Media Consumption

0.595

0.743

0.148**

Television not Main Media Consumption

0.58

0.586

0.006

Positive Feelings Towards Media

0.504

0.593

0.089*

Negative Feelings Towards Media

0.63

0.647

0.017

Primarily Consume International News

0.629

0.441

-0.188**

Primarily Consume Local, National News

0.575

0.633

0.058*

Conservatives

0.59

0.702

0.112**

Non-Conservatives

0.579

0.569

-0.01

Republicans

0.578

0.76

0.182***

Non-Republicans

0.585

0.558

-0.027

Christians

0.565

0.675

0.11**

Non-Christians

0.603

0.554

-0.049

Take the Bible Literally

0.551

0.655

0.104**

Do Not Take the Bible Literally

0.601

0.592

-0.009

Negative Feelings Towards Racial Minorities

0.481

0.655

0.174**

Positive Feelings Towards Racial Minorities

0.607

0.635

0.028

*** =p≤0.01   ** = p ≤ 0.05   *=p≤0.10

 

In the following analysis I test my second two hypotheses by examining the effect of othering by 1) media habits, 2) political characteristics, 3) religious characteristics, and 4) racial attitudes. I will test based on media habits because they evaluate how much a person is exposed to othering and how much they trust the news. Also, certain groups within these characteristics are more likely to identify Muslims as the outgroup which would lead to othering having an impact on their views. The results from all of the tests are shown in Table – 2.
Figure – 3


In Figure 3, I present confirmation of my second hypothesis. Overall, about 55% of Americans rely on television for their news (Gallup, 2013). Within my survey, people who self-reported their primary media consumption as television were 14.8% more likely to call the incident terrorism in the treatment group than in the control group. Approximately 75% of the treatment group called the attack terrorism compared to 60% in the control group. The statistical significance of these results among those who watch television suggests that those who primarily consume television are more susceptible to othering. Yet those who did not report television as their primary media consumption were almost equally as likely to call the attack terrorism regardless of which question they received. The result from this survey shows that those who put more trust in television as their primary form of media consumption are more impacted by othering than those who rely on other forms of media.
Those who felt more warmly towards the media are more likely to trust it, giving the media, and any othering that occurs, more credibility in their perspective. Another way to measure trust is evaluating attitudes towards the media. Using a feeling thermometer to measure attitudes toward the media, respondents were asked to rate their feelings about a variety of groups and entities, including the media, on a scale from 0 to 100. These responses were coded into those who felt cold towards the media (0-50) and those who felt warmly (51-100). Those who reported feeling more warmly towards the media on a feeling thermometer were 9% more likely to call the attack terrorism in the treatment group than those in the control group. In contrast, those who felt cold toward the media were only approximately 1.7% more likely to call the attack terrorism in the treatment group when compared to the control group which was statistically insignificant. In summary, for those who do not watch TV or have negative attitudes towards the media, what the television communicates is less important.
Media consumption also matters when looking at the scope of news because the likelihood of a source othering Muslims varies. Respondents were asked how much of their information intake is from either local, national, or international news. The response options were either more or much more with each of the three scopes. These respondents were dichotomized into those who consumed either more or much more international news compared to all of the other response options. I hypothesized that Americans who watch international news will be less likely to identify the attack as terrorism. Those who reported watching either more local, or national news were 6% more likely than those who watch international news to call the attack terrorism in the treatment group, this was not statistically significant. On the other hand, those who reported watching primarily international news were 19% less likely to call the described attack terrorism in the treatment group. This suggests that people who watch mostly international news receive different narratives in the media that do not associate Muslims with terrorism as often. Because Muslims are a minority in the United States, they are more likely to be perceived as an outgroup, but internationally their outgroup status is less pronounced, (Pew Research Center, 2018b; Pew-Templeton Global Religious Futures Project, 2010) signifying that the othering of Muslims is not as likely in the international media. Thus, when the media mentions that the attacker is Muslim, labeling the attack as terrorism is not the public’s first reaction.
These results support my second hypothesis regarding the influence of media habits on the efficacy of othering. This shows that for othering to have an impact, people need to be frequently exposed to the media and inclined to trust the othering that occurs. The results also add to the effects that I hypothesized by showing that othering is more effective in those who report consuming television the most compared to other types of media. This is important because it shows that how people consume media is just as important as the content of the media. Those that consume certain types of media, such as television, as well as those who feel more warmly towards the media are more impacted by othering. While those who consume more international news are less likely to be impacted by the othering of Muslims.
Figure – 4


Similar to the dynamic of international news altering the outgroups status of Muslims, I also found support for my third hypothesis. This hypothesis stated that Americans that belong to groups that are more likely to view Muslims as outgroups will be more likely to identify an attack committed by a Muslim as terrorism. These results are shown in Figure 4 as well as Table 2. Political ideology was shown to have an effect on respondent’s definitions of terrorism. This question’s responses were placed on a scale ranging from one to seven. One being extremely liberal, and 7 meaning extremely conservative. Liberals and moderates, or non-conservatives were 1% less likely to call the attack terrorism if they were in the treatment group compared to the control group. While these results were not statistically significant, identifying the assailant as a religious minority makes them less likely to consider the attack terrorism. However, 70% of those who identified as conservative defined the attack as terrorism in the treatment group while 59% did in the control group. This 11% difference was statistically significant.
People who self-identified as Republican followed a similar pattern. This was based on a scale ranging from strong Democrat to strong Republican. Those who identified as either strong or not so strong Republicans were coded into one group while Democrats and Independents were coded into another group. 76% of the treatment group said they would define the question as terrorism while 58% of the control group did. The 18% difference among Republicans is statistically and substantively significant. On the other hand, those who did not identify as Republican (Democrats and Independents) were 3% less likely to call the attack terrorism in the treatment group, this was statistically insignificant. Overall, othering led Republicans and conservatives to be more likely to identify an attack as terrorism. While not statistically significant, there is also suggestive evidence that othering in this way made Democrats and liberals less likely to label the attack as terrorism.
Religion also significantly influenced the efficacy of othering. Self-Identified Christians, those who responded that they were Protestant or Catholic, who were in the treatment group were 11% more likely to call the incident terrorism than those in the control group. People were also asked what they believed about the Bible with the options that it was the actual word of God and should be taken literally, the Bible was the actual word of God but had multiple interpretations, that it was inspired by God, that it is an ancient book of Fables, and no opinion (Campbell et al., 2018). Likewise, people who viewed the Bible as the actual word of God and believed it should be taken literally were 10% more likely to call the attack terrorism in the treatment group compared to the control group. The average treatment effect was statistically significant for both Christians and those who take the Bible literally.
Again, those more likely to see Muslims as an outgroup responded to the othering differently than those likely to see Muslims in the ingroup. Non-Christians were 5% less likely to call the incident terrorism in the treatment group. Likewise, people who did not view the Bible in that way also were only slightly less likely to call the incident terrorism in the treatment group. While neither of these differences are statistically significant, this pattern demonstrates the efficacy of othering. Because this experiment features othering of a religious nature, those who identified as Christian or as taking the Bible literally had those differences highlighted and referred to the attack as terrorism. 
Finally, negative feelings towards racial minorities were also shown to influence othering’s influence in this instance. This was measured through taking the average feelings thermometer responses towards African Americans and Hispanics then the averages were coded into 0-50 (those who felt cold towards racial minorities) and 51-100 (those who felt warmly). Those who reported feeling less warmly towards racial minorities were 17% more likely to call the attack terrorism in the treatment group. Those who reported feeling more warmly towards racial minorities were 3% more likely to label the attack as terrorism, but the difference between the treatment and control groups was not statistically significant. Those who felt negatively towards racial minorities were more likely to have the differences between them and Muslims highlighted compared to those who felt more warmly towards racial minorities. I found very strong support for my third hypothesis. People in groups that were likely to view Muslims as an outgroup were more likely to be affected by the othering. This is further evidence of othering’s separation between the ingroup and the outgroup.
Based on these divergent trends, there is a further reason to believe that othering works based on the identity of both an ingroup and outgroup. In this instance for groups that were either also outgroups or had Muslims in the ingroup, the othering primed them to be less likely to label an attack as terrorism. Those who identified as liberals and/or Democrats as well as non-Christians and those who did not see the Bible as the literal word of God were less likely to see the attack as terrorism. This suggests that those who are in the same ingroup as Muslims are aware of and rejecting the stereotype that Muslims are terrorists. The same pattern held among those who watch primarily international news and felt more negatively towards the media. This is evidence of the above groups feeling that Muslims were an ingroup. Because these groups felt this way, they were less likely to label an attack as terrorism if the attacker was labeled as Muslim, because they were able to identify with Muslims as part of their ingroup. While often these groups did not have statistically significant results the contrast between these results and the results of those viewing Muslims as an outgroup makes the results substantively significant.
Discussion
The goal of this study was to evaluate the effects that othering has on the public perception of terrorism. I hypothesized that if an attacker was connected to the Muslim faith people would be more likely to call the attack terrorism. Additionally, I hypothesized that both media habits and demographic groups would have an impact on perceptions of terrorism. I hypothesized these effects because previous literature showed exceptionally negative American views towards Muslims in addition to the differences in how the media treats Muslims when compared to any other group.
To test my hypotheses, I used an original survey experiment. While the t-test of the entire population of respondents did not reveal a statistically significant difference in how respondents defined the attack, testing among specific groups showed that media habits impact people’s susceptibility to othering. Those that primarily consume television, and those with warm feelings towards the media were more likely to call an attack committed by a Muslim terrorism in comparison to those who did not watch television or felt cold towards the media. On the other hand, those that reported watching more international news were less likely to call an attack committed by a Muslim terrorism in comparison to those who reported watching either more local or national news. How people consume news media and how they feel towards it influences their susceptibility to othering. Those who do not consumeothering where it frequently occurs and those who do not feel warmly towards the media are less likely to be impacted by what the media says.
Testing the impact among different demographic groups showed that othering also primes identities. Conservatives, Republicans, those with negative feelings towards racial minorities, Christians, and those who take the Bible literally were more likely to call the attack terrorism if they received the treatment question with the mention of the Mosque compared to those in the control group. Overall, identity proved to be a strong determining factor in how people responded to the othering.
These results show how effective othering is in activating identity which shapes public opinion. This is important in today’s society because of the media’s history of othering minority groups (Douai &Lauricella, 2014; Hoewe et al., 2014; Powell, 2011). Othering changes certain group’s ideas regarding an identical event. The judgements of superiority that othering can create, in addition to the association with Muslims and terrorism, create an environment where Muslims are viewed more negatively by manydemographicgroups. It is also important because if an attack is popularly defined as terrorism it tends to get a stronger response (Gallup, 2017; Huddly& Feldman, 2011; Kam & Kinder, 2007), so if that label is applied unequally it becomes more difficult to prevent further attacks.
Reporting specific details to put a person either in the in or out group is influential on the public discourse and can be used by journalists in reporting on acts of mass violence. This study shows the importance of evenly using the word terrorism when it applies, regardless of which demographic groups the attacker is in. While certain aspects of the definition of terrorism are still debated there is nothing preventing networks from using the term evenly. This will allow all attacks to be responded to in the way that they should and not given more or less weight depending on how the word terrorism has been applied. Doing so also has the long-term potential of removing or significantly lessening the connection between Islam and Terrorism because it will be evenly associated with people of all demographic groups.
Going forward, there are many possibilities for further research. Similar experiments featuring different forms of othering, including highlighting mental illness, and other minority groups, could be supportive of this hypothesis because it could show othering can affect people in more identities. There is reason to believe that if a different outgroup was highlighted, different demographic groups would react in a similar manner shown in this study. Further research regarding how people define terrorism and why they do so is also important; it is apparent from this study that people’s individual definitions of terrorism vary significantly. Determining other factors behind perception of terrorism is important because the popular definition is a large determinant in the consequences of the attack. Furthering our understanding of intergroup relations remains important because of the crucial role they play in many conflicts worldwide.

 

 

References
Barnett, B. (2003). Guilty and Threatening: "Visual Bias in Television News Crime Stories." Journalism and Communication Monographs, 5(3): 104-155.
Berinsky, A., G. Huber, and G. Lenz. 2012. “Evaluating Online Labor Market for Experimental Research: Amazon’s Mechanical Turk.” Political Analysis, 20(3): 351-368.
Buhrmester, M., T. Kwang, and S. Gosling. 2011. "Amazon’s Mechanical Turk: A New Source of Inexpensive, Yet High-Quality Data?" Perspectives on Psychological Science, 6(1): 3-5.
Cainkar, L. (2002). “No Longer Invisible: Arab and Muslim Exclusion After September 11.” Middle East Report Online, 32: 22-29.
Campbell, D.E., Layman, G.C., Green, J.C., &Sumaktoyo, N.G. (2018). “Putting Politics First: The Impact of Politics on American Religious and Secular Orientations.” American Journal of Political Science, 62(3): 1-15.
Dervin, F. (2015). Discourses of othering. In K. Tracy (Ed.), The International Encyclopedia of Language and Social Interaction. Hoboken, NJ: Wiley. Retrieved from http://search.credoreference.com
de Vreese, C.H., &Boomgaarden, H. (2006). "News, Political Knowledge and Participation: The Differential Effects of News Media Exposure on Political Knowledge and Participation." Acta Politica. 41: 317-341.
Douai, A. &Lauricella, S. (2014). "The ‘terrorism’ frame in ‘neo-Orientalism”: Western news and the Sunni-Shia Muslim sectarian relations after 9/11." International Journal of Media & Cultural Politics, 10(1): 7-24.
Ellis, R., Fantz, A., McLaughlin, E.C., & Hume, T. (14 June 2016). Orlando Shooting: What Motivated a Killer?.Cable News Network. Retrieved from http://cnn.com/
Gallup. (2013, July 8). TV Is Americans’ Main Source of News (Data Set). Retrieved from http://news.gallup.com/poll/163412/americans-main-source-news.aspx?_ga=2.16787321.138632592.1517495122-153919168.1517495122
Gallup. (2017, October 30). Direction From the American Public on Gun Policy (Data Set). Retrieved from http://news.gallup.com/opinion/polling-matters/221117/direction-american-public-gun-policy.aspx?_ga=2.155880755.962551174.1517939370-153919168.1517495122
Gallup (2018, March 14). U.S. Preference for Stricter Gun Laws Highest Since 1993 (Data Set). Retrieved from http://news.gallup.com/poll/229562/preference-stricter-gun-laws-highest-1993.aspx?g_source=link_NEWSV9&g_medium=NEWSFEED&g_campaign=item_&g_content=U.S.%2520Preference%2520for%2520Stricter%2520Gun%2520Laws%2520Highest%2520Since%25201993
Hoewe, J., Bowe, B.J. &Makhadmeh, N. (2014). "Broadcasting Sharia: American TV News’ Illustration of Social Identity and the Emergence of a Threat." Journal of Media and Religion, 13(2): 67-81.
H.R.3162, 107th Cong., U.S. G.P.O. (2001) (enacted).
Huddy, L. & Feldman, S. (2011). "Americans Respond Politically to 9/11: Understanding the Impact of the Terrorist Attacks and Their Aftermath." American Psychologist, 66(6): 455-467.
Jamal, A. (2008). Civil Liberties and the Otherization of Arab and Muslim Americans. In Jamal, A., &Naber, N. (Eds.), Race and Arab Americans Before and After 9/11: From Invisible Citizens to Visible Subjects (114-130). Syracuse, New York: Syracuse University Press
Johnston, L. (2001). Ingroup and outgroup. In J. Michie (Ed.), Reader’s Guide to The Social Sciences. London, UK: Routledge.
Kalkan, K.O., Layman, G.C, &Uslaner, E.M. (2009). "“Bands of Others”? Attitudes toward Muslims in Contemporary American Society." The Journal of Politics, 71(3): 847-862.
Kam, C.D., Kinder, D.R. (2007). “Terror and Ethnocentrism: Foundations of American Support for the War on Terrorism.” The Journal of Politics, 69(2): 320-338.
Levay, K., J. Freese, and J. Druckman. 2016. “The Demographic and Political Composition of Mechanical Turk Samples.” SAGE Open, January-March: 1-17.

Little, D. (2008). American Orientalism: The United States and the Middle East Since 1945. Chapel Hill, NC: University of North Carolina Press
Mason, W. and S. Suri. 2012. "Conducting behavioral research on Amazon’s Mechanical Turk."Behavior Research Methods, 44(1): 1-23.
Morin, A. (2016). "Framing Terror: The Strategies Newspapers Use to Frame an Act as Terror or Crime." Journalism and Mass Communication Quarterly, 93(4): 986-1005.
Ogan, C., Wilnat, L., Pennington, R., & Bashir, M. (2014). "The Rise of anti-Muslim Prejudice: Media and Islamophobia in Europe and the United States." The International Communication Gazette, 76(1): 27-46.
Pew-Templeton Global Religious Futures Project (2010). Muslims (Data Set). Retrieved from http://globalreligiousfutures.org/religions/muslims
Pew Research Center. (6 November 2018a). Republicans Account for a Small but Steady Share of U.S. Muslims (Data Set). Retrieved from http://www.pewresearch.org/fact-tank/2018/11/06/republicans-account-for-a-small-but-steady-share-of-u-s-muslims/
Pew Research Center. (3 January 2018b). New Estimates Show U.S. Muslim Population Continues to Grow (Data Set). Retrieved from http://www.pewresearch.org/fact-tank/2018/01/03/new-estimates-show-u-s-muslim-population-continues-to-grow/

Pagliery, J., & Devine, C. (17 February 2018). School Shooter Showed Violence and Mental Instability at Home, Police Reports Reveal. Cable News Network. Retrieved from http://cnn.com/
Powell, K.A. (2011). “Framing Islam: An Analysis of U.S. Media Coverage of Terrorism Since 9/11.” Communication Studies, 62(1): 90-112
Schnell F. & Callaghan K. (2005). Terrorism, Media Frames, and Framing Effects: A Macro- and Microlevel Analysis. In Schnell F. & Callaghan K. (Eds.), Framing American Politics (pp. 123-147). Pittsburgh, PA: University of Pittsburgh Press.
Sides, J. & Gross, K. (2013). “Stereotypes of Muslims and Support for the War on Terror.” The Journal of Politics, 75(3): 583-598.
Tajfel, H. & Turner, J. C. (1979). An Integrative Theory of Intergroup Conflict. In W. G. Austin & S. Worchel (Eds.), The Social Psychology of Intergroup Relations. Monterey, CA: Brooks-Cole.
Wei, R. & Lo, V. (2008). “News Media Use and Knowledge about the 2006 U.S, Midterm Elections: Why Exposure Matters in Voter Learning.” International Journal of Public Opinion Research, 20(3): 347-362.
Wike, R. & Grim, B.J. (2010). Western Views Toward Muslims: Evidence From a 2006 Cross-National Survey. International Journal of Public Opinion Research, 22(1): 4-25.

 

Data available for replication by contacting the author. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.