While race-based and gender-based prejudice is prevalent and serious (e.g., Charlesworth & Banaji, Reference Charlesworth and Banaji2019; Willard, Isaac, & Carney, Reference Willard, Isaac and Carney2015), so is nation-based prejudice—negative attitudes toward the people of another nation (Bieber, Reference Bieber2018, Reference Bieber2022). Such nation-based prejudice is sometimes caused by collective responsibility, where people attribute responsibility to all citizens of this country (Khraban, Reference Khraban2023). During the recent Russia-Ukraine conflict, for example, many people in the West hold negative attitudes toward Russians, and some of them even direct their hate toward innocent Russian chefs and Russian children (Adam et al., Reference Adam, Bauerova, Rosenzweig-Ziff and Pitrelli2022). Likewise, during the Iraq war, many people outside the US hated Americans (Chiozza, Reference Chiozza2009).
Undoubtedly, collective responsibility is multiply determined. In this research, we focus on one potential contributing factor—the way media outlets report negative international events. Media outlets play a crucial role in shaping how the public perceives and understands international events (Steuter & Wills, Reference Steuter and Wills2010; Loewenstai-Barkai, Reference Lowenstein-Barkai2021; Merskin, Reference Merskin2004). For example, Baum (Reference Baum2002) observed that coverage of foreign policy crises in a soft news context garnered greater attention from politically uninvolved Americans. Wittenberg et al. (Reference Wittenberg, Tappin, Berinsky and Rand2021) discovered that the presentation of political information in video format as opposed to textual format resulted in different perceptions of authenticity and persuasion. Merskin (Reference Merskin2004) noted that after September 11, 2001, President George W. Bush’s media statements regarding the enemy image of Arabs as violent terrorists had a significant and lasting impact on the human rights of Arab Americans.
Media outlets often report negative international events using two prevalent framings. One is nation frame (e.g., “Russia invades Ukraine”) and the other is leader frame (e.g., “Putin invades Ukraine”). In this study, we aim to compare whether and how these two framings differently influence negative impressions toward people in the associated nation (e.g., Russians). We focus on negative events in this article but will discuss the implications for positive events in the General Discussion.
We predict a framing effect: relative to leader-framed news, nation-framed news leads readers to form worse impressions of the people in the nation. Note that our primary goal is to compare the negative impressions created by these two frames in media outlets, without endorsing the accuracy of either frame or recommending the use of either frame. In terms of literal meaning, the nation-framed version may be more accurate than the leader-framed version. For example, it is impossible for Vladimir Putin as an individual person to invade Ukraine, but it is possible for Russia as a nation, with its vast resources and military power, to invade Ukraine. Yet, the nation-framed version may produce the implication that the entire people of the nation are responsible for the negative event, thus prejudicing the reader against the entire people.
The aforementioned framing effect may operate through different cognitive processes, either intuitive, deliberative, or a combination of both. By intuitive, we mean that readers tend to process news framings intuitively and accept them at face value even when they know who is responsible for the reported event (Bago et al., Reference Bago, Rand and Pennycook2020; Evans, Reference Evans2003; Evans & Stanovich, Reference Evans and Stanovich2013; Hsee and Li, Reference Hsee and Li2022; Hsee, Imas and Li, Reference Hsee, Imas and Li2024; Pennycook & Rand, Reference Pennycook and Rand2019, Reference Pennycook and Rand2021; Ross et al., Reference Ross, Rand and Pennycook2021), and people are most likely to do so if the news is from apparently credible sources (Greifeneder et al., Reference Greifeneder, Schwarz, Jaffe and Newman2020). By deliberative, we mean that readers may consider framing as a conversational norm, and thereby consider different framings as reflecting different realities (Koch & Peter, Reference Koch and Peter2017) or different majority views (Bermúdez, Reference Bermúdez2022; Schwarz, Reference Schwarz1999; Sher et al., Reference Sher, McKenzie, Müller-Trede and Leong2022).
Without denying that the framing effect can be due to deliberate processing, we propose that it can also occur intuitively. Specifically, we propose that the negative effect of nation-framed news can occur even if readers know or believe that it is the leader, not the people, who is responsible for the reported event. Thus, when the news is in the nation frame, readers will intuitively associate the reported event with the people in the nation based on the face value of the nation frame (Bago et al., Reference Bago, Rand and Pennycook2020; Phillips et al., Reference Phillips, Fletcher, Marks and Hine2016).
Since the negative effect of nation-framed news is at least partly due to intuitive processing rather than solely due to ignorance of who is responsible, we predict that the negative effect will attenuate if the reader is prompted to mentally separate the leader from the people. This ameliorating effect will occur even if the reader is not given any new information (e.g., not told who is actually responsible). In other words, merely prompting readers to mentally separate the leader from the people will activate their existing knowledge (or belief) that the leader is responsible, thereby disrupting the intuitive association between the event and the people.
We focus on two specific ways to prompt readers to mentally separate the leader from the people. One is to show readers the nation-framed and leader-framed versions side by side and ask them which version is more accurate before asking about their impression of the people. The other is to first ask readers about their impression of the leader before asking about their impression of the people (Schwarz, Reference Schwarz1999; Schwarz & Clore, Reference Schwarz and Clore1983). We predict and show that both methods can attenuate the negative effect of nation-framed news on the reader’s impression of the people, by activating their existing knowledge (or belief) and mentally separating the leader from the people.
Study Overview
We tested these predictions in five studies (see Table 1 for an overview). Our studies were diverse—they used participants from different parts of the world, included both real-world data collected from Twitter and questionnaire data collected online, and employed both natural language processing techniques and standard experimental methods.
All studies were conducted in the context of the 2022 Russia-Ukraine conflict, and during the months immediately after the onset of the conflict, when the event featured prominently in the media. Most studies examined Western participants’ impressions of Russians during the Russia-Ukraine conflict. To test the generality of our findings, one study examined Chinese participants’ impressions of Americans during the conflict.
Study Overview
Study 1 (the Twitter study) was based on 149,099 Twitter replies. For all the remaining studies, we pre-set a target sample size of 150 participants per cell before any data analysis. We report all manipulations, measures, and data exclusions (if any). Study 4 was pre-registered at https://aspredicted.org/KKV_W4M. All data and analysis codes are available at https://osf.io/d5zkm/?view_only=6d4e2ee1ab674e1fb92e15a4f9a0f639.
Study 1
In Study 1, we conducted an investigation of existing Twitter data. We analyzed replies to news tweets concerning the ongoing Russia-Ukraine conflict from major media outlets. We tested whether nation-framed news tweets, relative to leader-framed news tweets, generated a higher percentage of negative responses to Russians.
Method
Our method consisted of two steps. First, we identified nation-framed and leader-framed news tweets about the Russia-Ukraine conflict. We included 10 major English-language media outlets accounts (New York Times, Wall Street Journal, Fox News, ABC, CNN, USA Today, NPR Politics, BBC World, BBC Breaking, and NPR Extra) representing both left- and right-leaning perspectives. To cleanly test our hypothesis, we used crawler code to collect all original news tweets and their replies from the above accounts that met the following criteria: (a) we used tweets published during the first three months (February 24 and May 24, 2022) of the conflict, because the event was featured prominently in the media during that period, (b) we used tweets containing either “Russia” or “Putin,” but not both, in order to better distinguish between nation-framed and leader-framed news, (c) we used tweets containing the word “invades” or its variants (e.g., “invasion,” “invading,” “invaded”), and the phrase “special military operation” (as officially termed by the Putin administration), and tweets containing “Ukraine.” The reason we included these terms was to ensure that the tweets were relevant to the ongoing Russia-Ukraine conflict. We found 876 nation-framed news tweets (which contained “Russia” but not “Putin”) and 85 leader-framed news tweets (which contained “Putin” but not “Russia”). The dominance of nation-framed news tweets is consistent with our observation that news media tends to use the nation frame rather than the leader frame.
Our next step—which was the main step of the study—was to analyze readers’ responses to the two types of news tweets. We report the results of the analyses below.
Results and Discussion
In total, these news tweets generated a total of 149,099 replies, with each tweet receiving at least one reply. Table 2 reports the results by type of news tweets (nation-framed versus leader-framed). Row A of the table reports the number of all replies, and Row B reports the number of negative replies about Russians, defined as replies that contained “Russians” or “Russian people” and had a negative sentiment. We measured sentiment with VADER, a lexicon/rule-based sentiment analysis tool that was developed specifically to measure sentiment in social media posts (Hutto & Gilbert, Reference Hutto and Gilbert2014) and has been used to identify hate speech and cyberbullying on social media (Gitari et al., 2015; Rodriguez et al., Reference Rodriguez, Argueta and Chen2019; Zhao & Mao, Reference Zhao and Mao2017). (Negative sentiment is closely related to hate speech [Bilewicz & Soral, Reference Bilewicz and Soral2020; Mohiyaddeen & Siddiqi, Reference Mohiyaddeen and Siddiqi2021; Schmidt & Wiegand, Reference Schmidt and Wiegand2017].) Examples of negative replies about Russians include “God hates Russians,” “Damn, Russians don’t have respect for anything they have become Taliban,” “Nah fuck the Russians, bunch of Nazis,” “We’re okay with racism as long as it’s against Russians,” and “May these brainwashed Russians rot in Hell!”
Study 1 Results. The results show that nation-framed news tweets, relative to leader-framed news tweets, generated a higher percentage of negative responses about Russians
Row C reports our key dependent variable, the percentage of replies that were negative replies about Russians. The percentage was almost twice as high for the nation-framed news tweets (1.34%) as for the leader-framed news tweets (0.69%), χ 2(1) = 65.72, p < 0.001, Cramer’s V = 0.021. The result supports our proposition: nation-framed news, relative to leader-framed news, generated significantly more negative replies about Russians, controlling for the number of news tweets and number of replies.
Rows D and E report additional results. Significantly more readers liked and retweeted negative replies about Russians when the replies were responses to nation-framed news than when they were responses to leader-framed news, t(1343.28) = 3.28, p = 0.001, d = 0.10 and t(1216.28) = 2.92, p = 0.004, d = 0.09, respectively.
Using field data, Study 1 provided preliminary evidence that nation-framed news about the Russia-Ukraine conflict generated more negative reactions toward Russians than leader-framed news about the conflict. Also, the negative reactions to nation-framed news (relative to leader-framed news) received more “likes” and generated more retweets from other readers.
Because Study 1 was not a controlled experiment, we could not ascertain the causality between the framing of the news tweets and readers’ responses. It is possible that the readers of the nation-framed news tweets tended to be more anti-Russian to begin with than the readers of the leader-framed news tweets. It is also possible that the nation-framed tweets contained more negative information about Russians than the leader-framed tweets. Due to these and other potential confounds, the results of Study 1 are only suggestive. In the remaining studies, we empirically manipulated framing to minimize confounds.
Study 2
In Study 2, we manipulated framing by showing one group of participants a set of nation-framed news headlines about the Russia-Ukraine conflict and showing another group the corresponding leader-framed news headlines. We predicted that participants who read the nation-framed versions would have worse impressions of Russians than participants who read the leader-framed versions.
Method
Participants were 301 US-based workers (149 females, M age = 41.01) recruited on CloudResearch. Participants were randomly assigned to either a nation-framed or leader-framed condition, and the randomization controlled for any possible baseline difference between the two conditions. Those in the nation-framed condition read the following six nation-framed news headlines, adapted from actual news headlines found online and each presented on a separate page:
Russia invades Ukraine
Russia’s aggression is unifying NATO countries
Russia takes censorship to new extremes: stifling war coverage
Russia renews widespread attack across Ukraine
Russia may use chemical and biological weapons in Ukraine
Russia could use nuclear weapons if existence threatened
The leader-framed condition was identical to the nation-framed condition except that the word “Russia” in each headline was replaced with “Putin.”
We used a single item to measure the dependent variable, and all participants were directed to another page after reading the headlines and were asked, “Overall, what is your impression of the Russian people?” They answered on a 9-point scale, anchored by very bad and very good. In our analysis, we coded the 9-point ratings from −4 to + 4 scale, with very bad coded as −4 and very good as +4.
Results and Discussion
As Figure 1 illustrates, participants who read the nation-framed news headlines had significantly worse (more negative) impressions of Russians (M = −2.17, SD = 1.97) than participants who read the leader-framed versions (M = −1.60, SD = 2.24), t(299) = 2.35, p = 0.019, d = 0.27. These results demonstrate the framing effect in a controlled experimental setting.
Study 2 Results. The results show that impressions of Russians were significantly worse in the nation-framed condition than in the leader-framed condition.
Study 3
Study 3 was a conceptual replication of Study 2 that tested the generality of the framing effect by using a non-Western sample and taking a non-Western perspective. Specifically, we asked Chinese participants to read either a nation-framed or leader-framed news headline about the US’s involvement in the Russia-Ukraine conflict, and measured their impressions of Americans. (The US involvement was viewed as a negative event in China.) We predicted that Chinese readers who read the nation-framed version would have worse impressions of Americans than those who read the leader-framed version.
Study 3 had two other design features that Study 2 lacked. First, we measured participants’ impressions of Americans before they read the headline (i.e., baseline impressions) as well as after, so we could test how their impressions changed after they read the headline. Second, we measured impressions of Russians as well as impressions of Americans (our primary dependent variable). We were curious whether reading the (manipulated) headline would have a spillover effect on impressions of the people of other nations who were involved in the international event but were not the focus of the headline.
Method
Participants were 300 workers (190 females, M age = 29.00) recruited on Credamo, an online survey platform in China. They were randomly assigned to either a nation-framed or leader-framed condition.
First, participants in both conditions were asked to report their impressions of Americans and their impressions of Russians, and they answered each question on a 9-point scale, anchored by very bad and very good. (As in the previous study, we coded very bad as -4 and very good as +4.) On the next screen, participants read, “When you woke up this morning, you saw the following headline:” In the nation-framed condition, the headline read, “The US pours fuel on the fire by continuing to send military aid to Ukraine.” (This headline was from an official Chinese Weibo account and was originally in Chinese: 
) In the leader-framed condition, we replaced “the US” with “Biden.” Finally, participants in both conditions were asked to rate their impressions of Americans and Russians again on the same 9-point scale.
Results and Discussion
Not surprisingly, there were no significant differences between the nation-framed and the leader-framed conditions in either participants’ baseline impressions of Americans or their baseline impressions of Russians, ps > 0.10.
Our primary dependent variable was how impressions of Americans changed after reading the nation-framed or leader-framed headline, namely, impressions of Americans after reading the headline minus baseline impressions of Americans. Figure 2A displays the results. Our analyses found that the changes were negative in both the nation-framed and the leader-framed conditions (Ms = −1.35 and −1.06, SDs = 1.24 and 1.09, t(149)s = 13.39 and 11.86, ps < 0.001, compared with 0), but the changes were significantly more negative in the nation-framed condition than in the leader-framed condition (F(1, 298) = 4.73, p = 0.030, η 2 = 0.016). It seems that reading the headline worsened participants’ impressions of Americans regardless of the frame of the headline, but this worsening effect was significantly greater when the headline was nation-framed than when it was leader-framed. This latter result replicated the framing effect found in the other studies. In other words, controlling for their baseline impressions of Americans, those who read the nation-framed headline had worse impressions of Americans than those who read the leader-framed version. (The framing effect still held, even if we did not control for baseline impressions, t(298) = 2.55, p = 0.011, d = 0.30.)
Figure 2B displays the results of our secondary dependent variable—changes in impressions of Russians from the baseline level to the after-manipulation level. Our analyses found no significant changes (Ms = −0.10 and 0.05, SDs = 1.07 and 1.08, ps > 0.10, compared with 0) and no significant framing effect (p > 0.10).
These results suggest that while both the act of reading the headline (about the US military aid to Ukraine) and the framing of the headline influenced Chinese participants’ impressions of Americans, neither factor had a spillover effect on impressions of Russians. It seems that reading a headline influenced only the reader’s impression of the focal people, and not their impressions of other people.
Study 3 demonstrated the generality of the framing effect by replicating it with a different participants pool (Chinese instead of Americans), and a different target people (Americans instead of Russians).
Study 4
Study 4 explored the underlying psychology by asking readers to judge the truthfulness of the nation-framed versus leader-framed news, and tested the ameliorating effect of separating the leader from the people. The study consisted of three framing conditions—nation frame, leader frame, and comparative frame. In the first two conditions, we showed participants either nation-framed or leader-framed news headlines about the Russia-Ukraine conflict; we asked them first to judge the truthfulness of the headlines, and then to report their impressions of Russians. In the comparative frame condition, we showed participants both the nation-framed and leader-framed versions side by side; we asked them first to judge which version was more truthful, and then to report their impressions of Russians.
We predicted that the headlines would be judged as similarly truthful in the nation-framed and leader-framed conditions because, without a direct comparison of the two versions, readers would accept the presented news at face value. We also expected to replicate the framing effect (i.e., worse impressions of Russians in the nation-framed condition than in the leader-framed condition). In the comparative-framed condition, however, we expected that participants would find the nation-framed headlines less truthful than the leader-framed headlines. We also expected that their impressions of Russians would be similar to the impressions of those who read only the leader-framed version, and better than the impressions of those who read only the nation-framed version. According to our theorization, the juxtaposition of the two versions should enable participants to mentally separate the responsibility of the Russian people from that of the Russian leader and should activate participants’ preexisting knowledge (or belief) that the leader, not the people, was largely responsible for the reported events. (For the purpose of this research, we are interested only in readers’ judgment of truthfulness; it is beyond the scope of this research to determine which version was actually more true.)
Method
Study 4 was pre-registered at https://aspredicted.org/KKV_W4M. Participants were 451 workers (230 females, M age = 27.07) recruited on Prolific, which reports that most of its workers are from OECD (Organization for Economic Cooperation and Development) countries. Participants were randomly assigned to one of three conditions: nation frame, leader frame, and comparative frame.
Participants in the nation-framed condition read the following six nation-framed news headlines, each presented on a separate page:
Russia invades Ukraine
Russia’s brutal attack on Ukraine is wrong and must stop
Russia’s aggression is unifying NATO countries
Russia takes censorship to new extremes: stifling war coverage
Russia may use chemical and biological weapons in Ukraine
Russia could use nuclear weapons if existence threatened
After reading each item, participants were asked whether it reflected the truth on a 9-point scale anchored by Definitely no and Definitely yes. We coded the ratings so that Definitely no = 1 and Definitely yes = 9.
The leader-framed condition was identical to the nation-framed condition except that we replaced “Russia” in each item with “Putin.”
In the comparative-framed condition, participants read the nation-framed and the leader-framed versions side by side, with each pair on a separate page. After reading each pair, participants were asked which version better reflected the truth on a 9-point scale (greater numbers meant that the leader-framed version better reflected the truth).
Finally, participants in all conditions completed the main dependent measure: “Overall, what’s your impression of Russians?”; they answered on a 7-point scale anchored by very bad and very good. (We coded the ratings so that very bad = −3 and very good = +3.)
Results and Discussion
Before reporting the results of our main dependent variable (impressions of Russians), we report the judged truthfulness of the news headlines. As we predicted, the nation-framed and leader-framed versions were judged as similarly truthful in the nation-framed and leader-framed conditions, Ms = 7.22 and 7.16, SDs = 1.07 and 1.25, t(300) = 0.40, p > 0.10. On the other hand, the leader-framed versions were judged as significantly more truthful than the nation-framed versions in the comparative-framed condition, M = 6.41, SD = 2.15, t(148) = 7.99, p < 0.001, compared with the midpoint of the scale. It seems that readers believed that the leader-framed versions were more truthful than the nation-framed versions, but this belief was not activated unless readers saw the two versions side by side for comparison (Hsee, Reference Hsee1996; Li & Hsee, Reference Li and Hsee2019).
We now report the main results—impressions of Russians, shown in Figure 3. As we predicted, participants in the nation-framed condition (M = −1.02, SD = 1.59) had significantly worse impressions of Russians than both participants in the leader-framed condition (M = −0.43, SD = 1.64), p = 0.001, and participants in the comparative-framed condition, (M = −0.36, SD = 1.51), p < 0.001, with no significant difference between the latter two conditions, p > 0.10.
Study 4 Results. The results (a) replicate the framing effect (the left bar and the middle bar), and (b) show that impressions of Russians in the comparative frame (the right bar) were similar to impressions of Russians in the leader frame (the middle bar), and better than impressions of Russians in the nation frame (the left bar).
Consistent with our theory, the results suggest that participants who read only one version accepted what they read at face value, so those who read the nation-framed version had worse impressions of Russians than those who read the leader-framed version. On the other hand, participants who read both versions side by side believed that the leader was largely responsible for the reported events; therefore, their impressions of Russians were similar to the impressions of Russians of those who read only the leader-framed version, and significantly better than the impressions of Russians of those who read only the nation-framed version (see Figure 3).
Study 5
While Study 4 tested the juxtaposition of nation-framed and leader-framed news headlines as one way to prompt readers to mentally separate the leader from the people, Study 5 tested another method: asking participants to report their impressions of the leader before asking about their impressions of the people. We predicted that readers who were first asked about their impressions of the leader would have less bad impressions of the people than readers who were not asked about the leader first. This prediction is inspired by classical literature showing the effect of a priming question on the attribution of subjective well-being (Schwarz, Reference Schwarz1999; Schwarz & Clore, Reference Schwarz and Clore1983).
Method
Study 5 manipulated both framing and question sequence; for completeness, it measured participants’ impressions of both Russians (primary dependent variable) and Putin (secondary dependent variable). Participants were 602 workers (328 females, M age = 27.92) recruited on Prolific. They were randomly assigned to one of four between-subjects conditions in a 2 (frame: nation vs. leader) × 2 (question sequence: people first vs. leader first) factorial design. Note that the framing manipulation was about the headlines and the order manipulation was about the dependent-variable questions.
Participants in the nation-framed conditions read the following six nation-framed news headlines, each on a separate screen:
Russia invades Ukraine
Russia plans to take whole of Ukraine but is failing
Russia wants to divide Ukraine
Russia may use chemical and biological weapons in Ukraine
Russia wants to create more fear
Russia’s war on Ukraine hurts us all
The headlines were the same in the leader-framed conditions except that “Russia” was replaced with “Putin.” After reading each headline, all participants were asked two questions on separate pages. In the people-first conditions, the first question was, “What’s your impression of Russians?” and the second was, “What’s your impression of Putin?” The order of the two questions was reversed in the leader-first conditions. Participants answered each question on a 7-point scale anchored by very bad and very good. As before, we coded very bad as −3 and very good as +3.
To test whether the leader-framed versions were perceived as more truthful, we recruited a separate sample from the same population as in the main study (N = 150, 82 females, M age = 25.73), showed them both versions of each headline, and asked them which version better reflected the truth on a 9-point scale (higher numbers meant that the leader-framed version better reflected the truth). For all six headlines, participants considered the leader-framed version more reflective of the truth, all ps < 0.001.
Results and Discussion
Figure 4A displays the results of the primary dependent variable—impressions of Russians. A 2 × 2 ANOVA yielded a main framing effect, F(1, 598) = 20.61, p < 0.001, η 2 = 0.033, and a main question-sequence effect, F(1, 598) = 81.71, p < 0.001, η 2 = 0.120. The results replicated the framing effect and demonstrated the question-sequence effect. Specifically, a comparison of the dark and light bars on the left reveals the framing effect: when asked about their impressions of Russians first, participants who read nation-framed headlines had worse impressions of Russians than those who read the corresponding leader-framed headlines, Ms = −1.75 and -0.93, SDs = 1.54 and 1.43, F(1, 598) = 25.10, p < 0.001. The framing effect disappeared if participants were asked about their impressions of Putin first, Ms = −0.41 and −0.18, SDs = 1.28 and 1.39, p > 0.10, as illustrated by the dark and light bars on the right.
Study 5 Results. The results (a) replicate the framing effect (the two left bars in Fig 4A), (b) show the question-sequence effect (the two dark bars in Fig 4A), and (c) show that framing and question sequence did not have the same effects on impressions of Putin as on impressions of Russians.
A comparison of the two dark bars reveals the question-sequence effect: participants who read the nation-framed headlines had less bad impressions of Russians if they were asked about their impressions of Putin before (vs. after) their impressions of Russians, Ms = −0.41 and −1.75, SDs = 1.28 and 1.54, F(1, 598) = 67.34, p < 0.001. The question-sequence effect was weaker for participants who read the leader-framed headlines, Ms = −1.18 and −0.93, SDs = 1.39 and 1.43, F(1, 598) = 21.01, p < 0.001, as shown by the two light bars.
The results suggest that participants who were not asked about their impression of Putin first took the nation-framed news headlines at face value, intuitively associating the events with Russians in general. By asking participants about their impression of Putin first, however, we activated the belief that Putin, rather than Russia, was primarily responsible, and participants mentally separated their impression of Russians from their impression of Putin.
Figure 4B displays the results of the secondary dependent variable—impressions of Putin. A 2 × 2 ANOVA found no main framing effect, p = 0.678, but a main question-sequence effect, F(1, 598) = 5.00, p = 0.026, η 2 = 0.008. The question-sequence effect on impressions of Putin was in the opposite direction from the effect on impressions of Russians: asking participants about their impressions of Russians before asking about impressions of Putin worsened impressions of Putin. The fact that the question-sequence manipulation had opposite effects on impressions of Putin versus impressions of Russians indicates that readers strongly believed that the reported events were largely driven by Putin rather than by the Russian people.
To summarize, Study 5 showed both the framing effect and the question-sequence effect. Readers of nation-framed news headlines had worse impressions of Russians than readers of leader-framed news headlines, and this effect attenuated if the readers were first asked about their impressions of Putin.
While the framing manipulation affected impressions of Russians, it did not affect impressions of Putin. Asking readers about their impressions of Russians before asking about their impressions of Putin even worsened their impressions of Putin. It seems that readers believed that Putin was largely responsible for the events reported in the news headlines, and the priming question about Russians activated that belief.
General Discussion
Different media outlets may report the same international events either in terms of the nation (e.g., “Russia invades Ukraine”) or in terms of the leader (e.g., “Putin invades Ukraine”). Focusing on the ongoing Russia-Ukraine conflict, and using both field and experimental data, we find that reading nation-framed (vs. leader-framed) news headlines about the conflict leads to significantly worse attitudes toward the people of the nation. We also find that prompting readers to mentally separate the leader from the people moderates the negative effect of nation-framed news. It should be noted that we are not asserting which framing is more accurate or should be adopted by the media. Although Study 4 shows that the leader frame is perceived to be more truthful than the corresponding nation frame, caution needs to be applied when using the leader frame, and it should not be used indiscriminately.
Open Questions
As mentioned in the Introduction, the framing effect may have multiple determinants. In addition to the intuitive and deliberate processes previously discussed, another potential explanation is that readers might interpret “Russians” as representing different groups depending on the frame used. They may perceive “Russians” as including both civilians and Putin when presented with the nation frame, while interpreting “Russians” as civilians excluding Putin when presented with the leader frame. We hope that future research will further differentiate these multiple explanations.
The present research has focused on negative international events rather than positive international events. Nevertheless, we conjecture that the framing effect observed in this research may also apply to positive international events: the nation-framed news about a positive international event may give readers a more positive impression of the nation’s people than the leader-framed news, even if the event is driven largely by the nation’s leaders. For example, “Saudi Arabia accepted two thousand asylum seekers” likely gives readers a more positive impression of the Saudi people than “King Salman accepted two thousand asylum seekers.”
The present research has focused on negative international events rather than negative domestic events. We suspect that the framing effect is less applicable to domestic events than to international events. Readers usually have a more stable attitude about the people of their own nation (which includes themselves) than about the people in other nations, so the framing of news about a domestic event is less likely to alter their impressions of the people in their own nation than the framing of news about an international event to alter their impressions of the people in the other nation. These are our speculations and await future research to verify.
While the current research has focused on negative international events, we are not saying that all negative international events are driven by the nation’s leaders. Rather, we recommend that when reporting an international event, media outlets should choose the frame that most accurately conveys responsibility instead of defaulting to the nation frame. At the same time, readers should be aware that many events reported in nation-framed news are actually the decisions or actions of the nation’s leadership, especially in nations where the leaders have supreme power. Next time you read nation-framed news such as “Russia…” or “North Korea…,” think about who in the country is doing the “…”
In conclusion, this research, focusing on the 2022 Russia-Ukraine Conflict and involving both field and experimental data, reveals the differential impact of nation-framed versus leader-framed news on the impressions of the people in the associated nation (Russians). Our findings indicate that the framing effect arises at least partly due to readers’ tendency to intuitively accept the face value of the news reports. Thus, the framing effect can be mitigated when readers are presented with both nation-framed and leader-framed versions simultaneously and asked to evaluate which version is more accurate. Additionally, the effect is further mitigated when readers are prompted to first report their impression of the leader before that of the people. The negative impression may be a double-edged sword. On one hand, it may foster xenophobia, motivate hate crimes, and escalate international tensions. On the other hand, if the entire people in a nation (not just the leaders) are responsible for a negative event, then an international negative attitude toward these people can be effective in ending the negative event (Small, Reference Small2002; Vanderheiden, Reference Vanderheiden, Vincent, van de Poel and van den Hoven2011; William A., Reference William1999). Considering the complex consequence of collective responsibility, we recommend that both the media and the readers be cognizant of the effect of framing on the reader’s impressions of foreign nationals.
Availability of data and materials
The data that support the findings of this study are openly available in OSF at https://osf.io/d5zkm/?view_only=6d4e2ee1ab674e1fb92e15a4f9a0f639.
Open access
