Psychophysiological audience responses to war journalism and peace journalism

The role of peace journalism (PJ) was summarized by Johan Galtung as ‘making audible and visible the subjugated aspects of reality’ (cited in Lynch and McGoldrick, 2005: 224, personal communication, 21 March 2003). Decades earlier, Galtung had shown, with Mari Holmboe Ruge (1965), that influences on news content are predominantly structural, manifest in a set of news values, or ‘tuning factors’, which arise from the economic and political interests of media industries.

These influences gave rise to conventions, perhaps the most salient of which, when reporting conflict, is ‘frequency’ – a bias, in news, in favour of event over process, with prominence accorded to manifest increments that begin and end in the interval between deadlines, hence the daily digest of bombs and battles in the reporting of conflict. Another widespread convention, Galtung and Ruge (1965) argued, was elite orientation: ‘hardly strange [because] the actions of the elite are, at least usually and in short-term perspective, more consequential … [which] applies to elite nations as well as to elite people’ (p. 68).

The operation of such conventions, over time, leads to certain identifiable perspectives, orders of priority and versions of events dominating in news discourse, while others are subjugated. So, for instance, ‘In an elite-centred news communication system, ordinary people are not even given the chance of representing themselves’ (Galtung and Ruge, 1965: 68). The continuing predictive capacity of the Galtung–Ruge newsworthiness model was confirmed in a later study by Harcup and O’Neill (2001), with ‘entertainment’ the only significant addition as a major theme of news.

PJ was conceived as a remedial strategy, a programmatic response to ‘the policy implications of the [1965] study’ (Lynch and Galtung, 2010: xii). One of its key precepts, of both news values and reportorial technique – disseminated worldwide in countless pedagogical initiatives – is on how to establish the consequentiality of non-elite people:

The challenge is to convey their connection with the conflict, and the outlook for its likely development, in the course of a news story … you need … to construct a framework of understanding, in which the relevance of new perspectives to the story about the bomb can be made clear. (Lynch and McGoldrick, 2005: 165–166, emphasis in the original)

Galtung and Ruge based their findings on a study of Norwegian newspaper reporting on crises in Cuba, Cyprus and the Congo. Today, however, the milieu in which the issues of representation they raised, and their potential political ramifications, are perhaps most widely discussed is the Israel–Palestine conflict. Philo and Berry (2004) showed how the operation of news reporting conventions produced impressions among audience members, of basic facts about this conflict, that were conducive to politically dominant perspectives in a US-allied country such as the United Kingdom, where the fieldwork was carried out while unsupported by evidence from the conflict itself.

At one point during their initial survey research, Philo and Berry found that British television viewers were, for example, more likely to believe it was the Palestinians who were occupying the Occupied Territories, than to realize it was the Israelis. They had heard the words, ‘occupied’ and ‘occupation’, but these were never adequately explained. Aspects of background and context, essential to understand the situation, would inevitably require an account of process, rather than a mere recitation of events, so would generally be excluded by the ‘frequency’ tuning factor. And the elite voices from Israel and the United States, who – Philo and Berry showed – dominated the choice of on-screen commentators presented as consequential in UK television news had no interest in dwelling on the illegality of Israel’s military occupation of territory it seized by force in 1967 or of the settlements it has established on that land.

The subjugation of these aspects in media framing found its direct correlative, Philo and Berry found, in audience framing:

There was no apparent rationale for Palestinian actions when they attacked Israelis. The Palestinians could thus be seen as ‘starting’ the violence. The Israelis were portrayed in the news as ‘responding’ to those attacks … the audience clearly took this message from what they saw [on television news]. (Philo, 2010: 412)

This is the background against which routine Israeli protestations of self-defence, in the military onslaughts on Gaza of recent years, are played out. The UN Human Rights Council Fact Finding Mission on ‘Operation Cast Lead’, in 2008–2009, said it was:

Designed to punish, humiliate and terrorise a civilian population, radically diminish its local economic capacity both to work and to provide for itself, and to force upon it an ever increasing sense of dependency and vulnerability.

No equivalent study to the one by Philo and Berry (2004) has been published on Australian television news or its effect on framings by Australian audiences. However, a national survey on Australian public understanding and opinions concerning the Israel–Palestine conflict found:

the largest plurality (37%) selected ‘Ancient hostility between Jews and Arabs’ [as] the main cause of the conflict … which is the dominant narrative in popular culture including films and television programmes but devoid of historical evidence [to support it]. (Han and Rane, 2011: 60)

This is the explanation that tends to ‘prevail by default’ (Lynch and McGoldrick, 2005: 63) in the absence of any explicit contextualizing material.

To make good on these deficits is, then, a task for the remedial strategy of PJ. PJ is a set of distinctions in the reporting of conflict, put forward originally by Johan Galtung (1998) and operationalized in subsequent research to yield evaluative criteria for content analysis (Dente Ross and Tehranian, 2008; Lee and Maslog, 2005; Lynch, 2004, among others). Surveying early examples of this research, Shinar (2007: 200) found that scholars identified, as PJ, reporting of conflict issues which:

For the experiment of which a part is presented and discussed in this article, these five observed characteristics were adopted as a set of headings for the codification of the war journalism (WJ) and PJ versions of familiar stories from television news, in four countries (Australia, the Philippines, South Africa and Mexico) between September 2010 and May 2012. The WJ versions were stories that had already been recently broadcast by a television newsroom in each country that acted as host and partner for the research. PJ versions were created by gathering new material (pictures and interviews) and re-editing and re-scripting in the same style and to approximately the same duration as the first (with some material from the WJ version shortened and/or omitted, to make room) and voiced by the same reporter who had recorded the original. In all, between the four locations, 21 stories were re-versioned in this way. In each place, a WJ and a PJ ‘bulletin’ were created by recording the stories through a studio, with ‘links’ (introductions) read from autocue by a local television news presenter.

Distinctions in the PJ model, arranged under Shinar’s five general headings (above), were fine-tuned to ‘capture’ differences of representation likely to prove ideational among audiences at the particular time and place of the experiment, with reference to the particular stories and conflicts chosen. Understanding of the issues at stake in audience framing was developed, in each country case, through a process attentive to the principles of Critical Discourse Analysis, ‘focusing on the role of discourse in the reproduction and challenge of dominance’ in an ‘unabashedly normative’ quest to create opportunities for ‘change through critical understanding’ (Van Dijk, 1993: 251–253).

In Phase 1 of the research, the PJ and WJ bulletins were played to different audiences in each country, with a range of techniques employed to generate both qualitative and quantitative data, to compare viewer responses. A total of 472 participants in the four countries were broken up into groups of equal size for simultaneous screenings in separate rooms of the WJ and PJ ‘bulletins’ in each country. These took place in October 2010 in Australia, February 2011 in the Philippines, April 2012 in South Africa and May 2012 in Mexico.

The Differential Emotion Scale (DES; Izard, 1977) was employed to test for differences between the groups before and after viewing, similar to the method used by Unz et al. (2008). The DES is a 30-item questionnaire, consisting of 10 fundamental emotions, each assessed by three items. The ten fundamental emotions in this scale are interest, enjoyment, surprise, distress, anger, disgust, contempt, fear, empathy and guilt. Each of these 30 items is rated on a 5-point Likert scale (1 = very slightly, not at all; 5 = extremely), yielding a score out of 5.

Before watching the bulletin, Phase 1 participants filled in the DES to measure baseline levels on each of the 10 subscales, which revealed no statistically significant pre-existing differences between any ‘pair’ of groups in any of the countries. During the bulletin, the video was paused at the end of each story for participants to fill in another DES, enabling changes in their self-reported emotional state to be tracked across the bulletin in response to each story, compared with the baseline.

During the pauses between stories, Phase 1 participants were also asked to perform a Thought-Listing Protocol (TLP), inviting them to simply write down any notes about thoughts or feelings prompted by what they had just watched (following the method pioneered by Coleman and Thorson, 2002; see also Russo et al., 1989; Shapiro, 1994). Data from these TLPs were themed according to a simplified version of Entman’s (1993) four functions of substantive framing:

Results from Phase 1 of the study are presented in Lynch and McGoldrick (2013), Lynch (2014) and McGoldrick and Lynch (2014). In general, the results showed that ‘cognitive responses shared by participants … in TLPs and focus groups, were led by emotional responses’ (Lynch, 2014: 160). Watching PJ generally ‘left people less angry and fearful, and more hopeful and empathic’ (Lynch and McGoldrick, 2013: 1041), with this response leading to an increase in receptiveness to non-violent conflict responses:

Where members of a particular group have been the subject of psychological distancing in political spectacle, presenting their perspective in their own words is an essential element of an approach leading to greater empathy, hope and happiness among TV viewers, and less anger and fear. If accompanied by suitable explanation of contexts and backgrounds, this in turn activates audience [cognitive] frames in which problem definitions are more likely to be structural than focused on individual malefactors, and treatment recommendations therefore focused more on the delivery of [social] justice … than on deterrence and punishment. (Lynch and McGoldrick, 2013: 1055)

Some of the strongest interactions between the independent variable (media framing) and dependent variable (audience responses) in Phase 1 of the study (McGoldrick and Lynch, 2014) came in response to two stories dealing with Australian government policy towards asylum seekers and a stage of the Israel–Palestine ‘peace talks’ brokered by then-US Secretary of State Hillary Clinton. Both featured a non-elite speaker or protagonist in a framework of understanding which, the Phase 1 results suggested, prompted and enabled viewers to appreciate the consequentiality of the perspectives they presented. In the story about responses to asylum seekers, Ali Jafaari, a refugee who had arrived by boat from Afghanistan, talked of fleeing persecution to make a new life and wanting to make Australia proud of him. In the Israel–Palestine story, Bishara Costandi, the Palestinian refugee now living in Sydney, invited local residents to imagine a journey in the occupied West Bank as the equivalent of setting out to travel from Marrickville to Glebe (two adjacent suburbs), only to have to negotiate 14 army checkpoints. Hence, these were the two stories selected for further data collection in Phase 2 of the study.

The Phase 1 findings must be regarded as subject to a number of methodological limitations. Self-reporting emotional questionnaires, such as the DES, have their limits in that people are not always aware of feelings – what they are and what they are experiencing in their body. They may have disassociated from their feelings or lack the linguistic dexterity to express their emotions and/or use the adjectives in the tests in different ways to those conceived by their designers. This could explain why the DES did not always demonstrate strong interactions for individual emotions, even when responses in TLPs and Focus Group Discussions (FGDs) seemed to indicate a pronounced response.

So, the aim in Phase 2 of the research was to gather more direct, less filtered indications of emotional responses. Since a large amount of brain activity is unconscious, a physiological measure of heart rate variability (HRV) was used. Second, the testing was slowed down, showing participants only two stories, rather than six, to ensure clearer measures of responses to these stories. And third, closer attention was paid to cognitive changes by measuring thoughts about the topics in these two stories one week before viewing and then again directly afterwards.

In addition to the physiological testing, another new protocol was devised, drawing on Han and Rane’s 2011 study of Australian public opinion on the Israel–Palestine conflict, to try to obtain indications of the nature and extent of any cognitive shifts brought about among viewers by watching the war or PJ framing of the Israel–Palestine conflict. Because the heart rate measure required the use of specialist equipment, only one participant would be tested at a time and a distraction was needed to ensure that the participant’s heart rate would not be artificially elevated by paying attention to the equipment and the activity around it. A cognitive task was created to provide such a distraction, involving two adaptations – one for each story – of the Australian public opinion survey.

Han and Rane (2011), like Philo and Berry (2004), explored the impact of media framing of the Israel–Palestine conflict on public understanding, using a set of ten questions:

To identify respondents’ understanding of the cause and nature of the conflict; perceptions of its global importance; prospects for its resolution; attitudes regarding Australian government involvement and policy preferences; opinions regarding settlements, refugees and Jerusalem; and sources of information. As the Israel–Palestine conflict is subject to multiple narratives, the wording of the questions and the options provided were designed to reflect pro-Israeli, pro-Palestinian and neutral narratives in order to determine the particular narrative with which respondents most closely identified.

The vast majority of the more than 1000 people surveyed by Han and Rane identified the news media as their main source of information on the conflict. Those surveyed online represented Australian citizens and residents, in a sample weighted for gender, age, state of residence and level of education. ‘The overwhelming majority (92 per cent) expressed support for Palestinian refugees’ right of return’, they found. ‘On the issue of Israeli settlements, over three-quarters (78 per cent) expressed opposition’. As Han and Rane argue, these findings ‘are also consistent with Philo and Berry’s (2004) observation of a widespread identification among audiences with ‘common or universal values based on a concern for human suffering as well as opposition to oppressive practices and the excessive use of force’ (Han and Rane, 2011: 66).

Would these opinions change – or perhaps be activated in different ways – compared with the baseline, as a result of watching a news item coded according to either war or PJ evaluative criteria? And would opinions change or be differently activated on the other story selected for this step of the study – also highly familiar from the Australian television news agenda – namely, on government handling of asylum seekers? If so, how? Clearly, participants in this study could not be given the questions on arrival in the room, then again ten minutes later after watching the stories, because they would remember their previous answers, which would risk distorting or even invalidating the experiment. So, participants received their opinion surveys a week before watching the differently framed TV news items and were not accepted into the study unless they had returned their pre-test questionnaires and demographics.

Results

Opinion survey results

First, the opinion survey results showed some strong cognitive effects as demonstrated by changes in participants’ answers to the questions after viewing each story. After watching the Israel–Palestinian peace talks, more WJ viewers thought ‘religion is the main cause of the conflict between Israelis and Palestinians’, compared with their pre-test answers, but fewer PJ viewers endorsed that statement after viewing than before. Similarly, the number of PJ viewers who thought ‘The Israeli–Palestinian conflict is about Israelis fighting for security against Palestinian terrorism’ went down after viewing, but WJ viewers became more likely to endorse the statement after viewing. And in response to a third point, ‘There is no solution to the Israeli–Palestinian conflict’, more WJ viewers endorsed after viewing than had done so beforehand, but there was no change after viewing among PJ viewers.

In conclusion, watching WJ made viewers more likely to believe that the conflict was about religion, that Israelis are fighting for security against Palestinian terrorists and that there is no solution to the conflict. But watching PJ made them less likely to see religion as the cause of the conflict and less likely to attribute it to Israelis fighting for security against Palestinian terrorists than they were before viewing.

There were equally some strong results from watching the asylum story, but not quite as expected. There was a small increase in PJ viewers agreeing that ‘Increasing strife and unrest in the world leading people to leave their homes is the main reason for asylum seeker claims here in Australia’. But many more WJ viewers endorsed it after viewing. In a more expected interaction, many more PJ viewers agreed that ‘Boat people are forced to take desperate measures to protect themselves and their families, and deserve our compassion and support’, compared with their responses to that question before viewing. And WJ viewers became less likely to endorse this statement after watching.

More WJ watchers agreed that ‘Australia’s policy on refugees and asylum seekers should be based more on a need to expand the population, to make a bigger market, and fill space’, compared with their pre-test responses. Fewer PJ viewers agreed after viewing. These results contain an apparent contradiction: the WJ viewers became more likely than before to endorse a statement that frames policy responses to asylum seekers in ‘selfish’ terms – what’s in it for Australia? But the standard political reaction to such calculations, across most of Australia’s political parties, is to devise means to limit the numbers arriving, not expand them as implied by the wording of this question.

ERQ empathy measures

There were strong interactions between the experience of watching news with war or peace framing and the extent of empathy among viewers. Students who saw PJ almost exclusively empathized with asylum seekers, whereas the WJ condition was more mixed. However, there appeared to be a saturation effect with the Israel–Palestine conflict, as several people did not have a preference from that story.

Psychophysiological results

The strategic purpose within the present research of applying the psychophysiological measures outlined above is to show some of the unconscious processes that go into the acts of meaning-making reflected in the results from cognitive and self-reporting emotional responses. As noted above, measuring heart rate may indicate when participants are responding with different portions of the autonomic nervous system: the SNS and PNS. ‘When a visceral organ is innervated by both the SNS and the PNS, the effects are antagonistic. For example, SNS neurons … accelerate the heart [whereas] … PNS neurons … slow the heart’ (Porges et al., 1994: 168).

A quickening of the heart rate may therefore be triggered by a perception of threat or risk, since – as noted above – the SNS is associated with the limbic ‘fright, flight and fight’ syndrome. A slowing of the heart rate may indicate that the PNS has, as it were, ‘gained the upper hand’ in the innervation of the heart, which may be triggered by the provision of soothing or agreeable stimuli. The vagal nerve is the dominant portion of the PNS and Porges et al. (1994) proposed measuring ‘vagal tone’ (VT) as an ‘autonomic correlate of emotion’ (p. 167).

The ‘primary vagal input to the sino-atrial node … regulate[s] atrial rate … and determine[s] heart rate’ (Porges et al., 1994: 170). Porges et al. propose ‘quantify[ing] respiratory sinus arrhythmia (RSA)’ as a way of ‘evaluating the vagal control of the S-A node (i.e. cardiac vagal tone)’.

Our data for HRV indicate RSA because:

heart-rate increase is associated with phases of inspiration, when respiratory mechanisms in the brain stem attenuate the vagal efferent action on the heart. The heart-rate decrease is associated with phases of expiration, when the vagal efferent influence to the heart is reinstated. (Porges et al., 1994: 170)

In the results set out below, therefore, the dip in heart period (increase in heart rate) among WJ viewers as they watched the asylum story (see Graph 1) may be significant, indicating a lesser degree of reinstatement for the vagal efferent influence to the heart. The change in VT (see Graph 2) may also be significant, with WJ viewers’ VT continuing to drop through the experience of watching both stories, whereas PJ viewers’ VT ‘recovered’ when the second story, about the Israel–Palestine conflict, came on. It suggests that both PJ versions contain material that viewers found soothing or reassuring, compared with the WJ versions, which supports the findings of the ERQ and PANAS, summarized above.

OPEN IN VIEWER

Graph 1.

Change in measured heart rate during exposure.

OPEN IN VIEWER

Graph 2.

Change in vagal tone during exposure.

Basic heart rate effects

Both groups had consistently different baseline values (p < .01) – this is most likely due to individual differences. The slowing of the heart over time (i.e. loss of heart period) is relatively consistent in both groups, with no relevant interactions reported overall. This is consistent with the nature of an experimental task in the first place (i.e. sitting still for a prolonged period). This was such that the overall contrast for ISPAL versus BASEL, common between PJ and WJ viewers, was significant (p = .002).

PANAS positive/negative overall

Participants reported significant contrasts overall between (a) both vignettes and (b) both the decrease in positive and increase in negative mood (all ps < .01). This is entirely consistent with the subject matter being unpleasant or confronting. The loss of VT overall during the vignettes, the increase in negative affect and decrease in positive affect are all consistent with the serious/conflictual nature of the stimuli.

ERQ results

ERQ scores were considered as individual difference variables as both raw and subscale scores.

ERQ versus VT

No correlations were significant in the ASYL condition between VT change and any ERQ measure. Overall VT change versus ERQ score was very marginally significant (p = .11) in the ISPAL condition; however, this was entirely the result of the WJ VT change versus ERQ correlation being comfortably significant (r = −.405, p = .018), while PJ VT change versus ERQ was unchanged (r = −.013; see Graph 3).

OPEN IN VIEWER

Graph 3.

Overall VT change versus ERQ score.

When the ERQ scores were split into sub-groups, empathy (r = −.334), distress (r = −.390) and injustice (r = −.349) scores were all at least close to significant. While this may support a central role for self-reported distress as the primary correlate of diminished VT, as the correlation is highest, the distress–VT correlation was not significantly greater than the other subscales (Fisher’s r-to-z transformation; distress > empathy, z = .46, p = .323; distress > injustice, z = −.56, p = .288; p values one-tailed).

Our hypotheses did predict decreased VT in parallel with increased ERQ score, but did not extend as far as the reasons for or structure of this relationship. Assuming that emotional regulation is involved, this perhaps reveals a qualitative difference between ‘strong’ and ‘overwhelming’ feelings. As a consequence, the ERQ scores were re-coded as the number of times participants reported ‘A lot’ or ‘Extremely’ sympathetic/upset/disturbed and so on (no. 5 or 7 on the Likert scale).

The correlation between VT change during viewing the WJ condition and the presence of any strong emotion was significant (r = −.475). The scatterplot shows the linear regression, which shows a steady decrease in VT with the amount of times a strong feeling was reported.

However, it might be argued that instead of a correlation being used, this represents a binary characteristic (i.e. presence vs absence of strong emotion), as many participants did not report any strong emotions (15 of the 32 included J participants) and thus a follow-up t test was conducted (two-tailed, equal variances not assumed), which was also significant (p = .024). Only two participants (out of 17) who reported the presence of a strong emotion increased in VT during the War Journalism (WJ) condition – that is, 15/17 decreased in VT. None of the equivalent overall or sub-group correlations in the Peace Journalism (PJ) conditions was significant.

Interpretation

War framing significantly changes the emotional valence, empathic content or similar of the ISPAL vignette. The result within the ISPAL data is quite clear and is significant in parallel analyses. This fits quite clearly with the overall place of HRV within biobehavioural models – namely, that it can be used as a correlate of emotional responding or regulation.