Setting the collective memory agenda: Examining mainstream media influence on individuals’ perceptions of the past

Abstract

Memory studies scholars tend to stress the significance of the media in shaping collective memories. This study offers a quantitative–empirical examination of this phenomenon. Applying a “memory-setting” research design, inspired by agenda-setting theory, the study examines correlations between the “media memory-agenda” and “public memory-agenda,” to illuminate the influence of the media on the shaping of collective perceptions of the past. Findings point at a significant correlation between media and public memory-agendas, one that increases during periods of heightened coverage of past events. On the individual level, the role of media exposure to commemorative content is significant, surpassing that of direct participation in public commemoration. At the same time, some of the findings point to the resiliency of the public memory-agenda. Therefore, the study’s findings offer a novel understanding of the role of mass media in shaping collective memory, as well as the limits to its influence.

Keywords

agenda-setting collective memory commemoration Israel media memory

As the field of memory studies continues to thrive, memory scholars increasingly point at the significance of the media as offering powerful representations of the past, and as a major source for knowledge and attitudes regarding the past (e.g. Edy, 2006; Kansteiner, 2002). The media are seen as having an “enormous influence” on the shaping of collective recollections (Huyssen, 2000: 29), due to their wide reach and pervasive messages, their adept use of narrative forms (Kitch, 2005), and their mnemonic collaboration with other social agents (Meyers et al., 2009). Yet decades of media effects studies have shown that the influence of the media on its audiences is often less direct and evident than expected. Most of the works that underscore the media’s role do not link public representations of the past with individual perceptions—the individual, in fact, is often absent in collective memory studies (Kansteiner, 2010). This study aims to further probe media’s role as a mnemonic agent, by quantitatively examining its influence on individuals’ perception of the collective past.

Within the field of “media memory” studies—that is, “the systematic exploration of collective pasts, narrated by the media, through the media and about the media” (Neiger et al., 2011: 1)—there is a prevalent tendency to focus on mass mediated representations of memory, such as news coverage or documentaries, while overlooking the memories of individuals. Media texts are often used as a proxy for the collective understanding of the past. Yet, even Halbwachs (1952/1992: 182) claimed that while memory relies on social frameworks, it is individuals who recollect. Thus, the relations between public representations of memory and the memories of individuals remain a key debate in the field (Kansteiner, 2002, 2010; Olick, 1999; Ricoeur, 2004).

The following study contributes to this debate by examining one possible area of media influence: how the past events most salient in the media may influence individuals’ perceptions of the most important events in history. To do so, this study borrows methods from media studies, and applies agenda-setting theory and methodology to the study of collective memory, using a “memory-setting” approach (Kligler-Vilenchik, 2011). The memory-setting research design incorporates two methodologies—a content analysis and a survey—in order to examine the correlation between the past events most salient in mainstream media and those perceived by individuals as most important. In order to enable increased causal interpretation of correlative data, this research design was applied at two time points: at a time of routine, during which the media covered the past marginally; and during the celebration of Israel’s 60th anniversary in May 2008, a time at which discussion of historical events was brought to the foreground of media discourse. Comparing the two points in time enabled us to examine how intensified media mentions of past events may affect individuals’ perceptions in the short term.

This design bridges between media scholarship and memory studies, applying methods and insights gained in media effects research to enrich the understanding of the processes of formation and change of collective memory. Empirically analyzing the associations between media and public memories can help us further establish the media’s role in shaping public recollection.

Collective memory and media memory

When Halbwachs (1952/1992) first introduced the notion of a social framework within which remembering occurs, he claimed that while individuals are those who remember, they obtain their memories in a social setting, and recollect in socio-cultural contexts. Conceptually, we can think of collective memory as consisting of two complementary aspects: on the one hand, the mnemonic signifiers found in our surroundings, such as memorial sites, museums, and media coverage. On the other hand, the term signifies the knowledge and perceptions of individuals in a society regarding its past (Bar-On, 2001; Olick, 1999). The two are seen as interrelated, but how?

For media memory scholars, the media are seen as a powerful tool in shaping perceptions of the past. Edy (1999) claims that the media’s crucial role can be attributed to their ability to simultaneously reach mass audiences, and provide interpretations of the past that are hard to avoid. Moreover, the representations of the past conveyed by the mass media appear in a narrative form (Kitch, 2005), stressing an emotional aspect, rendering them more resonant than structured learning. This is true for both fictionalized representations and news. While intuitively, we would expect news to focus solely on the present, the past is nearly omnipresent in journalism (Meyers, 2007; Zelizer, 2008).

Media representations of the past have often been critically analyzed as cultural artifacts, usually focusing on one specific event, yet they are rarely examined quantitatively. This may be partly due to the difficulty of operationalizing this complex phenomenon. In our quantitative content analysis scheme, we choose to focus on one aspect of media memory: the extent of mentions of past events. This approach enables to compare references to the past both synchronically, across different forms of media, and diachronically, in the same media outlet over time. Using the same categories and coding instructions to examine both the media and individual responses enables us to link the two types of data.

We conducted a content analysis to ask how prevalent the past is in the media, and how this prevalence changes over time. To answer these questions, media content was examined at two points in time. The first is a time of routine (T₁), in which no commemoration was expected according to the national calendar. The second point in time (T₂), set during Israel’s 60th anniversary celebrations, was hypothesized to be one of enhanced media coverage of past events, as the media partake in national commemoration efforts. This includes both ceremonial televised “media events” (Dayan and Katz, 1992) such as the official Independence Day ceremony, broadcast live on all broadcast channels, and coverage of celebratory entertainment content.

While Israel’s 60th anniversary was expected to garner particular media attention toward past events, this tendency is repeated, to a lesser extent, around every Independence Day. The annual reiteration of this period is accompanied by media commemoration of what are deemed as the major national past events. Hence, the 60th anniversary was expected to serve as a particularly salient example of the commemorative role the media takes upon itself every year. The more pertinent question was, how may this increased mediated commemoration shape individuals’ perceptions of the past?

Between individual and collective memories

The controversy over the relations between individual and collective memories (see, for example, Kansteiner, 2010; Olick, 1999; Ricoeur, 2004) may very well have begun with its founder. While mostly known as a student of Durkheim, Halbwachs was also taught by Bergson, and was influenced by his individualistic notions of memory (Coser, 1992). While Halbwachs generally accepted a Durkheiman view of society as the significant source of influence on memory, Bergsonian individualistic influences can still be detected in his work (Ricoeur, 2004; see also Misztal, 2003). Halbwachs saw individuals as recalling memories, though the groups to which the individual belongs provide the contents for her memory, notably through other people with whom the memory is shared.

Despite the extensive use of the term “collective memory,” there is a lack of empirical evidence regarding the extent to which large segments of the public share a common set of past events that are deemed most central. Methodologically and conceptually, analysis of public representations cannot serve to infer that a shared memory among groups of individuals exists (Bourdon and Kligler-Vilenchik, 2011; Schwartz and Schuman, 2000). Studies that do connect personal and collective memories rely mostly on in-depth interviews (e.g. Welzer, 2010) or focus groups (Volkmer, 2006). Yet, if we want to generalize and argue for a collective memory, shared by wide audiences, we need to implement methodologies that will enable us to reach these populations. A series of studies conducted by Howard Schuman and his colleagues represents one attempt at this approach.

In the first study, Schuman and Scott (1989) examined the effect of generational identity on the shaping of collective memory. The authors employed a survey, asking Americans—as an operationalization of the concept of collective memory—which national and global events they perceived as most important. A generational effect was found, as people perceived events from their youth as most important. At the same time, there was a sense of a widely shared memory: 64% of respondents mentioned one of the top three events. This study spurred a series of subsequent investigations, which examined the memories of individuals across different nations, generations, and events (for the Israeli context, see Schuman et al., 2003).

As part of the debate over the individual and collective aspects of memory, this method was met with some critique. For instance, Olick (1999) claims that “collected memory”—the aggregation of disparate individuals’ memories—“misses a great deal of what is going on” (p. 342). He contends that there are collective phenomena that cannot be predicted through the responses of individuals, as well as aspects which social institutions stress, even if individuals try to avoid them. His final point, however, is that individual and collective memory cannot be treated as separate.

The approach embraced by this study attempts a middle ground between Schuman’s approach and Olick’s cultural critique. Like Olick, we see public culture as a key arena for retelling the collective past and shaping the public memory. Like Schuman, we believe that collective memory can also be examined through the aggregation of individuals’ memories of past events. As Schwartz and Schuman (2000) claim, analyzing individuals’ memories and cultural representations of the past do not contradict but complement each other. We propose bridging these traditions not only because we see both as offering valuable ways to understand how societies perceive their past, but also because we see collective representations of memory and individual memories as potentially mutually shaping one another. To examine how this operates, this study borrows from agenda-setting theory, a media effects theory designed to analyze the media’s influence on its audience.

Applying agenda-setting toward the study of collective memory

Originating from the work of McCombs and Shaw (1972), agenda-setting theory posited that the press “may not be successful much of the time in telling people what to think, but it is stunningly successful in telling its readers what to think about” (Cohen, 1963: 13). In its basic version, agenda-setting theory seeks correlations between the public agenda and the media agenda. The media agenda examines the salience of issues mentioned in the media, usually obtained using a quantitative content analysis. The public agenda refers to the issues perceived as important by the public at a certain point in time, usually measured through surveys. The agenda-setting effect of the media on the public agenda “seems to hold under a variety of conditions, for a diversity of issues, and when explored with diverse research methods” (Dearing and Rogers, 1996: 92).

To apply agenda-setting theory to the study of collective memory, we examine the correlation between the media memory-agenda: the past events salient in the media, and the public memory-agenda: past events deemed as most important by the public. This process, which we call “memory-setting,” is based on several conceptual and epistemological assumptions which agenda-setting and memory studies share (Kligler-Vilenchik, 2011). Clearly, the richness of collective memory cannot be fully captured by mentions of past events alone. Yet referencing events cues wider narratives about the collective past, or, as Zelizer (1995) vividly argues, “events give memory a platter on which to serve historical accounting” (p. 231).

In fact, both agenda-setting and “memory-setting” can be thought of as functioning on two levels. In agenda-setting research, the distinction is made between “first-level agenda-setting,” which asks which issues are the most salient on the media agenda, and “second-level agenda-setting,” which analyzes the salience of attributes through which these issues are discussed (see, for example, Kiousis et al., 1999). Similarly, the media can be conceptualized as shaping public memory in two interrelated ways: first, by highlighting or marginalizing what is remembered, and, second, by shaping the nature of these memories. Much memory studies work attends to the second aspect, discussing the ways in which the media shape differing narratives of past events (e.g. Zelizer, 1992). These rich narrative aspects cannot be captured in our content analysis. However, the first stage of setting the memory-agenda—determining which events are salient in public memory—is in itself a key source of power and influence. This is particularly true when some past events have acquired insistent political meanings over time (Kligler-Vilenchik, 2011; Misztal, 2003).

Relying on the general concept of agenda-setting theory, if we assume that the media influence individuals’ memories, this would manifest in a correlation between media and public memory-agendas. Therefore,

H1: The media memory-agenda (a rank-order of the past events mentioned in the media) and public memory-agenda (a rank-order of the past events mentioned as important by individuals) will be significantly correlated.

Due to the expectation that the extent of media memory will increase in T₂, as the media partake in commemorating Israel’s 60th anniversary, it is assumed that the media’s influence on individuals’ memories will increase as well, resulting in stronger associations between the two:

H2: The correlation between the media memory-agenda and the public memory-agenda will increase in T₂, in comparison to T₁.

By randomly assigning the survey respondents to one of the points in time, T₁ or T₂, a quasi-experimental setting is enabled. The increased media coverage of past events expected between T₁ and T₂ can be seen as a natural manipulation, enabling to gauge the influence of increased media memory on collective memory perceptions among individuals. The aim of this research design is to increase our ability to infer causal relationships when using correlational data.

Despite this methodological addition, such a correlational analysis does not rule out the possibility that factors other than media are responsible for the effect. A second stage of analysis in this study includes individual-level regressions conducted on the survey data. These enable us to control for the effect of variables such as direct participation in commemorative rituals, while pointing to the specific role of the media in predicting the naming of certain past events as most important:

RQ1: Which individual-level variables best predict naming the past event most salient in the media?

Method

Content analysis

The goal of the content analysis was to quantify mentions of past events in the media, as an operationalization of media memory. In each of the study’s time points, content analysis was held for 7 days, prior to the administration of the survey. The corpus included television (Israeli channels 1 and 2, Israel’s leading public and commercial broadcast channels) and daily newspapers (Yediot A’haronot, the most popular daily newspaper at the time, and Ha’aretz, considered Israel’s leading elite newspaper).¹ Television content consisted of prime-time broadcasts (8:00–11:00 p.m.), including commercials. Newspaper content included the main news-section, and all supplemental sections, including all advertisements.

Coders identified mentions of past events according to a coding scheme list. This list consisted of past events mentioned as important in the survey, as well as additional past events that were salient in the media during preliminary coding. The final list included 19 events,² listed in Table 1. The unit of analysis was defined as a single mention of a past event (n = 4001 mentions), which could be explicit: the name of the event (including multiple possible names for some events) or an implicit mention of that event. For example, the statement “Israel celebrates its 60th anniversary” was considered as one (implicit) mention of the establishment of the State of Israel. An event could be mentioned multiple times in the same newspaper item or television program, and each mention was counted separately. The entire coding corpus consisted of 84 hours of television broadcasts (1267 mentions), and 22 newspaper issues (2734 mentions of past events). The content was coded by two coders. Inter-coder reliability was tested on 20% of the corpus. Nominal Krippendorff alphas were at over 0.89 for unitization (number of mentions per article or per 5 minutes of broadcast), and over 0.97 for the event mentioned. The media memory-agenda was defined as the rank-order of the aggregate frequency of past events mentioned in the media, according to the content analysis.

Table 1.

Frequency and percentages of past events mentioned in the survey and content analysis by time point.

Event	Survey		Newspaper coverage		TV coverage
	T₁ n (%)	T₂ n (%)	T₁ n (%)	T₂ n (%)	T₁ n (%)	T₂ n (%)
Establishment of the State of Israel	123 (27.7%)	125 (28%)	61 (14.7%)	594 (68.7%)	54 (24%)	813 (78%)
World War II	76 (17.1%)	74 (16.6%)	78 (18.8%)	83 (3.6%)	13 (5.8%)	22 (2.1%)
9/11	50 (11.3%)	40 (9%)	4 (3%)	3 (0.1%)	0 (0%)	0 (0%)
Holocaust	40 (9%)	49 (11%)	145 (34.9%)	239 (10.3%)	41 (18.2%)	45 (4.3%)
Six-Day War	13 (2.9%)	29 (6.5%)	12 (2.9%)	53 (2.3%)	14 (6.2%)	32 (3.1%)
Collapse of USSR	20 (4.5%)	18 (4%)	3 (0.7%)	5 (0.2%)	4 (1.8%)	0 (0%)
Second Lebanon War	17 (3.9%)	20 (4.5%)	45 (10.8%)	109 (4.7%)	32 (14.2%)	58 (5.6%)
Rabin’s Assassination	15 (3.4%)	21 (4.7%)	10 (2.4%)	49 (2.1%)	13 (5.8%)	3 (0.3%)
Inventions	15 (3.4%)	17 (3.8%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
Moon Landing	15 (3.4%)	12 (2.7%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
Peace with Egypt	12 (2.7%)	11 (2.5%)	4 (1%)	7 (0.3%)	0 (0%)	9 (0.9%)
Gaza Disengagement	12 (2.7%)	9 (2%)	5 (1.2%)	56 (2.4%)	36 (16%)	5 (0.5%)
World War I	8 (1.8%)	5 (1.1%)	2 (0.5%)	8 (0.3%)	0 (0%)	0 (0%)
Tsunami	6 (1.4%)	7 (1.6%)	3 (0.7%)	7 (0.3%)	0 (0%)	5 (0.5%)
Gulf War	9 (2%)	3 (0.7%)	11 (2.7%)	9 (0.4%)	3 (1.3%)	3 (0.3%)
Yom Kippur War	4 (0.9%)	3 (0.7%)	11 (2.7%)	63 (2.7%)	5 (2.2%)	23 (2.2%)
First Lebanon War	3 (0.7%)	3 (0.7%)	7 (1.7%)	13 (0.6%)	1 (0.4%)	8 (0.8%)
Operation Entebbe	4 (0.9%)	0 (0%)	11 (2.7%)	18 (0.8%)	8 (3.6%)	16 (1.5%)
Lebanon Withdrawal	2 (0.5%)	1 (0.2%)	3 (0.7%)	3 (0.1%)	1 (0.4%)	0 (0%)
Total	444 (100%)	447 (100%)	415 (100%)	2319 (100%)	225 (100%)	1,042 (100%)

USSR: Union of Soviet Socialist Republics.

Survey responses are for mentioning an event as one of the two most important events. The total of 100% for survey mentions is for responses that correspond to categories in the content analysis. Responses with no parallels in the content analysis (e.g. events from personal life) were omitted.

Survey

The aim of the survey was to establish the public memory-agenda: the aggregate frequency of past events mentioned by the public. The survey further elicited amount and content of media exposure over the past week, as well as demographic variables. The phone survey was administered to a representative sample of the adult Jewish Hebrew-speaking population³ by the University of Haifa’s Computerized Survey Center, beginning on the first Sunday following the week of the content analysis. The first wave, 3–6 March 2008, consisted of 251 respondents; and the second wave, 11–15 May 2008, consisted of 266 respondents.⁴ Of the 517 respondents, 46.8% were male and 53.2% female, with a mean age of 47 years. Israeli-born respondents made 66.7% of the sample, while 33.3% were born elsewhere. There were no significant differences on demographic variables between respondents at T₁ and T₂.

Measures

Most important past event

This variable served to establish the public memory-agenda. Building on the work of Schuman et al. (e.g. Schuman and Scott, 1989), the first question of the survey asked respondents: “Of all the events and changes that occurred in Israel and the world over the past 100 years, please mention the two events that seem to you as most important.” Respondents could mention any event (in fact, some mentioned purely personal events), and their responses were coded into the same list of events used for the media coding. Inter-coder reliability for 20% of the responses elicited almost full agreement (Krippendorff’s alpha = .99).

Naming the event most salient in the media served as the dependent variable, only for the second stage of the analysis (the logistic regressions). Based on the first survey question (the most important past event), this measure was coded as a binary, where (1) meant naming as the most important past event the same event that was overwhelmingly most salient in the media: the establishment of the State of Israel, and (0) meant choosing any other event.⁵

Amount of television exposure

This variable was determined by averaging two questions: “In the past week, how much television did you watch on an average weekday,” and “In the past weekend, how much television did you watch on average per day?” The questions refer to the past week, as this week was examined in the content analysis.⁶ Answer categories included, less than an hour, or none (1); between 1 and 2 hours (2); over 2 hours but less than 4 hours (3); and 4 hours or more (4) (M = 2.04, standard deviation (SD) = 0.91).

Amount of newspaper reading

Respondents were asked, “In the past week, how many times did you read newspapers?” If asked, respondents were told to include reading online news as well. Answer categories included, not at all (1); about once a week (2); two–three times a week (3); and four times a week or more (4) (M = 3.05, SD = 1.16).

Amount of television exposure related to Israel’s 60th anniversary

In this question, asked only in T₂, respondents were asked to estimate: “Out of the television contents you watched this week, how much was related to Israel’s 60th anniversary? For example, Independence Day celebrations, programs and documentaries about events in Israeli history, and celebratory entertainment shows?” Possible answer categories ranged from “none of the contents were related to Israel’s 60th anniversary” (1) up to “all of the contents were related to Israel’s 60th anniversary” (5) (M = 2.96, SD = 1.19).

Amount of newspaper reading related to Israel’s 60th anniversary

This question, appearing only in T₂, asked respondents, “Out of the contents you read this week on the newspaper, how much was related to Israel’s 60th anniversary? For example, content about Independence Day celebrations, non-fiction accounts of events in Israeli history.” The same categories were used as for the previous question (M = 2.76, SD = 1.00).

Participation in ceremonies commemorating Israel’s Fallen Soldiers

Respondents were asked, “This year, did you participate in ceremonies commemorating Israel’s Fallen Soldiers, conducted at a public venue or a school?” Answer categories were, yes (48.1%); I did not participate but watched a ceremony on television (29.3%); and no (22.2%). In the regression, since the variable of interest was direct participation, the variable was recoded with participation coded as 1, and no participation or watching on TV coded as 0.

Participation in public celebrations of Israel’s Independence Day

Respondents were asked, “Did you participate in public celebrations of Independence Day this year (e.g. attend a public ceremony, celebrations at a public venue, etc.)?” Of the respondents, 44% participated in public celebrations, 28.6% watched celebrations on television, and 27.4% did not participate. The variable was recoded (participation = 1; no participation or watching on TV = 0).

Demographic variables

Variables included were gender (male = 0, female = 1), age (in years), country of birth (Israel = 0, elsewhere = 1), and years of schooling (M = 14.5, SD = 3.38). Religiosity, income, and political views were omitted as they were not found significant predictors of the dependent variable, and their inclusion did not affect the pattern of results.

Results

Preliminary results: the media and public memory-agendas at T₁ and T₂

Table 1 presents frequencies and relative frequencies for the events mentioned by survey participants as one of their two most important past events at a routine point in time (T₁) and at the week of Israel’s 60th anniversary (T₂), as well as the frequency of mentions these events received in the newspaper and television coverage at these times, according to the content analysis. The table lists all events mentioned in the media; events mentioned by survey respondents that did not have a parallel in news coverage (e.g. personal events) were not included in the table. Relative frequencies were calculated for the events in the table only.

In the survey, the two events most frequently mentioned at both time points were the establishment of the State of Israel and World War II. One of the most interesting findings of the study emerges from these initial descriptive results: the consistency between the public agendas at the two time points is striking. Despite the fact that these are two different samples, and despite the differential media coverage and commemorative efforts during these points in time, there is a strong correlation between the public agenda at T₁ and T₂ (Pearson’s r = .983, p < .001).⁷ This points to the resiliency of collective memory, and the extent to which, at least among a Jewish-Israeli population, there is high agreement on the set of most important historic events. On the surface level, not detecting a higher percentage of survey respondents mentioning the establishment of the State of Israel given the greater media coverage of it at T₂ points at a lack of a media effect. As our further analyses will show, the media’s influence turns out to be more nuanced.

Turning to mainstream media coverage, we find a more diverse picture. In the newspapers, the events most frequently mentioned at T₁ were the Holocaust, World War II, and the establishment of the State of Israel. At T₂, the establishment of the State of Israel was the most salient event, followed by the Holocaust and the Second Lebanon War. In television coverage, the most frequently mentioned event at T₁ was the establishment of the State of Israel, followed by the Holocaust and the Disengagement from Gaza. At T₂, the establishment of the State of Israel was the most salient event, followed by the Second Lebanon War and the Holocaust.

Due to the commemorative role that the media undertakes (Neiger et al., 2011), we expected that the extent of media memory—the frequency of mentions of past events in the media—would increase from T₁, a time of routine coverage, to T₂, the week of Israel’s 60th anniversary. This should come as no surprise, yet this increase in the mention of past events needs to be noted, as it serves as the “natural manipulation” for our quasi-experiment.

In prime-time television broadcasts, past events were mentioned 225 times at T₁ and 1042 times at T₂, a statistically significant increase of 363% (χ²(1) = 527, p < .001). In newspaper coverage, mentions of past events increased from 415 at T₁ to 2319 at T₂, a statistically significant increase of 458% (χ²(1) = 1326, p < .001). Thus, as expected, during the week of Israel’s 60th anniversary, past events were much more salient on the media’s agenda, creating a natural condition of extensive media memory.

The memory-setting function of the media

The first hypothesis of the study posited that, as would be expected according to agenda-setting theory, a significant correlation will be found between the media memory-agenda: the rank-order of the aggregate frequency of past events mentioned in the media, and the public memory-agenda: the rank-order of the aggregate frequency of past events mentioned most often by the public. To test this, a correlation was conducted on the full list of 19 events mentioned (see Table 1). This hypothesis was supported. A strong correlation was found between the public memory-agenda and the television memory-agenda (Pearson’s r = .816, p < .001), as well as the public memory-agenda and the newspaper memory-agenda (Pearson’s r = .854, p < .001). These correlations present a first step in pointing to the media’s possible memory-setting effect.⁸

As previously discussed, however, analyzing correlations between the public and media memory-agendas at one point in time is not sufficient in order to infer a causal media influence. For this reason, the correlations were analyzed separately for T₁ and T₂. The increased volume of media memory at T₂, due to the celebration of Israel’s 60th anniversary, thus served as a natural manipulation. In this quasi-experimental design, an increased correlation between the media and public memory-agendas at the second time point could be interpreted as the result of increased presence of media memory.

As Table 2 shows, at T₁, the public memory-agenda and the media memory-agenda were moderately correlated, both in terms of the correlation between the public and television memory-agendas (Pearson’s r = .61, p = .005), and the correlation between the public and newspaper memory-agendas (Pearson’s r = .56, p = .014). At T₂, correlations between the public and media memory-agendas increased in magnitude to a strong correlation. This is true both for the correlation between the public and television agendas (Pearson’s r = .81, p < .001), and the correlation between the public and newspaper agendas (Pearson’s r = .84, p < .001). Based on a confidence interval around the difference between Z scores for both correlations, there is a significant increase in the magnitude of all three correlations between T₁ and T₂ (for α = .05). The correlation between the media and public memory-agendas was significantly stronger at T₂, at a time of increased media coverage, confirming our second hypothesis.

Table 2.

Correlations between the media memory-agenda and the public memory-agenda at T₁ and T₂.

	T1	T2	Confidence intervals for the differences
Television	r = .61, p = .005	r = .81, p < .001	0.24 ≤ p₁ – p₂ ≤ 0.53
Newspaper	r = .56, p = .014	r = .84, p < .001	0.37 ≤ p₁ – p₂ ≤ 0.63

The increased correlation between the media and public memory-agendas at T₂ may come as a surprise, given the similarity between the public agendas at T₁ and T₂. But which event mentions make for this tighter alignment at T₂? As Table 3 shows, through the shift from T₁ to T₂, the public converged around a number of “major” Israeli events: ones that were already at the forefront of public attention (the 6 most salient Israeli events, mentioned by over 10 survey respondents at both time points).⁹ At the same time, from T₁ to T₂ there was a decrease in the public’s mentions of the 6 “minor” Israeli events (e.g. Operation Entebbe) and of International events. This shift—an increase in the number of mentions of major Israeli events, and decrease in minor Israeli events and international events—is significant for the survey mentions (χ²(2) = 7.61, p = .022) as well as in the media (for newspaper coverage, χ²(2) = 172.92, p < .001; for television, χ²(2) = 100.33, p < .001).

Table 3.

Changes in categories of event mentions in the survey from T₁ to T₂.

Type of event	“Major” Israeli	“Minor” Israeli	International	Total
T₁	220	34	190	444
T₂	255	19	173	447
Total	475	53	363	891
Changes in categories of event mentions in newspaper coverage from T₁ to T₂
T1	277	48	90	415
T2	2051	162	106	2319
Total	2328	210	196	2734
Changes in categories of event mentions in television coverage from T₁ to T₂
T₁	154	54	17	225
T₂	960	55	27	1042
Total	1114	109	44	1267

Individual-level logistic regressions

In the second stage of our analysis, the research question asking which factors lead to agreement with the media memory-agenda was answered using a stepwise binary logistic regression, in which the dependent variable was naming the event most salient in the media—the establishment of the State of Israel—as one of the two most important past events.¹⁰ This analysis is presented only for T₂, in order to evaluate variables related to the commemoration of Israel’s 60th anniversary, which were present only at the second point in time.¹¹ The results of the regression are shown in Table 4.

Table 4.

Factors affecting the propensity to name the past event most salient on the media memory-agenda (logistic regressions) (N = 174).

	Model 1		Model 2		Model 3
	β (SE)	exp(β)	β (SE)	exp(β)	β (SE)	exp(β)
Age	.027* (.011)	1.027	.025* (.012)	1.026	.028* (.012)	1.028
Gender (female = 1)	−.478 (.320)	0.620	−.371 (.333)	0.690	−.371 (.333)	0.690
Country of birth (else = 1)	−.063 (.350)	0.939	.035 (.366)	1.036	.027 (.371)	1.028
Education	.147** (.057)	1.158	.163** (.062)	1.177	.160** (.062)	1.173
Amount of TV exposure			.025 (.199)	1.025	.030 (.200)	1.031
Amount of newspaper exposure			.128 (.246)	1.136	.156 (.251)	1.169
TV content related to 60th anniversary			.425** (.166)	1.529	.422** (.166)	1.525
Newspaper content related to 60th anniversary			−.309 (.194)	0.734	−.311 (.194)	.732
Participation in celebration of 60th anniversary					.236 (.351)	1.266
Participation in soldiers’ memorial ceremony					.095 (.343)	1.100
Intercept	−3.314** (1.040)	0.036	−4.515** (1.644)	0.011	−4.845** (1.709)	0.008
Likelihood ratio (χ²)	14.541		22.024		22.641
df	4		8		10
Pseudo R²	10.7		15.9		16.3

SE: standard error; df: degrees of freedom.

Estimates are standardized and unstandardized logic coefficients with standard errors in parentheses.

p < .05; **p < .01.

The first set of variables in the models is composed of demographic variables. Age and education were found to be significant factors affecting the propensity to agree with the media memory-agenda. Controlling for all other variables in the model, each year increase in respondents’ education increased the odds of naming the establishment of the State of Israel as the most important event by 16% (for p < .01). Ceteris paribus, each year increase in respondents’ age increased the odds of agreeing with the media memory-agenda by 2.8% (p = .01). Demographics alone were found to account for 10.7% of the variance in the dependent variable.

The second set of variables were media exposure variables, added in model 2. The variables relating to the amount of media exposure were found to be nonsignificant, showing that extent of any exposure to the media did not significantly affect the propensity to agree with the media memory-agenda. The type of television content, however, was found to be a significant contributor. On a five-point scale, respondents estimated how much of the content they watched on television the previous week was related to the celebration of Israel’s 60th anniversary. Ceteris paribus, each point increase on this scale increased the odds of mentioning the establishment of the State of Israel as the most important event (for p < .01) by 52.5%.¹² Newspaper content, however, was not found significant. Media variables were found to increase the explained variance of the base-line model by 5.2%.

The third set of variables related to direct participation in national commemorations (i.e. in Memorial Day ceremonies or in public events celebrating Israel’s 60th anniversary). Such events, examples of state-sponsored commemoration, are often hypothesized to serve as commemorative agents alternative to the media, and much memory scholarship is focused on such efforts (see, for example, Rosenberg et al., 2010). In model 3, however, the two variables were found nonsignificant in affecting the propensity to name Israel’s establishment as the most important event, when controlling for media exposure. Media influence thus surpassed the role of direct participation in public commemoration events.¹³

Discussion

The goal of this study was to empirically examine the influence of the media on collective memory through an examination of media coverage of Israel’s 60th anniversary and individuals’ perceptions of the most important past events. One of the most interesting findings is that, in contrast to the logic of our argument, despite the documented dramatic increase in the attention paid by Israeli mainstream media to Israel’s establishment toward Independence Day, public perceptions of the importance of that specific event remained unchanged. This might point to the resistance of collective memory to even the strongest efforts to modify it, confirming Schudson’s (1992) arguments about the “resistance of the past” (p. 205).

But does this mean that the media do not have any influence on the public’s perceptions of important past events? Several of our findings suggest that this is not the case. Consistent with our argument regarding a memory-setting effect of the media, our results documented a significant correlation between the media memory-agenda and the public memory-agenda, which was significantly stronger at T₂, a time of increased media memory. The stronger correlation between the media and the public in T₂ stems from the fact that at that time point, people were more likely to mention past events in Israeli history that were mentioned by the media as important. The coverage of Israel’s anniversary did not focus merely on the State’s establishment but rather on its history beyond its creation. Myriad Israeli past events moved to the forefront of media coverage during the week of Independence Day, parallel with an increase in the mentioning of major Israeli events among the public, “at the expense” of international events and minor Israeli events. Finally, the individual-level regressions showed that heavier consumers of televised media memory were the most likely to name the establishment of the State of Israel, and major Israeli events in general, as the most important events. It is worth noting that both these findings do not represent very strong effects. Yet, the small magnitude of this effect should be interpreted in light of the fact that media effects are generally not very powerful (McGuire, 1986).

While these findings establish a key step in supporting the claim of media’s influence on collective memory, they do not fully establish that the causal mechanism indeed works from media exposure to audience perceptions. Alternative explanations exist. First, we could be witnessing self-selection: it may be that the same people who already cared about the 60th anniversary were more likely to follow media coverage of the anniversary. Second, reverse causality is possible: we could conclude that it is the media agenda that has changed in the direction of the public agenda at T₂ rather than vice versa.

As we address these alternative explanations, we acknowledge that only a true experimental design (i.e. in a laboratory setting) could fully refute them. However, in the field of memory studies, we believe that context is key: we are interested in processes of memory formation based on real-life circumstances of “messy” commemoration, in which several mnemonic agents collide. We thus preferred the naturally occurring commemoration over the control that a laboratory setting would have enabled us. In future studies, examining media memory in controlled settings may help substantiate the causal claims of studies such as this, which err on the side of external validity.

The reverse causality argument can be addressed in several ways. First, our quasi-experimental design helps substantiate our causal interpretation. The correlations between the media and public memory-agendas increased in T₂, after Israel’s 60th anniversary received a heavy dose of media attention. Thus, when the salience of media memory increased, the media’s ability to set the public memory-agenda was stronger as well. Second, the causal direction of influence from the media to the public was empirically substantiated in agenda-setting research (using experiments, longitudinal designs, cross-lagged analyses, and other designs; see Dearing and Rogers, 1996), upon which we rest our theoretical argument.

Finally, the increased correlation between the public and the media memory-agendas during the “high holidays” of Israel’s national calendar might not result from the influence of one variable on the other, but rather from the enduring influence of Jewish-Israeli civil religion (Liebman and Don-Yehiya, 1983), especially during this specific time of the year. Durkheim may have seen this as the effect of a patriotic current shaping all of society (Wallace, 1977). In terms of sheer prominence, our study might suggest that mainstream Zionist ideology is still potent as an agenda setter for Israeli-Jewish society, influencing both the media and the public. At the same time, only complementary in-depth textual analysis of media contents and public interpretations can determine the narrative shaping of media memory, and whether shifts have occurred in how these frequently mentioned historical events have been interpreted by different groups throughout the years (Meyers, 2002).

The media’s role as a mnemonic agent, as limited as it emerged from our study, raises the question of the potential political consequences of the ability to shape public remembering. After all, “collective memory is always a mean for something else” (Zelizer, 1995: 226), often for legitimizing current regimes and social conditions. As seen in the increased extent of media memory as part of Israel’s 60th anniversary, in this case, the media assume a key role in disseminating state-sponsored commemoration, which has legitimizing functions. Other studies have shown other possible political effects of collective memory: the past events most salient to individuals have been found to influence the historical analogies they use to make sense of current events, thus shaping their political evaluation of these events (Schuman and Corning, 2012). It seems that indeed, “control of a society’s memory largely conditions the hierarchy of power” (Connerton, 1989: 1). Future studies can further investigate the ways in which this elusive form of power operates—for this purpose, qualitative approaches such as in-depth interviews or focus groups would be most fruitful.

Finally, this study’s findings contribute to the controversy over the interrelations between individual and collective memories. On the one hand, the study suggests that a “collective memory” can be said to exist, at least among relatively homogenous groups: Jewish-Israeli Hebrew-speaking adults generally agreed on what the most important past events are. Moreover, this sense of collective memory was arguably influenced by the media, corroborating the significance of analyzing public manifestations of collective memory. Yet individuals are far from powerless: some events were important to the public though they were not salient in the media, nor intentionally commemorated by the state (e.g. the 9/11 attacks were very salient for the public, while almost absent from mainstream media discourse). In his conception of collective memory, Halbwachs (1952/1992) gave individuals relatively limited agency in constructing their own understanding of the shared past. This study, while corroborating Halbwachs’ thesis on social frameworks of memories, points to the fact that individual memories cannot be ignored when studying collective memory. While public commemoration is often reflected by individuals, the unique aspects of individual memory merit the full attention of memory scholars.

Footnotes

Acknowledgements

The authors thank Motti Neiger, Jonathan Cohen, and Maxwell McCombs, as well as the two anonymous reviewers, on their insightful comments on previous versions of this article.

Notes

Author biographies

Neta Kligler-Vilenchik is a doctoral candidate at the Annenberg School for Communication and Journalism at the University of Southern California. Her current research focuses on changing modes of political engagement in the new media context.

Yariv Tsfati (Ph.D., University of Pennsylvania, 2002) is Associate Professor at the Department of Communication, University of Haifa, Israel. His research focuses on various facets of public opinion, in particular on trust in media, the third person effect, and campaign effects.

Oren Meyers (Ph.D., University of Pennsylvania, 2004) is Senior Lecturer at the Department of Communication, University of Haifa, Israel. His research interests focus on journalistic practices and values, collective memory and popular culture.

References

Bar-On

(2001) Smoking Borders: Studies in the Early History of the State of Israel 1948–1967. Jerusalem, Israel: Yad Ben-Zvi (Hebrew).

Bourdon

Kligler-Vilenchik

(2011) Together, nevertheless: television memories in mainstream Jewish Israel. European Journal of Communication 26(1): 33–47.

Cohen

(1963) The Press and Foreign Policy. Princeton, NJ: Princeton University Press.

Connerton

(1989) How Societies Remember. Cambridge: Cambridge University Press.

Coser

(1992) Introduction: Maurice Halbwachs 1877–1945. In: Halbwachs

(ed.) On Collective Memory. Chicago, IL: The University of Chicago Press, pp. 1–36.

Dayan

Katz

(1992) Media Events: The Live Broadcasting of History. Cambridge, MA: Harvard University Press.

Dearing

Rogers

(1996) Communication Concepts 6: Agenda Setting. Thousand Oaks, CA: SAGE.

Edy

(1999) Journalistic uses of collective memory. Journal of Communication 49(2): 71–85.

Edy

(2006) Troubled Pasts: News and the Collective Memory of Social Unrest. Philadelphia, PA: Temple University Press.

10.

Halbwachs

(1952/1992) On Collective Memory. Chicago, IL: The University of Chicago Press.

11.

Huyssen

(2000) Present pasts: media, politics, amnesia. Public Culture 12(1): 21–38.

12.

Kansteiner

(2002) Finding meaning in memory: a methodological critique of collective memory studies. History and Theory 41: 179–197.

13.

Kansteiner

(2010) Memory, media and Menschen: where is the individual in collective memory studies? Memory Studies 3(1): 3–4.

14.

Kiousis

Bantimaroudis

Ban

(1999) Candidate image attributes: experiments on the substantive dimension of second level agenda setting. Communication Research 26(4): 414–428.

15.

Kitch

(2005) Pages from the Past: History & Memory in American Magazines. Chapel Hill, NC: The University of North Carolina Press.

16.

Kligler-Vilenchik

(2011) Memory-setting: applying agenda-setting theory to the study of collective memory. In: Neiger

Meyers

Zandberg

(eds) On Media Memory. New York: Palgrave Macmillan, pp. 226–237.

17.

Liebman

Don-Yehiya

(1983) Civil Religion in Israel: Traditional Judaism and Political Culture in the Jewish State. Berkeley, CA: University of California Press.

18.

McCombs

Shaw

(1972) The agenda-setting function of mass media. Public Opinion Quarterly 36: 176–187.

19.

McGuire

(1986) The myth of massive media impact: savagings and salvagings. Public Communication and Behavior 1: 175–257.

20.

Meyers

(2002) Still photographs, dynamic memories: A study of the visual presentation of Israel’s past in commemorative newspaper supplements. The Communication Review 5(3): 179–205.

21.

Meyers

(2007) Memory in journalism and the memory of journalism: Israeli journalists and the constructed legacy of Haolam Hazeh. Journal of Communication 57(4): 719–738.

22.

Meyers

Zandberg

Neiger

(2009) Prime time commemoration: an analysis of television broadcasts on Israel’s Memorial Day for the Holocaust and the Heroism. Journal of Communication 59(3): 456–480.

23.

Misztal

(2003) Durkheim on collective memory. Journal of Classical Sociology 3(2): 123–143.

24.

Neiger

Meyers

Zandberg

(eds) (2011) On Media Memory: Collective Memory in a New Media Age. London: Palgrave Macmillan.

25.

Olick

(1999) Collective memory: the two cultures. Sociological Theory 17(3): 333–348.

26.

Ricoeur

(2004) Memory, History, Forgetting. Chicago, IL: The University of Chicago Press.

27.

Rosenberg

Dean

Granzow

(2010) Centennial hauntings: reckoning with the 2005 celebration of Alberta’s history. Memory Studies 3(4): 395–412.

28.

Schudson

(1992) Watergate in American Memory: How We Remember, Forget, and Reconstruct the Past. New York: Basic Books.

29.

Schuman

Corning

(2012) Generational memory and the critical period: evidence for national and world events. Public Opinion Quarterly 76(1): 1–31.

30.

Schuman

Scott

(1989) Generations and collective memories. American Sociological Review 54(3): 359–381.

31.

Schuman

Vinitzky-Seroussi

Vinokur

(2003) Keeping the past alive: memories of Israeli Jews at the turn of the millennium. Sociological Forum 18(1): 103–135.

32.

Schwartz

Schuman

(2000) The Two Meanings of Collective Memory (Newsletter of the Sociology of Culture). Washington, DC: American Sociological Association, pp. 1–3.

33.

Volkmer

(ed.) (2006) News in Public Memory: An International Study of Media Memories across Generations. New York: Peter Lang.

34.

Wallace

(1977) Emile Durkheim and the civil religion concept. Review of Religious Research 18(3): 287–290.

35.

Welzer

(2010) Re-narrations: how pasts change in conversational remembering. Memory Studies 3(1): 5–17.

36.

Zelizer

(1992) Covering the Body: The Kennedy Assassination, the Media, and the Shaping of Collective Memory. Chicago, IL: The University of Chicago Press.

37.

Zelizer

(1995) Reading the past against the grain: the shape of memory studies. Critical Studies in Mass Communication 12(2): 214–239.

38.

Zelizer

(2008) Why memory’s work on journalism does not reflect journalism’s work on memory. Memory Studies 1(1): 79–87.

Setting the collective memory agenda: Examining mainstream media influence on individuals’ perceptions of the past

Abstract

Keywords

Collective memory and media memory

Between individual and collective memories

Applying agenda-setting toward the study of collective memory

Method

Content analysis

Survey

Measures

Most important past event

Amount of television exposure

Amount of newspaper reading

Amount of television exposure related to Israel’s 60th anniversary

Amount of newspaper reading related to Israel’s 60th anniversary

Participation in ceremonies commemorating Israel’s Fallen Soldiers

Participation in public celebrations of Israel’s Independence Day

Demographic variables

Results

Preliminary results: the media and public memory-agendas at T1 and T2

The memory-setting function of the media

Individual-level logistic regressions

Discussion

Footnotes

Acknowledgements

Notes

Author biographies

References

Preliminary results: the media and public memory-agendas at T₁ and T₂