Stages and Information Sources in the Formation of Historical Consciousness

The study inquires into the process of formation of historical consciousness among the population of the Czech Republic. The process is examined in terms of biographical stages and information sources. First, the study draws upon the concept of sensitive period developed by Schuman and Corning. Second, it notes the increasingly mediated nature of socialization to historical consciousness. Data were gathered through a representative survey of the Czech popu­ lation, with events of February 1948 and August 1968 serving as historical reference points of his­ torical consciousness. Analysis identifies a formative pattern consisting of three stages. The study also identifies typical information sources corresponding to each formative stage. Further analysis employed regression analysis to assess the influence of basic sociodemographic characteristics on the stage in which individuals are socialized.


Introduction
The practices of memorialization and remembrance have undergone a series of deep transformations. As Paul Connerton [1989] demonstrated, the maintenance of social memory has traditionally been achieved through habitualized body gestures often used when performing rituals. However, modern societies produce conditions that rather facil itate forgetting [Connerton 2009]. The temporality of media is among the key elements constituting the conditions suitable for forgetting. Modern media gradually increase the volume of content and the speed with which it is processed by individuals. Moreover, the tangible substrate of media was transformed or entirely removed with the onset of digital media, making records ever more ephemeral [Connerton 2009: 79-88, 124].
The relationship between communication media and shared images of the past has also been emphasized by Jack Goody, who provides evidence on how written communication was crucial for the creation of "history" as a system of records of past events [Goody 1987: 132]. In following this course, knowledge of the past has been gradually decontextual ized to a point when authority is ascribed rather to mediated communication (through books, for example) rather than to personal facetoface communication [Goody 1987: 163]. As indicated by the growing body of research on the mnemonic aspects of various media [Edgerton 2001;Hoskins 2009bHoskins , 2009aLandsberg 2004;MayerSchönberger 2009], the direct transfer of knowledge from generation to generation has been disrupted by various other means of information dissemination including schooling, broadcasting, or digital networks.
Currently, the information sources constituent of historical consciousness in the Czech Republic have been explored by Jiří Šubrt and his team [Maslowski -Šubrt 2014;Šubrt 2010;Šubrt -Vinopal -Vávra 2013;Vávra 2013]. 1 This research endeavor represents a stepping stone for a more focused research into the matter at hand. However, its findings paint a very broad and static picture of production and reproduction of historical consciousness. The findings shed light on information sources on history in general, undif ferentiated with regard to particular historical events or stages of individual biographies. To tackle these points of interest, this study aims to account for the use of information sources in the process during which individuals, in the course of their lives, form their historical con sciousness and to do so with regard to particular events of modern Czech history.

Theory
Historical consciousness is a concept that revolves around the relationship between knowledge of the past and other temporal orientations of its bearers. It can be charac terized as an interplay between interpretations of the past, perception of the present and expectations of future [Rüssen 2006[Rüssen , 2008. According to Jiří Šubrt, whose empirical research we take as our starting point, historical consciousness can be analyzed in terms of four components: lived historical experience, ideologies, knowledge produced by histo riography, and collective memory (that is, shared representations of the past constitutive of group identity) [Šubrt -Pfeiferová 2010;Šubrt -Vinopal 2010: 10]. By involving both specialized and everyday knowledge of the past, the concept of historical consciousness is meant to be more general -in terms of both scope and abstraction -than that of collective memory, which has been defined in contraposition to historiography already since the times of Maurice Halbwachs [Halbwachs 1992;Olick -Robbins 1998: 110].
As indicated by the Czech historian Miroslav Hroch [Hroch 2014: 52], historical con sciousness is constituted as a broad set of unordered and often contradictory memories, facts, myths or other information mediated by various institutions such as school, media, historiography, or simply wordofmouth. Hroch then sees collective memory as a subset of historical consciousness that is deemed legitimate and worthy of remembering for a giv en collectivity [Hroch 2014: 52]. In this vein, memory is constituted through mediated dis semination of emotionally charged and mobilizing narratives [Kubišová 2014: 92] selected from the horizons of historical consciousness. On the other hand, historical consciousness is constituted by a circulation of objects and information that is wider in scope and could be likened to a common historical "archive" .
Historical consciousness also cuts through the established distinction between com municative and cultural memory by subsuming both everyday communication about recent past experienced by living generations and institutionalized symbolism referring to events constitutive of the collectivity. As Jan Assmann [2008: 110-112], the proponent of the distinction, indicated, it directly relates to the phenomenon of floating gap. The events of origin are, together with recent events, chiefly among the chapters of past that are collectively maintained [Vansina 1985: 23-24]. In contrast, the gap in between has a low density of events [Zerubavel 2003: 26] that are remembered. This not only points to the fact that certain periods are expected not to be commemorated, but also to the fact that these periods can only be determined with regard to a particular collectivity.
Referential collectivities of historical consciousness and collective memory have tra ditionally been represented by nationstates, which have been legitimized through natu ralized historical narratives since the very beginning of their existence [Anderson 2006;Hobsbawm -Ranger 1983]. 2 In this study, we intend to focus on age cohorts (and their biographical stages) within a nationstate as referential collectivities of historical con sciousness. Although these cohorts may exhibit common patterns with regard to historical consciousness, it is still an open question whether they are bound by a shared "generational consciousness" as argued originally by Karl Mannheim [1952: 299]. Such a question has recently been tackled by Raili Nugin in her research on postsocialist transformation in Estonia, while its various aspects have been explored with regard to postsocialist trans formation also by other authors [AarelaidTart 2016;Kõresaar -Jõesalu 2016;Nugin 2015Nugin , 2016Nugin -Jõesalu 2016: 207;Kalmus 2016: 325;Kalmus -Masso -Lauristin 2013: 16;Pozniak 2013]. The common assumption behind this research is that specific historical experiences or specific forms of mediation constitute the differences between generations with regard to the transformation. We will tackle the two analytically distinct sources of historical consciousness further.
While media have always been crucial to historical consciousness, some authors argue that the past is being increasingly mediated by specialized institutions and technologies [Hoskins 2009b] and the content of older media is being remediated by newer ones [Erll -Rigney 2009: 3-4]. Media serve as intermediaries of past events for cohorts that did not expe rience them directly. Moreover, they also constitute a scaffolding for the memory of those who lived through the mediated events [Hájek -Dlouhá 2011: 51], sometimes to a point when personal memories and media content blend into an undifferentiated image [Sturken 1997: 20]. As a result, it is necessary when studying the formation of historical consciousness to go beyond traditional personal interaction and involve cultural production of objects that are replicated and circulated on a mass scale [Guggenheim 2009: 44;Welzer 2008: 286-288].
In a model that Furnham and Stacey [2016: 9] proposed while studying the formation of children's understanding of the social world, they distinguish between intrapersonal, interpersonal and sociohistorical aspects of the process. According to the model, socializ ing content stemming from the sociohistorical context is filtered, modified, or negotiated by interpersonal mediation and intrapersonal factors, both of which affect the formative process in developmental stages of the individual biography. In other words, the exposure 2 In this regard, national holidays and other days of importance have been instrumental in building group identity through organized commemoration. Their importance is made visible in numerous struggles over their selection and definition [for the case of former Czechoslovakia, see eg. Gammelgaard 2011; Miháliková 2005]. Some research indicates that holidays can be seen as pauses in increasingly accelerated timeframes of current societies, allowing for synchronization of activities, attention and presence, but also differentiating various segments of societies based on participation in them [Vihalemm -HarroLoit 2019]. While interesting in its own right, the role of holidays in establishing communities with shared historical consciousness is out of scope of this study.
to content mediated and interpreted in various ways is anchored in the process of ageing. It is at a certain age that individuals begin to ask, to read, to be schooled, or to be allowed to control their overall media consumption and it is also at a certain age that individuals start exhibiting particular abilities of generalization or inference. As a result, the dynam ics of media and personal interaction within individual biographies can be described as a gradual process of growing into the shared historical consciousness. On the other hand, age cohorts are also distinctively affected by experiencing historical events as they unfold during their formative years. According to contemporary literature [de Regt et al. 2017;HartBrinson 2014;Ma -Kim 2015;Schuman -Corning 2000, 2012Tekcan et al. 2017], one of the most significant factors influencing individual biog raphy is the "sensitive period", or the socalled "critical years". It represents a time frame between the 10th and 30th year of life during which the process of formation of historical consciousness is most intense. The experience gathered during this period through various significant events is most likely to produce longitudinal generational effects. Schuman and Corning explain the mechanisms behind this crucial period as follows: National and world events that occur during the critical years should have a lasting impact on individuals because they are usually the first major national or world events experienced, and therefore serve as a baseline against which later events are compared. Individuals beyond their late 20s will ordinarily have experienced earlier events that seem to them at least equal or greater in importance, while for those <10 years, limited awareness of the world beyond their own family and neighborhood means that larger national and world events are not likely to have registered sufficiently to be retained strongly in later memory. Still earlier events learned about indirectly from school or media (e.g., World War II for young people today) cannot have the same emotional impact regardless of their objective significance. [Schuman -Corning 2017: 521] To summarize, the formation of historical consciousness involves two interlocking pro cesses. First, individual ageing brings about an increasing scope of information sources that quickly reach beyond the immediate surroundings of the family and the immediacy of the present. Second, a common experience with unfolding events during the sensitive period shapes the understanding of other significant events, future or past. To observe the formation of historical consciousness then means to identify patterns in age cohorts that relate the biographical stages of individual lives with experienced events and information intermediaries. To understand the phenomenon, the stages of individual biographies (not only the overall age) must be linked to the sources of experience and information.
To accommodate this premise, we designed our research instrument specifically to provide evidence not only on the current age and media use of respondents, but also on the age (and media used at that age) at which they established their understanding of the historical events. Furthermore, we chose two historical events as references for our survey questions: 1) August 1968 when the Soviet troops and most of their Warsaw Pact allies invaded Czechoslovakia and 2) February 1948 when the Communist Party assumed control over the government of Czechoslovakia. The events are comparable in their signif icance for the Czech national history while they also represent events that either almost all respondents had only mediated information about (February 1948) or part of them experienced the event during their youth or adult lives (August 1968). A research design that takes into account biographical stages of formation of historical consciousness and compares the evidence for events that were or were not directly experienced allows us to at least partially disentangle the effect of a common historical experience of a cohort from the effect of individual ageing.

Methodology
While we are generally interested in the role of biographical stages and information sources in the formation of historical consciousness, we attempted to address sever al specific questions in this study: 1) are we able to identify a typical age of historical consciousness formation in the individual biographies? 2) does the age vary for different age cohorts? 3) which sources of information are characteristic of the formative period? 4) what is the role of direct experience with the historical event visàvis mediated infor mation? We attempt to answer these questions using data from a quantitative survey of the Czech population.
A selfreport survey was developed in order to gather data through the CAPI tech nique (Computer Assisted Personal Interviewing) within Omnibus random data col lection that took place in Autumn 2016. Respondents answered questions concerning information on how they gathered information about two significant events of the Czech history mentioned earlier (August 1968 andFebruary 1948). For each of the events, respondents answered questions addressing sources from which they learned about the events for the first time and sources of information that were the most important for them in forming their knowledge of these events. Additionally, the questions also addressed the declared level of knowledge of the events, the age of the respondents when they first got to know about the events and the age when they formed their current understanding of the events.
The research sample included a representative sample of Czech population (N = 1.008). The sample was randomly selected based on quota for gender, age, education, and region of residence. The age of the sample ranged from 15 to 89 years with an average age of 45.3 years (SD = 17.05). 3 Most respondents had acquired secondary education either without a schoolleaving examination (37.6%) or with a schoolleaving examination (33%), fol lowed by university (16.2%) and elementary school graduates (13.2%). More than half of respondents (63.4%) were economically active.
Data preparation was conducted indicating missing values in the data set and mul tivariate outliers. Since there was no presence of serious missing values Little's Missing Completely at Random tests that justify the use of the EM algorithm for replacing miss ing values with predicted values was no need to test. Influential multivariate outliers were detected with Mahalanobis distance analysis. All outliers above the critical chi square value (p < 0.001) were removed from further analyses [Field 2005]. Presented analyses were conducted using IBM SPSS Statistics v. 24 including calculation of fre quencies and testing of hypothetical hypotheses by conducting Chisquare tests with reported effect size.

Results
In this section, we will first compare our descriptive analysis of information sources with the findings of the research carried out by Šubrt's team. Then we will provide an anal ysis focused on biographical stages in order to identify patterns in formation of historical consciousness. Finally, we will employ statistical modeling to test (binary logistic regres sion) for the influence of basic sociodemographic characteristics on the identified patterns.
By simply looking at the frequencies of particular information sources, we note that our results are consistent with previous research [Vávra 2013: 40] in that school teaching and audiovisual media (TV shows, documentary films or radio broadcasting) are declared most crucial for establishing historical consciousness. In Table 1. we provide a more elaborate over view of information sources that respondents designated either as conveying first information about an event or sources that were most important in establishing their understanding of an event. 4 The use of information sources is further differentiated with regard to particular events: February 1948, which most of our respondents did not experience during their lifetimes and August 1968, which took place in the course of life of a significant portion of respondents.  Note: n i = absolute frequency; f i (%) = relative frequency (%) There are also significant differences in comparison to the previous research results. For example, the role of family members (aggregating parents, grandparents and other family members) is much more significant in our results. This could be explained by framing the survey questions with particular historical events that took place during the lifetime of respondents or their older relatives. In contrast, the questions in Šubrt's survey were aimed at sources of information about history in general, which points to much older events, not belonging to the biographies of other family members. Another difference that could be explained by a different historical time frame of the survey is the high frequency with which radio broadcasting was selected as the first source of information about the events of August 1968. Given that radio receivers were present in many households at the time and that radio broadcasting monitored the events as they unfolded [Končelík -Večeřa -Orság 2010: 190], it seems plausible that the spike in our data is caused by respondents who listened to the broadcast. A similar claim could be made with regard to television broadcasting as a first source of information about August 1968.
Our results also provide more granularity with regard to the roles various information sources play in formation of historical consciousness. For instance, the Internet has already come out as an ambivalent source of information in Šubrt's research, having significant portions of respondents (characterized by age and education) that declare not using it at all for the purposes at hand [Vávra 2013: 41]. Our survey question was formulated specifically to filter out the content that originated in other media and is remediated by the Internet (see Appendix I.). The resulting findings suggest that without remediated content, the Internet is not very significant with regard to formation of historical consciousness. Fur thermore, it seems that audiovisual media are not playing the role of first sources with regard to events that the respondents did not live through, while at the same time, they are seen as important sources of information for establishing the understanding of events. On the other hand, school teaching and family members are often declared to both provide first information and affect the formation of a deeper understanding. However, to elab orate and assess these findings, we need to relate the influence of information sources to biographical stages in the formation of historical consciousness.
The two tables below provide an overview of age periods, in which respondents declare to form their best understanding of the events. The tables further relate the age periods with information sources that, according to respondents, contributed most to the for mation of their understanding at that time. 5 Taking the concept of critical period as our reference point, we can aggregate the age periods into three phases: 1) precritical period (up to 10 years of age), 2) critical period (11-29 years) and 3) postcritical period (30 years and more). The information sources are aggregated into four groups according to their relatedness to socializing institutions or agents, a categorization that is employed by existing research on historical consciousness [Labischová 2012: 155]. Such aggregation of age periods and information sources allows us to observe the most significant patterns and their variation with regard to the respective historical events.
In both cases, most responses indicate the formation of understanding during the crit ical period between the age 11 and 29 (64.18% in the case of February 1948 and 73.33% in the case of August 1968). Taking a closer look at the data, the age period to which most responses point is situated between the age 11 and 15, when senior classes of elementa ry school are attended. This age interval represents an early phase of the critical period, during which the formal educational system is most influential. Within this age period, school attendance plays a crucial role with regard to knowledge of the two events (26.78% in the case of February 1948 and 15.68% in the case of August 1968). The second most frequently indicated is the precritical period. Its role is more significant with respect to February 1948 as 30% of responses indicate formation of understanding at the age of less than 10 (8% were even less than 6 years old). The same holds for almost 20% of responses in the case of August 1968. Only a very small part of responses is to be found after the critical period (5.53% in the case of February 1948 and 7.25% in the case of August 1968).  Looking further, we might observe that the influence of various information sources in the respective age periods is structured similarly for both events, forming an overall staged pattern. In the precritical period, family members (mainly parents) are the most influ ential sources of information. Almost 20% of responses declared this influence as most significant for their understanding of February 1948. In the case of August 1968, there are 10% of responses declaring this. The beginning of critical period brings the influence of a variety of other sources. Between the ages 11 and 15, the dominant source is the formal educational system, aggregating the influences of teaching, peers, or books and textbooks. A significant portion of responses indicates the formation of understanding in this peri od with the help of mass media. In another phase of the critical period (16-29 years), the role of formal education gradually vanes (it accounts for 4-5% of all responses). The understanding formed at this age is informed predominantly by mass media such as tele vision and radio broadcasting or the press. Around 12% of responses regarding February 1948 can be found at the intersection of this age and the aforementioned sources. With regard to August 1968, the number reaches 18% of all responses.
The results of the above analysis are further validated when we look at the relationship between information sources and the current age of respondents. In the case of February 1948 (Table 4), the dominant sources (TV shows, documentary films, school classes and parents) are consistent across all age cohorts. It is also true of all cohorts that none of the contains a significant number of individuals who experienced the events directly. In this regard, there is a key difference in the case of August 1968. Here, the oldest age cohort (60+) departs from the established pattern by listing radio broadcasting, printed media or peers as sources forming their understanding of the events. These sources are usually used to inform about events as they unfold and indeed, the cohort is old enough to have experienced the events directly. Moreover, the difference is especially fitting the staged pattern identified above as members of the cohort were 12 or older during the event. At this age, according to the pattern, the sources shift from the family environment to school and media. This is why members of the cohort 45-59 who were largely born before the events, but were younger than 12 reported the same sources as cohorts which were born after the events took place.

Factors Influencing Belonging to Particular Formation Period
For further exploration of our data, we use binary logistic regression that helps us to identify main factors influencing the likelihood that respondents 6 form their best under standing of the historical events (February 1948/August 1968) in one of the previously determined formation periods: (a) The precritical period up to 10 years (models M1 and M2); (b) the early critical period from 11 to 15 years (models M3 and M4); (c) the late critical period from 19 to 29 years (models M5 and M6).
All models used standard sociodemographic variables (gender, age, education, and size of residence location) as independent variables. The results of relevant statistical tests assessing the significance of constructed models are listed in the tables below (see Tables 6 -8.). The results of statistical modelling point to the following findings.
The structure of factors influencing the likelihood that respondents form their best understanding of a historical event in the precritical period is very similar with regard to both February 1948 and August 1968s (see Table 6). The variables play the same roles in both models (M1 and M2) and especially age and size of residence location made a unique statistically significant contribution to the models. However, their power of influence var ies slightly.
The main difference between them can be found in the strongest predictor. In the M1 model the age category 60+ results in an odds ratio of 4.6. This indicated that respon dents over 60 years were about five times more likely to form their understanding of February 1948 in the precritical period than some of the younger age cohorts. However, the power and direction of this category changed in the M2 model. In this case, respon dents over 60 years are less likely to form their best understanding of August 1968 before the age of 11. For every extra year of age, the odds of belonging to the period decrease by a factor of 26. Moreover, respondents in age between 15 and 59 are generally more likely to belong to a precritical period in the model for August 1968 than in model for February 1948. Furthermore, the impact of whether people live in large cities as mea sured by the number of inhabitants has also increased slightly in the model for August 1968. The odds of belonging to the precritical period increase with the growing size of residence location. The results of statistical modeling for the second formation period (models M3 and M4 in Table 7), the early critical period, are very similar to the previous periodage and size of residence location are the most significant variables. In the M3 model, respondents belonging to older age cohorts are less likely to form their understanding of February 1948 in the period from 11 to 15 years when compared to the youngest age cohort (15-29 years). The likelihood of belonging to this formation period decreases with growing age. The influence of age is still significant in the M4 model, but not as much as it was in the previous model. In comparison to older age cohorts, the youngest cohort is still more likely to form understanding of August 1968 in the early critical period. In this regard, respondents aged 15-29 years are three times more likely to form their under standing of August 1968 than respondents aged 30-44 years, five times more likely than respondents aged 45-59 years, and seven times more likely than respondents aged 60+. The likelihood of belonging to this formation period decreases with growing age. Statistical models dealing with factors influencing the formation of historical con sciousness in the late critical period (models M5 and M6 in Table 8) point to education as having statistically significant influence. According to the M5 model, the probability of belonging to the period is increasing with every further educational level beyond pri mary education. In comparison to those who attained primary education (ISCED 2), respondents having secondary education with state exam (ISCED 3a, b) are almost twice as likely to form understanding of February 1948 in the age of 16-29 years; respondents with any form of tertiary education degree (ISCED 5, 6) are 2.6 times more likely to belong to this period. In the M6 model, the level of education is even more important because odds ratios are slightly higher. In the case of August 1968, we can add another important factor -the age of respondents. In the M6 model respondents aged 60+ were about 3.5 times more likely to report acquiring the best information about the events during the late critical period when compared to the youngest cohort included in the survey. However, the cohorts aged 30-44 and 45-59 did not make a unique statistically significant contribution to the model.

Discussion
The staged pattern identified in our analysis corresponds to some of the classical theo ries of socialization. In particular, the distinction between primary and secondary social ization as a function of transition from the influence of significant others to that of a gen eralized other fits the pattern quite well [Berger -Luckmann 1966: 153;Mead 1972: 154]. The primary phase takes place predominantly within the family through socialization by significant others early in the biography of the individual. The secondary phase, on the other hand, takes place later in the biography and involves agents from outside of family such as those found in the school environment or within the media landscape, represent ing the influence of a generalized other. According to our analysis, the boundary between primary and secondary socialization with regard to formation of historical consciousness is to be found around the 10th year of age.
However, the staged pattern also shows that socialization often begins earlier than the concept of sensitive period would lead us to expect. Given that the broadest age spectrum of the sensitive period begins with the age of 10, it is the socialization mediated predom inantly by family members which undercuts this boundary. Considering the secondary socialization phase involving agents outside of the family environment, the age spectrum defined by the sensitive period seems to hold. This has been validated by relating infor mation sources to the objective age of respondents with regard to August 1968 (Table 5).
Focusing on the other end of the age spectrum, the findings correspond to the concept of sensitive period in that the understanding is generally formed by the 30th year of age. This seems particularly interesting given that the survey questions focused on key events associated with the socialist regime, which were redefined after 1989 in the official dis course. Should such a reinterpretation take place also at the individual level, the data would indicate formation of historical consciousness in later stages of life at least for some age cohorts. This finding could indicate a discrepancy between official discourse and informal knowledge to be present already in socialization during socialism. Moreover, February 1948 no longer represented an event of origin after 1989 (and it ceased being celebrated as a national holiday), while August 1968 was overshadowed by November 1989, resulting in low need of reinterpretation at the individual level. This finding could thus be explained by the existence of a floating gap in collective remembering after 1989.
The results of statistical modeling measured the influence of basic sociodemographic factors on the probability of respondents' belonging to one of the identified formative stages. While looking at age, the models largely confirm what can be deduced from the characteristics of the sample. For example, the decreasing likelihood of older age cohorts to belong to the precritical stage with regard to August 1968 simply reflects the fact that the event took place when a significant part of the cohorts was older, i.e. they could not be socialized in the precritical period, because the event has not happened yet. Perhaps the most interesting factor identified is education, which had similar effect with regard to both historical events. Respondents with higher education are more likely to form their understanding in later stages of life based on their engagement with various media. Edu cation, as it seems, has an effect of prolonging the formative period while bringing in an additional wealth of information sources. The influence of residence location size, on the other hand, is associated more with the events of August 1968. Specifically, it was the strongest predictor of belonging to the precritical stage (the larger the residence location, the higher probability of belonging). It seems as if the events were much more impactful in larger cities, causing children to socialize earlier than expected.

Conclusion
In this article, we aimed to examine the process in which historical consciousness is formed. In particular, we were interested in which information sources respondents use at various stages of their life. Through quantitative analysis, we were able to identify three early life periods that together form a staged pattern: 1) the precritical period (age: >10, dominant set of sources: family members) 2) the early critical period (age: 11-15, dominant set of sources: school environment); and 3) the late critical period (age: 16-29, dominant set of sources: media environment). To not rely only on declarative statements about when in the course of a lifetime the information sources were used, we also validated the bound ary between some of the stages against the current objective age of the respondents. Statis tical modeling was employed to further examine the influence of basic sociodemographic characteristics on belonging to any of the identified formative stages. Most interesting in this regard was the role of education, having the effect of prolonging (in terms of age) and widening (in terms of information sources used) the formative process.
The study also has limitations that need to be accounted for. The abstract nature of the identified pattern seems suited for comparative research. However, to a fairly significant degree, the results will depend on the particular historical reference points used. This lim its the potential for making general inferences based on comparative surveys, especially given the nationspecific character of history. Another limitation of this study can be found in the assumption that respondents remember how they formed understanding of the events. Especially older respondents, having a significant time distance from the peri od the survey inquired about, may have difficulty remembering. The small age difference that respondents reported with regard to first sources of information as opposed to best sources may indeed indicate conflation and ambiguity. It may even be argued, that instead of the actual usage of information sources, the survey measured internal notions of how and when one should have learned about the events according to respondents. However, this does not negate the identified staged pattern and its association with various socio demographic characteristics. The structure of the pattern is consistent across age cohorts and changes only with regard to respondents who experienced the events directly. Further research focusing on respondents as they pass through the sensitive period is needed to corroborate the pattern and to clarify its nature.