The influence of multiperspectivity in history texts on students’ representations of a historical event

School history textbook narratives of a nation’s past often present limited perspectives, which may impede the aim of teaching history from multiple perspectives. Less is known about the influence of including multiple perspectives on students’ representations of the past. This study examines the extent to which students include multiple perspectives when processing a schoolbook text that includes multiple perspectives compared to a schoolbook history text containing fewer perspectives. Tenth grade students ( N = 104) in four schools were randomly assigned to read one of two texts on the Dutch Revolt and asked to make a summary. Multiperspectivity was analysed through the representation of actors, aspects of scale, dimensions and historiography. The students working with the text having high multiperspectivity showed more perspectives in their representations. In the summaries, these students included significantly more perspectives than did the students using the text with fewer perspectives. Moreover, these students’ representations of the main actors were more nuanced. The students using the text with high multiperspectivity situated the conflict in a broader international context and integrated more historiographical dimensions. The insights generated by these outcomes emphasize the important role of textbooks when aiming to teach history from multiple perspectives.


Introduction
A key issue in history education is the need for students to learn to discern multiple perspectives. Multiperspectivity implies the admission of perspectives of different historical actors, historians, or contemporaries and as a consequence the admission of possible alternative narratives-each with their own narrational voices and perspectives (Munslow, 2016). Stradling (2003) stressed the necessity to relate and compare different perspectives to each other to enable a deeper understanding of historical relationships between nations, majorities and minorities in and outside national boundaries. He argued that multiperspectivity can enhance historical thinking and promote democratic citizenship. Multiperspectivity in historical thinking and reasoning helps students acknowledge that historical narratives are open to interpretations based on social-cultural diversity instead of being representations of a monocultural view (Council of Europe, 2018). This is all the more important at a time when abundant sources of all kinds including hoaxes and fake news are available to students at the click of a button (Wineburg, 2018). In the classroom, students can learn to think about multiperspectivity in many ways, for example, through whole-classroom discussions or specific tasks. An obvious, close-at-hand starting point, however, is the history textbook. In many countries, the history textbook is by far the most accessible source of information for students. The texts therein can be considered as historical narratives that follow traditions in academic historiography (Foster, 2011;Sakki, 2014). History textbooks represent a narrative in which specific historical actors, events, developments and perspectives are represented. However, it is argued by several researchers that the narratives about the nation in textbooks often represent a limited, nationalistic perspective (Sakki, 2014;Van der Vlies, 2017). This certainly applies to narratives about the origin of the nation-state. With respect to the reproduction and perpetuation of national narratives, Wertsch (2008Wertsch ( , 2012 stated that schematic templates according to abstract categories function as organizers of specific narratives. In these templates embedded in the social-cultural setting of modern states, episodic and configurative dimensions fuse together in collective memory and support identity formation (Grever, 2006). In history textbooks, specific narratives are presented that conform to certain narrative templates. For the Netherlands, Grever (2006) provided a schematic narrative template as "small country bravely fighting for its freedoms". Carretero et al. (2012) discuss several features of master narratives of the national past, such as the mechanism of inclusion and inclusion (e.g. the use of "us"), the presence of heroic characters and motives, the search for freedom or territory as a characteristic theme, moral orientations and presenting the nation and nationals as pre-existing political entities.
In recent historical scholarship, nationalist master narratives are de-constructed and alternative narratives include actors with different perspectives (for example, related to different social positions, interests, or gender), situate events in a broader context of international developments and mutual influence and bring more balance in attention to political, socio-economic and cultural aspects. For example, Heerma van Voss et al. (2018) in their Wereldgeschiedenis van Nederland [World history of the Netherlands] discussed the history of the Netherlands as part of global history.
Textbooks presenting a more closed narrative do not seem in line with the goal of history education to stimulate the exploration of multiple perspectives. In the Netherlands, as in other countries, secondary school students are expected to develop their historical thinking and reasoning abilities, including using multiple perspectives. Students are expected to understand the interpretative nature of periodization, historical explanations and narratives about the past and the positionality of historical actors. By doing so, students learn that a historical narrative is not a "given" but a construct about the past (Carretero, 2017;Lopez et al., 2014;Van Nieuwenhuyse et al., 2015). Scholars have argued that multiple perspectives are needed to transform essentialist narratives about a nation's past into narratives that better express the interpretative character of history (Barton, 2012;Carretero, 2017;Peck et al., 2011). Nevertheless, this specific element of history is difficult for students to master (Lee & Howson, 2009;Monte-Sano, 2011201120112011). One of the possible influencing factors could be the manner in which narratives of the past are presented in history textbooks. However, little is known about how the availability of multiple perspectives in textbooks bears upon the multiperspectivity of the students' representations. In this study is explored what the influence is of multiperspectivity in history texts on students' representation of a historical event. Wansink et al. (2018) introduced a temporal model of multiperspectivity and made a distinction (1) between the perspectives of subjects in the past, (2) of subjects between the past and present and (3) of subjects in the present. The perspectives of students can be considered as belonging to the subjects in the present. Like the historian Bergmann (2000), they consider the perspectives of agents in the past. But in addition to Bergmann, they stipulate the mediating role of what they call the subjects between past and present, especially the perspectives of historians. These perspectives of historians who think and write about the past-in time separated from the present and the past-constituted the second form of perspectivity. The perspectives that historians present are therefore also constrained in time and place. For example, the perspective on Columbus and the "Discovery of America" has changed over time (Carretero et al., 2012;Grever & Adriaansen, 2019). Likewise, history textbooks offer narratives wherein various perspectives can be presented. Just as historians in the past, these authors are also constrained in their own positionality as they perceive the past and history in the present. Wansink et al. (2018) thus draw particular attention to historiography, the role of historians and the historicity of representations of the past in addition to the positionality of historical actors. The focus brought to historiography and the role of historians can help to understand history as an interpretation of the past from multiple perspectives. Whereas Wansink considered multiperspectivity over time, from past to present, Grever (2020) and Stradling (2003) and others situated multiperspectivity more in the same time layer of the past. Within this layer, multiperspectivity could take shape by (a) providing different temporal and geographical scales (via plots or multiple local and/or international perspectives) (Grever, 2020); (b) the selection of historical actors and their experiences and perspectives determined by social class, gender, age, ability, race and ethnicity (Bergmann, 2000;Grever, 2020;Grever & Van Boxtel, 2014;Stradling 2003;Weiner, 2018); and (c) emphasizing particular dimensions (e.g. political, social, economic, cultural, military and religious history) (Grever, 2020;Grever & Van Boxtel, 2014;Stradling, 2003). These dimensions can be understood as forms of historiography wherein the choice of a particular form of history is of influence on the represented agency, plotlines and order of events (Grever, 2020).

Multiperspectivity
These ways to express multiperspectivity (actors, scale and historiographical dimensions) relate to the choices made by a historian when composing a narrative. Historians do more justice to the past if they pay attention to the actions, intentions and experiences of different actors (e.g. actors with different interests). When paying attention to more scale levels and dimensions, more aspects of an event are illuminated, other actors also can become part of the narrative, and even the plot can change. In addition to these forms of expressing multiple perspectives, there is another one, that is more on the meta-level: explicitly stating that a historian gives a certain interpretation (Paxton, 2002;Lee, 2013;Myskow, 2018;Wansink et al., 2018).
With respect to this latter aspect, Crismore (1989) argued that metadiscourse, which reflects an author's presence in a text, is underestimated in the production of textbooks. Paxton (1999) stated that "muted thought about the authors of historical texts" (p.333) hinders students from evaluating the trustworthiness of their textbooks and the perspectives taken therein. In later research, Paxton (2002) found that when using a text wherein the author was visible, students showed more thinking about the discussed history and the authorship of the text. In history textbooks, it is not always clear for students that the presented narratives are interpretations of the past and that more than one interpretation is possible. Lee (2013) found that different narratives on the same topic explicitly interpreted by different historians/authors prompted students to engage in more historical reasoning than conventional texts did.

Text comprehension and students' representations of history texts
Research has shown that history textbooks have an authoritative status for students (Perfetti et al., 1994;Rouet et al., 1997;Wineburg, 1991). However, in previous research, it was found that Dutch students' representations stayed well within the framework of the dominant narrative without giving much attention to possible other perspectives , even when a teacher addresses other perspectives . This is in line with the findings of researchers in other countries (Carretero, 2017;Tutiaux-Guillon, 2017;Van Havere et al., 2017). Therefore, it is important to know how students might develop a representation of a historical event when reading a text that contains more multiperspectivity compared to a text that contains less multiperspectivity. Kintsch and Rawson (2005) described the underlying processes of text comprehension. At the most basic linguistic level, the reader has to process the words and phrases in the text itself. The combination of words in the microstructure of interrelated propositions in the text forms the meaning of the text. However, the meaning of the text is not only formed by the meaning of the words and the interrelated propositions but also enhanced by the global structure of the text. Interrelated propositions and global structure taken together form the textbase, which represents the actual meaning of the text. To reach a deeper understanding beyond the mere reproduction of the text, the reader has to construct a mental model of the situation described in the text, i.e. a situation model. The latter requires blending the textbase, relevant prior knowledge and the goals of the reader (Kintsch & Rawson, 2005).
With regard to history education, Wineburg (1994) called the situation model an event model. Wineburg (1994) and others (e.g. Voss & Wiley, 2006) argued that in history education, inferences based upon the event model may help readers to construct a subtext about the intentions and motivations of the authors involved, including their positionality and perspectives. However, this refers to the use of multiple texts and source materials, not to the representation of an event based upon a single textbook text (Ter Beek et al., 2020). Schoolbook texts provide background information that is needed to perform historical contextualization, an important aspect of historical thinking and reasoning. When building a historical context to better understand a historical document or image, a historical event, or actions of people in the past, students need information about long-term developments; the social, economic, political and cultural features of the specific time; and historical locations (Endacott & Brooks, 2013;Huijgen et al., 2017;Van Boxtel & Van Drie, 2012). When students use a text with a high degree of multiperspectivity, they can construct an event model which may better support historical contextualization.
The event model of a historical event that students construct on the basis of a history textbook text is influenced by several elements such as prior knowledge, personal beliefs and the associations and features of the text (Wineburg, 1994;Voss & Wiley, 2006). Macrostructural information is crucial for the comprehension of informative texts. When students master macrostructures (e.g. comparison, sequence, cause and effect and problem and solution), they produce more coherent and richer texts and are more likely to retain the main ideas in the text (Turcotte & Caron, 2020). However, it is often difficult to identify a single structure in a text. In texts about historical topics, macrostructures are common, such as cause and effect and comparison. Multiperspectivity can also structure a text. This is also the case with respect to the perspectives incorporated in an event model. For example, at a macro level, the Dutch Revolt can be described from a primarily regional, political and religious perspective, whereas at a micro level, from the socioeconomic perspective, the daily lives of peasants and citizens during a civil war can become part of the model. However, little is known about the students' representations of multiple perspectives when they use a specific history text. Based upon research findings about the construction of an event model, one would expect that in their event model, students would follow the structure of the text wherein several perspectives are presented (Kintsch & Rawson, 2005;Voss & Wiley, 2006;Wineburg, 1994). On the other hand, such an event model is not only constructed on the basis of the suggested relations in the text. Different perspectives in a text can be less obvious to students than causes and consequences or differences and similarities. Presenting perspectives of different actors and particularly discussing interpretations of the same event by different historians can make a text more complex. Complexity is, however, also determined by sentence length, the structure of the text, domain-specific terms used, implicit and explicit (causal) relationships between events and developments and linguistic features (Berendes et al., 2018).
There is ample research on how students make mental representations of texts. Several researchers have argued that due to the lack of an adequate understanding of the concepts and ideas conveyed by the text, students may encounter obstacles (McKeown & Beck, 1994;McKeown et al., 1992;Van Silfhout et al., 2014). However, little is known about what kind of text representations students make when they use a text that presents multiple perspectives and how these might differ from those students make when they use a text with fewer perspectives. To what extent do students bring forward in their representations the various perspectives that are offered in the text? Gaining a better understanding of the event models that students construct of schoolbook history texts high or low on multiperspectivity is important because these mental representations will likely affect the historical narratives they remember and use in new, also extracurricular, situations.

Aims and research questions
In this study, the aim is to analyse how the degree of multiperspectivity in history textbooks affects students' representation of the historical phenomena presented in the text. When students process a text that contains a variety of perspectives, does their representation of the event also reflect this multiperspectivity? Do they include more perspectives than students who engage with a text limited in multiperspectivity? To this end, the following research question was posed: what is the degree of multiperspectivity in students' representations when they engage in a text processing assignment based upon a schoolbook history text that contains multiple perspectives compared to a schoolbook history text containing fewer perspectives?
The expectation is that in their historical representation, the students will present more perspectives if they make use of a history text with high multiperspectivity than if they make use of a history text with low perspectivity.

Method
To answer the research question, an experimental study was conducted in which students' text representations were compared in two conditions to which they were randomly assigned. In the first condition, students were given a text containing multiple perspectives (text HP), while in the other condition, the students received a text low in perspectivity (text LP). All students were asked to fulfill three text processing assignments: underlining the text, making a summary and making a poster. In this study, the analysis on the presence of multiperspectivity is restricted to the summaries that the students were asked to make.

Participants
Participants included 104 10th grade history students, who were following the intermediate track in preparation of higher professional education (in Dutch: havo). This group represents the largest number of secondary education students taking a central examination in history. The students were from four different classes and four different schools. To establish a more representative sample of the student population, the schools were spread over different regions of the country (both (sub)urban and rural). One school (School 1) was situated in a socio-economic weak urban area with a high degree of ethnic diversity. Schools 2 and 3 were situated in two small towns. The last school (School 4) was a school with an explicit Christian signature in a rural environment in the middle of the Netherlands. The other three schools had no specific religious profile. The students' ages varied between 15 and 18 years, with a mean of 16 years. All students gave active consent to participate in this study. No students dropped out, as the experimental study required only one lesson. Students had some prior knowledge of the topic of the text, as the topic was part of the curriculum in primary and lower secondary education.
Students from each class were randomly assigned to one of the conditions: 53 students in the HP condition (27 males, 26 females) and 51 students in the LP condition (27 males, 24 females). From eight students, data were missing, which resulted in 48 summaries for each condition (N =96).

Two versions of a schoolbook text
Two versions of a schoolbook text were designed on the topic of the Dutch Revolt . This topic is part of the compulsory program in the Netherlands in the upper levels of havo and is defined as "the conflict in the Low Countries that resulted in the founding of a Dutch state". Dutch history teachers consider the history of the Dutch Revolt especially salient for teaching students aspects of multiperspectivity (Wansink et al., 2017).
The texts were written for Dutch students in the upper levels of havo (4 havo). The texts were comparable in length (1306 vs 1325 words). Both texts were written by the first author, who has experience as a writer of history textbooks. Guidelines for teachers and syllabus designers were used when writing the texts (Fulcher, 1997). To ensure the ecological validity and equivalency of the texts, the texts were first discussed with the other authors, who both had ample experience in developing and editing textbook materials. Second, keeping in mind that the texts were to be used in the upper levels of havo, two focus groups were asked-one of expert teachers and another of novice teachers-to comment on the content and complexity level of these texts. Some minor text revisions, such as the incorporation of headers to each section, seemed necessary. Third, the readability of the texts was confirmed by an independent educational expert with expertise in the Dutch language. Fourth, the readability of both translated texts was additionally confirmed by running the Flesch Reading Ease Level test and the Flesch Kincaid test on-line (Sourceforge, 2007, retrieved 5th of June 2019). The tests indicated that the texts were suitable for 15/17 year olds (9th and 10th graders) (Flesch, 1948;Kincaid et al., 1975); this age group corresponds to the age group in the upper levels of havo.

Differences in perspectivity
Both texts aimed to explain the backgrounds of the revolt in the Low Countries against the Spanish king Philip II and how this revolt proceeded in its main stages during the second half of the sixteenth century. Common themes were the resistance against centralization and tax policies, the rise of radical Calvinism and the Iconoclasm in 1566 and the subsequent attempts to repress heresy and troubles by Spanish troops and reactions to it. Both texts described the realization of the Pacification of Ghent and ultimately the recognition of the independence of the Netherlands in the Act of Abjuration. The difference between the two versions lay in the level of multiperspectivity in each text. As four different ways were distinguished in which multiperspectivity can be presented in textbooks (actors, scales, dimensions and historiography), the texts were written accordingly, each with a different number of perspectives of actors, scales, dimensions and historiography. The LP and HP text did not differ much on the other actors. Both texts included the personages of William of Orange, Philip II, Margareth of Parm, the Duke d' Alba and Charles V. The texts, however, do both have an actor who appears only in that text; in the LP text, it is Balthasar Gerards (probably known to students from previous lessons on the topic) and in the HP text Catrijn van Leemput, an actor who is most likely not known because the role of women in the methods is underexposed. In the case of the HP text, it is therefore interesting to investigate how many students include her in the summary.
The text that aimed explicitly for multiperspectivity is referred to as text HP, as it was high on multiple perspectives. In this text, the revolt is more framed in the context of developments in the realm of Philip II and the main problems he encountered. In text HP, political developments were embedded in a more international context seen. For example, how the Turkish empire threatened the balance of power and locked troops of Philip II in the Mediterranean area. Economic and social dimensions were mentioned in several instances that gives another perspective than the strict political. For example, the experiences of the common people are described, what they went through during sieges, quartering of soldiers, inundation of their farmlands, or the looting of soldiers without pay. Also the example of the people of Utrecht who had to deal with the Spanish mercenaries was introduced. Especially the role of Catrijn van Leemput-a prominent female citizen of Utrecht-was described in the text by the historian Els Kloek (2013) in order to incorporate women's voices into the narrative framework and to enhance the recognizability of the role of women in history for both boys and girls. Historiographical perspectives were expressed by staging three historians in order to show that the textbooks were part of the historiographical debate. For example, the historian Parker (1977) emphasized the international dimensions of the conflict.
The second text was low in multiple perspectives. This text is referred to as text LP, as it presented fewer perspectives. It described the conflict from the perspective of William of Orange and his successors. The events followed the traditional narrative that is also present in existing Dutch school history texts . The conflict was presented as the result of the religiously inspired political and military deeds of individuals or collectivities such as Calvinists. Margareth of Parm was presented as the Governess of the Netherlands during the sixties of the sixteenth century. In this text, Catrijn van Leemput was absent, but the murderer of William of Orange, Balthasar Gerards, was introduced as he is commonly present in Dutch textbooks and is part of the dominant narrative of the revolt in the Low Countries. The socioeconomic dimension was restricted to the increase of taxes by Duc d' Alba. The experiences of common people were not addressed. The focus in this text was on the developments in the Low Countries with attention to the role of provinces and the fate of Dutch towns; no comparison with or references to international developments were made. The perspectives of different historians were left out.

Assignments
All students received three text processing assignments to fulfill during a full lesson time of 50 min: first, to underline the text, second to summarize the text and third to make a poster based upon the text. The students worked individually and used only the provided text, which they were allowed to keep with them during working on the assignments. However, not all students followed that order. As a consequence, not all students (8 out of 104) handed in a summary.
The summary of the students can be seen as an indicator of the "event model" which they have constructed (Bogaerds-Hazenberg et al., 2021). Making a summary reactivates the text's macrostructure (Turcotte & Caron, 2020). In a summary, students skip details and secondary events, make inferences from prior knowledge and present the essential information in their own words (Kintsch & Van Dijk, 1978). Students were invited to write what they found important for properly understanding the Dutch Revolt.
In the Appendix, typical examples of a summary of text LP and text HP can be found.

Procedure
The assignments were conducted in spring 2018 during standard history lessons. The three assignments took a full lesson time of 50 min to complete. A research assistant introduced the assignment and answered procedural questions. The teacher was present to take care of classroom management issues when necessary.

Analysing multiperspectivity
Based on earlier research on perspectives in schoolbook texts , we analysed the following aspects of multiperspectivity: actors, scale, dimensions and historiography (see Table 1). First, the summaries were coded if a historical actor was included. Agency of the actor was coded for if this was expressed in an active or passive form. The agency of the actor was included in the case it was the subject of the sentence, or the (in)direct object. The agency of the actor was excluded in the case it was used as a genitive. For example, in the sentence "Philip continued the policy of Charles V", Philip was coded as an actor, but Charles V was not. We focused on the five actors that were mentioned in both texts (William of Orange, Philip II, Margretha of Parm, Duc d'Alba and Charles V) and on Balthasar Gerards in text LP and Catrijn van Leemput in text HP. The LP and HP text did not differ much on the other actors. The texts, however, do both have an actor who appears only in that text; in the LP text, it is Balthasar Gerards (probably known to students from previous lessons on the topic) and in the HP text Catrijn van Leemput, an actor who is most likely not known because the role of women in the methods is underexposed. In the case of the HP text, it is therefore interesting to investigate how many students include her in the summary. Because Philip II is the more central actor in text HP, we were curious if the students in this condition would also pay more attention to him in their summaries. Furthermore, we examined the percentage of students who included a certain actor in their representation. However, we have to take into account that the sheer mentioning of an actor more often-for example, Philip II or William of Orange-does not necessarily say much about whether the student gives more consideration for the perspectives of these actors in this struggle or what kind of goals they were pursuing. Therefore, we added a qualitative analysis of what is said about these actors to determine whether multiple perspectives of actors are present.
Second, the number of different scales that students included in their representation (regional, interregional and European) was coded. As three elements of scale (see Table 1) were distinguished, one point for each element was assigned. The scores for scale thus ranged from 0 to 3. We considered the presence of more than one element of scale as an indicator for incorporating multiple perspectives. With the scale "regional", we coded those aspects in the summaries which included developments inside the Low Countries without mentioning Spanish or other foreign involvement. With the scale "interregional", we coded those elements in the summaries which exclusively highlighted the relationship between Spain and the Low Countries. With de scale "European", we coded those (part of) sentences that contained references to other European/global territories than Spain or the Low Countries. We conducted an analysis of variance to determine whether the number of different scales was significantly different between the two conditions. Third, the number of different dimensions that were included in students' representations (political-military, socioeconomic and cultural-religious) were coded. As three dimensions in the coding (see Table 1) were distinguished, one point for each element of perspectivity was assigned (range 0-3). We considered the presence of more than one dimension as an indicator for incorporating multiple perspectives. A (part of a) sentence was coded as containing a political-military dimension if it included a reference to the (opposition of) exercise of power, wielding of force or administered rules of law. It was coded as containing a socioeconomic dimension if the sentence included a reference to aspects of production, (the relation of) social groups, finance, income and taxes. And it was coded as containing a cultural-religious dimension if the sentence included a reference to aspects of culture or religious customs or convictions. An analysis of variance was used to determine whether the number of different dimensions was significantly different between the two conditions. Fourth, the presence of historiographical interpretations (the mentioning of historians or implicit indications of the interpretative character of history by using terms such as "probably" or "although") was coded. It turned out that there were hardly any references to historiographical interpretations made in the summaries (only seven in condition HP). Therefore, these results will be described in a qualitative way.
The interrater reliability was established on the coding of first author and a researcher not participating in this project by each coding the work of four students-two for each condition-of each class . The interrater reliability turned out to be substantial to good (Cohen's kappa Actors: .81, Scale: .68, Dimensions: 1.00). The main difference in the coding for the elements of scale was the distinction between regional and interregional. As historiographical interpretations were hardly present in the products of the students (7 in the summaries of text HP), interrater reliability could not be calculated.
Tests of normality (Kolmogorov-Smirnov) showed that actors, scale, and dimensions did not follow a normal distribution. Therefore, a nonparametric test (Mann-Whitney) was performed to make a comparison between the two conditions.

Results
In this section, the results are presented for students' inclusion of actors, scales, dimensions and historiography in their written summaries. The summaries were comparable in length (condition HP mean length of 170.8 words, LP 160.9 words).  Table 2 shows the percentages of students who included the selected historical actors in the summary for both conditions. The inclusion of William of Orange differed significantly by condition (χ 2 (1) = 17.68, p < .001). Table 2 shows that relatively more students using text LP mentioned William of Orange in their summaries, compared to students using text HP. In condition LP, William of Orange was positioned as the main actor in the conflict. For example, student 228 framed the summary of text LP around the person of William of Orange, opening with the sentence "William of Orange died as a result of the attempt on his life by a spy in Spanish service, Balthasar Gerards Willem was seen as the father of the fatherland" 1 (see the example summary LP in the Appendix). The students in condition LP described William of Orange in a positive manner. For example, one student stated, "William of Orange thought that people could believe in what they wanted, strove for freedom, later this led to a Republic" (student 107, text LP, summary). Student 233 (text LP) wrote the following in the summary: "He [William of Orange] took care for the independence of the Netherlands from Spain". Philip II figured foremost as the opponent of William of Orange and is also described in more negative terms. For example, a student stated, "Citizens saw how wrong Philip was" (student 352, text LP). Students using text LP described the conflict in the Low Countries as a conflict of individual actors. A considerable number of students using text LP include Balthasar Gerards (38%).
The inclusion of Philip II did not differ by condition (χ 2 (1) = .295, p = .587). In both conditions, almost all students included Philip II (Table 2). But whereas in the summaries of the LP condition Philip II figured as the opponent of William of Orange, he was the main actor in the summaries of the students in condition HP. For example, student 367 framed the summary with "Philip II had a large empire from 1555, … Catholic altar inspired him in exercising power and solving 2 major problems -> enormous lack of money and political unrest". And William of Orange figured only in the phrase "Civil war between followers of William of Orange and troops of Philip II" (see the example summary HP in the Appendix). The students in condition HP paid more attention to the awkward situation in which Philip II found himself and the conflict in the Low Countries was seen less as a conflict of individuals. Some students in condition HP (17%) even mentioned no other actors than Philip II in their summaries.
Catrijn van Leemput (in text HP) was included in 13% of the summaries (Table 2). For example, "The castle where the mercenaries were housed was later demolished under Catrijn's leadership" (student 4101, text HP, summary). This mentioning of Catrijn added to the multiperspectivity of the summaries of the students wherein she represented the perspective of women and ordinary people.

Geographical scale
The representation of the perspectives was analysed on the presence of the three elements of scale: Low Countries, Spain vs the Low Countries and European. Tables 3 and 4 present the results on the level of scale.
In their summary, students using text HP (Mdn = 2) included significantly more perspectives related to scale than the students using text LP (Mdn = 2): U = 1433, z = 2.39, p = .017, and r = .25; there was a small effect size.
Students who had read text LP described the conflict in territorial terms and as being confined to the Low Countries, as can be seen in example in the summary of student 228 (Appendix). They mentioned the consequences of the conflict primarily in local terms, for example, the taking of Brielle or the relief of Leiden. In frequency of inclusion, next following the references to local circumstances were those related to the conflict with "Spain", but these references were not further defined. The conflict viewed as part of broader European development or as related to global events was not brought forward by these students. A considerable number of students-for example, student 367 (Appendix)-who had read text HP incorporated not only the Low Countries scale, as shown in Table 4, but also the European scale of the conflict. They referred to European events, such as the conflicts of Charles V with other European sovereigns, the position of the Turkish Sultan and the influence of the treasure fleet. For example, a student (364) wrote "The biggest problems came from outside his empire: the Turks threatened his power around the Mediterranean" and remarked that "all the inhabitants of the Low Countries felt that they could regulate the affairs of the provinces themselves".

Dimensions
The representation of three dimensions was analysed: the political-military, socioeconomic and religious-cultural dimensions. In Tables 5 and 6, the results are presented.
In their summary, the students reading text HP (Mdn = 3) scored significantly more dimensions than the students reading text LP (Mdn = 3): U = 1584, z = 3.5, p = .000, and r = .36; the effect size was medium.
Students in both conditions emphasized the political-military dimension of the conflict. Nevertheless, students reading text HP included more elements related the socioeconomic dimension than students reading text LP (see Table 6). For example, student 112 wrote "Trade and industry suffered greatly from the many wars, and poor harvests led to rising food prices. This meant that raising taxes to compensate for the lack of money soon led to political and social unrest" or student 498 who wrote "The population suffered greatly from all military operations such as looting. In 1575, Philip II's troops had not been paid for two and a half years, so they started plundering Antwerp".  For students in the LP condition, one of the main causes of the conflict was the religious strife between Philip II and William of Orange. Students in the HP condition also mentioned the religious-cultural dimension, but they mentioned it more often in combination with other dimensions.

Historiography
No students in condition LP cited a historian or referred implicitly to the interpretative character of history in their summaries. Of the students in condition HP, 7 (15%) cited one of the historians mentioned in the text or made a remark about the interpretative character of history.
Although the work of historians was explicitly presented in text HP, the majority of the students addressed these interpretations as "facts", which they mostly did by leaving out the author of the interpretation. For example, student 242 (text HP), who stated that "The greatest problems to Philip arose from outside his empire", left out the part of the text stating that "Historians have pointed out that…". A good example was student 116 (text HP, summary) who explicitly explained that history has an interpretative character: "…Wars can be described from different angles, one-sided information [emphasis added] can be provided, so Philip II is sometimes seen as a problem maker". Contrary, we found no indications of history as an interpretation constructed by historians in the LP condition, even not by using words such as "probably" or "although". In sum, the vast majority of the students used both texts not as interpretative narrative representations but as factual descriptions.

Conclusion and discussion
Multiperspectivity is a key issue in history education, but little is known about how the presence of multiple perspectives affects the representations students construct of a historical event. We investigated whether the level of multiperspectivity in a text affected 10th grade students' inclusion of perspectives in their representations of the historical event. An analysis was made of the representations of actors, aspects of scale, dimensions and historiography. The results show that students' representations are affected by the level of multiperspectivity of the texts. This is in line with earlier findings, for example, by Lee (2013) who showed that the features of texts affect students' thinking and representations. First, with respect to actors, it was found that students using text LP more often presented  William of Orange as the main actor, whereas students using text HP more often presented Philip II as the main actor. Although students in both conditions mentioned different actors, this does not implicate that these mentions represented different perspectives. After all, these actors (with the exception of Catrijn van Leemput) represented the dominant perspective of the narrative of the conflict in the Low Countries. Second, students using text HP showed significantly more perspectives on scale than those who had read text LP. The latter included the Low Countries and-to a lesser extent-the Low Countries versus Spain scale, whereas students using text HP also incorporated phenomena on a European scale. Third, students using text HP included significant more dimensions compared to the other group, as they incorporated more socioeconomic and religious-cultural dimensions in their representations. However, in both conditions, the political-military dimension was dominant. This can be explained by the fact that both texts were an elaboration of the topic offered in the compulsory history program that focused on the origins of the Dutch state and thus on politics. It might also be the result of students' experiences with textbooks in which the political-military dimension is dominant  and, therefore, students being accustomed to paying attention to this dimension. Lastly, only a few students from condition HP incorporated historiographical elements. Students tended to leave out references to historians and to represent interpretations as fact. The students in this study read only one text. Studying several different texts could yield more attention for different perspectives. Several researchers demonstrate the potential of tasks in which students use multiple documents (Ter Beek et al., 2022;Wineburg & Reisman, 2015). Furthermore, it is possible that students' epistemological views on the nature of history influence the way they used the texts (Maggioni et al., 2006;Stoel et al., 2017). However, the epistemological views of the students were not part of this study. Although students in both conditions differed in the number of perspectives included in the summaries, they also ignored perspectives mentioned in the texts. One of the reasons might be that students are accustomed to concentrating on national, political history. This kind of history dominates the collective memory and narrative about the origins of the Dutch state. For example, Balthasar Gerards-only present as part of the first sentence in text LP as the murderer of William of Orange-is a known figure in the national narrative and was frequently mentioned by the students in condition LP. In contrast, it was probably the first time that the students read about the female actor Catrijn van Leemput, prominently present in text HP, but ignored by most of the students. However, there are perhaps other reasons why the perspective of ordinary people who were engaged in the revolt-for example, Catrijn van Leemput-was ignored. In general, the understanding of the macrostructure of a text enhances text comprehension (Turcotte & Caron, 2020). In this case, it could be that students did not mention Catrijn van Leemput because, to them, this kind of agency was not part of the macrostructure of the text. This raises the question of how domain-specific elements, such as a detailed description presented from a specific perspective in the case of Catrijn van Leemput, are "admitted" to the macrostructure of the text in the understanding of students. Turcotte and Caron (2020) suggested that combined teaching of macrostructure (e.g. the perspectives of ordinary people) and microstructure (e.g. the perspective of Catrijn) might lead to better text comprehension, especially at the macrostructural level. As a consequence, students might become more aware of the implicit information offered in a text, enabling them to understand historical narratives as more or less open to multiple perspectives. Future research is necessary to shed more light on the question of the relations between the macro-and microstructural qualities of a history text and how thinking about multiperspectivity is fostered among students.
Because students were randomly assigned to the two conditions, we assume that the HP and LP groups were equal with respect to potential confounding variables, such as prior knowledge about the topic or their ability to comprehend written text. We did not measure students' prior knowledge and reading comprehension ability. Our study does not provide insight into the influence of prior knowledge on the extent to which students adopt or process multiple perspectives from a text. We do not know to what extent students with more prior knowledge pay attention to new perspectives, or ignore them, for the reason that these do not fit into the dominant narrative they have appropriated. Compared to topics such as the Second World War, the role of the Netherlands in the slave trade or decolonization, the topic of the Dutch Revolt is much less present in public debates and other elements of historical culture. As a result, there may have been less of a standard narrative stuck with the students. More research is needed on how students' prior knowledge affects the processing of multiple perspectives.
We are aware that statistical analysis was possible only for actors, scale and dimensions and not for historiography. The absence of historiographical perspectives in the students' summaries seems to suggest that students regarded both texts primarily as providing factual information. This underlines all the more the need to pay explicit attention to epistemological beliefs when teaching history as interpretation, also when it concerns a text in a history textbook (Stoel et al., 2017).
Future studies can pay attention to students' reading skills; however, it should be taken into account that this skill can also be described as a domain-specific skill (Ter Beek et al., 2022;Van Driel et al., 2022). For example, information about the author's perspective can be gained by investigating the language and metaphors used by an author . The same can be said about the domain-specific reading ability when it comes to identify different types of causes and effects or diverse, different perspectives. Therefore not only reading ability of students as measured by general reading ability tests is important, but also their domain-specific reading ability. If students are trained in identifying perspectives of different actors, characteristics of the spatial context in a more narrow sense (e.g. local or national), but also in a broader sense (e.g. inter-regional, European) and different dimensions (e.g. social, economic, political and cultural), they might construct richer representations of texts.
In this study, we focused on students' representations of a historical text on the topic of the Dutch Revolt by constructing a summary. More research needs to be done to investigate whether other topics would yield comparable results. In addition, a summary is only one way of gaining insight in students' representations. Future research could in include other representational forms, such as visual representation (i.e. a poster) or writing an historical account. It would also be interesting to include a delayed posttest to investigate long-term effects of the provided text. Another limitation of focusing on representations through a summary is that we cannot draw any conclusions on whether students who included more different perspectives also gained a better understanding of the event. Although we offered the students texts that are comparable in length and readability and experts deemed both texts adequate to the level of the students, we cannot exclude-even though we checked the texts for ecological validity and equivalence-that students experienced the text HP as more difficult to construct an event model than the students who used text LP. Therefore, the HP students may have presented fewer perspectives than might have been possible. To support the construction of an event model of a text high on perspectivity, different experiences and perspectives of actors could be made more explicit as also the perspectives scale, dimensions and historiography. Furthermore, it could be helpful for students if the author's voice was made more explicit in the texts (Paxton, 2002;Myskow, 2018). The teacher might explicitly show which perspectives are present in a text. Further research could explore how different types of assignments, classroom activities and different topics stimulate students to include multiple perspectives in their representations of national history. For future research, it would beneficial if is to take into account the sociocultural background of the students. How does students' cultural, ethnic and social background affect the processing of perspectives included in a text.
In conclusion, the number and types of perspectives offered in the texts were reflected in the representations of the students. This confirms the authoritative stature of historical schoolbook texts and is in line with the findings of earlier research (Foster, 2011;Foster & Crawford, 2006;Perfetti et al., 1994;Rouet et al., 1997;Wilkens, 2011;Wineburg, 1991). Moreover, this outcome stresses the importance of including multiple perspectives in history textbooks, as students are inclined to follow the text in the textbook. Therefore, textbook authors should support teachers and students in learning and reasoning about multiperspectivity by incorporating more and diverse perspectives.