This article was downloaded by: [University of Tennessee At Martin]On: 06 October 2014, At: 12:58Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Curriculum StudiesPublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/tcus20

What do pupils and textbooks do witheach other?: Methodological problemsof research on socialization througheducational mediaVERONIKA KALMUSPublished online: 20 Feb 2007.

To cite this article: VERONIKA KALMUS (2004) What do pupils and textbooks do with each other?:Methodological problems of research on socialization through educational media, Journal ofCurriculum Studies, 36:4, 469-485, DOI: 10.1080/00220270310001630670

To link to this article: http://dx.doi.org/10.1080/00220270310001630670

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the“Content”) contained in the publications on our platform. However, Taylor & Francis,our agents, and our licensors make no representations or warranties whatsoever as tothe accuracy, completeness, or suitability for any purpose of the Content. Any opinionsand views expressed in this publication are the opinions and views of the authors,and are not the views of or endorsed by Taylor & Francis. The accuracy of the Contentshould not be relied upon and should be independently verified with primary sourcesof information. Taylor and Francis shall not be liable for any losses, actions, claims,proceedings, demands, costs, expenses, damages, and other liabilities whatsoever orhowsoever caused arising directly or indirectly in connection with, in relation to or arisingout of the use of the Content.

This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Journal of Curriculum Studies ISSN 0022–0272 print/ISSN 1366–5839 online © 2003 Taylor & Francis Ltdhttp://www.tandf.co.uk/journals

DOI: 10.1080/00220270310001630670

J. CURRICULUM STUDIES, 2004, VOL. 36, NO. 4, 469–485

Veronika Kalmus is a research fellow in the Department of Journalism and Communication,University of Tartu, 18 Ulikooli St., 50090 Tartu, Estonia; e-mail: vkalmus@ut.ee. Herresearch interests centre on the school textbook as an agent of socialization, the ‘hiddencurriculum’ of educational media, and the emerging media society in Estonia. She haspublished in Trames: A Journal of Humanities and Social Sciences and Discourse & Society.

What do pupils and textbooks do with each other?:Methodological problems of research on socializationthrough educational media

VERONIKA KALMUS

School textbooks are assumed to be important socializers. However, because of thecomplexity and methodological stumbling blocks involved in the subject matter, the impactof specific textbooks and the interaction between pupils and textbooks have seldom beenstudied. I discuss the methodological problems in explorations of the role of schooltextbooks as agents of socialization, and suggest methods for exploring and understandingthe dynamic interaction between textbooks and their readers.

The volume of academic and applied research in textbook analysis andcriticism (Woodward et al. 1988, Johnsen 1993, Mikk 2000) indicates thatschool textbooks and other educational media1 are seen as importantinstruments for transmitting knowledge and values to the young generation,and thus for reproduction or transformation of the social order. The researchinterest in school textbooks has been based on the more or less implicitassumption that educational media influence, persuade, or mould theirreaders. This assumption probably rests on the fact that textbooks aredesigned to teach pupils what educators believe ought to exist—in otherwords, textbooks ‘tell children what their elders want them to know’(FitzGerald 1979: 47). More often than not, adults with an instrumental-rational predisposition take it for granted that textbooks fulfil thisfunction.

Such a position is probably grounded on another tacit assumption—thatchildhood equals dependency (Barker 1993). This assumption has alsostimulated a great part of research on children’s encounters with the massmedia, particularly TV, which has been presumed to be forbiddinglypowerful, and potentially dangerous. Almost always, children have beentreated as subjects in the search for ‘effects’ and as naıve victims in the handsof all-powerful media.

In other words, the assumption of the influence of textbook content onthe readers’ mind undergirds a significant part of a basic thrust in textbookstudies, i.e. textual research. This approach is guided by yet another implicit

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

At M

artin

] at

12:

58 0

6 O

ctob

er 2

014

470 V. KALMUS

assumption: that meaning is determined by the text itself. The authors oftextual studies tacitly grant objective status to the text, and in effect claim atthe same time that their interpretation of the text is the ‘preferred’ one(Taxel 1989: 32). That thrust in textbook studies can also be called the‘ideological research tradition’ (Johnsen 1993: 65–153). Work within thattradition is aimed at discovering political or cultural influence of textbooksby their attentive reading. The most typical representation of such researchis seen in the work of the Georg Eckert Institute in Braunschweig, Germany,particularly in the 1950s to the mid–1980s, steered by the founder’s ‘beliefin the ability of textbooks to promote tolerance and the desire for peace’(Johnsen 1993: 71).

In a smaller number of studies, textbooks are treated as agents ofsocialization, or socializers. In these studies socialization is seen as anongoing dialectical process, a continuous interplay and interaction betweentwo sets of actors, the individuals being socialized and the agents ofsocialization. Such studies (e.g. Heraclides 1980, London 1984, Gruene-berg 1991, Knain 2001) assume that socialization is a cumulative processinfluenced by various agents, events, experiences, and discourses, of whichtextbook discourse is merely one. They also presume that the interpretationof textbook information, and the ways in which it will be accepted, rejected,or added to a pre-existing structure of beliefs, will depend on such factors asthe similarity to existing knowledge and attitudes, the credibility andauthority of the medium, and the context of the situation (Grueneberg1991: 17). As a result, it is difficult to measure the exact influence oftextbooks in the process of socialization. Nevertheless, assuming thattextbooks are part of the communication process which may mould socialbeliefs, attitudes, and values, there are studies which have sought to displaythe knowledge and values in textbooks through textual analysis.

This paper offers an overview of the principal methodological issues inresearch on the role of school textbooks in the process of socialization, andconsiders some research strategies for overcoming those difficulties. To avoiddiffuseness, I focus on political, cultural, and ideological socialization. Bypolitical socialization, I mean the process by which the individual acquiresattitudes, beliefs, and values relating to the political system of which he orshe is a member and to his or her role as a citizen within that political system(Greenberg 1970, cited in Gallatin 1980: 344). Cultural or ideologicalsocialization comprises the acquisition of values, attitudes, and beliefsregarding one’s own group, other groups or nations, war and peace, religion,science, as well as cultural values such as individualism and collectivism.

Problems in studying textbooks as agents of socialization

Assumptions

The tacit assumptions underlying conventional, mainly textual, textbookresearch are problematic for several reasons. First, they neglect the reality ofinterpretation of a textbook’s text by pupils. Depending on social contextand socio-psychological variables such as ethnicity, gender, social class,

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

At M

artin

] at

12:

58 0

6 O

ctob

er 2

014

PUPILS AND TEXTBOOKS 471

personality, and psycho-pathogenesis, individual pupils may reach vastlydifferent interpretations or decodings of one and the same text (Anyon1981). In a process of decoding, various kinds of symbolic appropriation andresistance (Stray 1994: 6) may take place. In Soviet-era Estonia in the 1980s,for instance, some pupils would add moustaches, glasses, etc. to theportraits of Lenin in their textbooks. Second, the assumptions ignore therelationship between pupils and textbooks (as a part of the educationalsystem), and the resulting and different strategies pupils use in encounterswith those educational media. The strategies vary, again depending on socialcontext and socio-psychological variables, and range from the total disuse ofa textbook to learning parts of its content by heart. Observational studies(Anyon 1981, Willis 1981) have shown the many creative, often contra-dictory and at times self-defeating ways in which pupils accept, reject,modify, and transform the various dimensions of the school curriculum.

In other words, by ignoring pupils’ capacities for creative interpretationand strategic handling of the curriculum, researchers underestimatechildren’s roles as active, resistant, and sometimes cynical readers. Severalmedia researchers (Buckingham 1993, 1996, Gauntlett 1995, 1996, Tobin2000) have abandoned this assumption—and found children to be activerather than passive readers or viewers, capable of discussion and criticism of,for instance, various TV programmes.

However, we have to note that, for educational media, the tacitassumption of a direct media influence can be justified, at least in part.School texts are more closed than open in Eco’s (1979) sense: they aim for aparticular and specific reading. Educational texts are institutionally defined,and they are often used in a context where there is one person who knowsthe correct answer (i.e. the teacher) and others (i.e. pupils) who arecontrolled and tested for the correct answer (Selander 1995a: 12). The readerof such texts is not any reader, but one who is role-defined (Selander 1995b: 152);the texts in school textbooks normally have greater ‘authority’ (Olson 1989)than texts which are more open for meaning-making, with the implicationthat the potential for different interpretations is ‘much less than infinite’(Buckingham 1993: 270). Moreover, ideological influence is generallyfacilitated ‘if recipients have no alternative opinions’ (van Dijk 1998: 246)because they lack the relevant social and political knowledge, which is oftenthe case with pupils and textbooks. Furthermore, in some fields ofknowledge, textbooks constitute the sole, and trusted, source of informa-tion, and in such cases they are most likely to affect pupils. In addition,educational texts can provide frameworks to support and develop everydayunderstanding, through which pupils interpret other texts. Like particularTV programmes, textbooks may organize pupils’ frames of thinking inspecific ways, ‘creating dichotomies and oppositions, forming associationsand connotations’ (Livingstone 1990: 30).

Finally, a considerable part of the socializing message in textbooks is ahidden, or unplanned, curriculum (e.g. social representations of genderroles, ethnic minorities, etc.), which may escape the attention of pupils’critical consciousness as it passes through the route of peripheral cognitiveprocessing (Petty et al. 1981). For such processing, source credibility is veryimportant. Because school textbooks possess institutionally defined author-

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

At M

artin

] at

12:

58 0

6 O

ctob

er 2

014

472 V. KALMUS

ity and reliability, their hidden messages may just be taken for granted bypupils.

Textbook research in the light of socialization theories

The problems that flow from these tacit assumptions within textbookresearch require that research on texts and other educational media begrounded on a theory that takes account of the relationships between amedium, i.e. a textbook, individuals, i.e. pupils, and the social contextwithin which the medium is used. Socialization theories serve thispurpose.

Socialization may be defined as ‘the comprehensive and consistentinduction of an individual into the objective world of a society or a sector ofit’ (Berger and Luckmann 1991: 150). Although this definition carries anidea of the guidance by someone who sets ‘the rules of the game’ (Bergerand Luckmann 1991: 154), socialization does not imply simple passivity.Individual characteristics, e.g. cognitive capacity, epistemic motivation, andpreviously-stored knowledge (Bar-Tal and Saxe 1990) as well as ascriptivevariables such as nationality (Gallatin 1980: 362–363) and gender(Westholm et al. 1990) play their parts in the interaction between the personsbeing socialized and the agents of socialization.

One of the most complete models of interactive political socializationhas been proposed by Pawelka (1977). The central element in this model isthe individual being socialized. The second element is the agents of politicalsocialization: the primary group (family), the school, media, peer groups,churches, the employment system, and the political system. The agents ofsocialization embrace various socializers—persons, groups, organizations,objects (e.g. textbooks), events, etc. The third element consists of thedominant subsystem of the society, which, in turn, can be divided into asubsystem in power and a subsystem in opposition. The final elementcomprises the peripheral subsystems of the society (e.g. subcultures), whichhave some significance for the individual being socialized. Pawelka (1977)sees mutually influential relationships among these four elements: betweenthe individual being socialized and the agents of socialization as well as theactive socializers within this larger group of socializing agents; among thedifferent agents of socialization and as well as among their respectivesocializers; between the dominant subsystems and the agents of social-ization; between the peripheral subsystems and the agents of socialization;and between the dominant subsystems and the peripheral subsystems.

Keeping the relationships in Pawelka’s model in mind, we may think ofthe agents of socialization and the societal subsystems as sites for discourses,which penetrate and influence each other and constitute interaction andsocial practice in society (van Dijk 1997b). (Indeed ideological socializationlargely takes place through discourse [van Dijk 1998].) However, while theconcept of ‘discourse’ is central in theories of cultural reproduction andsocial constructivism—which many socialization theories rely upon—it is aquite narrow concept in its emphasis on ‘progression of communications’(Biocca 1991: 45) and language use. Therefore, following Bourdieu (1991,

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

At M

artin

] at

12:

58 0

6 O

ctob

er 2

014

PUPILS AND TEXTBOOKS 473

1998), I suggest that the agents of socialization and the societal subsystemsshould be viewed also as fields—structured social spaces with dominant anddominated social agents and unequal power relations, which are constantlystruggled over. These fields are discursively interrelated. The field of school,for instance, is discursively related to the fields of the political system, themedia, family, and peer-groups (and those fields are interrelated with eachother) when, e.g. a TV debate on a policy of interethnic integration and itsviewing by the pupils’ parents and siblings are discussed in a (multicultural)classroom after the pupils have read the chapter on ethnic minorities in theircivics textbook.

The assumptions revised

Menck (2000, 2001) suggests the conceptualization of all classroom work asfundamentally a process of ‘interpretation’, a process by which two worlds—the symbolic world of textbook text and pictures and the life-worlds ofpupils—are brought together. This conceptualization generalizes the under-standing of reader-response theories, which see literary works as cominginto being in the ‘transaction’ that occurs when reader meets text (Taxel1989: 32). Pupils are seen as active, creative, and dynamic readers engagedin a process of meaning-making. And without ‘meaning’ there can be no‘effect’ (a central assumption in reception analysis [Jensen 1991: 135]).Each pupil interacts with the literary text, and constructs his or her ownmeaning in the social process of reading.

Put in a different way, textbooks are ‘multiply encoded and can bemultiply decoded’ (Stray 1994: 6). Indeed some socio-cognitive researchers(e.g. Wodak 1996) assume that there is no general ‘text’ which is valid for allreaders. Text and context interact, creating various, often contradictory,readings. However, as Menck (2000) also suggests, pupils, as role-definedreaders, can never be totally free in their meaning-making and it may beuseful to assume that meaning-making and, eventually, socialization throughthe interaction of textbooks and pupils is a site of negotiation between twosemi-powerful sources (Livingstone 1990). On the one hand, the ideologicalpower of textbooks is constrained by pupils’ developing ‘critical’ readingskills; on the other hand, the interpretive power of pupils is limited by theclosure and institutional authority of textbooks. In terms of the dialectic ofthe socialization process ‘structure works through agency, and agencythrough structure’ (Buckingham 1993: 273).

The intricacy of such interaction and the complexity of the relationshipsbetween various socializers makes it difficult to estimate the role of specificagents such as school textbooks in any process of socialization. Thus thecognitive-developmental tradition in political socialization research hasenvisioned the entire process as being extremely complex and, as a result,has generally stopped short of trying to study directly the impact ofindividual agents of socialization. Instead, the tradition has explored, forinstance, the growth of political thinking per se (Gallatin 1980).

Nevertheless, the questions around the political and cultural social-ization that takes place through the medium of school textbooks merit

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

At M

artin

] at

12:

58 0

6 O

ctob

er 2

014

474 V. KALMUS

serious interest. And with those questions in mind, the assumptions Ioutlined above set a demand that any analysis of socialization throughtextbooks should focus on both the content of educational media and theprocess by which pupils acquire knowledge, values, and attitudes from thosemedia. In other words, research on textbooks as agents of socialization musttake the form of reception analysis, or audience-cum-content analysis, with itsdistinction between potential and actualized meanings (Jensen 1991:135–137)—in Eco’s (1979) terms, between the ‘virtual’ and the ‘realized’text. And for analyses of ‘effects’, we have to distinguish potential and actualeffects. The latter are, of course, difficult to measure.

Problems and studies into the effects of school textbooks

The methodological discussions I will develop in this section derive largelyfrom my review of research on political socialization, although I alsoconsider studies from social and educational psychology and media studies.This scope reflects my concern for an interdisciplinary approach in textbookresearch.

Investigators interested in the effects or role of textbooks in politicalsocialization have typically employed one of the following methods: pretest-posttest studies assessing pupils before and after they have read a specifictextbook; or comparing pupils who have read a textbook with those whohave not (Gallatin 1980: 360). There are several problems with such studies.Klapper (1960: 5; emphasis added) has noted that media should beregarded ‘as influences, working amid other influences, in a total situation’.The relationships between the influences of textbooks and other influencesof the school, let alone the impact of other fields of socialization, are intense,but hardly measurable. Socializers function concurrently, are linked to oneanother, influence one another, and function in the context of differentsocial structures, cultures, and processes (Dekker 1991). It is impossible toisolate textbooks from other socializers, such as the teacher, the classroomclimate, and peers, in a research design.

A countermeasure researchers can use is to keep other factors constant.For instance, Lichter and Johnson (1969) used a pretest-posttest design,controlling for the teacher, the classroom, the school, and the reading abilityof subjects, to measure the influence of a multi-ethnic reader on theattitudes of white 2nd-grade children toward people of colour. The 34 pupilsin the experimental groups used the multi-ethnic reader, which includedcharacters from several different racial and ethnic groups, for 4 months,whereas the 34 pupils in the control groups used the regular reader thatincluded only whites. Use of the multi-ethnic reader resulted in markedpositive change in the subjects’ attitudes toward people of colour.

Another strategy is to neutralize the influence of other importantsocializers by reducing it to the minimum. The teacher effect, for instance,can be neutralized by many practising teachers in, say, a teaching-practiceschool (Ahonen 1990). Yet another possibility is to measure the impact ofother agents of political socialization, e.g. by asking young people themselvesdirectly who or what they think has exerted the most influence on their

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

At M

artin

] at

12:

58 0

6 O

ctob

er 2

014

PUPILS AND TEXTBOOKS 475

political knowledge and attitudes with respect to some specific politicalissue. For instance, in the Netherlands, de Hond (1981; cited in Dekker1991: 31) asked 18/19-year-olds what channels had led them into contactwith politics, and which channels had influenced them the most. Pro-grammes on radio and TV were claimed to be the most influential; schoolcame in second place, followed by newspapers and magazines, anddiscussions about politics at home with parents. However we have toconsider a possibility that what pupils say has happened and what has in facthappened are not necessarily identical (Dekker 1991).

It should also be noted that in interview or observational studies wenormally engage one pupil at a time. However, political behaviour isrelational in nature, i.e. it refers to relationships between persons. ‘Thisforces the investigator . . . either to attempt to impose some commoncontext on the individuals under study, or to make unchecked ceteris paribusassumptions about the immediate social environment of the behaviorobserved’ (Chaffee et al. 1977: 230). A solution is to use serial sampling byselecting ‘nests’ of pupils (i.e. school classes) and studying all individuals inthe nest. This approach to sampling enables data collection around the nestas such (e.g. sociometric relationships between pupils) and analysis of theinfluence of classmates and teachers on individual pupils. Obviously, thistechnique sets limits to the sample size.

A further complication centres on the difficulty in identifying cause andeffect. Even with adequate measures, the statistical techniques available arelargely correlational. Correlation, however, does not necessarily implycausation. While common sense tells us that behaviour is obviously theproduct of certain persuasive stimuli (i.e. a textbook) and innate character-istics, it is difficult to demonstrate that this is indeed the case (Gallatin1980). The conformity of the political viewpoints of the textbook and pupilscan be due to the influence of other socializers. The absence of conformity ofopinion may be in fact an unexpected result of interaction with the textbook:the individuals being socialized may have formed totally different, reac-tionary opinions (Dekker 1991). One, and perhaps the only technique forovercoming this fundamental difficulty is, again, the use of more direct andintensive approaches: asking the young people themselves who, and what,has exerted the most influence on their political knowledge and attitudes.

Research on textbooks as agents of socialization also involves sample-related difficulties. Children and adolescents as research informants causecommon as well as specific problems of validity and reliability. In terms ofvalidity, a universal question is whether the knowledge, opinions, andattitudes that have been measured are the most relevant ones for discoveringwhat the researchers actually want to know (Dekker 1991). A specificproblem of reliability springs from the fact that pupils are usually tested orinterviewed in an institutional context (i.e. a school), which increases theprobability of obtaining socially desirable answers, the so-called ‘answeringfor an A effect’. Saying ‘this is not a test’ is of little significance: for pupilsany activity in school which has to do with giving answers obviously is a test(Connell 1969; cited in Gallatin 1980: 355). (My experience supports thisassumption: some 9th-formers firmly requested a mark for completing aquestionnaire!) Because of social desirability, there is a tendency to respond

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

At M

artin

] at

12:

58 0

6 O

ctob

er 2

014

476 V. KALMUS

positively in any situation where evaluation is called for (Blanchard andPrice 1971). A partial work-around to inhibit socially acceptable answers topolitics-related questions is testing in classes other than civics or socialstudies.

Another specific phenomenon, and the converse of social desirability,may reduce reliability. Pupils (adolescent boys in particular) may ‘play atrick’ on a researcher by deliberately giving wrong or funny answers, or bychoosing forced-choice items at random. Vaillancourt (1973) surveyedgroups of children and younger adolescents (age 9–15) on three differentoccasions several months apart. She discovered that the correlationsbetween the sets of answers were only moderately positive, leading her tosuspect that many of the subjects had simply been expressing their non-attitudes on the issues. My own experience supports this kind of suspicion:several 15–16-year-old boys, by answering a pilot questionnaire item, ‘Who,according to the Constitution, has the supreme power in Estonia?’,surprised me with ‘witty’ answers such as ‘the underworld’, ‘Me andNATO’, ‘Me and X’, etc.

Studies intending to measure the influence of a textbook on pupils’political knowledge (i.e. ‘how-much-pupils-have-learned studies’) have toacknowledge a further methodological caution. When the same test itemsare used in a pretest-posttest design, inflated growth rates become an issuebecause of repeated exposure. But Westholm et al. (1990) explored thatpossibility by including in the second-wave sample a control group that hadnot previously participated in the study. The results showed that effects dueto repeated exposure existed but were trivial (on average, the second-wavescores were only one percentage point higher than they would have been fora fresh sample).

The solution of some of these methodological problems (e.g. thereliability problem related to subjects ‘playing tricks’) is very difficult, if notout of reach. Other problems can be mitigated by the use of complex designstrategies.

Research and design strategies

An interdisciplinary approach

The field of textbook studies, let alone research on textbooks as agents ofsocialization, would benefit if it developed into a truly interdisciplinaryenterprise. This proposition, seemingly obvious or even commonsensical,has not lost its actuality: the overwhelming majority of textbook studies arestill carried out in the framework of a single approach—predominantlypedagogy.

All too often a mono-disciplinary perspective inhibits understanding ofmethodological difficulties, not to mention ways of overcoming thoseproblems. For instance, it may be difficult to recognize the problems of theunderlying assumptions I reviewed above, or the complexity of relationshipsbetween pupils, agents of socialization, and societal subsystems. Studies thatignore these intricacies can, at best, provide partial insights.

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

At M

artin

] at

12:

58 0

6 O

ctob

er 2

014

PUPILS AND TEXTBOOKS 477

What is needed is an approach that combines the perspectives of severaldisciplines—pedagogy, sociology, media studies, political science, cognitivescience, semiotics, discourse studies, and reception analysis,2 to name themost important. Cognitive science, for instance, has enriched research onpolitical socialization and media reception with its concept of schema—acognitive structure, constructed by an individual, that organizes previouslyacquired information and has an impact on remembering and retrievinginformation and using it for solving problems (Torney-Purta 1990). Theconcept and the related theory should be more extensively used in researchon pupils’ interaction with textbooks.

In a model study within communications research, Livingstone (1990)draws on the concepts of schema, active viewer, and the open text for ananalysis of audience interpretation of soap operas. Her theoretical founda-tions come from social psychology (specifically, social cognition and socialrepresentations), discourse theory, semiotics, and media studies. Shecombines structural literary analysis of three soap operas (Dallas, CoronationStreet, and EastEnders) with multidimensional scaling analysis of viewers’similarity judgements and free sorting of soap characters into similar groups(a typical research design in experimental social psychology). One of hermain conclusions is that viewers’ social knowledge of real-life people and thestructure of soap operas as text both determine viewers’ representations ofsoap characters—the resultant representation depends on the nature ofinteraction or negotiation between the text and the viewers’ knowledge.

Livingstone’s research design and conclusions would have been hardlyimaginable within a mono-disciplinary approach. But in textbook researchthe potential resources of cognitive science, discourse studies, and receptionanalysis are, to a great extent, still unemployed. The study of textbooks asagents of socialization will be enriched by borrowing from several relevantdisciplines in terms of theories, concepts, and methods.

Combination of different methods

Several reviews of research (e.g. Dekker 1991) have noted that aggregatetechniques (tests and written questionnaires, in particular) are the methodsused most often for obtaining data in research on political socialization.These methods are relatively inexpensive, they allow for large representativesamples, and the data obtained can be processed relatively quickly. Howeversuch methods are often based on problematic assumptions, and may involveall the problems discussed above. The basic concern is that tests andquestionnaires do not allow the identification of cause and effect with anycertainty, nor do they permit the researcher to understand what goes on inthe pupils’ heads when they interact with the textbook. To overcome thoseshortcomings, critics have suggested the supplementary use of more directand intensive methods, such as questionnaires with open-ended questions,longitudinal follow-ups, extended or in-depth interviews, systematic (class-room) observation, experimental research, and autobiographical storytelling(Gallatin 1980, Dekker 1991). Additional approaches may include focusgroups, records of group encounter sessions, ‘think-aloud problem-solving

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

At M

artin

] at

12:

58 0

6 O

ctob

er 2

014

478 V. KALMUS

techniques’ (Torney-Purta 1990), and the analysis of pupils’ productions(essays, drawings, etc.).

The combination of methods is not a novelty in the fields of politicalsocialization and media reception. For instance, in an early study of politicalsocialization of 4th- through 8th-grade children, Greenstein (1965) used apaper-and-pencil questionnaire, loosely structured interviews, extensiveinformal contact with children and their teachers, and secondary analysis ofliterature on children’s social development. The major source of findings wasnevertheless the questionnaire: the interviews served as a preparation todesigning the questionnaire and as a check on its reliability and validity.Greenstein, however, acknowledged the need to develop a wide range ofinstruments and measures for studies of children’s political learning.

In the field of children’s media reception, Palmer (1986) combinedloosely structured interviews with individual children, lengthy observationof their TV viewing in the home environment, and a questionnaire survey.The interview component of the study showed that the verbal responses ofchildren can tell a researcher what they really think and feel about the mediain greater detail and with much greater precision than a set of questionnaireresponses. The survey was mainly used to test the generalizability of certainfindings.

Each of these methods has its own strengths and weaknesses, and eachof them is tailored for particular problems and aspects of research. Itherefore suggest the combination, or triangulation,3 of different methods inany single study of textbooks as agents of socialization—presuming that theadvantage of one method compensates for the shortcoming of another.Research designs that combine quantitative and qualitative or statistical andintensive approaches obviously have the greatest potential. For instance, mycurrent study of civics textbooks as an agent of political socialization(Kalmus 2002, 2003) includes quantitative as well as qualitative analysis ofthe textbooks’ text, a panel questionnaire survey of pupils (a pretest-posttestdesign), the analysis of pupils’ essays written after the encounter withrelevant parts in the textbook, and semi-structured interviews with a smallsample of the essay-writers and the teachers. Quantitative content analysisenabled me to obtain the exact frequencies of different civic and politicalactivities depicted in the textbooks. Qualitative text analysis was instru-mental in mapping the interpretative schemata and general ideologyprovided by those educational media. The panel survey was indispensablefor measuring changes that took place in pupils’ political knowledge andattitudes in the course of one school year. The pupils’ essays allowed themto demonstrate their knowledge of and attitudes towards timely politicalproblems, and provided the written interpretations of the myriad of textsthey had encountered in the process of political socialization, thus enablingme to trace the intertextual references to different discourses, which can beassumed to reflect the relevance of the corresponding agents of socializationas sites for those discourses. The essay method proved to be reliable; a widerange of viewpoints on sensitive political issues, instances of politicalincorrectness, and the use of slang can obviously be interpreted as a sign ofpupils’ frankness. The ‘effect of answering for mark A’ in terms of theexpressed attitudes seemed to be rare. Finally, the interviews gave me a

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

At M

artin

] at

12:

58 0

6 O

ctob

er 2

014

PUPILS AND TEXTBOOKS 479

chance to ask pupils directly whence they might have derived the thoughtsthey had expressed in the essays, how they used their civics textbooks, etc.The interviews shed light on the relative importance of various agents in theprocess of political socialization (the role of the civics textbook appeared tobe rather modest) and helped me to interpret the similarities anddiscrepancies between the textbook discourse and the pupils’ attitudes. As aresult, the combined research design mitigated the main problems describedabove. Such studies, with their attempts to grasp the totality of a segment ofpolitical socialization, approach a kind of a research cycle—the next strategyto be suggested.

A cycle of research

The theoretical and methodological discussion I have been outliningprompts rethinking the role of school textbooks in transmitting knowledgeand values to the young. Thus I argued, first, that we presume that schooltextbooks are but one socializer among many others in the fields ofsocialization, which interact with each other as well as with the individualbeing socialized. In the course of that interaction pupils gradually constructtheir images of the political scene or the value system of the society. Thesocializing effect of a textbook depends on the knowledge and beliefsexisting in pupils’ minds as well as on the availability and authority ofalternative discourses. But while every pupil interacts with the textbook toconstruct his or her own meaning in the process of reading and learning, theauthoritative institutional context of the situation lends its power to thetextbook text, which constrains the range of possible interpretations.

I argued, second, that the interaction between the pupil and thetextbook ought to be seen as negotiation between two semi-powerful agents.Both, obviously, deserve attention in research on textbooks as agents ofsocialization. This proposition, and the methodological problems aroundstudies into the effects of particular socializers, set a need for researchapproaches that combine sophisticated insights into the discourse oftextbooks and the process of interaction between the pupils and educationalmedia. We must unite the methodologies of systematic and detailed textualanalysis and reader-response studies into a cycle of reception analysis, inorder to secure comparative insight to the textbook discourse and thediscourse by pupils which results from their interaction with the textbook.Such an approach should also include analyses of the production of textbooksas institutional texts which are intended to be effective.

Such an approach is virtually non-existent in textbook research. An earlystudy by Litt (1966) is one of the very few examples. Litt studied the processof political indoctrination through civic education programmes in threedifferent socio-economic communities. He used three measures of civiceducation norms: content analysis of all textbooks used in the civiceducation programmes in the three schools; interviews with ‘potential civicand educational influentials’ (i.e. school administrators, teachers, presidentsof major civic groups, etc.); and a questionnaire given to the three civiceducation classes and their control groups before and after a semester’s

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

At M

artin

] at

12:

58 0

6 O

ctob

er 2

014

480 V. KALMUS

exposure to the civics course. His main findings were that the civicsprogramme did not simply reinforce the prevailing opinions and politicalclimate of the community, nor were attitudes about political participationand activities influenced by courses in civic education. Litt’s study isremarkable for its emphasis on the socio-economic and political context ofsocialization. It is, however, imperfect in its lack of attention to pupils’interpretations of the civics programme and textbooks.

In a recent study of pupils’ interpretations and use of science textbooks,Knain (2002) combined critical text analysis of a sample of texts on thetopic of energy in two textbooks and semi-structured interviews with 1216-year-old students in two different tracks (academic and vocational) intwo upper-secondary schools in Norway. The interview guide was structuredto identify relationships between the students’ interests, their views on thepurpose of school science, the use of science textbooks, and the inter-pretations of the texts in the sample. It appeared that both the academic-track and the vocational-track students were satisfied with their textbooks,but for different reasons. The academic-track pupils drew on the schoolcontext before personal interests or everyday life; they talked about thetextbook in terms of how effective it was for learning purposes at school andfor further studies. For the vocational-track pupils, the textbook as such wasnot important. The interviews suggested that their perspectives on the worldwere not addressed in the textbook culture of school science, but the pupilsdid not ask that those perspectives be addressed. This study, however, doesnot explain how and why science textbooks are produced with such asocializing potential.

To answer such questions, Taxel (1989) has suggested that analyses oftexts and inquiries into the responses of readers should be joined into abroad ‘cycle of research which would account for the creation of the textitself by an author and its subsequent production and distribution by thepublishing industry’ (p. 35). At the culmination of the cycle of research therewould be ‘the analysis of precisely what occurs when readers, especially childreaders, actually read books. What is sought is an understanding of thedynamics of the transaction which takes place when individuals interact withliterary works’ (p. 39). Three phases can be identified in Taxel’s cycle:analysis of the creation of texts; analysis of texts; and analysis of response totexts.

Although advocating Taxel’s ‘cycle of research’, I would urge that it becomplemented with concepts from the relevant disciplines. A research cycle(especially the first two phases) can be directed by the principles of criticaldiscourse analysis, in particular by those of the discourse-historical methodof the Vienna School (Wodak and her associates, see Titscher et al. 2000,Wodak 2001). In advocating an interdisciplinary approach, Wodak4 hasdeveloped a socio-psycholinguistic ‘theory of text planning’ that can informthe first procedure in a cycle, the analysis of the creation of the textbook textby an author. The theory helps to identify the (ideological) intentions ofauthors and the extralinguistic (contextual) factors in text production: timeand place, sociological variables (i.e. group membership, age, professionalsocialization), and psychological determinants (i.e. experience, routine,etc.). Frames and schemata for the structuring of reality and for the concrete

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

At M

artin

] at

12:

58 0

6 O

ctob

er 2

014

PUPILS AND TEXTBOOKS 481

realization of a text derive from this socio-psychological preconditioning(Titscher et al. 2000).

The research cycle’s second phase—the description of texts—must beaccomplished as precisely as possible at all linguistic levels. The discourse-historical method makes a distinction between three analytical dimensions:content, argumentation strategies, and forms of linguistic implementation(at text, sentence, and word levels). The analysis of text itself should beunderstood ‘not as a sequence of separate operational steps but as a cycle inwhich the three analytical dimensions are systematically and recursivelyrelated to the totality of contextual knowledge’ (Titscher et al. 2000:158).

The concept of context is of central importance in critical discourseanalysis. Context should be recorded as accurately as possible. In the field ofschool textbooks, the description of context has to take into account:

� the text-internal co-text and the local interactive processes (theclass and the classroom situation);

� the intertextual and interdiscursive relationships between thetextbook texts and other texts, genres, and discourses;

� the text-external social/sociological variables and institutionalframes (the school); and

� the broader socio-political and historical context in which thepractices of textbook production and consumption are embedded,as well as the history to which the textbook topics are related(Titscher et al. 2000: 157).

The research cycle’s third phase aims to understand what takes placewhen readers (i.e. pupils) interact with the text—in terms of politicalsocialization, how they interpret, understand, and internalize the politicalcontent of the textbook. For this task the integrative model of textcomprehension developed by the Vienna School of critical discourse analysisoffers some useful analytical tools. Wodak and her colleagues contend thattext comprehension is also dependent upon the socio-psychological influ-ences that are important in text production (Titscher et al. 2000). All pupilsperceive new texts on the basis of their previous knowledge, personalqualifications, and culture, gender, and class membership. The degree ofconcurrence between such ‘reader’s schema, or organized knowledge of theworld’ (Armbruster and Anderson 1984: 181) and the text’s schemadetermines how well the pupil will understand the content of the textbook,and to what extent he or she internalizes it. Moreover, pupils approach thetext with various strategies that, in being, again, dependent upon the socio-psychological characteristics of pupils, sometimes ‘work in an unconscious,irrational and emotional way’ (Titscher et al. 2000: 156). Therefore, thethird phase of a research cycle should encompass both the cognitive and theemotional aspects of the process of socialization. The methods suitable forthat purpose coincide with the intensive approaches listed above (i.e. in-depth interviews, the analysis of pupils’ productions, etc.).

The overall aim of such a cycle of research in political socialization is tounderstand as deeply as possible how pupils come to understand thepolitical world and act in it, and the role that textbooks play in that process.

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

At M

artin

] at

12:

58 0

6 O

ctob

er 2

014

482 V. KALMUS

As a form of reception analysis, a cycle of research can be seen as acomparative analysis of textbook discourses and pupils’ discourses, ‘whoseresults are interpreted with emphatic reference to context, both thehistorical as well as cultural setting and the “con-text” of other mediacontents’ (Jensen 1991: 139). That kind of analysis should enable aresearcher to get a grasp of ‘what pupils and textbooks do with each other’in the process of socialization.

Coda: problems never end

The measures and strategies suggested above mitigate or, hopefully,surmount many of the methodological problems involved in research onschool textbooks as agents of socialization. However, the scope andcharacter of the research strategies, particularly those around a cycle ofresearch, create new problems, mainly technical and ethical. First, theimplementation of these strategies demands great amounts of time andlabour. Second, most of the intensive methods suggested set limits to samplesizes that, in turn, create a problem of generalizability. We, therefore, have tomake a special effort to select the most typical group, or ‘nest’, of pupils.Third, intensive measures, such as classroom observation, experimentalresearch, and records of group-encounter sessions, may affect the naturalprocesses of learning and socialization, and thus give biased results andcreate ethical problems. (Classroom observers, however, are usuallyconvinced that their presence in the classroom does not influence theteaching [Johnsen 1993].) Finally, intensive and time-consuming methodsgive rise to the problem of motivating teachers and pupils to participate ina study. Although teachers may find motivation in an opportunity toimprove their teaching on the basis of research findings, pupils see nocompensation for their work. Explanations of the type ‘results are used toimprove the textbooks and, in a long run, society’ are hardly convincing forsceptical teenagers. Grades obviously cannot be given.5 In other words,there are methodological problems in social sciences that always remainpartly unsolved! However, such problems do not mean that significant, andintricate, sociological phenomena, such as socialization through textbooks,are not worthy of serious exploration.

Notes

1. By ‘educational media’ I mean school textbooks in particular, but also the whole range ofpossible texts, pictorial illustrations, films, computer programs, etc., produced foreducational purposes, i.e. to inform and convince the reader/viewer/listener that thepresented information and perspectives are correct (Selander 1995a).

2. Some of the areas I mention (e.g. reception analysis and discourse studies) are themselvesundergoing interdisciplinary development (Jensen 1991, van Dijk 1997a).

3. The ‘combination of methods’ means the use of different methods in one and the samestudy, i.e. the different methods are used for the analysis of the different aspects of thephenomenon being studied. ‘Triangulation’ means the comparison of results on the basisof different data (e.g. qualitative and quantitative) and using different methods (Titscheret al. 2000).

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

At M

artin

] at

12:

58 0

6 O

ctob

er 2

014

PUPILS AND TEXTBOOKS 483

4. Drawing on cognitive science, Wodak and her associates view the discourse-historicalmethod, as hermeneutic and interpretative, to be part of the research background tosociolinguistics and text linguistics (Titscher et al. 2000).

5. A strategy I have used to motivate pupils to write substantial political essays is to give‘pluses’ or ‘small marks’ (which are summed into real marks at the end of the term by theteacher) for the length and quality of argumentation, regardless of the content anddirection of statements. The strategy justified itself: I received several solid essaysrepresenting a wide variety of opinions on sensitive political issues. The ‘effect ofanswering for mark A’ in terms of the expressed attitudes was seen in only a few cases. Noone, however, can say with any certainty this effect would not have been there, if nomotivating stimuli had been used.

References

AHONEN, S. (1990) The form of historical knowledge and the adolescent conception of it. ResearchReport 80, Department of Teacher Education, University of Helsinki.

ANYON, J. M. (1981) Social class and school knowledge. Curriculum Inquiry, 11(1), 3–42.ARMBRUSTER, B. B. and ANDERSON, T. H. (1984) Structures of explanations in history

textbooks, or so what if Governor Stanford missed the spike and hit the rail? Journal ofCurriculum Studies, 16(2), 181–194.

BARKER, M. (1993) Sex, violence and videotape. Sight and Sound, 3(5), 10–12.BAR-TAL, D. and SAXE, L. (1990) Acquisition of political knowledge: a social-psychological

analysis. In O. Ichilov (ed.), Political Socialization, Citizenship Education, and Democracy(New York: Teachers College Press), 116–133.

BERGER, P. L. and LUCKMANN, T. (1991 [1966]) The Social Construction of Reality: A Treatisein the Sociology of Knowledge (London: Penguin).

BIOCCA, F. (1991) The analysis of discourses within the political ad. In F. Biocca (ed.),Television and Political Advertising, Vol. 2: Signs, Codes, and Images (Hillsdale, NJ:Erlbaum), 45–59.

BLANCHARD, E. B. and PRICE, K. C. (1971) A developmental study of cognitive balance.Developmental Psychology, 5(2), 344–348.

BOURDIEU, P. (1991) Language and Symbolic Power, ed. J. B. Thompson, trans. G. Raymondand M. Adamson (Oxford: Polity Press).

BOURDIEU, P. (1998) On Television, trans. P. P. Ferguson (New York: New Press).BUCKINGHAM, D. (1993) Children Talking Television: The Making of Television Literacy (London:

Falmer).BUCKINGHAM, D. (1996) Moving Images: Understanding Children’s Emotional Responses to

Television (Manchester, UK: Manchester University Press).CHAFFEE, S. H., JACKSON-BEECK, M., DURALL, J. and WILSON, D. (1977) Mass

communication in political socialization. In S. A. Renshon (ed.), Handbook of PoliticalSocialization: Theory and Research (New York: Free Press), 223–258.

DEKKER, H. (1991) Political socialization theory and research. In H. Dekker and R.Meyenberg (eds), Politics and the European Younger Generation: Political Socialization inEastern, Central and Western Europe, European Studies, 1 (Oldenburg, Germany:Bibliotheks- und Informationssystem der Universitat Oldenburg), 15–58.

ECO, U. (1979) The Role of the Reader: Explorations in the Semiotics of Texts (Bloomington, IN:Indiana University Press).

FITZGERALD, F. (1979) America Revised: History Schoolbooks in the Twentieth Century (NewYork: Vintage Press).

GALLATIN, J. (1980) Political thinking in adolescence. In J. Adelson (ed.), Handbook ofAdolescent Psychology (New York: Wiley), 344–382.

GAUNTLETT, D. (1995) Moving Experiences: Understanding Television’s Influences and Effects(London: John Libbey).

GAUNTLETT, D. (1996) Video Critical: Children, the Environment and Media Power (Luton, UK:University of Luton Press).

GREENSTEIN, F. I. (1965) Children and Politics (New Haven, CT: Yale University Press).

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

At M

artin

] at

12:

58 0

6 O

ctob

er 2

014

484 V. KALMUS

GRUENEBERG, R. (1991) The Representation of Minorities in Israeli Civics Textbooks. Master’sdissertation, University of Amsterdam, Amsterdam, The Netherlands.

HERACLIDES, A. (1980) Socialization to conflict. Greek Review of Social Research, 12(38),16–42.

JENSEN, K. B. (1991) Reception analysis: mass communication as the social production ofmeaning. In K. B. Jensen and N. W. Jankowski (eds), A Handbook of QualitativeMethodologies for Mass Communication Research (London: Routledge), 135–148.

JOHNSEN, E. B. (1993) Textbooks in the Kaleidoscope: A Critical Survey of Literature and Researchon Educational Texts (Oslo: Scandinavian University Press).

KALMUS, V. (2002) Do textbooks teach democratic values: a reception study of a ninth gradecivics textbook. In J. Mikk, V. Meisalo, H. Kukemelk and M. Horsley (eds), Learningand Educational Media: The Third IARTEM Volume (Tartu, Estonia: Tartu UniversityPress), 152–159. http://www.leeds.ac.uk/educol/documents/00002108.htm (visited 10July 2003).

KALMUS, V. (2003) ‘Is interethnic integration possible in Estonia?’: ethno-political discourseof two ethnic groups. Discourse & Society, 14(6), 667–697.

KLAPPER, J. T. (1960) The Effects of Mass Communication (New York: Free Press).KNAIN, E. (2001) Ideologies in school science textbooks. International Journal of Science

Education, 23(3), 319–329.KNAIN, E. (2002) Students’ interpretations and use of science textbooks in school science

discourses. In J. Mikk, V. Meisalo, H. Kukemelk and M. Horsley (eds), Learning andEducational Media: The Third IARTEM Volume (Tartu, Estonia: Tartu University Press),55–61.

LICHTER, J. H. and JOHNSON, D. W. (1969) Changes in attitudes toward negroes of whiteelementary school students after use of multiethnic readers. Journal of EducationalPsychology, 60(2), 148–152.

LITT, E. (1966) Civic education, community norms, and political indoctrination. In E. Litt(ed.), The Political Imagination: Dialogues in Politics and Political Behavior (Glenview, IL:Scott, Foresman), 486–494.

LIVINGSTONE, S. M. (1990) Making Sense of Television: The Psychology of AudienceInterpretation (Oxford: Pergamon Press).

LONDON, G. (1984) Finnish American primers: the ABCs of conservative and socialistideologies. Oulun Yliopisto Historian Laitos Erikoispainossarja, 111, 169–205.

MENCK, P. (2000) Looking into Classrooms: Papers on Didactics (Stamford, CT: Ablex).MENCK, P. (2001) The formation of conscience: a lost topic of Didaktik. Journal of

Curriculum Studies, 33(3), 261–275. http://faculty.ed.uiuc.edu/westbury/JCS/Vol33/Menck.html (visited 15 May 2003).

MIKK, J. (2000) Textbook: Research and Writing (Frankfurt am Main: Peter Lang).OLSON, D. R. (1989) On the language and authority of textbooks. In S. de Castell, A. Luke

and C. Luke (eds), Language, Authority and Criticism: Readings on the School Textbook(London: Falmer), 233–244.

PALMER, P. (1986) The Lively Audience: A Study of Children Around the TV Set (Sydney,Australia: Allen & Unwin).

PAWELKA, P. (1977) Politische Sozialisation (Wiesbaden, Germany: AkademischeVerlagsgesellschaft).

PETTY, R. E., OSTROM, T. M. and BROCK, T. C. (eds) (1981) Cognitive Responses in Persuasion(Hillsdale, NJ: Erlbaum).

SELANDER, S. (1995a) Pedagogic texts—educational media. Internationale Schulbuch-forschung, 17(1), 9–23.

SELANDER, S. (1995b) Research on pedagogic texts: an approach to the institutionally andindividually constructed landscapes of meaning. In P. Skyum-Nielsen (ed.), Text andQuality: Studies of Educational Texts (Oslo: Scandinavian University Press), 152–169.

STRAY, C. (1994) Paradigms regained: towards a historical sociology of the textbook. Journalof Curriculum Studies, 26(1), 1–29.

TAXEL, J. (1989) Children’s literature: a research proposal from the perspective of thesociology of school knowledge. In S. de Castell, A. Luke and C. Luke (eds),Language, Authority and Criticism: Readings on the School Textbook (London: Falmer),32–45.

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

At M

artin

] at

12:

58 0

6 O

ctob

er 2

014

PUPILS AND TEXTBOOKS 485

TITSCHER, S., MEYER, M., WODAK, R. and VETTER, E. (2000) Methods of Text and DiscourseAnalysis, trans. B. Jenner (London: Sage).

TOBIN, J. (2000) ‘Good Guys Don’t Wear Hats’: Children’s Talk About the Media (New York:Teachers College Press).

TORNEY-PURTA, J. (1990) From attitudes and knowledge to schemata: expanding theoutcomes of political socialization research. In O. Ichilov (ed.), Political Socialization,Citizenship Education, and Democracy (New York: Teachers College Press), 98–115.

VAILLANCOURT, P. M. (1973) Stability of children’s survey responses. Public OpinionQuarterly, 37(3), 373–387.

VAN DIJK, T. A. (1997a) The study of discourse. In T. A. van Dijk (ed.), Discourse as Structureand Process: Discourse Studies: A Multidisciplinary Introduction, Vol. 1 (London: Sage),1–34.

VAN DIJK, T. A. (1997b) Discourse as interaction in society. In T. A. van Dijk (ed.), Discourseas Social Interaction: Discourse Studies: A Multidisciplinary Introduction, Vol. 2 (London:Sage), 1–37.

VAN DIJK, T. A. (1998) Ideology: A Multidisciplinary Approach (London: Sage).WESTHOLM, A., LINDQUIST, A. and NIEMI, R. G. (1990) Education and the making of the

informed citizen: political literacy and the outside world. In O. Ichilov (ed.), PoliticalSocialization, Citizenship Education, and Democracy (New York: Teachers College Press),177–204.

WILLIS, P. E. (1981 [1977]) Learning to Labour: How Working Class Kids Get Working Class Jobs(Aldershot, UK: Gower).

WODAK, R. (1996) Disorders of Discourse (London: Longman).WODAK, R. (2001) The discourse-historical approach. In R. Wodak and M. Meyer (eds),

Methods of Critical Discourse Analysis (London: Sage), 63–94.WOODWARD, A., ELLIOTT, D. L. and NAGEL, K. C. (1988) Textbooks in School and Society: An

Annotated Bibliography and Guide to Research (New York: Garland).

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

At M

artin

] at

12:

58 0

6 O

ctob

er 2

014