This article was downloaded by: [Kungliga Tekniska Hogskola]On: 07 October 2014, At: 05:51Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Research Papers in EducationPublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/rred20

Socratic questioning in the PaideiaMethod to encourage dialogicaldiscussionsMaree Daviesa & Anne Sinclaira

a Faculty of Education, University of Auckland, Auckland, NewZealand.Published online: 14 Nov 2012.

To cite this article: Maree Davies & Anne Sinclair (2014) Socratic questioning in the PaideiaMethod to encourage dialogical discussions, Research Papers in Education, 29:1, 20-43, DOI:10.1080/02671522.2012.742132

To link to this article: http://dx.doi.org/10.1080/02671522.2012.742132

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

Socratic questioning in the Paideia Method to encouragedialogical discussions

Maree Davies* and Anne Sinclair

Faculty of Education, University of Auckland, Auckland, New Zealand

(Received 11 December 2011; final version received 17 October 2012)

This study focused on the impact of using Socratic questioning, based on thePaideia Method, on the nature of middle-schools students’ patterns of interactionand on the cognitive complexity of their discussions. The hypothesis is that anexperimental group will increase in both interaction focus and complexity at T3,which is the face-to-face seminar when compared to T1 (baselines) and that thisincrease will be above normative increases compared to a control group. Aquasi-experimental method was employed because, although the Paideia Methodwas not controlled by the researchers, the researchers did have some controlover when to measure the outcome variables. Using SPSS 18.0, a series oft-tests and ANOVAs were conducted to analyse data first for interaction focusand then for complexity to test for differences between the experimental controlgroups. The study was conducted in 12 experimental and 12 control classroomsacross six schools, in New Zealand, totalling 720 students (ages 11–13). Resultssuggest that the experimental group increased in student-to-student focus andcomplexity of discussion above a normative increase with the greatest levelbeing in the Paideia Seminar (T3).

Keywords: Paideia Method; Socratic questioning; dialogical discussion

Introduction

As with many countries which have promoted curriculum reform, the principles ofthe newly gazetted 2007 New Zealand Curriculum put ‘students at the centre ofteaching and learning, asserting that they should experience a curriculum that engagesand challenges them and is forward looking and inclusive’ (Ministry of Education2007, 9). Within this document, eight effective pedagogies are outlined which furtherillustrate teaching practice, which places students at the centre of their learning. Oneof these pedagogies is ‘Facilitating shared learning’. Teachers are advised that

Students learn as they engage in shared activities and conversations with other people,including family members and people in the wider community. Teachers encouragethis process by cultivating the class as a learning community. In such a community,everyone, including the teachers, is a learner; learning conversations and learning part-nerships are encouraged. (Ministry of Education 2007, 34)

The New Zealand Curriculum document was largely uncontested by schoolsand some principals of schools began to seek out pedagogy which would be in

*Corresponding author. Email: mj.davies@auckland.ac.nz

Research Papers in Education, 2014Vol. 29, No. 1, 20–43, http://dx.doi.org/10.1080/02671522.2012.742132

� 2012 Taylor & Francis

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

keeping with its intentions. A timely anecdotal incident prompted one principal toapproach the researchers for advice on such student-centred pedagogy. The presentproject partly arose from a practicum visit to an Intermediate school (Years 11–13)to observe University Graduate Primary Students, where the principal mentioned tous his concern at the level of teacher talk within the classrooms. Whilst on a recentschool trip to Japan, a student had voiced to the principal how ‘bored’ he was thisyear because ‘the teacher talked too much’. The proposition tested in this paper isthat an intervention, the Paideia Method, could assist teachers in providing anopportunity for more active participation by students through dialogic opportunitiesin class discussions. In doing so, the shift in pedagogy towards student-initiated dia-logue would be meeting the ‘spirit’ of the New Zealand Curriculum document andincreasing complexity of thinking.

Opinions currently differ on the theoretical underpinning of dialogical discus-sions. Social constructivist and sociocultural theory are routinely invoked to explainthe role of discussion in promoting students’ understandings of discourse and text(Murphy et al. 2009). However, Alexander (2006) does not think that ‘dialogic’ fitseasily with Vygotskian theory in a broad ‘sociocultural’ paradigm. He argues that itis Socrates who we should be looking at as a way of explanation of dialogic.Alexander claims that Socrates’ objection to writing appears to stem from a view ofthinking as essentially tied to the context of face-to-face dialogue such that writingwords down to carry beyond the context in which they are spoken will destroyliving thought, leaving only a kind of shadow or ‘ghost’ of thought (Alexander2006). Similarly, Rupert Wegerif has had a change of heart when explaining thetheoretical underpinning of dialogism. He developed and proposed a socioculturalapproach to teaching and thinking through teaching ways of talking (Wegerif andDawes 2004 in Wegerif 2010). Now, however, he no longer thinks that themetaphor of learning to use cultural tools that the sociocultural approach (Rogoff,Gauvain, and Ellis 1991) relies upon can understand the creative kind of thinkingwe learn through engaging in dialogue. He, too, proposes that we need a newmetaphor that captures something of what Socrates was saying about the primacyof the dialogic relationship, but develops the original dialogic account of thinkingin a way that can account for the role of technology in mediating dialogue (Wegerif2010).

Scaffolding, which is a term first coined by Wood, Bruner and Ross in the1970s (Wood, Bruner, and Ross 1976) in the context of mother–child interactionand now more commonly applied to what goes on – or does not – in classrooms.This is the use of carefully structured interventions to bridge what Vygotsky calledthe ‘zone of proximal development’, or the gap between the child’s existing knowl-edge and ways of solving problems unaided and the understanding, which can beattained only with the guidance of the teacher or a ‘more capable peer’ (Alexander2006). That is, one’s reasoning is necessarily a response to what has been said orexperienced as well as an anticipation of future social experiences (Murphy,Wilkinson, and Soter 2011). Mercer (2000) suggested we turn Zone of ProximalDevelopment (ZPD) with reference to dialogical discussions into a more open andmultidirectional ‘intermental development zone’ (IDZ) where ‘interthinking’ canoccur between peers without the assumption of a teacher leading a learner. This isclearly a move in a more dialogic direction, but the notion of ‘dialogic space’ goesfurther again, in that it is not primarily conceptualised as a ‘mediating means’ sup-porting cognitive development but as an end in itself. The point of education is not

Research Papers in Education 21

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

to use dialogue to achieve something other than dialogue, as notions of ZPD orIDZ imply, but to enter more deeply and fully into dialogue (Wegerif 2010).

Dialogic teaching has increasingly been the subject of increasing discussion inthe last few years (Lyle 2008). There have been, internationally, a number of uni-versity-led attempts to promote dialogical discussions: these projects have differedin their emphases and in the age range for which they are intended, but all givecentral importance to harnessing and building on students’ interests, encouraginginquiry through practical investigation and the critical use of textual material, andthe adoption of a dialogic stance to the co-construction of knowledge (Wells 2007).It is worth mentioning that, within these studies, some key findings have influencedthinking and writing around dialogical discussions. In the 1970s, Barnes and Todd(Barnes and Todd 1977) undertook an important early study in how children talkwhile working together in school. It involved secondary age children but theirinsights have informed much other research since, including primary research.Barnes and Todd (1995) suggest that pupils are more likely to engage in open,extended discussion and argument when they are talking with their peers outsidethe visible control of their teacher and that this kind of talk enabled them to take amore active and independent ownership of knowledge.

Rojas-Drummond et al. (2006) study in Mexico on writing a joint text foundthat the more dialogic relationship in groups thinking better together could be seenbest not through the use of logical connectors such as ‘because’ and ‘therefore’ butin the increasing prevalence of admissions of uncertainty, asking for advice andindividuals changing their minds in the face of evidence. Rojas-Drummond et al.(2006) found that, in collaborative reasoning discussions where a text was dis-cussed, the teacher’s rate of talk is on average less than his or her rate of talk dur-ing conventional classroom discussions. These studies, (Chinn, Anderson andWaggoner 2001; Murphy et al. 2009) have also shown that the student rate of talk-ing almost doubles during collaborative reasoning discussions, as compared to base-line discussions in the same classrooms. In addition, there is evidence of morecognitive processes associated with improved learning and problem-solving, includ-ing a significantly higher rate of providing explanations and elaborating ideas bylinking them to prior knowledge and ideas.

An intensive line of research by Mercer and his colleagues (Wegerif, Mercerand Dawes 1999) established that exploratory talk is the most effective type of con-versation to solve problems through cooperation. This type of talk demonstrated itsvalue for promoting knowledge construction, given that it promoted reasoningthrough language, facilitating understanding and problem-solving. In addition,exploratory talk is associated to ‘educated’ discourse because it gives rise to con-structive criticism and the formulation of well-argued proposals, which are valuedin many academic and social contexts in many cultures. Groups that became moresuccessful at thinking together after teaching the ground rules of exploratory talkshifted away from initial fixed identity positions, where either individuals identifiedwith a self-image and sought to win the argument (disputational talk) or they identi-fied with a harmonious group image and resisted any kind of questioning or criti-cism (cumulative talk) towards something new, which was described as identifyingwith the space of dialogue itself (Wegerif and Mercer 1997). This was a feature ofboth Barnes and Todd (1995) and Mercer’s (2000) definition of exploratory talk, asit was less about explicit reasoning than about an improved quality of their grouprelationship (Wegerif 2010).

22 M. Davies and A. Sinclair

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

The researchers chose to use the Paideia Method because a meta-analysis ofinterventions, which attempted to promote dialogical discussions, namely Philoso-phy for Children, Collaborative Reasoning and the Paideia Method signalled thatwithin a critical-analytic stance, Paideia Seminars showed moderate effects on stu-dent talk, teacher talk and scriptally implicit comprehension in multiple-group stud-ies (Murphy et al. 2009). It was deemed by the researchers that the Paideia Methodcould be used across curricula and was not so bound by text as many of the otherdialogical discussion interventions were. The researchers were cognizant of utilisingan intervention, which would not alter greatly the impact on the prescribed curricu-lum studies of the school. The teachers in intermediate schools in New Zealandteach the students most of the curriculum subjects, with the exceptions being Tech-nology subjects, so it was possible for the teachers to imbed the Paideia Method into a curriculum area which did not disrupt the normal programme.

What was of interest to the researchers was the tension between allowing free-dom within the dialogical discussions for the students and the level of interventionfrom the teacher. They were conscious of dialogical discussion as a concept of edu-cation, which as opposed to say indoctrination or vocational training, implies morethan just the acquisition of knowledge; it also implies some growth in the intellec-tual freedom of the learner (Biesta 2006). Interventions suggested by some research-ers are quite specific, such as Nystrand’s (2006) study, which suggests quitespecifically the teacher’s role during dialogical discussions. He contends that, thatin order for a discussion to shift from a ‘recitation’ discussion in to the dialogicand then maintaining these discussion to what he deemed ‘dialogic spells’ involvedspecific intervention from the teacher such as uptake questioning: uptake question-ing involves follow-up questions that incorporate students responses. However, itwas felt by the researchers that the Paideia Method as an approach is complex andthey did not want to ‘bombard’ the teachers with too much information.

The researchers decided to look at the general use of Socratic questioning withinthe Paideia Method as a means of fostering the dialogic within the discussions. Priorresearch has provided commentaries on the value of Socratic questioning in develop-ing these critical thinking skills and enriching thinking through a dialectical approachof dialogue with peers (Billings and Fitzgerald 2002; Haroutunian-Gordon 1991,1998; Orellana 2008; Philgren 2008; Robinsson 2006; Robinson and Lai 2006). Cen-tral to Socratic questioning is the provision of a thought-provoking, open-endedquestion, which promotes inquiry and allows ideas to be probed, grappled with andtested (Adler 1983; Philgren 2008). It is not about arriving at a ‘right answer’ butrather having students focus explicitly on the process of thinking and, in turn, exam-ine their own thinking processes. Historically, Paideia Seminars have predominantlybeen examined in research through reading comprehension, involving the use oftexts, which all students receive. More recently, there has been a refocusing on Pai-deia with resources provided by the National Paideia Centre (Roberts and Billings1999), the notion of ‘rich texts’ has been expanded to include the work of localauthors, a mathematics problem or a piece of artwork. These ‘texts’ are rich to thedegree that they are challenging and allow the development of critical ideas.

Of course, the promotion of dialogical discussions is not new. ‘Discussion isbringing various beliefs together; shaking one against another and tearing downtheir rigidity. It is conversation of thoughts; it is dialogue-the mother of dialectic inmore than the etymological sense’ (Dewey 1916, 194–5). So why is it that schoolshave been slow to take them on board? One of the problems that stand in the way

Research Papers in Education 23

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

of schools’ acceptance of the dialogic principle is the typically unrecognised ambi-guity of the noun ‘knowledge’. In public debates about what should be learned inschools, the term knowledge is understood as ‘what is known’, that is to say, aswhat is taken to be true (Wells 2007). This is not to suggest that the decision toadopt a dialogic mode of interaction in the classroom is easy, either for the teacheror for the students. For the teacher, it means partially relinquishing control of theflow of discussion, giving up the habit of evaluating each student contribution andallowing students to initiate when they have something that they consider relevantto contribute. And for students, it means treating their peers’ contributions as wor-thy of careful consideration and making their own as clear and to the point aspossible (Wells 2007).

In the Bristol study of language development, for example, a representativesample of 32 children was followed from home to first school, it was found thatnot only did children almost cease to ask ‘real’ questions at school, but teachersalso rarely invited them to express and explain their beliefs and opinions – at leastwith respect to the official curriculum (Wells 1989). And in the later years of mid-dle and high school, the situation is not very different. The vast majority of lessonsconsist of teacher lecture, followed by episodes of what Tharp and Gallimore(1988) call the ‘recitation script’, or by individual seatwork; rarely is there negoti-ated group work or open-ended discussion of ideas put forward by students (Galtonet al. 1999; Nystrand 1997). In other words, while there is a great deal of mono-logic discourse, intended to ensure the handing on of basic skills and approvedknowledge, there is very little discourse that is truly dialogic in either form orintent (Wells 2007).

Exchanges between the teacher and the students follow an Initiate, Response,Evaluates (Mehan 1979) or Initiates, Response, Feedback (IRF) (Sinclair and Coult-hard 1975) pattern in which the teacher initiates a topic by asking a question, thestudent responds, and the teacher evaluates or gives feedback regarding the studentsresponse research. In recitation, the teacher holds interpretive authority and controlsthe talk. In recitation, the teacher typically talks almost 70% of the time (Cazden inAlexander 2008; Littleton and Howe 2010; Murphy, Wilkinson, and Soter 2011).The reason that IRF are deemed to result in a dialogue of a rather limited kind ismainly because of a tendency on the part of the teachers to use closed initiatives(Alexander 2004; Galton et al. 1999; Mercer and Littleton 2007), that is initiatives(typically questions) that permit a single answer, such as ‘what is the German forcat?’ Closed initiatives do not necessarily constrain contributions to a single student.The whole class could respond in chorus, choral responses are actually a well-docu-mented feature of contemporary classrooms (Alexander 2001; Pontefract and Hard-man 2005).

Gage (2009) concluded that the ‘model embodies something profoundly funda-mental in the nature of teaching’ (75). He noted several factors that could help toaccount for the longevity of the teacher-centred method of teaching including: (a)its traditional form and intergenerational qualities; (b) its apparent adequacy andsuccess in educating the populace; (c) the relative failure of alternative methodssuch as progressive education and discovery learning theories; (d) the failure of theIT revolution to change structural aspects of the classroom; (e) the reality of theconditions of teaching and the professional demands made upon teachers; and (f)the lack of incentives and completion to drive significant alterations in educationaldelivery.

24 M. Davies and A. Sinclair

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

Gellatly (1997) monitored the implementation of Paideia Seminars and foundthat one of the most powerful findings was how important it is for students to beable to talk in school and be heard. He noted that consideration and value placedon student input were the major variables that distinguished the Paideia Seminarexperience from the rest of the class experiences.

Because the Paideia Seminar is not a teacher-led instructional method, it is anopportunity for students to interact and talk with each other, and construct and de-con-struct ideas together within the classroom environment (Billings and Fitzgerald 2002).Additionally, the Paideia Seminar encourages dialogue as a ‘group-think’ (Philgren2007) where students explore ideas together in a dialogic discourse to come to a sharedunderstanding, rather than having one person’s ideas as the winner (as in a debate).

Research literature (Billings and Fitzgerald 2002; Haroutunian-Gordon 1991,1998; Orellana 2008; Philgren 2008; Robinsson 2006; Robinson and Lai 2006) pro-vides commentaries on the value of Socratic questioning in developing critical think-ing skills and enriching thinking through a dialectical approach of dialogue withpeers. Central to Socratic questioning is the provision of a thought-provoking, open-ended question which promotes inquiry and allows ideas to be probed, grappled withand tested (Adler 1983). It is not about arriving at a right answer but rather havingstudents focus explicitly on the process of thinking and, in turn, examine their ownthinking processes. In the process of cooperative interlocution, no statement is trea-ted as true or false without examination and it is in the flow of exchanges and thecollaborative interactivity between the individual and the question which leads theparticipants closer to a better solution or possibility (Lindström 1995).

The Paideia Method features three complementary teaching techniques or col-umns of instruction (Roberts and Billings 1999) which include: didactic instruction,the Coached Project and the Paideia Seminar. The major part of interest in thisresearch is the Paideia Seminar, which is based on Socratic questioning. This studyaims to evaluate an intervention which considered these theoretical propositionswithin classrooms of early adolescents. The Paideia Method was chosen due to itsstrong philosophical links with students learning through dialogic exchanges.

The Paideia Seminar

Paideia Seminars were defined by Adler (1982), as a method of teaching intendedto engage students in discussion of ideas and values, involving the use of ‘richtexts’ which all students received. Over the past decade, however, there has been arefocusing of the Paideia Seminar and with resources provided by the National Pai-deia Centre, the notion of rich texts has been expanded to include the work of localauthors, a mathematics problem or a piece of artwork. These texts are rich to thedegree that they are challenging and allow the development of critical ideas. Theteacher becomes the facilitator of dialogue, providing open-ended questions or pro-vocative statements to promote thinking, but refrains from making judgements orevaluating student comments. The rich texts are in line with Adler’s ‘Great Ideas’(1982) such as truth, beauty, liberty, equality and justice.

A history of Paideia

Adler (1982) argued that, although most children experience equal amounts of timespent in school, they are not receiving a sufficiently high quality of education. He

Research Papers in Education 25

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

stated that, unless we managed to offer all children the same high-quality education,then democracy itself was in danger. In order to maintain a democratic society,Adler contended we must simultaneously institute much higher academic standardsand render that intellectual rigour accessible to all students. He proposed the PaideiaMethod as a model that might rectify that inequality.

‘The Paideia Proposal: an Educational Manifesto’ (Adler 1982) offered a sys-tematic critique of American public education and was dedicated to three well-known educators who had a profound effect on his thinking: Horace Mann, JohnDewey and Robert Maynard Hutchins. Horace Mann (1796–1859) was the earlyAmerican educational reformer who articulated the connection between effective‘common’ schools and democratic well-being. Dewey became the ‘liberal’ influenceon Adler that balanced Hutchins’ focus on traditional academic rigour. RobertHutchins stressed the need for academic rigour based on the intellectual traditionsof the human community and became the ‘conservative’ influence on Adler’s think-ing, leading to the call for academic standards, which have been at the core of thePaideia philosophy since the early 1980s.

Paideia structure

The Didactic Stage of the Paideia Method provides an opportunity for students togain domain and strategic knowledge for them to participate in the seminars frominformed positions (Alexander and Judy 1998; Moore and Young 2001). Domainknowledge is defined as all types of knowledge including declarative, proceduraland conditional knowledge acquired in a specific field of study and has an impor-tant role in developing expertise (Alexander 1992). To be expert in a specific area,learners not only need to know how to deal strategically with the information theyencounter, but they also need to have a considerable amount of information aboutthe area (Alexander 1992).

Domain knowledge appears to be what distinguishes expert from novice learn-ers, according to Alexander and Judy (1998). Moving students from surface to deeplearning is predicated on Hattie’s (2009) claim that ‘you need surface [knowledge]to have deep [knowledge] and you need to have surface and deep knowledge tohave an understanding in a context of domain knowledge’ (2009, 29). Expert learn-ers have enough information and background knowledge about the area of theirexpertise that allows them to consolidate the newly learned information with moresophistication.

The ‘Coached Project’ stage of the Paideia Method requires students to gain thenecessary skills to be able to participate in a Paideia Seminar. Biggs and Collis(1982) allude to a tension between students who believe the goal is to memorisefacts and teachers who believe that the goal is to enhance deep learning. To suc-cessfully participate in a Paideia Seminar, students are expected to go beyond themere regurgitation of facts and extend themselves by hypothesising, analysing,explaining and evaluating. This requires that they organise their facts during theCoached Project stage and make links between various spheres of science, history,personality and context. ‘When students can move from idea to ideas and thenrelate and elaborate on them we have learning – and when they can regulate ormonitor this journey then they are teachers of their learning’ (Hattie 2009, 29).Hattie found that many students become disengaged from lessons when they areencouraged to learn at only surface levels. For the purposes of this study, the

26 M. Davies and A. Sinclair

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

five-stage SOLO taxonomy (Biggs and Collis 1982) was used to categorise the nat-ure and complexity of the discussions. To this end, the research question was:

What happens to the nature of the interaction and the complexity of the discussionwhen students are encouraged to use Socratic questioning in a Paideia Seminar invarying socio economic classrooms?

The teachers in the control and the experimental classes, which were aligned,included in their approach a variety of rich texts to act as the catalyst for discus-sion, such as: ‘The Rime of the Ancient Mariner’, using the proverb: ‘We are theauthors of our own disasters’ and asking the questions: ‘What do you think thisreally means?’ ‘How does this relate to the Ancient Mariner, and also to otherdisasters that you know about?’ Another set of control/experimental teachers pre-sented to their classes a YouTube clip of ‘Plato’s Cave’ to provide ideas around‘identity’, with the statement ‘How can this relate to cultural identity in a changingworld?’ In a different school, the story of the Lorax was used as motivation for dis-cussion, using the question: ‘Are humans parasites?’ Another class explored theidea that ‘Expression is a Risky Business’ using quotes from Gandhi: ‘To believe insomething, and not to live it, is dishonest’ and Winterson: ‘What you risk revealswhat you value’.

The three socio-economic groups were chosen to compare results becausethough New Zealand 15-year-old students’ overall reading performance inProgramme for International Student Assessment scores is substantially higher thanthe average for the 34 Organisation for Economic Co-Operation and Developmentcountries; New Zealand has a substantial achievement disparity between studentslinked to social class and ethnicities, and is at the very top in international rankingsin this respect. Thus, a large number of students leave school without the necessaryqualifications. There is high interest in New Zealand in studies, which may affectoutcomes for students in the lower socio-economic group.

Methodology

A quasi-experimental method was employed because, although the Paideia Methodwas not controlled by the researchers, the researchers did have some control overwhen to measure the outcome variables.

Participants

Six schools were selected deliberately for a range of socio-economic demographiccompositions and because of the willingness of principals to participate. The studywas conducted in 12 experimental and 12 control classrooms across six large stateschools in New Zealand, totalling 720 students (ages 11–13). The principals of theschools were asked to choose four classes of Year eight students (years 11–13) whichwere not considered either accelerant or low achieving, had similar numbers of gen-der groups and were of similar student attributes, i.e. with no extreme behaviouralneeds. The researchers did not select the teachers in this study, rather the principalswere asked to select the teachers based on the criteria above. The teachers, rangedfrom a beginning teacher to teachers with vast experience. There were three maleteachers and nine female teachers in the experimental classrooms, and four male

Research Papers in Education 27

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

teachers and eight female teachers in the control classrooms. Two of the classeswould be randomly assigned experimental and two would be assigned control.

Importantly, the four classes needed to be doing the same topic at the same timeof the year. Large state schools were chosen because the researchers knew thatschools of this size worked in teams of teachers who taught the same classes formost of the day and who worked closely in teams. It was essential that the fourclasses from each school were learning the same concepts. Schools were chosenfrom two low, two middle and two high socio-economic levels. Ethics approvalwas gained through the University of Auckland and informed consent was obtainedfrom the principals, the parents of the participants, the teachers from both the exper-imental and control classrooms, and assent forms were obtained from the studentsdue to their age. On average, 95% of the students were given consent by their par-ents. All teachers from both the experimental and control groups gave their consent,as did the principals of each of the schools. The students who did not receive con-sent for this research study went to the library with work related to the study andwere supervised by another adult or teacher. The parents from the middle socio-eco-nomic schools were the most likely to not give consent, while the parents from thehigh socio-economic schools gave the highest levels of consent. The parents did notstate why they did or did not give consent.

Process or approach

The researchers conducted two professional development days for the 12 teachersinvolved in the research study. The researchers acknowledged the different stagesthe teachers were at, in terms of dialogical discussions and took into account thework of Bransford, Donovan and Pelligrino (1999) who identified the importance ofhaving a deep foundation of knowledge in order to learn and so the teachers wereexposed to ideas in many different forms and at different times in the project. Theteachers were given historic and current literature about the Paideia Method to readbefore the professional development days. The professional development daysincluded a background on Paideia and instruction on the three complementaryteaching techniques or columns of instruction: the Didactic stage of teaching, theCoached Project stage and the Paideia Seminar. In addition, the teachers werereminded of the principles underpinning the choice of an open, contentious or pro-vocative statement to begin the discussion in order to engage the students. Thestatement needed to address essential human concerns or big ideas in order to pro-voke different responses from a variety of students. If the statement was not conten-tious enough, the case study ran the risk of losing the democratic ideal whereby allstudents were expected to participate. In line with Adler’s (1982) philosophy, thetopic needed to be thought provoking and address ideas of complexity and ambigu-ity, which could not be disposed of by simply agreeing or disagreeing.

The teachers were explicitly taught Socratic questioning to help shift the com-plexity of these discussions from surface to deep thinking. Examples of Socraticquestioning which probe students to provide reasons and evidence such as ‘Whatdo you mean by’; questions which probe reasons and evidence such as ‘Could youexplain your reasons?’; and questions which probe to provide implications and con-sequences, such as ‘What are you implying by that?’ The teachers received trainingin the use of Moodle, the online coaching device, to prepare the students for theseminars. They were given ideas on how to introduce their students to working in

28 M. Davies and A. Sinclair

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

an academic online site, in contrast to the students’ familiarity with a social net-working site. By using an asynchronous online environment, the students would beable to manage their own time to a large extent and reflect upon what they werelearning. (Garrison, Anderson, and Archer 2003). The two professional developmentdays and the ongoing professional guidance in the form of an online discussionforum for the teachers, emails, phone calls and many visits to the schools gave boththe researchers and teachers a chance to come to a shared understanding of thegoals and the processes involved in the principles of the Paideia Method.

Following the Didactic stage of the Paideia Method (which involved the stu-dents engaging with domain knowledge of their unit of study in the form of guestspeakers, journals, research articles, powerpoints, DVDs, Web 2.0, interviews andInternet), the teachers set up the Moodle online ‘classroom’. The Moodle classroombecame the Coached Project stage for the Paideia classes, where the students hadopportunities to discuss online the various provocations from the teachers. The pur-pose of the Moodle classroom was to allow students to practise the use of Socraticquestioning and to practise the art of expressing views with justification and evi-dence. The researchers had access to all of these online discussions throughout thestudy. The students were explicitly taught the skills of Socratic questioning andexamples of this questioning were posted on the Moodle (online) classroom. Thestudents were expected to discuss the provocation agreed to by the class. The onlinediscussion would be the same provocative statement that they would be expected todiscuss together during the Paideia Seminar (face to face). Students were able toparticipate in these online discussions outside of school hours if they wished. Oncethe students had sufficient time to practise discussing their thoughts online, eachteacher sets up a Paideia Seminar in their classroom. This involved the students sit-ting in a circle facing each other and discussing the various provocations for30min. The provocations had been provided either by the teachers or in manycases, the students and the teacher had agreed on a statement, which would provideample contention and ambiguity.

During this study, the control classrooms were studying the same topics as theexperimental classes but without following the Paideia Method. They were asked tocontinue teaching the topics with their usual ‘practice’. The researchers observedteachers during this time using worksheets such as cognitive organisers, whichincluded venn diagrams and PMI – positive, minus and interesting. The teachersused a variety of other strategies such as individual writing, peer discussions andgroup discussions. The students in the control classrooms appeared engaged in theirwork and the teachers were of ‘equal’ ability to the experimental teachers as deter-mined by the principals for this study. The researchers were impressed by the qual-ity of the teaching and the discussions held in these control classrooms.

Data gathering

The data for the study were collected at three points in time T1, T2 and T3 for bothcontrol and experimental classes. The timeframe varied slightly across the schoolsbut was predominately over a 12-week period. Time one (T1) involved gatheringnormative practice data of a normal class discussion for both the control and theexperimental classrooms. Each class was filmed and audio taped for 20min. Thetopics for both the experimental and control classrooms were the same in eachschool. The topics varied across schools but not within schools.

Research Papers in Education 29

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

Specifically within these varying types of interactions, data were also collectedto find if these interactions included the students asking the teacher a question, stu-dent responding to another student with a question, student responding to anotherstudent with a challenge, student responding to student with expansion of theirideas (piggy backing), student expanding on their own statement, student respond-ing to the teacher with an answer, student responding to a student with an answer,student responding to a student with further information and a student changing thesubject. The data were coded with a nominal scale.

The data, which gathered information on the complexity of these interactions,were determined by the use of SOLO taxonomy (Biggs and Collis 1982) with anordinal coding system as the numbers represented hierarchy in thinking from sur-face to deep. The descriptions of classroom discourse and interaction derived fromthe transcripts of the videotapes were divided into two main categories (Table 1):

Time two (T2) involved gathering transcripts from the Moodle discussionsonline for the experimental classes. This was at the midpoint or just thereafter ofthe unit study for the experimental classes (week six or seven). The students weregiven a provocative or ambiguous statement to discuss online. This same provoca-tive or ambiguous statement was given to the control classes. The researchers gath-ered the data from the experimental classrooms online but went back to the schoolsto film and audio tape the normal classroom discussions in the control classrooms.The control classrooms were filmed and audio taped for 20min.

Time three (T3) was in weeks 11 or 12 for both the experimental and controlclassrooms. The experimental classrooms engaged in a face-to-face Paideia Seminar.

Table 1. Explanation of SOLO taxonomy (Biggs and Collis 1982).

Example (Brown, Irving, and Keegan2008) Why does it get dark at night?

Ordinalcoding

Prestructural – students acquireunconnected pieces of information,which have no organisation and donot make sense

Because the earth spins around 1

Unistructural – simple but obviousconnections are made but theirsignificance is not grasped

Because the earth is spinning and thesun stays in our place

2

Multistructural – a number ofconnections may be made but themeta-connections between them aremissed, as is their significance

Because the earth is spinning and thesun stays in one place

3

Relational – in the relational stage,students can internalise differentideas from other sources and makeconnections

Because the earth spins round each dayand the sun stays in the one place, anyone point on the earth faces towards thesun for about 12 h and faces away forthe other 12 h. The darkness comesbecause the point at which we are onthe earth has spun away from the sun

4

Extended abstract – students makeconnections not only within thegiven subject area but also beyond it

Spherical shapes rotate around an axisin order to stay in balance. This meansthat, relative to any fixed point in space(e.g. a star), points on the sphere mustsystematically face and turn away fromthat point

5

30 M. Davies and A. Sinclair

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

The desks were pushed aside and students sat in a large circle facing each other.The teacher sat as a member of this large circle and began the seminar by restatingthe topic of discussion. The students discussed the same topic, which they had beendiscussing on line. The control classrooms engaged in a normal classroom discus-sion. The students in the control classrooms were also asked to discuss the sametopic as in Time 2.

The hypothesis is that the experimental group in the middle and high socio-eco-nomic schools will increase in both interaction focus and complexity at T3, whichis the face-to-face seminar when compared to T1 (baselines) and that this increasewill be above normative increase compared to a control group. The experimentallow socio-economic schools may not have a significant increase. There is a lowerlevel of computer exposure, particularly in the homes of those in lower socio-eco-nomic areas (McCloskey 2006; Wilcox 2002). This may result in less facility withthe technology, and thus not lead to the full potential of the use of technology inschools by these students. A quasi-experimental method was employed because,although the Paideia Method was not controlled by the researchers; the researchersdid have some control over when to measure the outcome variables. Using SPSS18.0, a series of t-tests and ANOVAs were conducted to analyse data, first for inter-action focus and then for complexity, to test for differences between the experimen-tal and the control groups. ‘When students can move from idea to ideas and thenrelate and elaborate on them we have learning – and when they can regulate ormonitor this journey then they are teachers of their learning’ (Hattie 2009, 29). Hat-tie found that many students become disengaged from lessons when they areencouraged to learn at only surface levels. For the purposes of this study, the five-stage SOLO taxonomy (Biggs and Collis 1982) was used to categorise complexityof discussion and nature of interaction.

Focus group

A focus group discussion for the teachers was held at the end of the project. Theywere given these questions prior to the focus group: ‘Following your involvementwith the Paideia Method, what do you think if anything needs to change about howyou teach (in general)’; ‘What do you consider are the successes, gaps, and failuresof using Socratic questioning within the Paideia Method?’; ‘How do you know ifthe sessions had any impact on the student learning – what evidence have you gotto show this?’; ‘Which kinds of students were better/not so good at learning thisway?’; ‘Were there any surprises?’; ‘If other teachers were to adopt this system,what needs to be in the training to optimise it, to get to the outcomes faster andmore effectively?’; and ‘What was helpful in terms of your training?’.

Data analysis

The descriptions of classroom discourse and interaction derived from the transcriptsof the videotapes were subdivided into two main categories: complexity ofdiscussion and nature of interaction. For the first category, it was decided to usethe SOLO taxonomy developed by Biggs and Collis (1982) to determine thecomplexity of the discussion, and to illustrate and analyse what surface and deeplearning looked like. The five stages are: prestructural; unistructural; multistructural;relational; and extended abstract. At the prestructural stage, students acquire

Research Papers in Education 31

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

unconnected pieces of information, which have no organisation and do not makesense. At the unistructural stage, simple but obvious connections are made but theirsignificance is not grasped. At the multistructural stage, a number of connectionsmay be made but the meta-connections between them are missed, as is their signifi-cance. In the relational stage, students are able to appreciate the significance of theparts in relation to the whole, and can internalise different ideas from other sourcesand make connections. At the extended abstract level, students are able to makeconnections not only within the given subject area but also beyond it. Theresponses involve the student going outside the known, and being able to elaborateand transfer the principles and ideas underlying a specific instance. ‘Relational’ and‘elaborative’ processes involve a change in the quality of thinking that is cogni-tively more challenging than surface learning. The implications are that active learn-ing and deep-level processing are central to success and to the transfer ofinformation where the learner is active in the process of learning.

The audio tapes were professionally transcribed and the resulting protocols werethen analysed by two research assistants working independently. The transcripts ofthese seminars were analysed according to SOLO taxonomy. These assistants wereasked to analyse every identifiable interaction that occurred using this taxonomy asa coding tool. Both individuals had had extensive experience in the use of the tax-onomy, and were blind to treatment group allocation. They were asked to code thecomplexity of the interchange using the taxonomy, from prestructural to extendedabstract. These codings were reviewed by the research team and disagreementsbetween the raters were then reconciled through discussion and consensus. Timewas spent with the coders to ensure a mutual understanding of the complexities ofthe coding tables. The overall agreements were divided by the agreements plus dis-agreements. This resulted in an overall inter-observer agreement of 84%.

Codings were also done by classifying each interaction in terms of whether itwas: (a) a teacher-to-student interaction; (b) a student responding to a teacher; or(c) a student responding to a student. Furthermore, interactions were coded as towhether they were questions, expansions of ideas, challenging others’ views oranswering questions. Codings to analyse the information from the focus groups withteachers were simply categories, which determined the successes, gaps and limita-tions of the use of Socratic questioning within a Paideia Seminar.

Results

In all, the raters reported a total of 3859 codings, of which 2035 stemmed from thetraditional classes, and 1824 from the Paideia classes. Of the total, 2023 codingsstemmed from the baseline data, and 1836 related to the final seminar periods. Theraw tallies, broken down by treatment group and SOLO level are depicted as a pyr-amid panel graph in Table 1, with seminar time as the panel variable (Figure 1).

For analysis purposes, it was decided to collapse the SOLO levels into two lev-els: surface and deep. The researchers were interested in whether or not the use ofSocratic questioning within the Paideia Seminar shifted thinking from surface todeep. Surface-level responding was determined by aggregating levels 1, 2 and 3,while deep-level responding was determined by summing levels 4 and 5. The fre-quencies generated by this procedure can be seen in Table 1 which also depicts thebreakdown concerning the nature of the interactions (teacher-to-student [TS]; stu-dent-to-teacher [ST]; and student-to-student [SS]) pattern as coded.

32 M. Davies and A. Sinclair

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

For purposes of data description, the available frequency data were interpretedin terms of the percentage of interactions coded at the deep-processing level, sincethis constitutes a meaningful dimension. Table 2 depicts the percentage of respond-ing coded at the deep level across the two time periods within each of the two treat-ment groups (Figure 2).

As can be seen in Table 2, in the final seminars for the Paideia classes, 175 of1008 responses (17.4%) were at the deep level, in contrast to the traditional classeswhere 61 out of 816 responses (7.5%) coded at this level. This represents a signifi-cant difference between the two groups, X2 (1) = 39, p< 0.01. It was apparent that,at the baseline, the frequency level of deep responses evident within Paideia classeswas not significantly higher than the traditional classes (7.5 against 5.5%)

Figure 1. Frequencies of SOLO level codings across two groups and two time periods.

Table 2. Over-time data comparing traditional and Paideia classes’ total interactions/deepinteractions for each type of interaction.

Traditional classes Paideia classes

Initial seminar Final seminar Initial seminar Final seminar

Total Deep Total Deep Total Deep Total Deep

Overall tallies 1207 66 828 62 816 61 1008 175TS 601 31 322 21 405 43 167 6ST 526 32 288 17 367 14 154 20SS 79 3 210 24 38 4 686 149

Note: ‘Total’ refers to the total number of ratings made, whereas deep refers to the frequencies made atSOLO levels 4 and 5 combined. The second, third, and fourth row will naturally tally with the level ofthe first row except in the case of 24 missing values where raters failed to agree on the interactionfocus dimension.

Research Papers in Education 33

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

(p= 0.07). However, the relative increase across sessions within the Paideia classes,from 7.5 to 17.4% (Fisher exact test, p< 0.001), was significant while this was notwithin the traditional classes, changing from 5.5 to 7.5% (Fisher exact test,p= 0.07).

Figure 1 also shows the frequency breakdown according to the type of interac-tion pattern. The data are shown in Figure 3 in terms of the three categories withbaseline data in the top panel, and final seminar data in the lower panel.

As can be seen in Figure 3, in the final Paideia Seminar, for SS discussions,686 responses were tallied, of which 149 (22%) were coded at the deep level. Incontrast, in the traditional group, there were 24 deep responses within SS discussionout of 210 responses (11.4%). Thus, the level of SS interaction at the deep levelwas significantly higher in the case of the Paideia classes, X2(1) = 10.9, p< 0. 01.Further, it was apparent that in the Paideia classes, the level of ST interaction at thedeep level (at 13%) was significantly greater than for the traditional classes (6%),X2 (1) = 6.6, p= 0.01. However, the difference between groups was not significantwhen considering TS frequencies (p= 0.18). Hence, the significant overall differ-ences in deep-level discussions between the two types of classes appear to be dueto student-initiated (SS and ST) discussions, and not teacher-initiated interactions(TS).

A 2� 3 Chi-square test found a significantly different pattern to types of interac-tions (TS, ST and SS), between the Paideia and traditional classes, X2 (2) = 58,p< 0.01. This result suggests that the impact of Paideia clearly differed in accordwith the type of interactions, with far stronger impacts being evident in the case ofthe student-initiated response categories.

The final set of analyses compared nature and type of classroom interactionsacross the three school socio-economic level classifications: low, middle and high.The percentage figures associated with these levels are shown in Figure 3. In the

Figure 2. Percentage of responses codes at a deep level across groups and seminar session.

34 M. Davies and A. Sinclair

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

initial baseline data, effects associated with socio-economic level did not emerge assignificant. However, in the final seminar, significant differences between the tradi-tional and Paideia classes were evident at the high socio-economic level (X2(1) = 40,p< 0.01), and at the middle socio-economic level (X2 (1) = 8.3, p< 0.01), Althoughthe low socio-economic students did not make a significantly greater percentage ofdeep responses in the final seminar Paideia classes than the traditional classes (X2

(1) = 1.6, p= 0.2), the low socio-economic students did significantly increase theirpercentage of deep responses in the final seminar relative to the percentage shownin the initial classroom discussion, X2 (1) 3.7, p= 0.05 (Figure 4).

This graph shows the nature of interaction, which occurred during Episode threefor both groups. Episode three for the traditional classrooms was a ‘normal’ class-room discussion based on the same unit study as the Paideia Group. Episode threefor the Paideia Group was a Paideia Seminar. Students sat in a circle facing eachother and discussed the provocative statement.

As previously identified, the Paideia Seminar resulted in higher SS interactions.This graph identifies specifically which type of SS exchanges seemed to generatedeep thinking (Figure 5).

Here is an extract of a Paideia Seminar from the group whose rich question was‘Expression is a risky business’. The students were seated in a circle, facing eachother and the teacher had stated the topic at the start of the seminar. This extract isfive minutes into the seminar and demonstrates that the nature of the interaction is

Figure 3. Percentage of responses coded at deep level as a function of interaction focus.

Research Papers in Education 35

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

mostly SSAE (Student responds to student with expansion of their idea), where stu-dents agree and then expand on their ideas with further information. This was ‘typi-cal’ in that the teacher’s voice is not apparent and the dialogue continues for sometime between students.

Student Three: I want to build on Jill’s opinion because for example in North Korea ifyou say something bad about their leader or government you will prob-ably going to be killed or punished or something. (SSAE – Studentresponds to student with expansion of their idea, Relational, manyideas)

Student Four: I agree with Won about the culture, like some, it usually comes and likesome people will get this cause they’re like different culture and howthey express themselves, through their culture. (SSAE – Studentresponds to student with expansion of their idea, Multistructural, manyideas)

Student Five: The people that went to Vietnam, they wanted to help the people inVietnam, like the Eye Clinic, and it was very cheap. But they also gotkilled at the end because they were reading a bible or something. Andthis, expression is a risky business with religion as well. In some coun-tries like in India they are Muslims and Christians fighting. (SSAE –Student responds to student with expansion of their idea, Relational,many ideas)

Student Six: It also depends on whom you’re dealing with. So if it is in World WarII and your expression was to disagree with Hitler, then you wouldprobably get killed and punished. (SSAE – Student responds to studentwith expansion of their idea, Relational, many ideas)

Figure 4. Percentage of responses at deep level as a function of school socio-economiclevel.

36 M. Davies and A. Sinclair

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

Student Four: I’m thinking back to Joe’s point, I definitely agree that education oneveryone’s belief and culture and the country and everything, so thatwas how people expressed themselves, so they would keep their emo-tions inside, they were scared or something. (SSAE – Student respondsto student with expansion of their idea, Relational, many ideas); (SSN–Student to Student with New Ideas)

The following is an example of SSDC interactions (SS with a disagreement andthen challenging the student with why they disagree). This exchange occurred dur-ing the Paideia Seminar when the topic was ‘How can this can relate to culturalidentity in a changing world?’ (Plato’s Cave).

I disagree with Josh’s point when he said like the woman was wearing yellow to stopwith the heat and everything. I think it’s just a metaphor, like staying with the dark-ness of the cave and then the brightness of the outside world.

Data from these students in their normal classroom discussion had indicated thatalthough students might have disagreed with a fellow student they did not elaborateon why they disagreed.

Transcript data analysis showed that not only were more questions generatedfrom student-to-student (SSQ) but they were at a higher complex level during thePaideia Seminars than in the normal classroom discussion. For example, one studentposed the following question: ‘But do you agree that America was the start of ste-reotypes in movies and television?’ This question, posed by the student, not the tea-cher, stimulated a robust discussion between the students.

(SS) = student to student SSQ = question

SSDC = disagreement & expansion SSAE = SSDE =

SSDE = expands

SSA = SSN = SSD =

SSAN = with answer SSAG =

answer & expansion of ideasresponds to disagreement & expands on own statement

responds to challenge &

agree new idea responds with disagreement respondsresponds with agreement

Figure 5. Type of SS exchanges which generated deep thinking.

Research Papers in Education 37

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

Students who were interviewed as part of a focus group experience indicatedthat they found that the opportunity to express their ‘voice’ was motivating and apositive experience, similar to the following comments by students:

It’s a lot friendlier and a lot more interactive with all of your peers. It wasn’t juststraight out of a textbook. (Pasifika male, low socio-economic)

The students (us) we drove the conversations and we understand better from eachother than if a teacher were to stand up there and get us to speak. (Asian female, highsocio-economic)

Another student noted how the seminars changed the dynamics in the classroom:

A lot of people in our classroom don’t usually speak to each other, like there is agroup of boys and a group of girls and then a group of both and they don’t reallycommunicate with each other but I think the seminar was a great way of sharing ideasbetween different people. (European male, high socio-economic)

In terms of fairness and ‘voice’, two students made these observations:

We all respect each other and we respect each other points of view. Even if you dis-agreed with someone it was good to be able to understand where they are comingfrom and just to be able to see two points of view is important, even if you disagree.(European female, high socio-economic)

When people don’t understand what people are saying, other people come in and help.(Maori male, low socio-economic)

A student in the low socio-economic school compared the difference between a nor-mal classroom discussion and a Paideia Seminar:

If Mrs. Earnshaw asks us a question in a class discussion, then the students just sayyes, no, yes, no with no reason and that’s why we aren’t learning. But with the semi-nar we still say if we agree or and not agree, but we have to say why.

The results of the focus group with teachers found the following repeated theme.Successes with the use of Socratic questioning within the Paideia Seminar wereidentified: a low socio-economic teacher whose class was English for Speakers ofOther Languages remarked that this method allowed and encouraged the children todig deeper into the many layers of an idea.

The thinking and the speaking feed off and enrich each other with many childrenentering into the discussion offering lots of different thoughts which dug deeper anddeeper into the original idea, exposing multiple layers of connected and related ideaswhich would never have been considered or even thought of during an ordinary class-room discussion.

A middle socio-economic teacher concurred by adding that she believed the qualityof what they were saying was at a much higher level when they started to respondto each other, rather than to her. ‘They started to question each other, which meantstudents had to justify their thinking’. Several teachers commented that they noticed

38 M. Davies and A. Sinclair

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

the students needed to listen to each other as the thinking was at a deeper level.‘Class discussions were richer, because they were listening to each other and notjust me’. Another middle socio-economic teacher identified that Socratic questioninghad shifted in to other curricula:

We are currently doing Literacy Circles in reading and the students are continuing touse the Paideia Language when they interact. They always want to add to each other’sideas or question and disagree. It’s great that they are more willing and able to justifytheir ideas.

A drawback over the use of Socratic questioning was that one of the teachers in thelower socio-economic school believed it suited only the students with good self-management as they liked having some control over what direction their learningwas taking. A middle socio-economic teacher was surprised that some of her ‘topend kids’ struggled with the freedom that this type of learning entailed and theyreally seemed to flounder. She observed: ‘Obviously after years of trying to guesswhat the teacher wants or is thinking they were stuck when expected to think ontheir own’.

Discussion and conclusion

The main findings showed the nature of the interactions changed from predomi-nantly teacher-initiated interaction in a traditional classroom discussion, to student-initiated interaction during the Paideia Seminars. As demonstrated by the results inthe study, the complexity of the discussions was deeper during the Paideia Seminarsthan in a traditional classroom context where students were working at a greaterrelational and extended abstract (deep) level. Significantly, the types of interactionswhich generated higher complexity of thought were: students agreeing with eachother and then expanding with further information; students disagreeing with eachother and then expanding on why they disagreed with each other; students respond-ing back to the student who had disagreed with them and explaining themselvesfurther and students asking another student a question. The greatest shift in thinkingfrom surface to deep occurred in the high socio-economic classrooms. Interventionfrom the teachers in the high socio-economic classrooms was minimal, and the dia-logue almost predominately student initiated. Perhaps, this indicates that the stu-dents who were from the lower socio-economic classrooms require greater teacherinput as Nystrand (2006) recommends. Of course, it is not that all students in thelower socio-economic classrooms were low-achieving students and that all studentsin the high socio-economic classes were high achieving but in general the literacylevels for the low-achieving classes were considerably lower than the students inthe high socio-economic classrooms. Nystrand (2006) argues that what is difficult isnot only shifting discussions in to the dialogic but then keeping them at the dialogicwith what he terms ‘dialogic spells’.

Though the use of Socratic questioning during the seminars was relatively low,more complex discussions were held throughout the seminars than the traditionalclassroom discussions within the control classes. A ‘spin-off’ from the teachersexplicitly teaching the students Socratic questioning appeared to be that the studentsrecognised the importance of having evidence and justifications for their thoughts.Without prompting from the teachers, the students not only initiated discussions

Research Papers in Education 39

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

with their peers but also invariably made a statement and supported this statementwith evidence and justification. These exchanges resulted in a change in the natureof the interactions, i.e. more SS dialogues and the change in the complexity of theseinteractions meant they engaged in a greater percentage of deep-level discussions.

Gaining skills in the use of Socratic questioning is likely to take far longer thanthe duration of this project. For many students, it was a major shift: to be givenautonomy to have student-initiated discussions and to be expected to provide evi-dence and justification for their statements. The positive outcomes of this study,however, indicate that teachers explicitly teaching Socratic questioning and provid-ing opportunities for students to practise this type of questioning and reasoning areworth pursuing. This study showed that there were differences in the students’ lev-els of complexity within the discussions between the socio economic groups.

In contrast, results of (Murphy et al. 2009) meta-anaylsis of dialogical discus-sions interventions suggested that the approaches exhibited greater effects forstudents of below-average ability than for students of average or above average abil-ity. But what was similar to this study was their suggestion that the students in thehigh socio-economic groups appeared to need little intervention from the teachers tokeep the discussions at the ‘dialogic’ (Murphy et al. 2009). This finding was inter-preted to mean that the higher ability students might be able to read a text and think-ing independently about the nuances of meaning even without participating indiscussion (Murphy, Wilkinson, and Soter 2011). In general, the low socio economicstudents dialogical discussions returned more quickly to that of recitation discussionsthan the high socio economic students dialogical discussions. Therefore, theresearchers wish to pursue greater degrees of explicit intervention from the teacherin the lower socio economic schools. Specifically, one researcher wishes to furtherinvestigate the work of Quality Talk (summary) (Wilkinson, Soter, and Murphy2010) with students within the lower socio-economic schools. Quality Talk involvesthe following steps:

(1) Use ground rules to establish the norms of productive talk.(2) Use authentic questions and follow-up, uptake questions to give students

opportunities to engage in productive talk.(3) Use informal assessment strategies during discussions, listening for evidence

of the elements of talk that indicate higher level thinking.

One limitation of this study was the inequitable access to the Internet for thestudents in the low socio-economic schools. Many of these students did not havethe benefit of being able to go online and discuss the provocative statement for thePaideia Seminar out of school hours and had limited exposure to the use of comput-ers when they were at school. The teachers within the low socio-economic schoolslacked confidence and expertise in using the online discussions with their studentsand so the students had less exposure to the online discussions than the middle andhigh socio-economic students. Ensuring a closer alignment with exposure to the useof the online discussions for all socio-economic groups of students could possiblyhave affected the outcomes of this study. There is a lower level of computer expo-sure, particularly in the homes of those in lower socio-economic areas (McCloskey2006; Wilcox 2002).

The results of this study show a significant shift in student-initiated dialogue,from teacher-initiated dialogue in normal classroom discussions to increased stu-

40 M. Davies and A. Sinclair

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

dent-initiated dialogue in the Paideia Seminar. Furthermore, these shifts generated ahigher complexity of thinking. What emerges from this study is the potential of thePaideia Method to increase depth of thinking and achieving greater autonomy forearly-adolescent students. The increase in student-initiated dialogue demonstratesthat this pedagogical approach can give rise, not only to student voice, but opportu-nity for student voice, which may help to realise these young people’s potential.All students, not just those from marginalised groups, seem more eager to enterenergetically into classroom discussion when they perceive it as pertaining directlyto them (Hooks 1994, 87).

Authentic help means that all who are involved help each other mutually, growingtogether in the common effort to understand the reality, which they seek to transform.Only through such praxis – in which those who help and those who are being helped,help each other simultaneously – can the act of helping become free from the distor-tion in which the helper dominates the helped (Freire 1994, cited in bell hooks, 54).

AcknowledgementsThe authors would like to thank the school staff and students who were involved in thisproject. The North Shore Teachers Trust Fund funded the study. The authors would like tothank Professor John Hattie, Associate Professor Richard Hamilton and Senior LecturersGreg Yates and Tony Hunt for their assistance and support.

Note1. A professional DVD has been made of this research project with ethics approval from

the University of Auckland. The full DVD is available for purchase from CEDD@auck-land.ac.nz.

Notes on contributorsMaree Davies is a senior lecturer in the School of Learning, Development and ProfessionalPractice, University of Auckland, Faculty of Education, 74 Epsom Avenue, Epsom,Auckland. The author is currently continuing to research dialogical discussions both onlineand face to face. The research focuses on why the psychological mechanisms of ‘qualitytalk’ and evidence-based argumentation affect the quality of thinking. The research is quasi-experimental and is being conducted in secondary schools in Auckland, New Zealand.

Anne Sinclair is an honorary research fellow in the School of Learning, Development andProfessional Practice, University of Auckland, Faculty of Education, 74 Epsom Avenue,Epsom, Auckland. She has expanded the Paideia research and is working on developingadaptive expertise amongst teachers and students, to promote higher metacognitive thinkingand self-regulation both online and face to face. She is considering the effects of large classteaching at the University level and how this could be mitigated through developing theadaptive capacity to utilise effective pedagogies and the principles of Paideia in this newapproach.

ReferencesAdler, M.J. 1982. The Paidea proposal: An educational manifesto. New York, NY: Collier

Book Macmillan.Adler, M.J. 1983. Paideia problems and possibilities (83–5). New York, NY: Macmillan.Alexander, P. 1992. Domain knowledge: Evolving themes and emerging concerns. Educa-

tional Psychologist 27, no. 1: 33–51.Alexander, R. 2004. Towards dialogic teaching: Rethinking classroom talk. Cambridge:

Dialogos.

Research Papers in Education 41

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

Alexander, R. 2001. Culture and pedagogy: International comparisons in primary education.Oxford: Blackwell.

Alexander, R. 2006. Towards dialogic teaching: Rethinking classroom talk . 3rd ed. Thirsk:Dialogos UK.

Alexander, R. 2008. Culture dialogue and learning, notes on an emerging pedagogy. Explor-ing talk in school. London: Sage.

Alexander, P., and J. Judy. 1998. The interaction of domain-specific knowledge in academicperformance. Review of Educational Research 58, no. 4: 375–404.

Barnes, D., and F. Todd. 1995. Communication and learning revisited. Portsmouth, NH:Heinemann.

Barnes, D., and F. Todd. 1977. Communication and learning in small groups’. Abingdon:Routledge and Kegan Paul.

Biesta, G. 2006. Beyond learning: Democratic education for a human future. Boulder, CO:Paradigm Press.

Biggs, J., and K. Collis. 1982. Evaluating the quality of learning, the SOLO taxonomy. NewYork, NY: Academic Press.

Billings, L., and J. Fitzgerald. 2002. Dialogic discussion and the Paideia seminar. AmericanEducational Research Journal 39, no. 4: 907–41.

Bransford, J.D., S.M. Donovan, and J. Pellegrino. 1999. How people learn: Bridgingresearch and practice. Washington, DC: National Academy Press.

Brown, G.T.L., S.E. Irvina, and P.J. Keegan. 2008. An introduction to educational assess-ment, measurement, and evaluation: Improving the quality of teacher-based assessment.2nd ed. Auckland: Pearson Education.

Chinn, C.A., R.C. Anderson, and M. Waggoner. 2001. Patterns of discourse in two kinds ofliterature discussion. Reading Research Quarterly 36: 378–411.

Dewey, J. 1916. Essays in experimental logic. Chicago, IL: University of Chicago Press.Freire, P. 1994. Pedagogy of the oppressed. 3rd ed. New York, NY: Continuum.Gage, N. 2009. A conception of teaching. New York, NY: Springer.Galton, M., L. Hargreaves, C. Comber, D. Wall, and A. Pell. 1999. Inside the primary class-

room: 20 years on. London: Routledge.Garrison, D.R., T. Anderson, and W. Archer. 2003. Critical thinking, cognitive presence and

computer conferencing in distance education. American Journal of Distance Education15, no. 1: 7–23.

Gellatly, A. 1997. Cognitive processes and theory development: A reply to Spencer and Kar-miloff-Smith. Human Development 40, no. 1: 55–8.

Haroutunian-Gordon, S. 1991. Turning the soul. Teaching through conversation in the highschool. Chicago, IL: The University of Chicago Press.

Haroutunian-Gordon, S. 1998. A study of reflective thinking: Patterns in interpretive discus-sion. Educational Theory 48, no. 1: 33–6.

Hattie, J. 2009. Visible learning: A synthesis of over 800 meta-analyses relating to achieve-ment. London: Routledge.

hooks, b. 1994. Teaching to transgress. New York, NY: Routledge.Littleton, K., and C. Howe, eds. 2010. Educational dialogues understanding and promoting

productive interaction. New York, NY: Routledge.Lindström, L. 1995. Den dialogiska relationen [The dialogic relationship]. Didactica

Minima.35 årgån 9, no. 4/95: 38–53.Lyle, S. 2008. Dialogic teaching: Discussing theoretical contexts and reviewing evidence

from classroom practice. Language and Education 22, no. 3: 222–40.McCloskey, P.J. 2006. The blogvangelist. Teacher Magazine 18, no. 2: 22–9.Mehan, H. 1979. Learning lessons. Cambridge, MA: Harvard University Press.Mercer, N. 2000. Words and minds. London: Routledge.Mercer, N., and K. Littleton. 2007. Dialogue and the development of children’s thinking: A

sociocultural approach. London: Routledge.Ministry of Education. 2007. The New Zealand curriculum. Wellington: Learning Media.Moore, M.G., and M. Young. 2001. Knowledge and the curriculum in the sociology of edu-

cation: Towards a reconceptualisation. British Journal of Sociology of Education 22, no.4: 445–61.

42 M. Davies and A. Sinclair

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014

Murphy, K.P., I.A.G. Wilkinson, and A.O. Soter. 2011. Instruction based on discussion. InHandbook of research on learning and instruction , 1st ed., ed. R.E. Mayer and P.A.Alexander, 382–407. New York, NY: Routledge.

Murphy, P.K., I.A.G. Wilkinson, A.O. Soter, M.N. Hennessey, and J.F. Alexander. 2009.Examining the effects of classroom discussion on students’ comprehension of text: Ameta-analysis. Journal of Educational Psychology 101, no. 3: 740–764. doi: 10.1037/a0015576.

Nystrand, M. 1997. Opening dialogue: Understanding dynamics of language and learningin the english classroom. New York, NY: Teacher College Press.

Nystrand, M. 2006. Research on the role of discussion as it affects reading comprehension.Research in the Teaching of English 40, no. 4: 392–412.

Orellana, P. 2008. Maieutic frame presence and quantity and quality of argumentation in aPaideia Seminar. Chapel Hill, NC: The University of North Carolina.

Philgren, A.S. 2008. Socrates in the classroom. Rationales and effects of philosophizing withchildren. Stockholm: Stockholm University.

Pontefract, C., and F. Hardman. 2005. Classroom discourse in Kenyan primary schools.Comparative Education 2: 87–106.

Roberts, T., and L. Billings. 1999. The Padeia classroom: Teaching for understanding.Larchmont, NY: Eye on Education.

Robinsson, D. 2006. The Paideia Seminar: Moving reading comprehension from transactionto transformation. Lowell, MA: University of Massachusetts.

Rogoff, B., G. Gauvain, and C. Ellis. 1991. Development viewed in its cultural context. InLearning to think, ed. P.Light, S. Sheldon and M. Woodhead. London: Routledge.

Robinson, V., and M.K. Lai. 2006. Practitioner research for educators: A guide to improv-ing classrooms and schools. Thousand Oaks, CA: Corwin Press.

Rojas-Drummond, S., M. Fernandez, N. Mazon, and R. Wegerif. 2006. Collaborative talkand creativity. Teaching Thinking and Creativity 1: 84–94.

Sinclair, J., and M. Coulthard. 1975. Toward an analysis of discourse: The english used byteachers and pupils. London: Oxford University Press.

Tharp, R., and R. Gallimore. 1988. Rousing minds to life: Teaching, learning and schoolingin school context. New York, NY: Cambridge University Press.

Wegerif, R. 2010. Dialogue and teaching thinking with technology opening expanding anddeepening the ‘inter-face’. In Educational dialogues understanding and promotingproductive interaction, ed. K. Littleton and C. Howe, 304–22. Oxon: Routledge.

Wegerif, R., and N. Mercer. 1997. A dialogical framework for researching peer talk. InComputers and talk in the primary classroom, ed. R. Wegerif and P. Scrimshaw. Clevedon:Multilingual Matters.

Wegerif, R., N. Mercer, and L. Dawes. 1999. From social interaction to individual reasoning:An empirical investigation of a possible sociocultural model of cognitive development.Learning & Instruction 9, no. 6: 493–516.

Wells, G. 1989. Language in the classroom: Literacy and collaborative talk. Language andEducation 3: 251–73.

Wells, G. 2007. Semiotic mediation, dialogue and the construction of knowledge. HumanDevelopment (0018716X) 50. , no. 5: 244–74doi: 10.1159/000106414.

Wilcox, B.L. 2002. Integrating technology. English Leadership Quarterly 25, no. 2: 1–2.Wilkinson, I.A.G., A.O. Soter, and P.K. Murphy. 2010. Developing a model of quality talk

about literary text. In Bringing reading researchers to life: Essays in Honor of Isabel L.Beck, ed. M.G. McKeown, and L. Kucan. New York, NY: Guilford.

Wood, D.J., J.S. Bruner, and G. Ross. 1976. The role of tutoring in problem solving. Journalof Child Psychiatry and Psychology 17: 89–100.

Research Papers in Education 43

Dow

nloa

ded

by [

Kun

glig

a T

ekni

ska

Hog

skol

a] a

t 05:

51 0

7 O

ctob

er 2

014