Primary Children’s Collaborative Cartography: Communication and Mapping Processes. David Owen [email protected] Sheffield Hallam University Division of Education and Humanities Sheffield United Kingdom

ABSTRACT Simple but powerful computer drawing tools give primary school pupils the opportunity to act as creative cartographers and to collaborate in the construction of electronic maps. Use of such technology in schools is promoted by the Ordnance Survey and by the Qualifications and Curriculum Authority, the body responsible for the delivery of the National Curriculum in English primary and secondary schools. However, there remains much work to be done in developing effective teaching and learning approaches that make the most of both new technology and well established cartographic principles. This paper presents data and analysis from an ongoing PhD research project. The research focuses on the collaborative processes used by children aged between seven and ten years old to solve mapping problems. During the current school year, 80 children from one primary school worked in pairs to make collaborative maps. They used a computer drawing application to represent nominal and ordinal data in the creation of a map of the locality surrounding their school. Data presented includes examples of the children's maps and the system used to categorise the maps, transcripts of the children's talk showing the evidence for peer support during the map-making process, and the results of interviews with the children. The initial analysis of this data explores possible relationships between the quality of mapping and the processes of collaboration. The paper concludes with a discussion of the next steps in the research and an interim exploration of the implications of the study for teachers, software developers and researchers interested in children and cartography.

INTRODUCTION The aim of this study is to explore the process of children’s collaborative mapping using electronic tools. Since 1989, map reading and map making have been part of children’s statutory entitlement to geography between the ages of 5 and 11. Geographical Information System (GIS) software and generic drawing software are now increasingly being used to create maps from digital location data as well as drawing maps using computer aided design techniques. These technologies are now available in the primary school, and are promoted by computing and GIS education agencies as appropriate and effective tools for learning (Ordnance Survey, 2002; ESRI, 2003). The study focuses on the use of computer based mapping tools by primary children aged seven to ten years. There is an increasing quantity of curriculum guidance that recommends the use of such tools for primary children, but there are few empirical studies that have explored the potential of electronic mapping technologies with children of this age group (although see Keiper, 1999 for an example). The study has three main objectives. These are to:  characterise the content and form of children’s digital maps ;  characterise the collaborative map making process by exploring how children collaborate, and whether better talk makes for better maps;  identify implications for teaching and primary mapping software design. This paper builds on previous pilot studies (see Owen, 2003; Owen, 2004) and reports on the first stage of data collection in the main research project. Ten pairs of children took part during December 2004. A further thirty pairs of children will complete the data collection phase in June and July 2005.

LITERATURE REVIEW Children’s ability to draw maps has been a major research focus for developmental and environmental psychologists, geographers and geography educators for many years (Liben, 2002). Within this research, there is disagreement over children’s skill at both representing spatial relations and symbolising space. Nativists, such as Spencer and the late Jim Blaut, argue that mapping is a ‘cultural universal’ through which children as young as three can understand the language of location and symbolisation (Blaut et al, 2003). Constructivists, such as Liben, build on the work of Piaget (Piaget et al, 1956 and 1960) to argue there is a recognizable developmental progression in both spatial and symbolic understanding. In this study, children’s spatial and symbolic understanding as shown on maps is analysed in terms of map content, map structure and children’s approaches to map symbolisation. Classification systems for freehand sketch maps have focused on the degree of structure shown in the maps. Hart and Moore (1973) and Moore (1976) proposed a three stage model of how both adults and children structure environmental knowledge. These stages were the undifferentiated egocentric (organised egocentrically from one viewpoint), differentiated and partially co-ordinated into fixed subgroups (where clusters of information on a map may show some internal accuracy but will not be related to other clusters of information on the map), and operationally co-ordinated and hierarchically integrated (based on a co-ordinated and abstract reference system). Deloache (2002) provides a developmental model of children’s ability to understand symbols, and Liben and Downs (1989) gave examples of how children aged 7-10 may hold misconceptions about the meaning of symbols. A symbol is “something that someone intends to stand for or represent something other than itself” (DeLoache, 2002, p. 207). Research on children’s symbolisation strategies has been also been carried out by Michaelidou et al. (2002) who investigated how children classified geographical features and also how children intuitively used visual variables (Filippakopoulu et al, 1999). Bertin (1983) proposed his basic set of ‘visual variables’ that a cartographer or graphic designer could draw upon to symbolically represent information selected from perceptions of the real world. These variables are shape, size, orientation, pattern, hue and hue value. The cartographer can therefore apply these resources to points, lines, areas and text on a map. Each of the categories can be used individually to draw attention to map features or they may be used in combination. The data to be mapped can be nominal data: information that is grouped into categories on the basis of qualitative judgements, such as distinguishing a road from a river; ordinal data: grouped by rank or hierarchy, such as low, middle or high ground; and interval data and ratio data: data that are ranked and include numerical data such as temperature or height. Bertin saw these visual variables as the basic building blocks of maps and diagrams, and devised rules to indicate how they might be used with different data. His classification has been challenged (Robinson et al, 1984; MacEachren, 1995) but remains influential to cartographers, web designers and graphic designers. When children are involved in environmental mapping, they are likely to be representing nominal data (such as roads, houses and parks) but may also be representing ordinal data in terms of places most or least visited, or places most liked or disliked. Of specific interest is the creation and classification of point symbols. Point symbols can be classified according to their level of abstractness (see figure 1).

Abstract ------------------------------------------------------------------------------------------------------ Pictorial Figure 1. A nominal point symbol to represent a sweet shop This continuum is of interest to those investigating children’s symbolisation strategies, as it has some commonality with Piaget influenced classifications of children’s sketch maps as devised by Hart and Moore (1973) and summarised by Harwood and Usher (1999). The research discussed above is based mainly on children’s use of symbols on paper, models or on film. There is much less research on the use of ICT for map symbolisation, but examples include Wiegand’s work on 11-14 year old students’ mental representation of thematic point symbols (Wiegand, 2002a) and the use of a simulated computer game to investigate symbology and teach children cartography (Filippakopoulu et al,1999). All aspects of map drawing, from content selection to digital design, can be thought of as individual choices to be made – the stance taken by developmental psychologists. However, a coherent argument has been developed that this meaning making is a socially negotiated activity (Rogoff, 1990; 2003). This approach has been applied to spatial and mapping collaboration (see Keiper, 1999; Wiegand, 2002b) and has exploited global positioning systems (GPS), computer mapping tools and electronic communication (Dimitracopoulou & Ioannidou, 2003). Investigating how peers talk to create maps also allows the researcher to access children’s thoughts about symbolisation, choice of map content and

potential reasons for map structure. Mercer (1995 and 1996) focused on the role of the teacher in structuring talk in a classroom situation based on such social constructivist principles. He identified three broad types of pupil talk, which are summarised below:  cumulative talk which reflects an orientation to share and understand each other but without any critical grounding of shared knowledge;  disputational talk where individuals treat dialogue as a competition which they seek to win;  exploratory talk which is oriented to sharing knowledge like cumulative talk but with the addition of critical challenges and explicit reasoning. Wegerif (2001) identified exploratory talk as most likely to support peer or group learning, but also claimed that such talk does not readily occur without overt teaching. Kress and Van Leeuwen (1995, 2001) and Lemke (in Kress et al, 2001) have further developed this work to include nonverbal forms of communication by introducing the concept of multimodal communication. In this framework, text only communication is seen as insufficient for communication in many day-to-day contexts. Lemke’s work on communication in science classrooms analyses classroom discourse as a combination of speech, gesture, use of models and written texts. Such an approach can usefully be applied to the analysis of children’s collaborative map-making as they can be seen to be communicating via talk, gesture, and the movement of graphic objects on the computer screen. This study therefore seeks continue initial research into children’s collaborative map making (such as Leinhardt et al, 1998; Bausmith and Leinhardt, 1998 and Wiegand, 2002a), but focus on the primary age group. The research will benefit teachers and curriculum planners as it will reveal what conceptions of symbolisation children bring to the mapping process, and how they operationalise these using ICT. It will allow strategies for scaffolding paired map-making to be developed, and to support autonomous learning aided by ICT. Finally, the research may be of use to software developers who create mapping software for primary children’s use.

METHODOLOGY The study can be positioned in the field of ‘exploratory cartography’ (as defined by Kraak & Ormeling, 1996, and van Elzakker, 2004). Exploratory cartography seeks to shed light on the processes involved in map use and creation. In this case, the focus will be on children who will complete a map design and construction task in friendship pairs. Research on collaborative learning suggests friendship pairs are appropriate in a learning situation and that talk between friends is likely to be easier than working with an unfamiliar partner (Cohen, 1994). All children taking part in the study (see Figure 2) have lived in the school’s catchment area for at least two years, so they are familiar with the area to be mapped. The sample has been selected from a population of approximately 180 year 3 and year 5 children from school. All children are volunteers. Age 7-8 9-10

Male 20 20

Female 20 20

Figure 2. The study sample

Design of mapping application In order to study the children’s mapmaking strategies in an ecological valid environment using an authentic task, a commonly used educational software tool has been used to support a motivating paired map-drawing task. The software tool to be used is Textease, a suite of primary ICT applications that includes a paint programme. The software allows the teacher to select the resources available to the learner, to set up a word bank of terms that can be dragged onto the screen, and to select a range of clipart symbols. The range of drawing, text and image facilities present in the software mirror the drawing tools available in geographical information systems (GIS) software and provide the tools that children requested in the exploratory survey (see Owen, 2003).The word bank is equipped with local road names and titles used by children in the exploratory phase. This allows children to focus on the size, font, and style of the lettering rather than struggle with spellings. The clipart bank is equipped with a range of symbols to symbolise landmarks and spaces that were selected by children in the exploratory project. The software also has the tools to draw lines, areas and points using a wide range of colours, and change the size and orientation of any line, point, area or text object.

The paired task The collaborative problem-solving task that the children have undertaken has some similarities to mapping tasks used by researchers wishing to access evidence of both children’s mapping ability and environmental knowledge ( as summarised by Matthews, 1992 and Spencer et al, 1989). The mapping task involves the children mapping a known area around the school and creating a route to be followed by other children. The children are asked to draw a map for a new family who have moved to the catchment area of the school. This map should show the local area around the school and a walking trail around that local area visiting four places of interest to the new children. The children are also asked to shade areas according to three levels of interest, and show parts of the route that are very busy, busy or quiet. The task has been designed so that the children have the opportunity to symbolise a variety of points, lines and spaces, as well as use text. The task reveals:  which places the children value within the local area;  the level at which the children can position and orient the landmarks in relation to each other;  whether the children map landmarks in proportion to one another;  the children’s approach to symbolising nominal data ( such as the places of interest );  their preference for abstract or pictorial symbols;  the children’s approach to symbolising ordinal data ( most, average and least interesting areas; busy, quite busy, not busy routes);  the level of detail they include on the map to facilitate use by the target user group;  how the children use text in conjunction with points, lines and areas.

Data recording and analysis Paired mapping sessions took take place in a small workroom adjacent to the children’s classrooms. The sessions began with an initial interview and a practice mapping activity. The purpose of the introduction is to acclimatise the children to the task, and to gather data on their experiences of mapping and computer use at home and in school. An Access database was created to hold records on each participant, and the data from the interview is recorded in this database, alongside school attainment data. The database can then be consulted when analysing individual maps or discourse. The main task - creating the trail map - was recorded using digital video. During the mapping session, field notes were taken to add detail and context. The researcher was in ‘classroom computer assistant’ role, answering technical queries about computer use, but not providing mapping advice. After the computer mapping was completed, the pair was interviewed about their maps at the computer. This interview was also recorded on video to allow access to gesture as well as talk during the interview. The interview schedule was designed to reveal the children’s thinking about the choices they made in constructing the map, their attitude towards working collaboratively, and their evaluation of the software used. A range of classification structures has been developed to describe the quality of children's maps (see Hart and Moore, 1973; Matthews, 1992). Harwood and Usher (1999) summarised these approaches and proposed their own system which graded maps using the following six criteria: spatial arrangements; scale and proportion; perspective; level of abstract/symbolisation; and amount of content In this study maps were graded using a modified version of Harwood and Usher's system which is shown below: Spatial arrangements (including trail) score A. Representation of spatial arrangements (position/orientation) 0 No recognisable arrangements, entirely random. 1 Beginnings of a recognisable arrangement, but unrelated to reality ( i.e. objects in lines and clusters). 2 A few places are roughly in correct order and spatial arrangement… the majority still inaccurately arranged. 3 A balanced mixture of accurate and inaccurate arrangements and order. 4 Accurate arrangements now dominant, but there are still a few inconsistencies. 5 Totally accurate arrangement. Proportion (nominal point symbols, lines) score B. Representation of proportion between nominal data 0 No consistency of proportion between graphic objects representing trail, roads, landmarks. 1 Some evidence of graphic objects in proportion, majority are out of scale. 2 Equal balance of similar symbols in proportion with each other and not in proportion. 3 Most features in appropriate proportion , but still occasional inaccuracies (roads of different width) 4 Majority of objects now in proportion. 5 All elements in correct proportion.

Mapping ordinal data score C. Representation of ordinal data 0 No attempt to map ordinal data. 1 Ordinal area data mapped as nominal point symbols. 2 Ordinal area data mapped using variation in colour (no correspondence to data). 3 Ordinal area data mapped using mixture of colour/value (some correspondence to data) – no key. 4 Ordinal area data mapping using mixture of colour/value (some correspondence to data) + key. 5 Ordinal area data mapped using variation in value (monochromatic or appropriate dichromatic variation) shown on key. Abstraction/symbolisation D. Degree of abstraction / symbolisation 0 No recognisable symbolisation of features. 1 Mixture of pictorial/abstract symbols – no pattern to selection. 2 Pattern to selection – mostly abstract symbols – no key. 3 Pattern to selection – mostly pictorial symbols - no key. 4 All places represented by pictorial symbols - no key. 5 All places represented by all abstract symbols/all pictorial backed up by use of key Amount of content score E. Amount of content ( everyday features likely to aid navigation) 0 No recognisable features. 1 Only school and trail destinations shown. 2 1-3 additional features shown. 3 4-6 additional features shown. 4 7-9 additional features shown. 5 >10 additional features shown. Use of text Score Use of text on map 0 No use of text. 1 Title present. 2 Title + location of school. 3 Title + appropriate road names + important spaces and landmarks – inappropriate location. 4 Title + appropriate road names + important spaces and landmarks + text located in appropriate place. 5 As 4 + Size/colour of text differentiated to show different categories of features. Figure 3: the map classification system All maps were analysed and assigned scores based on the above criteria. In order to increase levels of validity, an independent observer will also grade a random selection of the final 40 maps after all data collection has been completed. Talk and action during the main map-making activity was transcribed after viewing the digital video. The transcripts were then coded using an adaptation of Pilkington’s DISCOUNT scheme (Pilkington, 1999). This scheme was devised to record the structure and content of an argument as speakers talk with each other. The adaptation is shown in Figure 4. The transcripts have been analysed to provide quantitative summaries of the level of interaction between the pairs and the frequency of different types of discourse. Differences in the discourse of older and younger children can be assessed from this data. The relationships that can be derived from this data are listed below:    

number of words and conversational turns for the two age groups; level of symmetry of discourse for the two age groups; percentages of functional discourse moves for the two age groups; relationship between map structure classification and functional discourse types for each age group.

Coding categories Inform (I) makes observations and states facts about the map Direct (D) instructions for map-making Reason (R) reasoned mapping activity Question (Q) question inviting verbal response Other (O) organisational statements

Example The road is there draw it in the corner; Use a house symbol Use a green line as the path is green Where is it? Shall we use a blue line? Let's take turns

Support (S) Challenge (C)

positive comment with no justification negative comment with no justification

Yeah, like that; uh huh No way; that’s rubbish

Figure 4. Classification of functional talk Pilkington’s scheme reveals the function of language in achieving the children’s mapping goals. Once all 40 maps and transcripts are analysed, a selection of cases will be analysed in more detail, to establish the geographic and cartographic content of the talk. These cases will shed light on the maps with highest and lowest scores in each age group, in order to identify further the role of talk in successful and less successful mapping. The post mapping interviews were transcribed and coded to establish key themes regarding map construction, collaboration and evaluation of the software used.

RESULTS This section presents initial analysis of maps, discourse and interview responses. The function of this section is to guide the design and analysis of the final data collection sessions in June and July 2005. The full data set will be analysed to reveal statistical relationships between the map scores and the types of discourse, as well as variations related to age and gender. Figure 5 below shows the scores for the ten pairs of children who have completed the mapping activity at the time of writing.

0 4 1 1 2 2 4 4 4 4

0 0 2 0 0 0 3 3 2 2

3 4 4 3 4 4 4 4 5 4

2 1 1 2 1 3 2 1 1 1

Total score Max=30

Use of text

Amount of content -

Abstraction symbolisation

2 3 3 3 2 2 4 3 4 5

Mapping Ordinal Data

7-8 7-8 7-8 7-8 7-8 9-10 9-10 9-10 9-10 9-10

Proportion

Spatial arrangements

A and Y K and J K and O S and E M and O M and M R and I S and J B and B E and J

Age

Name

Map categorisation

1 2 1 3 1 0 1 1 1 1

8 12 12 12 10 11 18 16 17 17

Figure 5. Mapping scores Initial analysis reveals higher overall scores from the majority of 9-10 year olds. The older children were able to arrange the features and route of their trail more accurately than the younger children were, and their point, line and area features were organised in a more co-ordinated and integrated manner. The children’s attempts to map ordinal data were interesting. The more successful mapped ordinal area data using a variation in colour (red: high; yellow: medium; light grey: low) which could be generously be interpreted as approaching a ‘correct’ use of colour value according to Bertin’s visual variables. However, the majority of the pairs who did map ordinal data used a ‘traffic light’ schema in which green, yellow and red were used to represent the level of interest at a particular area (see Figure 6). In general, all children preferred pictorial symbols to abstract symbols when representing their trail. The younger children choose symbols (they referred to them as pictures) which most clearly resembled the referent. Examples can be seen in Figure 5. In general, all the pairs did not add on much extra detail to their maps such as key landmarks or routes, but instead focused on the construction of the trail itself. Text was used sparingly, mainly to give a title to the map and name features to be visited.

Figure 5. Map by 7-8 year old pair (S & E)

Figure 6. Map by 9-10 year old pair (J & E)

Analysis of mapping talk

Challenging

TOTAL MOVES

37.2 24.7

5.3 1.3

9.0 15.6

17.0 15.6

9.0 1.3

1.6 0

188 77

S and E (7-8)

S: 834

E: 821

20.1

35.8

0

16.7

13.0

11.0

4.0

299

M and O (7-8)

M: 328

O: 379 16.1

41.6

0

10.0

14.7

5.3

8.0

149

R and I (9-10)

R: 819

I: 1022 20.7

28.9

1.8

15.5

13.1

11.9

5.8

328

B and B (9-10)

BR:939 BE:783

30.2

26.8

7.5

8.3

14.7

6.4

3.4

265

E and J (9-10)

E: 869

23.8

34.4

3.8

9.6

6.1

14.2

8.4

261

J: 777

Other

18.6 38.9

Question

J: 473 O: 234

Reason

K: 456 K: 3O1

Direct

K and J (7-8) K and O (7-8)

Names + approx .age

Supporting

Inform % of total moves

Total number of words

Figure 7 shows a summary of the discourse for seven pairs of children. The table shows total number of words spoken by each participant (which can be used to judge symmetry of conversation) and the percentages of the functional discourse moves ( such as inform, direct etc.) alongside the total number of moves in each mapping session.

Figure 7. Functional discourse summaries There were no major examples of asymmetrical paired working . In general, the child who was not using the mouse spoke more than their partner did. All the groups relied heavily on direct and inform moves to meet their mapping goals, which reflect Wiegand’s findings working with secondary aged pupils (Wiegand, 2002b). The older children engaged in more reasoned statements, justifying why they were carrying out mapping activity. Older children were also engaged in a higher percentage of ‘challenging’ moves. If this pattern were to be found in the whole sample, it could be evidence of the older children being involved in more exploratory talk. It may be that a higher level of exploratory talk is associated with higher quality maps. This will be explored when all the forty pairs have completed the mapping activity. Figure 8 gives an example of a typical exchange whilst mapping.

Spoken : Child A

Non . Sp.

1.

2. 3. 4. 5. 6.

Spoken : Child B The Brownie pack or the Youth Club…let have the Brownies yes… no… shall we have… that one

Non – spoken Choosing from symbol palette Points to choice

Right then it’s too big

Drags symbol figure

Yeah You can smallen it Right - oh yeah - I forgot its bigger than the actual place

(Laughs)

7.

There

Resizes and positions figure

Figure 8. Choosing and locating a point symbol In this short extract, the children are choosing the symbol for the hall in which their Brownie pack meets. Child B has the mouse, opens the symbol palette, and moves the cursor over a choice of two symbols. Child A agrees with her choice. They then collaborate over resizing the symbol. This exchange illustrates two common features of the transcripts, first, that symbol choice to represent nominal point features was made quickly and often without much discussion, and second, that more collaboration typically occurred when positioning and resizing symbols to ensure accurate location and proportion in relation to the referent being mapped.

Post mapping interviews Key themes in the post mapping interviews are outlined below. Most of the pairs found it difficult to explain why they had selected particular symbols for the trail. Pictorial symbols were chosen ‘because it’s real’, ‘because it’s a really good picture’, ‘because pictures are easier to understand’. The children used ‘picture’ and ‘symbol’ interchangeably when discussing symbol choice. The children who mapped areal ordinal data explained the significance of traffic light colours in selecting red, green and yellow for areas. In general, all children were keen to talk about the significance of the referents in their map design, but less willing or able to verbalise their mapping strategies. All the pairs saw working collaboratively as a distinct advantage, so long as they were working with a friend or classroom work partner. They felt they were able to assist each other with locational decisions and with guidance on how to use the software. They were positive about the software tool and felt they had a wide range of symbols to choose from, and had access to a wide range of colours. Two pairs commented on difficulties in selecting and deleting objects on the map.

DISCUSSION Full analysis and discussion of results cannot take place until all the children have completed the mapping activity. However, analysis of results gathered so far point to the following trends:         

it may be that there is a relationship between quality of map and the presence of higher percentages of exploratory talk; it is likely that there is a relationship between quality of map and age; there is no evidence so far that gender is a significant factor in map quality or the frequency of different functional language types; all the children prefer pictorial rather than abstract symbols when mapping nominal point data, but they may be some evidence that older children prefer symbols that are more abstract than the younger children’s choice; the selection of an appropriate nominal symbol is generally completed quickly and without extended discussion or justification; more extended discussion occurs when attempting to resolve questions of spatial location and proportion between point , area and line data; older children have strategies for mapping ordinal data but these strategies often do not correspond with the guidance provided by Bertin (1983) and Robinson et al (1984) in relation to appropriate use of visual variables; few extra landmark features are added to the maps to aid navigation for the final users; children may be unsure of the role of map based text in communicating the purpose of their map.

Further work is now needed to analyse selected case studies. It is planned to select the following cases: the best map (highest score in grading); the best talk (judged by percentage of reason moves, most symmetry between pairs); the best map/talk combination; reliance on gesture and computer movement (high percentage of non-verbal moves); assymetrical pairs (one talker, one doer; academically able paired with less able). Analysis of these case studies will draw upon Wiegand’s work on the analysis of ideational discourse content ( Wiegand, 2002b); van Elzakker’s map use matrices (van Elzakker, 2004), and Lemke’s analysis of multi-modal communication science education (Lemke, 2001).

CONCLUSIONS It is possible to suggest tentative implications for teachers, software developers and researchers based on the research to date. Teachers need to develop strategies to ensure children can engage in exploratory talk, talk in which children are able to justify their decisions and respectfully question the decisions of others. Such programmes have been developed by Dawes et al (2002) and have been positively evaluated. Curriculum developers may also need to ensure that children can access specific geographical and cartographic language so that the children have the verbal ‘tools’ to collaborate over decisions of symbolisation, location and spatial relationships. Such language has an overlap with mathematical language (rotate, enlarge, reduce) and art/design language (pictorial, abstract, colour, shade) so in some contexts will already be part of children’s junior school experience. Software developers could improve children’s mapping applications by using Bertin’s graphic variables as a framework for hue and colour value selection. At present most examples of primary children’s mapping software provide children with a wide range of choice for colour hue, but do not lead the children to make choices according to colour value. If this was changed, children’s ability to apply the visual variables (reported by Filippakopoulou et al, 1999) could be further developed. Finally, this work opens up a research focus on the role of the teacher in supporting paired or group based computer mapping. The present study should reveal what children aged 7-10 can do in a peer-supported situation, so it would seem logical to plan further research on how adults can support and develop such peer interaction.

REFERENCES Bausmith, J.L., and Leinhardt, G. (1998), Middle-School Students' Map Construction: Understanding Complex Spatial Displays, Journal of Geography, 97, 93-107. Bertin, J. (1983). Semiology of Graphics: Diagrams, Networks and Maps. Madison: University of Wisconsin Press. (French Edition, 1967). Blaut, J., Stea, D., Spencer, C., and Blades, M. (2003), Mapping as a Cultural and Cognitive Universal. Annals of the Association of American Geographers, 93 (1), pp.165-185. Cohen, E.G. (1994), Restructuring the classroom: Conditions for productive small groups. Review of Educational Research, 64 (1), pp.1-35. Dawes, L., Mercer, N., and Wegerif, R. (2002), Thinking Together. Birmingham: Questions Publishing. DeLoache, J. S. Early development of the understanding and use of symbolic artefacts. In U. Goswami (2002) Blackwell Handbook of Childhood Cognitive Development, pp. 206-226. Blackwell :London. Dimitracoupoulou, A. and Ioannidou, I. (2003), Design of distributed collaborative activities for young children related to map use and construction. Paper presented to CSCL Conference, 2003. ESRI (2003), A road map for schools and libraries, URL: http://www.esri.com/industries/k-12/roadmap.html. Visited 10.02.03. Filippakopoulou, V.; Michaelidou E.; and Nakos, B. (1999), The Use of Primary Graphic Elements in Map design by first and Second grade Students. Proceedings of the Joint ICA Seminar of the Working Group on Cartography and Children and the Commission on Gender and Cartography. pp58-63 Montreal, Canada, August 10-12, 1999. Hart, R.A. and Moore, G.T. (1973), The development of spatial cognition: a review. In R.M. Downs and D.Stea (Eds.). Image and Environment, Chicago: Aldine. Harwood, D & Usher, M. (1999), Assessing progression in primary children's map drawing skills, International Research in Geographical & Environmental Education. 8 (3), 222-238. Keiper, T.A. (1999,) GIS for Elementary Students: An Inquiry into a New Approach to Learning Geography. Journal of Geography, 98, pp.47-59. Kraak, M, J, & Ormeling, F. (2003), Cartography. Visualization of Geospatial Data. 2nd Edition. London: Prentice Hall. Kress, G. & Van Leeuwen, T. (1995), Reading Images. The Grammar of Visual Design. London: Routledge. Kress, G. & Van Leeuwen, T. (2001), Multimodal Discourse. The Modes and Media of Contemporary Communication. London: Arnold. Leinhardt, G., Stainton, C., Bausmith, J.M. (1998), Constructing Maps Collaboratively. Journal of Geography, 97, pp.19-30.

Lemke, J. (2001), Multimodality. In Kress, G. Jewiit, C., Ogborn, J., and Tstatsarelis, C. Multimodal Teaching and Learning. London: Continuum. Liben, L.S. Spatial development in childhood: Where are we now? In U. Goswami (2002), Blackwell Handbook of Childhood Cognitive Development, pp. 326-348. Blackwell :London Liben. L. and Downs.R. (1989), Understanding maps as symbols: the development of map concepts in children, in Reese, H. (Eds) Advances in Child Development, New York: Academic Press. 145-201. MacEachren, A.M. (1995), How Maps Work. Guildford Press, New York. Matthews, M.H. (1992), Making Sense of Place. Harvester Wheatsheaf, Hemel Hempstead. Mercer, N. ( 1996), The quality of talk in children's collaborative activity in the classroom. Learning and Instruction, 64, pp.359-379. Mercer, N. (1995), The guided construction of knowledge: talk amongst teachers and learners, Clevedon: Multilingual Matters. Michaelidou, E., Filippakopoulou, Â., & Nakos, B. (2002), “Exploring children’s ability to categorize and symbolize”. Proceedings, Joint Seminar of ICA, Diamantina, Brazil, 4-7 August 2002. Moore, G.T. Theory and research on the development of environmental knowing In G.T. Moore and R.G. Golledge (Eds.) (1976) Environmental knowing: Theories, Research and Methods. Stroudberg, PA: Dowden, Hutchinson and Ross. Ordnance Survey (2002), The GIS Files, URL: http://www.ordsvy.gov.uk/gis-files/ . Visited 10.02.03. Owen, D. (2003), Collaborative electronic map creation in a UK 5-11 primary school: Children's representation of local space and the role of peer and teacher scaffolding in this process. 21st International Cartographic Congress, August 10-16, Durban, South Africa. Owen, D. (2004), Computer Supported Collaborative Mapmaking and Children’s Talk: Developing a Research Focus. In Catling, S. and Martin, F. Researching Primary Geography.pp. 109-119 London: Register of Research in Primary Geography. Piaget, J., Inhelder, B. (1956), The Child's Conception of Space . Translated by F.J. Langdon and J.L. Lunzer . London: Routledge and Kegan Paul. Piaget, J., Inhelder, B., Szeminska, A. (1960), The Child's Conception of Geometry. Translated by E.A. Lunzer, London: Routledge and Kegan Paul. Pilkington, R.M. (1999), Analyzing Educational Discourse: The DISCOUNT Scheme. Version 3, January 1999. CBL Technical Report No. 99/2 Leeds: University of Leeds Computer Based Learning Unit. Robinson, A. H. , Sale, R. D., Morrison, J.L., and Muehrcke, P.C. (1984), Elements of Cartography ( 5th ed.) New york: Wiley. Rogoff, B. (1990), Apprenticeship in Thinking: Cognitive Development in Social Context. New York: Oxford University Press. Rogoff, B. (2003), The Cultural Nature of Human Development. New York: Oxford University Press. Spencer, C., Blades, M., Morsley, K. (1989), The Child in the Physical Environment. Chichester: John Wiley. Van Elzakker, C.J.P.M. (2004), The use of maps in the exploration of geographic data. Netherlands Geographical Studies 326. Utretch: International Institute for Geo-Information Science and Earth Observation. Wegerif, R. (2001), Applying a Dialogical Model of Reason in the Classroom, in Joiner, R., Faulkner, D., and Littleton, K. (Eds.) Rethinking Collaborative Learning, Free Association Press. Wiegand, P. (2002a), School students' mental representations of thematic point symbol maps. Cartographic Journal , 39 (2), pp.125-136. Wiegand, P. (2002b), Analysis of Discourse in Collaborative Cartographic Problem Solving, International Research in Geographical and Environmental Education, 11,2, 139-159.

BIOGRAPHY David Owen is Primary Programme Leader in the Division of Education and Humanities, Sheffield Hallam University. He teaches geography and information and communications technology (ICT) to undergraduate and postgraduate teacher education students.