HOW DO STUDENTS AND TEACHERS

DEAL WITH MODELLING PROBLEMS?

Werner Blum and Dominik Leilj

University of Kassel, Germany

Abstract-In this paper, we shall report on some of the work that has been, and is being, done
in the DISUM project. In $1, we shall describe the starting point of DISUM, the SINUS
project aimed at developing high-quality teaching. In $2, we shall briefly describe the DISUM
project itself; and in $3 we shall present and analyse a modelling task from DISUM, the
“Sugarloaf‘problem. How students dealt with this task will be the topic of $4, the core part
of this paper. How experienced SINUS teachers dealt with this task in the classroom will be
reported in $5. Finally, in $6, we shall briefly describe future plans for the DISUMproject.

1. THE CONTEXT: QUALITY DEVELOPMENT

The teaching of mathematics in school is aimed toward supplying students with

knowledge, skills, competencies and attitudes so that they are able to use
mathematics in a well-founded manner when solving mathematical or real-world
problems. We know from educational research that the desired effects of
mathematics teaching can only (at most) be achieved if mathematics teaching obeys
certain non-trivial criteria for “high-quality teaching”. The following set of quality
criteria constitutes our definition of “Good Mathematics Teaching” and is the basis
of all our research and development activities (for details see Blum & Leifi, 2006):
I. Demanding orchestration of the teaching of mathematical subject matter
That means in particular: Providing manifold opportunities for learners to
acquire competencies, especially modelling ability, and creating manifold
connections, within and outside mathematics.
II. Cognitive activation of learners
That is: Stimulating permanently cognitive activities of students, including
metacognitive activities (that is, a conscious use of strategies and reflections
upon one’s own activities), and fostering students’ self-regulation and
independence.
III. Effective and learner-oriented classroom management
That means, for instance, separating learning und assessing consciously and
using pupils’ mistakes in a constructive way as learning opportunities.
In all aspects, the teacher has a crucial role to play. We can speak, in the words of
Weinert (1 997), of “learner-centredand teacher-directed teaching.
”
 Blum and Leirj 223
We all know that everyday teaching is usually far from “good teaching” in this
sense - also and particularly in Germany. The unsatisfactory TIMSS results induced
the German government (both federal and states) in 1998 to establish a reform
project that aimed at improving the quality of mathematics (and science) teaching,
called SINUS. It started in 1998 with 180 schools; at present, more than 1500
schools are involved. The guiding principles of SINUS are:
0 The “new culture of teaching”: Following consequently those quality criteria in
all teaching activities.
0 The “new culture of tasks”: Treating (“what?”) a broad spectrum of
competency-oriented tasks, notably modelling tasks, (“how?”) in ways obeying
the quality criteria.
0 The “new culture of cooperation ”: The new ways of teaching mathematics are
embraced and supported by the whole mathematics staff of a school, not only by
individual teachers.
However, classroom observations showed numerous shortcomings even in the
ambitious SINUS programme. Some of these shortcomings are definitely not due to
a lack of practical realisation of existing knowledge by the SINUS teachers, but
rather to a lack of knowledge both
0 of the actual procedures and difficulties of students when solving cognitively
demanding tasks, and
0 of appropriate ways for teachers to act when diagnosing students’ solution
processes and when intervening in case of students’ difficulties;
that is, a lack of corresponding research. It is really surprising how little we know
about the micro-structure of students’ and teachers’ dealing with such cognitively
demanding tasks.

2. THE FRAMEWORK: THE DISUM PROJECT

This lack of knowledge was the starting point for the research project DISUM
(“Didactical intervention modes for mathematics teaching oriented towards self-
regulation and directed by tasks”). DISUM is an interdisciplinary project between
mathematics education and pedagogy at the University of Kassel (see LeiB, Blum &
Messner, 2004). It investigates how students and teachers deal with modelling
problems, mainly in grade 9. A lot of activities have already been carried out in
DISUM, especially:
1. The construction of appropriate modelling tasks.
2. Detailed cognitive and subject matter analyses of these tasks (constructing the
“task space”, based on the modelling cycle).
3. A detailed study and theory-guided description of actual problem solving
processes of students in laboratory situations (involving pairs of students,
sometimes with and sometimes without a teacher).
4. A detailed study and theory-guided description of actual diagnoses and
interventions from teachers in these laboratory situations.
5 . A detailed study of regular Zessons with such modelling tasks, taught by
experienced teachers from the SINUS project, and a theory-guided description
of these lessons, emphasising our quality criteria.
224 How do Students and Teachers Deal with Modelling Problems?
6. The construction of various instruments to measure students’ achievements and
attitudes.
7. The construction of manageable and promising tools for
a) the training of students in strategies for solving modelling problems, and
b) the training of teachers in “well-aimed coaching” of modelling problems
How the DISUM project will be continued will be reported in $6.

3. AN EXAMPLE FROM DISUM: THE “SUGARLOAF” TASK

Sugarloaf
From a newspaper article:
The Sugarloaf cableway takes
approximately 3 minutes for its ride from
the valley station to the peek of the
Sugarloaf Mountain in Rio de Janeiro. It
runs with a speed of 30 km/hr and covers a
height difference of approximately 180 m.
The chief engineer, Giuseppe Pelligrini,
would very much prefer to walk - as he did
earlier on, when he was a mountaineer, and
first ran from the valley station across the
vast plain to the mountain and then climbed
it in 12 minutes.
How big is the distance, approximately, that Giuseppe had to run from the
valley station to the foot of the mountain? Show all your work.

A global cognitive analysis yields the following ideal-typical solution, oriented

towards the modelling cycle (Figure 1):
 Blum and LeiB 225

First, the text has to be read and the problem situation has to be understood by the
problem solver; that is, a so-called situation model has to be constructed, supported
by the photo (Figure 2):

Unknown:
running distance valley station - foot of mountain

foot of mountain Giuseppe valley station

Pelligrini

Figure 2. A situation model of the “Sugarloaf” task.

226 How do Students and Teachers Deal with Modelling Problems?

Then the situation has to be simplified, structured and made more precise, leading to
a real model of the situation (Figure 3):

Unknown:
length of the horizontal distance

peak

180 m

foot of mountain hoizontal distance valley station

Figure 3. A real model of the “Sugarloaf’ task.

Mathematisation transforms this real model into a mathematical model, basically a
right-angled triangle where one side length is unknown. Then mathematical tools are
activated, in the basic version of the solution process mostly the Pythagorean
Theorem, yielding the mathematical result: b = d1.5’- 0.1 8’ = 1.49 [km]
This result has to be interpreted in the real world as the real result: The distance
from the valley station to the foot of the mountain is approximately 1.5 km.The
important next step is a validation of this result: Is it reasonable? Is the accuracy
appropriate (taking into account Keynes’ well-known aphorism: “It is better to be
roughly right than precisely wrong!”)? First: For the result 1.5 km, no application of
Pythagoras would have been necessary; in a triangle where the length of one side is
approximately only of the length of another side, the length of the third side is
approximately equal to the length of the longer side! Second: The simplification
that the mountain has no width is certainly inappropriate, so something like 1.4 km
will be a more reasonable answer. Third: The rope of the cableway is certainly not
straight. Fourth: The speed of the cableway is definitely smaller near the beginning
and near the end (which results in a considerably smaller distance). So one might go
round the loop in Figure 1 several times. In any case, the whole process ends with
an exposition of a final answer to the original problem (see Figure 4):
This version of the modelling cycle has been influenced by various sources,
among others by the cognitive theories of Reusser (1998) or Verschafel, Greer &
deCorte (2000). We would like to emphasise that this version is more oriented
towards the problem solving individual than the versions usually found in the
ICTMA context, for instance in Blum (1995). We are convinced that it is a better
model of what problems solvers actually do. In particular, step 1, reading and
 Blum and LeiD 221
understanding, is strongly individually shaped, that means the resulting situation
model is an idiosyncratic construction of the problem solver (according to
constructivism and situated cognition; see, for example, de Corte et al, 1996). This
version of the modelling cycle has proved extremely helpful for our purposes. It
provides a better understanding of what students do when solving modelling
problems and it gives teachers a better basis for their diagnoses and interventions.

Figure 4. A pupil’s solution of the “Sugarloaf’ task.

Of course, actual individual problem solving processes are usually not as linear
as it is suggested by this model. For instance, the second step, simplifying and
structuring, is already influenced by the mathematical tools available to the problem
solver, and often the process goes several times back and forth between the real
world and mathematics. It is very interesting to identify students’ “modellingroutes”
and to compare these (see Chapter 5.5, Borromeo Ferri).

4. “SUGARLOAF” IN THE LAB: STRENGTHS AND DIFFICULTIES OF

STUDENTS

As mentioned in part 2, we have observed, videographed and interviewed

9‘hgraders solving modelling tasks, including the “Sugarloaf’ problem. We have
selected pairs of students of each of four “competency levels”, from weak
Hauptschule students to strong Gymnasium students.
The most obvious strength exhibited by students was - not surprisingly in
Germany - the procedural part of the process, that is the use of the Pythagorean
Theorem and the calculation of the unknown length, often to a ridiculous degree of
accuracy: (“b = 1489.16 m”). This presupposes, however, that the students got to
this point. Some students got lost before; they failed already in step 1 or step 2, also
because the text contains an unnecessary datum (the 12 minutes). There is a lot of
228 How do Students and Teachers Deal with Modelling Problems?
evidence from research, as we know, that reading a text and understanding both
situation and problem is a considerable cognitive barrier for students. In the
“Sugarloaf’ example, this was the most diflcult part of the task. Excerpt 1 shows an
example of this difficulty:
Excerpt 1 (Hauptschule students)
0.: How many kil ... How long that takes with the three minutes, that is when it drives 30
W h , that thing.
P.: Yes.
0.: And is over here in three minutes - how much does it cover then? How many kilometres
or how many metres? How should I know that?
P.: Wait, wait. That is ...
0.: 15 km, with 15 km/h it takes halfan hour.
P.: Rule of three!
0.:Rule of three!
The pupils calculate with the ”rule of three ” @roportions) that the distance is 1,5 km.
0.: 1.5 km.
P.: How long does it take ...
0.: 1500 metres. That is the distance that it takes approximately.
P.: That is it takes him 1.5 km.
0.: Yeah but what - perhaps the 12 minutes mean something as well.
P.: Here they ask how far he has run and not how long!
0.: 1.5 km, 1500 m.
P.:Yes.
The students can easily calculate by the “rule of three” s = 1500 m from the
given v and t and then they are convinced they are completely done. For a moment
Osman realises that they have not used the given “12 minutes” in their calculation,
but Pascal’s they ask how far he has run, not how long” stops the solution process.
This is a well-known phenomenon, a typical strategy especially for German
students: “You don’t have to understand the situation, just use the data of the task in
some way”. In other words: The students do not construct an appropriate situation
model.
Excerpt 2 (see p229) shows another section in the modelling process.
Katherina would like to apply the Pythagorean Theorem, but Christoph realises
that the situation model contains no right-angled triangle. Katherina obviously feels
puzzled. After some discussion, Christoph suggestion “Shouldn’t we do it
approximately like that, with a right angle?” helps Katherina to overcome her
cognitive barrier. She realises that “the mountain is very steep”, gives a reason for
the assumption of a right angle and thus accepts Christoph’s simplification (that is
the common real model). Immediately thereafter, the students draw a triangle and
begin to work successfidly within the mathematical model.
These two examples demonstrate how cognitively demanding the first two steps
are. So we should not simply and globally say (as it is often done) that the
mathematisation part of the modelling process is difficult but look upon the process
a bit more carehlly.
 Blum and Leilj 229

Excerpt 2 (Gymnasium students)

K.: Hight difference. That has nothing to do with width.
C.: No, but if you want to do this Pythagoras now, you need a right angle here and you don’t
have it.
K. : Why not?
C.: Is that a right angle here, if you ...
K.: Height difference, they speak of the height difference. That’s here.
C.: No, height difference is no right angle.
K.: Oh, the mountain is 180 m high.
C.: But you want to calculate the distance here.
K.: Yes, but now one can perhaps - yeah! You only have to know how wide the mountain is.
C.: It says “approximately”. Shouldn’t we do it like it says, approximately, with a right angle?
The pupils laugh.
C.: I think that’s sufficient.
K.: Yeah, it says the mountain is very steep, and if the mountain is very steep then it’s not so
much what is more here, then we can do it like that.
C.: That’s what I’m saying!
K.: Super!

Another uniform shortcoming was the lack of validation and of substantial

reflection. We know from research into teaching and learning how important it is to
look back and to reflect on one’s own problem solving process, that is - in the words
of Reusser (1998) - “to extract the relevant conceptual-schematic and processual-
strategic characteristics of a problem solution in an abstracting way” and, thus, to
contribute to the hrther development of meta-cognitive knowledge. More generally,
no conscious use of problem solving strategies by students was recognisable; in
particular, no student seemed to have the modelling cycle as a guiding tool at his or
her disposal. As a consequence of the absence of validation, no student tried to
improve his or her solution, they were all satisfied when they had reached any
solution whatsoever.

5. “SUGARLOAF” IN LESSONS: STRENGTHS AND DIFFICULTIES

OF TEACHERS

As we said in part 2, we have also observed and videographed lessons with our
modelling tasks, in all types of schools, from Hauptschule to Gymnasium. In
particular, in the context of a “Best Practice Study”, several experienced teachers
from the SINUS project included some of our tasks in their regular lessons, among
others the “Sugarloaf’ problem. Here are a few results of our observations.
First of all, most lessons clearly stood out positively from typical German
lessons. Looking more thoroughly at these lessons with our “quality glasses” shows
that (see Blum & LeiR, 2006)
0 the students had opportunities to model, to argue, to communicate;
0 mental activities were stimulated;
0 students, for the most part, could work independently;
0 the atmosphere was tolerant towards mistakes and free of assessment;
0 there was a discussion in the end, as an element of reflection.
230 How do Students and Teachers Deal with Modelling Problems?
None of these aspects is self-evident, on the contrary. For instance, the TIMSS
Video Study in six high performing countries has revealed that even in these
countries there was normally no reflection at all in the end of solution processes. All
teachers, including those in the Hauptschule, succeeded in helping the students to
find a solution. This is again an example of the important and often neglected
distinction between working independently, with support from the teacher, on the
one hand, and working totally on one’s own, on the other hand. Thus, the crucial
role of the teacher also in more student-centred learning environments is highlighted
once again.
Another important feature was to realise during the lessons how imperative it is
that teachers have a detailed and intimate knowledge of the modelling cycle, both as
a solid basis for diagnoses and as a rich source for supporting interventions. In one
of these lessons, a teacher succeeds in leading his Hauptschule class to an
appropriate real model, as an important intermediate stage between the situation
model and the mathematical model (T.: “It only matters what is mathematically
relevant, no trees or so”, S.: “Hence a triangle”).
A problem reflected even in those Best Practice teachers’ actions was the
absence of further reflections on the solution processes. In particular, no teacher
considered the question of how certain assumptions influence the solutions and how
accurate the results actually could then be. Often, also the teachers were satisfied
when the students had any solution. Equally, the counterpart of the above-mentioned
absence of problem solving strategies on the students’ side was the absence of any
stimulation of such strategies by the teachers. Such stimulation seems not to be a
part of teachers’ everyday repertoires. Consequences for teacher education are
obvious.

6. PROSPECTS FOR DISUM: THE NEXT PHASE

We are convinced that both

- a conscious use of problem solving strategies by students, and
- well-aimed and independence-supportingcoaching by teachers
will have positive effects on students’ achievement. However, this has never been
examined systematically and quantitatively in the context of demanding modelling
tasks. We will compare the effects of teachers’ training in well-aimed coaching with
the effects of “normal” instruction or of no teaching at all (students working totally
on their own). The notion of “well-aimed coaching” has a definite meaning; it
consists of five components, one being the DISUM strategy set for students (see
Blum & LeiB, 2006): goals/volition/organisation/strategy/evaluation.
We shall measure the effects on achievement by a classical pre-tesdpost-test
design. Our tests contain modelling tasks as well as tasks from the current curricular
topic (in our study: the Pythagorean Theorem and its context) and anchor items from
PEA. In addition, we shall measure the effects on students’ motivation, emotions
and independence feeling by classical scales taken from PISA and from the German
project PALMA (Pekrun et al. 2004).
Our final aim with DISUM is, of course, to improve both everyday teaching and
teachers’ expertise by implementing our instruments and findings in school
classrooms and in teacher education programmes.
 Blum and Lei0 23 1
REFERENCES
Blum, W. (1995) Applications and Modelling in Mathematics Teaching and
Mathematics Education - Some Important Aspects of Practice and of Research.
In C. Sloyer, W. Blum and I. Huntley (eds): Advances and Perspectives in the
Teaching of Mathematical Modelling and Applications. Yorklyn: Water Street
Mathematics, 1-20.
Blum, W. and Lei& D. (2006) "Filling up" - The Problem of Independence
Preserving Teacher Interventions in Lessons with Demanding Modelling Tasks.
In M. Bosch (ed) Proceedings of the European Society for Research in
Mathematics Education CERME 4 (2005). IQS FUNDIEMI Business Institute
www.fundemi.url.edu, 1623-1633. ISBN 84 61 1 3282 3.
deCorte, E., Greer, B. and Verschaffel, L. (1996) Mathematics Teaching and
Learning. In Berliner, D. and Calfee, R. (eds): Handbook of Educational
Psychology. New York: Macmillan, 49 1-549.
LeiS, D., Blum, W. and Messner, R. (2004) Sattelfest beim Sattelfest? Analyse
ko-konstruktiver Losungsprozesse bei einer realiBtsorientierten
Mathematikaufgabe. In Beitrage zum Mathematikunterricht 2004. Hildesheim:
Franzbecker, 333-336.
Pekrun, R., Gotz, T., vom Hofe, R., Blum, W., Jullien, S., Zirngibl, A., Kleine, M,
Wartha, S. and Jordan, A. (2004) Emotionen und Leistung im Fach Mathematik:
Ziele und erste Befunde aus dem ,,Projekt zur Analyse der Leistungsentwicklung
in Mathematik" (PALMA). In J. Doll and M. Prenzel (eds): Bildungsqualitut
von Schule: Lehrerprofessionalisierung, Unterrichtsentwicklung und
Schulerfdrderung als Strategien der Qualitatsverbesserung. Munster: Waxmann,
345-363.
Reusser, K. (1998) Denkstrukturen und Wissenserwerb in der Ontogenese. In Klix,
F. and Spada, H. (eds): Enzyklopadie der Psychologie, Themenbereich C:
Theorie und Forschung, Serie II: Kognition, Band G: Wissenspsychologie.
Gottingen: Hogrefe, 115-166.
Verschaffel, L., Greer, B. and de Corte, E. (2000) Making Sense of Word Problems.
Lisse: Swets & Zeitlinger.
Weinert, F.E. (1997) Neue Unterrichtskonzepte zwischen gesellschafilichen
Notwendigkeiten, padagogischen Visionen und psychologischen Moglichkeiten.
In Bayerisches Staatsministerium f i r Unterricht, Kultus, Wissenschaft und Kunst
(ed.): Wissen und Wertefur die Welt von morgen. Miinchen, 101-125.