Front

Statistics Learning Part 1

◆ students' concepts of average

◆ big ideas in statistics
◆ pedagogical considerations

-1-
 reflect

reflection on the studies and teaching

concepts of average, is it simple?
development of concepts takes time
concept of mode, median; not just definition and use
link to other concepts: variability, shape
research and teaching
students generated examples
doing and undoing
similarities to teaching of other topics: area,
factorization, ...
questioning, probing and task design
ambiguity in the curriculum

-2-
 CDI

Key Stage 3 Data Handling Dimension

http://www.edb.gov.hk/index.aspx?nodeID=4905&langno=1 CDI (1999) p.26

-4-
 Russell, Mokros

sented by that number. Statistically literate read-

ers can think about such a statement as "the me-
Russell, Susan Jo and Mokros, Jan
(February, 1996). What Do Children
dian price of a house is $150 000" or "the average
family size is 3.2" in terms of what it tells or does
not tell them about the distribution of the data.
Most students are unable to imagine what kind

Understand About Average? Edited by

of data an average might represent. One student,
who had had plenty of experience calculating the
mean, said, "I know how to an average, but I
don't know how to get the numbers to go into an

Donald L. Chambers. Teaching Children

What Do Children
Understand about
Average?
he statistical idea we come across most fre-
quently is the idea of average. Children in fourth
grade and beyond fairly easily learn to apply the
algorithm for finding the mean, but what do they
understand about the mean as a statistical idea?
'
Many students do not have opportunities to learn
about various kinds of averages as statistical con-
cepts. They view an average as a number found by a
particular procedure rather than as a number that
represents and summarizes a set of data. Students
may leam to find a mode, median, or mean - which
technically are all averages even though average
often refers to the mean - but they do not necessarily
know how these statistics relate to the data being
represented.
average, from an average."
Asking students to imagine what the data could
be for a given average yields interesting insights
into students' thinking about the relationship be-
Mathematics, 360-364.
tween data and the average of the data. They do not
find it easy to use the memorized algorithm to
construct data. They have to think about how the
average represents the data. You might want to try
the following problem, one we gave to students,
before reading about how students solve it.
We took a survey of the prices of nine different From: http://www.learner.org/courses/learningmath/data/overview/readinglist.html
brands of potato chips. For the same-sized bag,
the typical or usual or average price for all
brands was $1.38. What could the prices of the
nine different brands be?
We use the language "typical or usual or aver-
age" to keep the conversation open to any ways
that students have to think about an average. When
they show us one way, we ask them for other ways
so that we get a view of the range of their thinking.
Since these problems are administered in an inter-
view, we are able to interact, ask questions, and
probe students' thinking.

Avera,e as a
Statistical Idea
To investigate students' understanding
Journalfor Research
average, we lise what we call 1995,
Average as mode
In interviews with fourth graders, many students
Mokros, J. & Russell, S.J. (1995)
consistently associated the "typical or usual or
of the inidea
of average"
Mathematics Education value with the mode. In construction
"construction" prob- problems, they produced the data set by making all
Vol. 26, No. 1, 20-39

lems. Instead of asking students to find the average or most of the values the same as the average
for a given set of numbers, we give students an value. They might have made a few adjustments to
average and ask them what could be theCHILDREN'S data set it the data when pushed, but despite probing for
CONCEPTS OF AVERAGE
represents. This kind of problem is similar to situa- other approaches, these students stuck to a view of
tions we often encounter in life; we read about AND REPRESENTATIVENESS
Children's concepts of average and
a the average as the most frequent piece of data. As
median or mean in the newspaper or come across it in one fourth grader explained, "Okay, first, not all
our work and need to interpret what might be repre- chips are the same, as you told me, but the lowest
JAN MOKROS,chips I everTERC,
saw wasCambridge,
$1.30 myself,Massachusetts
so since the

representativeness.
SUSAN JOtypical RUSSELL,
price is TERC,
$1.38, I Cambridge, Massachusetts
just put most of them at
SusanJo Russell directs a K-5 curriculum project called Investigations in Numher, Data, and Space. $1.38, just to make it typical, and highered the
Her work focuses on how practicing teachers can learn more ahout mathematics and children's prices on a couple of them, just to make it
mathematical thinking. Jan Mokros codirects TERC's Math Center and recently Whenever
wrote athe need
book forarises to describea set of datain a succinctway, the issue of mathematical
realistic."
parents called Beyond Facts and Flashcards: Exploring Math with Your Kids (Portsmouth, N.H.:arises.The goal of this researchis to understandthe characteristicsof fourth
representativeness
Heinemann Publishers). througheighthgraders'constructionsof "average"as a representative numbersummarizinga data
Edited by Donald L. Chambers, Wisconsin Center for Education Research, University of
Average as mediaR
set. Twenty-onestudentswere interviewed,using a seriesof open-endedproblemsthatcalledon
childrento constructtheir
Another groupofof
own notion representativeness. Five on
students relied more basic constructionsof repre-
reason-

Journal for Research in Mathematics

Wisconsin-Madison, Madison, Wi 53706
sentativenessareidentifiedandanalyzed.These approachesillustratethe ways in which students
ableness in constructing the data from the average.
The action-research ideas in this article were prepared by Donald Chambers. are (or arenot) developinguseful, generaldefinitionsfor the statisticalconcept of average.
They drew on what was realistic in their own lives

Education. Vol.26, No.1, pp.20-39.

TEACHING CHILDREN MATHEMATICS
One objectiveof statisticsis to reduce large,unmanageable,anddisorderedcol-
lections of informationto summaryrepresentations.The need to summarizedata
is presenteven among young children.For example, in the surveys conductedby
primary-gradestudents,we see movementfrom focusing on individualpieces of
data("Ihave one brother")to highlightingand summarizingthe datain some man-
ageableform("Mostof the class membershaveonly one brotheror sister").As soon
as thereis the need to describea set of datain a more succinct way, the notion of
representativenessarises:Whatis typical of these data?How can we capturetheir
range and distribution?
The word average often emerges during children's discussions about data.
Youngerchildrenuse this wordin an informalway to referto typical,usual,or mid-
dle. Olderchildrenalso use the wordto indicatethe mean,median,or mode-terms
they have learned in school. The connections that children make-or fail to
make-between theirown ways of describingdataandthe "averages"thatthey are
learningaboutin mathematicsclass is a focus of this article.

BACKGROUND

Althoughchildrenandadultsalike have underdevelopednotions of average,we

know very littleaboutthe guidingconceptionsandmisconceptionsfromwhichchil-
drenbuild theirmodels of representativenessin a data set. It is often assumedby
educatorsandtextbookpublishersthataverageis simplyanotherapplicationof divi-
sion and thatchildrenwho understandthe notion of a fair or equal sharewill also
understandthe notion of average.

The work reportedin this paperwas supportedby NationalScience FoundationGrant

#MDR-8851114. All opinions, findings, conclusions, and recommendationsexpressed
herein are those of the authorsand do not necessarilyreflect the views of the funder.

- 12 -
 potato chips

Consider how students might respond to

this task ...
Construction Task: potato chips

We took a survey of the prices of nine

different brands of potato chips. For
the same-sized bag, the typical or
usual or average price for all brands
was $1.38. What could the prices of
the nine different brands be?

http://www.learner.org/courses/learningmath/data/session10/part_b/index68.html
- 13 -
 Untitled

The potato-chip task was presented to students in individual interviews to

research students' understanding of average. Here are some of the
students' responses:

1. Some students would put one price at $1.38, then one at $1.37 and one at
$1.39, then one at $1.36 and one at $1.40, and so forth.
2. One student commented, “Okay, first, not all chips are the same, as you told me,
but the lowest chips I ever saw was $1.30 myself, so, since the typical price is
$1.38, I just put most of them at $1.38, just to make it typical, and highered the
prices on a couple of them, just to make it realistic.”
3. One student divided $1.38 by nine, resulting in a price close to 15¢.When asked
if pricing the bags at 15¢ would result in a typical price of $1.38, she responded,
“Yeah, thatʼs close enough.”
4. When some students were asked to make prices for the potato-chip problem
without using the value $1.38, most said that it could not be done.
5. One student chose prices by pairing numbers that totaled $2.38, such as $1.08
and $1.30. She thought that this method resulted in an average of $1.38.

How do these students reason differently about average?

http://www.learner.org/courses/learningmath/data/session10/part_b/index68.html
- 14 -
 framework

Conceptual Framework
1. investigate children's notions of
representativeness
2. how new math knowledge grows from
informal understanding
3. examine how children construct and
describe data sets -> how they
understand average
4. average is a tool for summarizing,
describing a data set, comparing data
sets
Mokros & Russell, 1995
- 15 -
 background

background
◆ 21 students (grade 4, 6, 8) were
interviewed
◆ all taught how to compute an average as
part of their regular math class
◆ most of the students' exposure to average
consisted of practice with the algorithm,
using examples such as finding the
average score on a test
◆ the students had no special experience
with data in their math classes

Mokros & Russell, 1995

- 16 -
 approaches

students' approaches
◆ average as mode
◆ average as algorithm
◆ average as reasonable
◆ average as midpoint
◆ average as mathematical point of
balance

Mokros & Russell, 1995

- 17 -
 fig 1

With experience, students can begin envisioning the variety of

data sets that might be represented by a median value of, for
example, 48 inches.

Russell & Mokros, 1996, p.361

- 18 -
 fig 2

... introduce the idea of asymmetrical balancing by asking what

would happen if one family has 8 people. In this situation, the
move from 4 to 8 cannot be balanced by a comparable move
on the left-hand side of the mean ...

Russell & Mokros, 1996, p.363

- 19 -
 compare data

http://education.uncc.edu/dkpugale/maed5040_spring2006/Proportional%20Reasoning%20Lesson%20from%20Research%20in%20Data%20and%20Chance.pdf

J A N E M. W A T S O N AND J. M I C H A E L S H A U G H N E S S Y

Watson, J.M. & Shaughnessy, J.M. (2004).

Proportional Proportional reasoning: lessons from research in
Reasoning: data and chance. Mathematics Teaching in the
Lessons from Research in Data and Chance
Middle School. 10(2) pp.104-109.

P
RINCIPLES AND STANDARDS FOR SCHOOL In our view, the goal of the middle school curricu-
Mathematics (NCTM 2000) places propor- lum is for students to recognize, without being told,
tionality among the major concepts con- that proportional reasoning is fundamental to prob-
necting different topics in the mathematics lem solving across the curriculum, particularly in
curriculum at the middle school level (p. 217). the areas of data and chance. We will discuss two
What concerns us about many of the problems pre- contexts where proportional reasoning is needed for
sented to students, however, is that they are often a thorough understanding of the problem. The first
posed purely as a ratio or proportion from the start. task presents two data sets of different sizes in
Often the statement of a problem is a giveaway that graphical form along with a question about which a
a proportion is involved. For example, the question data set represents better performance by a class of
“If 15 students out of 20 get a problem correct, how students. In this setting, the focus is on the use of
many students in a class of 28 would we expect to the arithmetic mean as a formal tool reflecting pro-
get the problem correct?” does not tap the depth of portional reasoning, supported by visual intuitions
proportional reasoning that is required for mean- of proportion. The second task involves proportional
ingful problem solving. reasoning as students draw repeated samples from
a mixture of colored objects and either attend to or
JANE WATSON, Jane.Watson@utas.edu.au, is a teacher ignore variability when predicting the results of the
educator and mathematics education researcher at the repeated samples. Proportional reasoning
JANE M. WATSON is funda-B. MORITZ
and JONATHAN
University of Tasmania, Australia. She is interested in all mental to making connections between populations
aspects of statistics education, particularly in students’ and samples drawn from those populations; in the
growth in statistical literacy and in students’ understand- later grades, it provides a basis for statistical infer- INFERENCE:
THE BEGINNING OF STATISTICAL
ings of statistics in the media. MIKE SHAUGHNESSY, COMPARING
ence. It is our belief that teachers TWO DATA
and students maySETS
mike@mth.pdx.edu, is a teacher educator and mathemat- overlook opportunities to make connections with ap-
ics education researcher at Portland State University in plications of proportional reasoning in tasks such as
Oregon. He has a long-standing interest in students’ and comparisons of data sets or predictions of the com-
teachers’ understandings of probability and statistics and position of samples.The
ABSTRACT. Both of these
development typesstudents’
of school of tasks re-
understanding of comparing two data

Watson, J.M. & Moritz, J.B. (1999) The

in the development of conceptually engaging statistical quire comparisons
sets is exploredof ratios,
through thus,ofproportional
responses rea-interview settings. Eighty-eight
students in individual
students in grades 3 to 9 were presented with data sets in graphical form for comparison.
tasks for both students and teachers. soning. The word lessons in our title has a double
Student responses were analysed according to a developmental cycle which was repeated
in two contexts: one where the numbers of values in the data sets were the same and
104 MATHEMATICS TEACHING IN THE MIDDLE SCHOOL the other where they were different. Strategies observed within the developmental cycles
Copyright © 2004 The National Council of Teachers of Mathematics, Inc. www.nctm.org. All rights reserved.
werevisual, numerical, or a combination of the two. The correctness of outcomes asso-

Beginning of Statistical Inference: Comparing

This material may not be copied or distributed electronically or in any other format without written permission from NCTM.
ciated with using and combining these strategies varied depending upon the task and the
developmental level of the response. Implications for teachers, educational planners and
researchers are discussed in relation to the beginning of statistical inference during the
school years.

Two Data Sets. Educational Studies in

The need to develop an understanding of statistical inference is acknow-
ledged in recent national curriculum documents which include chance and/
or data handling as part of the mathematics curriculum. In Australia for
example, A National Statement on Mathematics for Australian Schools
(Australian Education Council [AEC], 1991) has a separate heading un-

Mathematics. vol. 37, pp.145-168.

der Chance and Data for the middle school years onward titled ‘Statistical
Inference.’ While leading to the ‘use of estimates of population paramet-
ers and confidence intervals’ in the senior years (p. 185), in earlier years
it is suggested that students ‘draw inferences and construct and evaluate
arguments based on sample data’ (p. 179). Other countries make similar
suggestions (e.g., Department for Education and the Welsh Office [DFE],
1995; Ministry of Education, 1992).
There are many points in the statistics curriculum where the idea of
inference can be introduced, for example using summary statistics to com-
pare a data set with an hypothesised model, or looking at the relationship
of two variables with correlation. Usually in formal statistics courses, basic
hypothesis testing begins with a single data set and an hypothesised pop-
ulation model rather than with a comparison of two data sets to look for
differences in populations. This is likely to be due to the relative complex-
ity of the formulae involved in testing hypotheses in the two situations. For
young students, however, it may be more sensible and relevant to compare
two groups with each other than to compare one against an hypothetical
model. Being close to many out-of-school applications, comparisons of

Educational Studies in Mathematics 37: 145–168, 1999.

© 1999 Kluwer Academic Publishers. Printed in the Netherlands.

Article: EDUC 779/DISK Pips nr. 202353 JM.WEB2C (educkap:humnfam) v.1.15

educ779.tex; 1/06/1999; 19:34; p.1

- 20 -
 part a

Task: Comparing Two Data Sets

Two schools are comparing some

classes to see which class is
better at quick recall of 9 math
facts. Consider the blue and red
classes.
...
Did the two classes score equally
well, or did one of the classes
score better? Explain how you
decided.

Watson & Shaughnessy (2004)

- 21 -
 part b

Task: Comparing Two Data Sets

Two schools are comparing some

classes to see which class is
better at quick recall of 9 math
facts. Consider the green and
purple classes.
...
Did the two classes score equally
well, or did one of the classes
score better? Explain how you
decided.

Watson & Shaughnessy (2004)

- 22 -
 part c

Task: Comparing Two Data Sets

Two schools are comparing some

classes to see which class is
better at quick recall of 9 math
facts. Consider the yellow and
brown classes.
...
Did the two classes score equally
well, or did one of the classes
score better? Explain how you
decided.

Watson & Shaughnessy (2004)

- 23 -
 part d

Task: Comparing Two Data Sets

Two schools are comparing some

classes to see which class is
better at quick recall of 9 math
facts. Consider the pink and black
classes.
...
Did the two classes score equally
well, or did one of the classes
score better? Explain how you
decided.

Watson & Shaughnessy (2004)

- 24 -
 4 cases

Task: Comparing Two Data Sets

In each case, was there a class that you think did

better or did the classes do equally well—and
why? What do you think your students would say
for these comparisons? How do you think they
would decide which class did better?

Watson & Shaughnessy (2004)

- 25 -