Making Statistics Learning Fun! - University of Texas at El Paso

advertisement
welcome to…
“Setting the Tone from Day One”
2012 American Society of Engineering Education Gulf Southwest Annual Conference
based on Lesser/Kephart paper in Nov. 2011 J. of Statistics Education,
which has all references (won “best paper” award at 2012 Sun Conference);
a related (Jan. 2012) webinar (also open-access) is archived at causeweb.org;
(partial support from NSF DUE-0618861 “Cultivating Authentic Discourse for the
2020 Engineer”, PI: L. Everett; included developing inquiry-based models based on
“counterintuitive concepts”)
Larry Lesser, Professor
The University of Texas at El Paso
Day 1 Matters
(Lesser & Kephart, 2011)
• Brouillette & Turner (1992): “Students will never
be more receptive…”
• Dorn (1987): high correlation of student evals
on Day 1 or 2 with those at end of the term
• Wilson & Wilson (2007) randomly assigned
students to + or - first day experience: the
former reported higher motivation for most of the
course, and ended with higher grades (p < 0.05)
Day 1 Likes & Dislikes
Lesser & Kephart (2011) has much more Day 1 lit review
• Faculty like: set positive atmosphere,
communicate objectives, introduce
oneself, preview content
• Students like: practical information,
grading standards, work required
• Students don’t like: covering new material,
HW assigned, using full class time
For a spring 2009 class, I wanted to….
• Use that first meeting (3 hours: a whole
week!) in a compelling way, beyond
discussing the syllabus and taking roll
• Set the tone (for a course about which
students have anxiety & misconceptions)
• Build classroom community/norms
• Model process for active learning
• Let them (and me) see that they have prior
knowledge/intuition to build on
instructional cycle*
Lesser & Kephart (2011)
Good problem presented
Initial Individual Reflection reduces anxiety, prepares students to share in
groups, especially ELLs (Fischer & Perez, 2008; Gibbons, 1998)
Small Group (3-4 students each) discussion
makes it more likely students will share in whole class; creates contexts for meaning (Rosebery, Warren &
Conant, 1992); active learning called for by ASA (2010)
Whole Class Discussion (with instructor as facilitator and co-learner)
optional: further individual reflection
* variation on “think-pair-share” (Lyman, 1981) or “1-2-4-whole group” (Minich, 2010) structures
plan to allow about 5 minutes per stage, but be prepared to modify as needed
a “liberating structure”
“We call [this] a ‘structure’ because it is a constraint imposed on the
participants. We call such a structure ‘liberating’ because it also
unleashes people to engage, in pairs and quartets, in conversations
and exchanges that would not happen going directly into the whole
group discussion.
People in pairs automatically talk to each other; this immediately
creates engagement of all participants. Groups of two create several
spaces for speaking up that are much safer than the whole group.
Quartets deepen the pairs’ exchanges in mostly safe spaces. These
additional conversations will frequently lead to significantly different
outcomes.
LS do not create a ‘free-for-all’ environment; rather the facilitator
maintains a well-defined but minimal structure, and freedom flourishes
within its confines.” -- Lipmanowicz & McCandless (2010, p. 9)
Henri Lipmanowicz & Keith McCandless “Liberating Structures: Innovating by Including
and Unleashing Everyone” March 2010 E&Y Performance, 2(4), 6-19)
a good problem...
•
•
•
•
•
•
Has context relevant to student backgrounds
Has minimal (math) prerequisites
Is efficient to pose
Is efficient to compute (e.g., simple numbers; Lesser, 2011)
Will be revisited/deepened later in course
Is open to multiple approaches,
representations, interpretations
• Stimulates habits of mind/questioning
• Shows statistics as “numbers with a context”
• May reveal a misconception
methodology
• DESIGN: case study; discourse analysis of
lesson
• DATA: transcribed video of all day 1 group
and class discussions
• RIGOR: transcripts checked and analyzed by
multiple researchers
• DELIMITATION: not designed to make claims
about student learning, but rather to identify
features of inquiry-based class
enough
theory/background/literature:
Let’s see some
problems!
“Student Test Performance”
Who did better, Amy or Bob?
fall
fall
spring
spring
TESTS
Amy
Bob
Amy
Bob
passed
3
1
2
3
taken
8
3
3
5
Who did better, Amy or Bob?
fall
fall
spring spring
full year
full year
Amy
Bob
Amy
Bob
TESTS
Amy Bob
passed
3
1
2
3
5
4
taken
8
3
3
5
11
8
3/8 > 1/3 ;
2/3 > 3/5 ;
but
5/11 <
4/8
Whole-class debrief
• Clarify what “doing better” means:
Amy had higher test passing % each
term, but Bob did for the year;
for number of tests passed, Bob won in
spring, but Amy won for fall and overall year
• Explore when comparisons can reverse
upon aggregation (Lesser, 2001), but delay explicit
label “Simpson’s Paradox” because that can
inhibit student exploration (Harper & Edwards, 2011)
• Make decisions carefully of when/how to group
Let’s try another problem….
that’s been in the news a lot
(e.g., recent front page of
El Paso Times)
“Average Class Size” Exploration (Lesser, 2009, 2010a)
185 students are divided among 7 rooms as:
20, 20, 20, 25, 30, 35, and 35.
What would you say is the ‘average class size’?
we can use a simpler substitute dataset:
20 students divided among 4 rooms as:
3, 3, 4, and 10
what’s Average Class Size?
Room 1: 3 kids
Room 2: 3 kids
Room 3: 4 kids
Room 4: 10 kids
Give answers your students would likely give.
Go ahead, call one out!
what’s Average Class Size?
Room 1: 3 kids
Room 2: 3 kids
Room 3: 4 kids
Room 4: 10 kids
answers I usually get:
5 (mean), 3.5 (median), 3 (mode),
and sometimes 6.5 (midrange)
Average per what?
Class
Student
mean
5
?
median
3.5
?
mode
3
?
Preceding answers were on per-class basis: {3,3,4,10}.
Now, have students use a “per-student basis” with:
{3,3,3, 3,3,3, 4,4,4,4, 10,10,10,10,10,10,10,10,10,10}
Al,Bob,Carl,Dee,Ed,Flo; Gil,Hal,Ivy,Jo; Kay,Lia,Mo,Ned,Olga,Pat,Qing,Ray,Sue,Ted
Average class size per…..?
mean
median mode
Class
5
3.5 3
Student 134/20 = 6.7 (10+4)/2 = 7 10
{3,3,4,10} for per-class basis
{3,3,3, 3,3,3, 4,4,4,4, 10,10,10,10,10,10,10,10,10,10}
for per-student basis
whole-class debrief
• Ambiguity of the word “average”: need to
specify not only which average, but also
basis unit over which you are averaging!
• Which basis results in a larger number?
• Which basis is more useful for consumers?
• How to count auditing students, online
courses, lab/recitation sections, etc.?
• Connection to “student-teacher ratio”?
some observations from transcripts
(see Lesser & Kephart 2011 for detail)
• Students often made assumptions explicit:
“My assumption was that each test has the
same weight” (Kara) [‘who did better?’]
• Professor was facilitator/co-learner, eliciting
multiple responses and finding something to
validate in each
• Technique: Professor asked class: “How do
you all think Kara came up with 3?” [‘avg. class size’]
Time for Discussion!
You’re invited to:
• Ask questions
• Share your insights & experiences
• Propose a new good opening problem
Thanks for attending this talk!
May you set a good tone
with your next class!
Professor Lesser
related archived webinar:
http://www.causeweb.org/webinar/activity/2012-01/
Which of these 5 countries did best at the
2008 Summer Olympics? (Isaacson, 2011)
Nation
Gold Silver
Bronze
China
51
21
28
USA
36
38
36
Zimbabwe
1
3
0
Bahamas
0
1
1
Australia
14
15
17
How to adapt for:
• Shorter periods
• Larger classes
• Other days of the course
• Other courses
Scaling to shorter periods
(Lesser & Kephart, 2011)
• Shorten periods in the instructional cycle
Individual, Small Group, Whole Class, Further Individual
• Move individual post-reflection to out of class
• Use problems even more streamlined, such as:
On a high-stakes test where minimum passing
score is a 7, which teacher’s class did better?
Mr. Jones’ 5 students’ scores were: 2,3,7,7,7.
Ms. Gomez’ 5 students’ scores were 4,5,6,6,8.
Scaling to larger classes
(Lesser & Kephart, 2011)
• Have more groups (but keep each group
size the same: 3 or 4 per group)
• You may not have time for every group to
report out, but you can have only groups
report out that had different results or
approaches than the first group reporting
Other days of the course
• When introducing a new concept/unit
• Day after test or holiday break (when
students did not do HW/reading since last
class meeting)
• Anytime you sense that they need a
reminder about the importance of
identifying and questioning assumptions
Setting the tone on the last day
(Hulsizer & Woolf)
• Have students write letters of accomplishment
& advice to future students
• Assess increase in learning & reasoning
• Present certificates of accomplishment
• Parting e-mail of appreciation to the class
• Tyrrell (2003) gave out individual bags of chips
and had students brainstorm how to apply all
tools they had learned to their supply of chips
How I started Day 1 for my intro
stat(literacy) class, spring 2012
• Showed many stories in that day’s front
page section and had them discuss how
they thought statistics was used in them
[ bar graph for GOP primary poll, graph of child obesity rates by state, fingerprints database upgrade,
redistricting maps(using Census data), projections for Alzheimer’s, wait times on El Paso bridges ]
•
•
•
•
Top 10 facts about me
Survey to learn about them
“Opening Intention” reading (Lesser, 2010)
Collection of anonymous data for repeated
future use
Download