Investigating And Visualizing
Measurement Error
for Novice “STEM” Learners
Scott A. Sinex
Department of Physical Sciences and Engineering
Prince George’s Community College
Largo, MD
Presented at ASEE Conference 2014 in the Best Zone Paper Competition
held in Indianapolis, IN on 15-18 June.
…want students to be able to
•
•
•
•
•
•
Know how good a measurement is
Graph and transform data
Derive mathematical models
Judge goodness-of-fit
Use simulations to explore multivariable models
Make predictions, interpret results, draw
conclusions
• Develop conceptual understanding
2
Students…
Calculate %error
List some possible
causes
Propagation
of error
Change the mindset of students
about how to think about errors
Faculty…
think like a novice and not an expert!
craft an experiment with careful scaffolding
3
Developed four experiments…
ruler
Stacking cookies
Stacking nested Styrofoam cups
balance
Nuts & Bolts of extrapolation
Nuts & Bolts of let’s make an error
Each has data collection, model development, and simulations with error analysis
4
…and what do students do?
So let’s examine the stacking cookies experiment
5
Just add data and presto you get a model
Enter data
Model (by linear regression)
Simulation to
manipulate a variable
Predictions
http://academic.pgcc.edu/~ssinex/excelets/NC/cookies_stack.xls
6
For the Stacking Cookies Experiment
How would you characterize the use of the “just add data” interactive Excel
spreadsheet for this activity? Scale: very easy/easy/so, so /difficult/very difficult
Fall 2006 (n = 19 students)
3.5 out of 5
89% no difficulty
Fall 2007 (n = 10 students)
3.4 out of 5
100% no difficulty
Fall 2008 (n = 15 students)
3.6 out of 5
93% no difficulty
Fall 2010 (n = 15 students)
3.5 out of 5
80% no difficulty
Did you happen to notice that the rulers did not have the zero centimeter mark at the
end? (all rulers had the zero mark off from the end)
Yes
No
Fall 2006
53% noticed error
Fall 2007
40% noticed error
Fall 2008
33% noticed error
Fall 2010
67% noticed error
If you did (circled yes in question 2), did you correct for it? Be honest!
Fall 2006
37% of total corrected error
Fall 2007
20% of total corrected error
Fall 2008
27% of total corrected error
Fall 2010
27% of total corrected error
Yes
No
7
…and what about ruler error?
Here we
have
induced a
constant
error
http://academic.pgcc.edu/~ssinex/excelets/Ruler_error.xls
8
Nuts & Bolts of Extrapolation
Mtotal
Mtotal = Mnut* n + Mbolt
Verify model by
just add data
spreadsheet
Mass bolt with
from one thru
five nuts on it
Enter data into
Google form
for discussion
Mass of bolt
by
extrapolation
Student data
…and they whine over
the finding the equation
Determine mass
of the bolt on a
balance - % error
n
9
The extrapolated value of the bolt
10
The pedagogical approach…
…so now we can extent concepts using simulations
11
Suppose we used magnesium nuts?
http://epublications.bond.edu.au/ejsie/vol5/iss1/3/
12
Why change?
• NRC Framework  Next Generation Science
Standards
• Common Core Standards in Mathematics
• AMATYC - The Right Stuff
• MAA CRAFTY College Algebra
• ASEE Innovation with Impact
• Enhance 21st-century technology in classroom
(ISTE, Educause…)
13
…and students
•
•
•
•
•
•
Are actively engaged
Are using accessible technology
Are involved in science process
Must think
Must develop an understanding of the model
See many multivariable systems
• More interactive Excel spreadsheets, see
http://academic.pgcc.edu/~ssinex/excelets/matsci_excelets.htm
14
Questions and more info…
• Scott Sinex
– See the paper in proceedings
http://www.asee.org/documents/sections/middleatlantic/spring-2013/Investigating-and-VisualizingMeasurement-Error-for-Novice-STEM-Learners.pdf
– Contact me - ssinex@pgcc.edu
• My thanks to the ASEE Zone 1 selection committee!
• This work was supported by the NSF Division of
Materials Research Partnership for Research and
Education in Materials (PREM) Grant Numbers
DMR-0611595 and DMR-1205608.
Thanks for coming today!!!
15