Keynote at CUE.NEXT Workshop on providing computing education for other-than-CS-majors' needs. https://cue.northwestern.edu/

1. Inventing computing education to meet
all undergraduates’ needs
Mark Guzdial 1

2. Story
▪ The state of K-12 computing education in the US
▪ Serving the CS Education needs of non-CS students
▪ Story #1: Designing a CS course for non-STEM students
▪ Media Computation
▪ Story #2: Teaching for Computational Literacy, not Software
Development
▪ Story #3: How computing education will need to change to reach
everyone
▪ Conclusion: Win-Win
2

3. STATE OF HIGH SCHOOL CS
EDUCATION IN US
3

4. Few students take a high school CS class
4
% of High Schools with a CS class % of High School Students who
take a CS class
Georgia 47% 1%
Indiana 30% 2.5%
California 30% 3%
Texas 45% 3.76%

5. Few students take a high school CS class
4
% of High Schools with a CS class % of High School Students who
take a CS class
Georgia 47% 1%
Indiana 30% 2.5%
California 30% 3%
Texas 45% 3.76%

6. Few students take a high school CS class
4
% of High Schools with a CS class % of High School Students who
take a CS class
Georgia 47% 1%
Indiana 30% 2.5%
California 30% 3%
Texas 45% 3.76%

7. Few students take a high school CS class
4
% of High Schools with a CS class % of High School Students who
take a CS class
Georgia 47% 1%
Indiana 30% 2.5%
California 30% 3%
Texas 45% 3.76%

8. Few students take a high school CS class
4
% of High Schools with a CS class % of High School Students who
take a CS class
Georgia 47% 1%
Indiana 30% 2.5%
California 30% 3%
Texas 45% 3.76%

9. Data and Visualization from Barbara Ericson
and Willa Hua5

10. 6
Data and Visualization from Barbara Ericson
and Willa Hua

11. Consider the scale
7

12. Consider the scale
▪ AP CS A had 66K exam takers in 2018.
7

13. Consider the scale
▪ AP CS A had 66K exam takers in 2018.
▪ AP CS Principles had 76K.
7

14. Consider the scale
▪ AP CS A had 66K exam takers in 2018.
▪ AP CS Principles had 76K.
7

15. Consider the scale
▪ AP CS A had 66K exam takers in 2018.
▪ AP CS Principles had 76K.
▪ AP English Lit has 580K
7

16. Consider the scale
▪ AP CS A had 66K exam takers in 2018.
▪ AP CS Principles had 76K.
▪ AP English Lit has 580K
▪ AP Calculus has 305K
7

17. Consider the scale
▪ AP CS A had 66K exam takers in 2018.
▪ AP CS Principles had 76K.
▪ AP English Lit has 580K
▪ AP Calculus has 305K
7

18. Consider the scale
▪ AP CS A had 66K exam takers in 2018.
▪ AP CS Principles had 76K.
▪ AP English Lit has 580K
▪ AP Calculus has 305K
▪ There are 15.1 million high school students in US.
7

19. Consider the scale
▪ AP CS A had 66K exam takers in 2018.
▪ AP CS Principles had 76K.
▪ AP English Lit has 580K
▪ AP Calculus has 305K
▪ There are 15.1 million high school students in US.
Likely 90% of US high school students NEVER see any CS
7

20. MEDIA COMPUTATION
Story #1: A Learner-Centered Design of Computing Education
8
(22)
http://mediacomputation.org

21. Requiring CS for All
▪ Fall 1999:
All students at Georgia Tech must take a course in computer
science.
• Considered part of General Education, like mathematics, social science,
humanities…
▪ 1999-2003: Only one course met the requirement.
▪ Overall pass rate: 78%
9

22. Disaggregating by major
10
Success Rates in CS1 from Fall 1999 to
Spring 2002 (Overall: 78%)
Architecture 46.7%
Biology 64.4%
Economics 53.5%
History 46.5%
Management 48.5%
Public Policy 47.9%

23. Contextualized Computing Education
▪ What’s going on?
• Research results: Computing is “tedious, boring,
irrelevant”
▪ Since Spring 2003, Georgia Tech teaches three
introductory CS courses.
• Based on Margolis and Fisher’s “alternative paths”
▪ Each course introduces computing using a context
(examples, homework assignments, lecture discussion)
relevant to majors.
• Make computing relevant by teaching it in terms of
what computers are good for (from the students’
perspective)
11

24. Design Process
12With Andrea Forte

25. Design Process
▪ Focus groups with liberal arts, architecture, and business students enrolled in CS1.
▪ “What would you want to do with computing?
12With Andrea Forte

26. Design Process
▪ Focus groups with liberal arts, architecture, and business students enrolled in CS1.
▪ “What would you want to do with computing?
12With Andrea Forte

27. Design Process
▪ Focus groups with liberal arts, architecture, and business students enrolled in CS1.
▪ “What would you want to do with computing?
▪ Advisory board with faculty from across campus.
▪ “What do you want your students to know about computing?”
▪ Considered building on pockets of computational practice, but.…CS1.
12With Andrea Forte

28. Design Process
▪ Focus groups with liberal arts, architecture, and business students enrolled in CS1.
▪ “What would you want to do with computing?
▪ Advisory board with faculty from across campus.
▪ “What do you want your students to know about computing?”
▪ Considered building on pockets of computational practice, but.…CS1.
▪ Started from the homework projects.
Taught the computing to make those.
Added to reach “CS1”
12With Andrea Forte

29. Open-ended, contextualized homework
in Media Computation CS1 & CS2
13
Sound collage
Involving
the artists:
Bryn Mawr,
Ball State
University

30. Open-ended, contextualized homework
in Media Computation CS1 & CS2
13
Sound collage
Involving
the artists:
Bryn Mawr,
Ball State
University

31. Open-ended, contextualized homework
in Media Computation CS1 & CS2
13
Sound collage
Involving
the artists:
Bryn Mawr,
Ball State
University

32. Open-ended, contextualized homework
in Media Computation CS1 & CS2
13
Sound collage
Involving
the artists:
Bryn Mawr,
Ball State
University

33. Open-ended, contextualized homework
in Media Computation CS1 & CS2
13
Sound collage
Involving
the artists:
Bryn Mawr,
Ball State
University

34. Change in Pass Rates
14
Change in Success rates in CS1 “Media Computation” from
Spring 2003 to Fall 2005
(Overall 85%)
Architecture 46.7% 85.7%
Biology 64.4% 90.4%
Economics 54.5% 92.0%
History 46.5% 67.6%
Management 48.5% 87.8%
Public Policy 47.9% 85.4%

35. Asking women:
What is CS to you?
▪ MediaComp
15
• Existing CS1
Rich, Perry, Guzdial, SIGCSE 2004

36. Voices from Media Computation Students
▪ Intl Affairs student (female): “I just wish I had more time to play around with that
and make neat effects. But JES [IDE for class] will be on my computer forever, so…
that’s the nice thing about this class is that you could go as deep into the homework
as you wanted. So, I’d turn it in and then me and my roommate would do more after
to see what we could do with it.”
▪ “I dreaded CS, but ALL of the topics thus far have been applicable to my future
career (& personal) plans—there isn't anything I don't like about this class!!!"
▪ "Media Computation is a CS class but with less severity. The media part of the class
is extremely visually interesting. I would only take another CS class if it were Media
Computation."
X

37. 38
Survey One Year Later
• 19% of respondents had programmed since class ended
• “Definitely makes me think of what is going on behind the scenes of such
programs like Photoshop and Illustrator.”
• “I understand technological concepts more easily now; I am more willing
and able to experience new things with computers now”
• “I have learned more about the big picture behind computer science and
programming. This has helped me to figure out how to use programs that
I've never used before.”
16

38. 38
Survey One Year Later
• 19% of respondents had programmed since class ended
• “Definitely makes me think of what is going on behind the scenes of such
programs like Photoshop and Illustrator.”
• “I understand technological concepts more easily now; I am more willing
and able to experience new things with computers now”
• “I have learned more about the big picture behind computer science and
programming. This has helped me to figure out how to use programs that
I've never used before.”
16
Used today at UCSD, West Point, RMIT — 50K downloads of most recent IDE.

39. BS in Computational Media
▪ After the Media Computation
class, created a joint degree
with liberal arts.
▪ Same CS classes as CS
major, but 1/2 as many.
▪ Add: Film theory, video game
design, studio design.
▪ Grew to 300 majors, and
increasingly female over time.
17

40. UCSD’s PI+PP+MediaComp Experiment (SIGCSE 2013)
▪ UCSD changed CS1 (quarter system)
in 2008 to:
• Peer Instruction
• Pair Programming
• Media Computation
▪ Tracked students from 2001.
• Increase retention of CS majors into second
year by 30% (from 51% to 81%)
X

41. TEACHING MORE THAN
SOFTWARE DEVELOPMENT
Story #2: Changing Pedagogy
18

42. Example 1: How sound works:
Acoustics, the physics of sound
19

43. X

44. X

45. X

46. X

47. X

48. X

49. Digitizing Sound:
How do we get that into bytes?
▪ We can do the same to estimate the sound curve with samples.
20

50. X

51. X

52. X

53. X

54. X

55. X

56. X

57. X

58. X

59. X

60. Reflection
▪ Prediction
21

61. Reflection
▪ Prediction
▪ 7 Lines
21

62. Reflection
▪ Prediction
▪ 7 Lines
▪ 1 Bit
21

63. Reflection
▪ Prediction
▪ 7 Lines
▪ 1 Bit
▪ Learning without writing a
Program
21

64. Reflection
▪ Prediction
▪ 7 Lines
▪ 1 Bit
▪ Learning without writing a
Program
Teaching CS for insight into
our world,
not software development.
21

65. HOW COMPUTING WILL
CHANGE TO REACH EVERYONE
Story #3
22

66. 23

67. Hughes Printing Telegraph Machine 1860
23

68. Hughes Printing Telegraph Machine 1860
23
From 1840 to 1868

69. For 30 years, this was the common keyboard
24

70. For 30 years, this was the common keyboard
We need to find
what makes the
great ideas of
computing
accessible.
24

71. For 30 years, this was the common keyboard
▪ We may still be waiting for our
QWERTY keyboard.
We need to find
what makes the
great ideas of
computing
accessible.
24

72. For 30 years, this was the common keyboard
▪ We may still be waiting for our
QWERTY keyboard.
▪ How much better would we all be if we
had adopted something even better
than QWERTY?
We need to find
what makes the
great ideas of
computing
accessible.
24

73. Most users don’t think about computing the way CS does
25

74. Most users don’t think about computing the way CS does
25
▪ “To understand a program you
must become both the machine
and the program” (Perlis, 1982)

75. Most users don’t think about computing the way CS does
▪ Katie Cunningham’s Study Participants.
25
▪ “To understand a program you
must become both the machine
and the program” (Perlis, 1982)

76. Most users don’t think about computing the way CS does
▪ Katie Cunningham’s Study Participants.
▪ “Yeah, I mean, it’s just like...it makes
me think like a computer. But I’m not a
computer. And it’s not that I can’t work
with the computer in tandem. I mean,
that’s why we have the computers.”
25
▪ “To understand a program you
must become both the machine
and the program” (Perlis, 1982)

77. Most users don’t think about computing the way CS does
▪ Katie Cunningham’s Study Participants.
▪ “Yeah, I mean, it’s just like...it makes
me think like a computer. But I’m not a
computer. And it’s not that I can’t work
with the computer in tandem. I mean,
that’s why we have the computers.”
25
▪ “To understand a program you
must become both the machine
and the program” (Perlis, 1982)
Her participants are part of “everyone”

78. What do students need to know to use computing effectively?
26

79. What do students need to know to use computing effectively?
▪ Goals:
Casual programming,
convivial programming,
conversational programming.
26

80. What do students need to know to use computing effectively?
▪ Goals:
Casual programming,
convivial programming,
conversational programming.
26

81. What do students need to know to use computing effectively?
▪ Goals:
Casual programming,
convivial programming,
conversational programming.
▪ Contrast with learning English Composition.
▪ We need journalists and novelists.
▪ Greatest social impact of writing is the everyday use.
26

82. 27
Rich, Strickland,
Binkowski, Moran,
and Franklin (ICER
2017) asked the
question:
What’s the starting
place for K-8 CS
learners?

83. Proposed:
What comes first when learning programming?
1. Precision and completeness are important
when writing instructions in advance.
2. Different sets of instructions can produce the
same outcome.
3. Programs are made by assembling
instructions from a limited
set.
4. Some tasks involve repeating actions.
5. Programs use conditions to end loops.
28

84. Scratch fluency doesn’t need that whole list
▪ Over 30 million users.
▪ Most Scratch projects are stories
that use…
▪ Only Forever loops
▪ No booleans
▪ Just movement and sequence.
29

85. Scratch fluency doesn’t need that whole list
▪ Over 30 million users.
▪ Most Scratch projects are stories
that use…
▪ Only Forever loops
▪ No booleans
▪ Just movement and sequence.
There is expressive
power in even a subset
of CS.
29

86. Bootstrap:Algebra doesn’t use all of that list
▪ Improves learning in algebra
▪ Students do not code repetition.
▪ Functional
30Schanzer, Fisler, Krishnamurthi, Felleisen, 2015

87. Bootstrap:Algebra doesn’t use all of that list
▪ Improves learning in algebra
▪ Students do not code repetition.
▪ Functional
There is
learning power
in even a subset of CS.
30Schanzer, Fisler, Krishnamurthi, Felleisen, 2015

88. Task-Specific Programming
Goal: Use programming* to enhance learning in high school and
university non-CS classes
• Using participatory design to result in adoptable programming
environments.
• Building task-specific programming environments to be highly-
usable
31
Social studies: Participatory design with social studies teachers about programming to create
visualizations.
Precalculus: Participatory design with precalculus teachers about programming to learn about
matrices (and wave functions).
* N.B. “programming” not “current programming languages.”

89. Pilot Studies with Social Studies Educators
32
Codap
JavaScript
VegaLite

90. Pilot Studies with Social Studies Educators
32
Codap
JavaScript
VegaLite

91. Teaching Vectors and Matrices by Making Image Filters

92. Teaching Vectors and Matrices by Making Image Filters

93. Teaching Vectors and Matrices by Making Image Filters

94. Teaching Vectors and Matrices by Making Image Filters

95. Teaching Vectors and Matrices by Making Image Filters

96. More Image Filters
37

97. Precalculus matrix manipulations

99. Results: Need to iterate
40

100. Results: Need to iterate
▪ Mathematics teachers: “Meh.” !
40

101. Results: Need to iterate
▪ Mathematics teachers: “Meh.” !
▪ They see disciplinary literacy.
40

102. Results: Need to iterate
▪ Mathematics teachers: “Meh.” !
▪ They see disciplinary literacy.
▪ They see concrete applications.
40

103. Results: Need to iterate
▪ Mathematics teachers: “Meh.” !
▪ They see disciplinary literacy.
▪ They see concrete applications.
▪ They don’t see a solution to their students’ learning problems.
40

104. Results: Need to iterate
▪ Mathematics teachers: “Meh.” !
▪ They see disciplinary literacy.
▪ They see concrete applications.
▪ They don’t see a solution to their students’ learning problems.
40

105. Results: Need to iterate
▪ Mathematics teachers: “Meh.” !
▪ They see disciplinary literacy.
▪ They see concrete applications.
▪ They don’t see a solution to their students’ learning problems.
▪ Key characteristics to emphasize as we change:
40

106. Results: Need to iterate
▪ Mathematics teachers: “Meh.” !
▪ They see disciplinary literacy.
▪ They see concrete applications.
▪ They don’t see a solution to their students’ learning problems.
▪ Key characteristics to emphasize as we change:
▪ Comparison
40

107. Results: Need to iterate
▪ Mathematics teachers: “Meh.” !
▪ They see disciplinary literacy.
▪ They see concrete applications.
▪ They don’t see a solution to their students’ learning problems.
▪ Key characteristics to emphasize as we change:
▪ Comparison
▪ Prediction
40

108. What students might learn in Task-Specific Programming:
What comes first when learning programming
1. Precision and completeness are important
when writing instructions in advance.
2. Different sets of instructions can produce the
same outcome.
3. Programs are made by assembling
instructions from a limited
set.
4. Some tasks involve repeating actions.
5. Programs use conditions to end loops.
41
Rich, Strickland, Binkowski, Moran, and Franklin (ICER 2017)

109. What students might learn in Task-Specific Programming:
What comes first when learning programming
1. Precision and completeness are important
when writing instructions in advance.
2. Different sets of instructions can produce the
same outcome.
3. Programs are made by assembling
instructions from a limited
set.
4. Some tasks involve repeating actions.
5. Programs use conditions to end loops.
42

110. Conclusion: Go for the win-win
43

111. Conclusion: Go for the win-win
▪ For the other-than-CS faculty:
43

112. Conclusion: Go for the win-win
▪ For the other-than-CS faculty:
▪ Don’t settle.
43

113. Conclusion: Go for the win-win
▪ For the other-than-CS faculty:
▪ Don’t settle.
43

114. Conclusion: Go for the win-win
▪ For the other-than-CS faculty:
▪ Don’t settle.
▪ For the CS faculty:
43

115. Conclusion: Go for the win-win
▪ For the other-than-CS faculty:
▪ Don’t settle.
▪ For the CS faculty:
▪ There are benefits in meeting their needs.
43

116. For the other-than-CS faculty
44

117. For the other-than-CS faculty
▪ Insist that computing be taught for your discipline
44

118. For the other-than-CS faculty
▪ Insist that computing be taught for your discipline
▪ Use Vega-Lite, Helena, Mathematica, MATLAB, R, and
new languages still be invented.
44

119. For the other-than-CS faculty
▪ Insist that computing be taught for your discipline
▪ Use Vega-Lite, Helena, Mathematica, MATLAB, R, and
new languages still be invented.
▪ Use the tasks and practices of the discipline
44

120. For the other-than-CS faculty
▪ Insist that computing be taught for your discipline
▪ Use Vega-Lite, Helena, Mathematica, MATLAB, R, and
new languages still be invented.
▪ Use the tasks and practices of the discipline
44

121. For the other-than-CS faculty
▪ Insist that computing be taught for your discipline
▪ Use Vega-Lite, Helena, Mathematica, MATLAB, R, and
new languages still be invented.
▪ Use the tasks and practices of the discipline
▪ Don’t need CS faculty to teach it well
44

122. For the other-than-CS faculty
▪ Insist that computing be taught for your discipline
▪ Use Vega-Lite, Helena, Mathematica, MATLAB, R, and
new languages still be invented.
▪ Use the tasks and practices of the discipline
▪ Don’t need CS faculty to teach it well
▪ Must know computing within the discipline
44

123. For the other-than-CS faculty
▪ Insist that computing be taught for your discipline
▪ Use Vega-Lite, Helena, Mathematica, MATLAB, R, and
new languages still be invented.
▪ Use the tasks and practices of the discipline
▪ Don’t need CS faculty to teach it well
▪ Must know computing within the discipline
▪ or be willing to invent it.
44

124. For the other-than-CS faculty
▪ Insist that computing be taught for your discipline
▪ Use Vega-Lite, Helena, Mathematica, MATLAB, R, and
new languages still be invented.
▪ Use the tasks and practices of the discipline
▪ Don’t need CS faculty to teach it well
▪ Must know computing within the discipline
▪ or be willing to invent it.
▪ Critical to know pedagogical content knowledge (PCK).
44

125. For the other-than-CS faculty
▪ Insist that computing be taught for your discipline
▪ Use Vega-Lite, Helena, Mathematica, MATLAB, R, and
new languages still be invented.
▪ Use the tasks and practices of the discipline
▪ Don’t need CS faculty to teach it well
▪ Must know computing within the discipline
▪ or be willing to invent it.
▪ Critical to know pedagogical content knowledge (PCK).
▪ Most CS faculty don’t: 87% vs. 10%
44

126. Meeting non-CS needs will advance CS needs
45

127. Meeting non-CS needs will advance CS needs
▪ Non-CS majors are much more diverse.
▪ We need to learn how to reach diverse users.
45

128. Meeting non-CS needs will advance CS needs
▪ Non-CS majors are much more diverse.
▪ We need to learn how to reach diverse users.
▪ Non-CS uses expect things to work, in ways that we still have to
figure out.
▪ Example: Stopify (thanks to Shriram Krishnamurthi for the slides)
45

133. Sometimes you want infinite loops…

135. Meeting non-CS needs will advance CS needs
▪ Non-CS majors are much more diverse.
▪ We need to learn how to reach diverse users.
▪ Non-CS uses expect things to work, in ways that we still have to
figure out.
▪ Meeting their needs advances CS software research.
▪ Non-CS programmers want different programming environments.
▪ Pushes us to advance PL and HCI research.
48

136. Summary:
49

137. Summary:
▪ Undergraduates across your campus are unlikely to have had high
school CS.
49

138. Summary:
▪ Undergraduates across your campus are unlikely to have had high
school CS.
▪ Story #1: Involve other students and faculty voices in design.
49

139. Summary:
▪ Undergraduates across your campus are unlikely to have had high
school CS.
▪ Story #1: Involve other students and faculty voices in design.
▪ Story #2: Teaching computing beyond software development involves
new pedagogy.
49

140. Summary:
▪ Undergraduates across your campus are unlikely to have had high
school CS.
▪ Story #1: Involve other students and faculty voices in design.
▪ Story #2: Teaching computing beyond software development involves
new pedagogy.
▪ Story #3: Be open to new tools and practices that meet different
needs.
49

141. Summary:
▪ Undergraduates across your campus are unlikely to have had high
school CS.
▪ Story #1: Involve other students and faculty voices in design.
▪ Story #2: Teaching computing beyond software development involves
new pedagogy.
▪ Story #3: Be open to new tools and practices that meet different
needs.
▪ We need to involve more voices and more diverse voices in design
of computing education.
49

142. Some of the Collaborators on This Work
▪ Barbara Ericson, Miranda Parker, Kathryn Cunningham, Amber Solomon, Bahare
Naimipour, Richard Catrambone, Lauren Margulieux, Betsy DiSalvo, Tom McKlin, Rick
Adrion, Renee Fall, Sarah Dunton, Brad Miller, Ria Galanos, Brian Dorn, and Briana
Morrison
▪ Media Computation (CCLI), Miranda Parker (GRFP), and Katie Cunningham (GRFP) were
funded by the US National Science Foundation
▪ http://computinged.wordpress.com
▪ http://guzdial.engin.umich.edu
50

143. Some of the Collaborators on This Work
▪ Barbara Ericson, Miranda Parker, Kathryn Cunningham, Amber Solomon, Bahare
Naimipour, Richard Catrambone, Lauren Margulieux, Betsy DiSalvo, Tom McKlin, Rick
Adrion, Renee Fall, Sarah Dunton, Brad Miller, Ria Galanos, Brian Dorn, and Briana
Morrison
▪ Media Computation (CCLI), Miranda Parker (GRFP), and Katie Cunningham (GRFP) were
funded by the US National Science Foundation
▪ http://computinged.wordpress.com
▪ http://guzdial.engin.umich.edu
50
Thank you!

144. Miranda Parker’s Model of Georgia High Schools
Regression Analysis: What predicts GA schools offering CS?
There is a statistically significant, strong correlation between CS enrollment across
years
Median income plays a small role, explaining 5.2% of the variance in CS
enrollment
If a school had CS in 2015, as well as median income and school enrollment, can
explain if a school has CS in 2016 (R2=55.8%)
51
Motivation Quantitative Factors Case Studies of Schools Implications & ContributionsDefining Computer Science

145. What explains the rest of the variance?
Case studies with four Georgia high schools
Teachers can make a difference, but it may be hard to adjust hiring priorities, find
qualified personnel, train a teacher from a different discipline, transition after
retirement
Community values matter, such as with schools around Fort Gordon offering
cybersecurity
There may be a disconnect in terms of what CS is, what courses are available,
and what careers are associated with CS
52
Motivation Quantitative Factors Case Studies of Schools Implications & ContributionsDefining Computer Science

146. Implications CS in US High Schools on CUE.NEXT
▪ Few students will have had high school CS.
▪ Computer science is a small discipline in US high schools.
▪ The students who see CS will be disproportionally male and
higher SES.
▪ Some of the factors that influence CS in schools are within our
power to influence:
▪ Explaining what CS is, and helping schools get started.
53