Thursday, July 23, 2020

In Favor of Adopting Open Textbooks

Legislators, school administrators, and other policy makers have encouraged and/or mandated the use of open source textbooks to ease the financial burden on students. An obvious counter-argument is that you get what you pay for.

There is no doubt that major publishing companies provide more faculty-pleasing bells and whistles than are found with open source textbooks. On the other hand, many students do not obtain their textbooks until after classes begin, often going days or weeks without required materials.  The benefits of the commercial textbook diminish or disappear entirely for students who do not have their textbooks, and this problem is exacerbated when students take more classes on-line and visit the campus bookstore less often.

Meanwhile, the faculty involved with the PreTeXt project (https://pretextbook.org/) are working to create textbooks superior to commercial textbooks.

The first obvious advantage is that the instructor can provide a link to the html version of the textbook before class begins. Next, unlike a pdf with fixed line breaks that may force the reader to scroll left and right, the html version is designed to be readable not only on computer screens but also on the small displays of mobile devices.

The PreTeXt project is a partner with the American Institute of Mathematics (https://aimath.org/) in the NSF-funded UTMOST Project (https://utmost.aimath.org/), seeking “to understand how students and faculty actually use textbooks in undergraduate mathematics courses and to use that understanding to produce textbooks that are more effective in promoting student learning.”

Two relatively mature textbooks authored in PreTeXt are Active Calculus (Matt Boelkins) and ORCCA (Open Resources for Community College Algebra, Ann Cary, Alex Jordan et al.). Both have associated videos corresponding to each lesson, and both have homework sets coded in the open source WeBWorK online homework system. (For schools that cannot host the free WeBWorK system themselves, there are options such as xyzhomework and Edfinity that offer low-cost alternatives to the publisher sponsored homework systems.)

Some of the textbooks written in PreTeXt provide embedded interactive exercises that the student answers in the book and gets immediate feedback. And via a collaboration with Runestone Academy (https://runestone.academy/), it is even possible to set up a course so that Runestone Academy will record the responses students make in the textbook without requiring the student to log into a homework system.

One talking point about open source textbooks is that faculty have the option of taking the parts they like, changing parts they dislike, and/or adding components they feel are missing, then publish their own versions. That does not happen very frequently in practice, but interested users can work with the textbook authors to create ancillaries to share with the community.

Thursday, March 29, 2018

Switching to Linux

I’ve recently switched from Windows to Linux.

The main impetus was that I’m trying to help Katherine Yoshiwara (my wife) code her math textbooks into PreTeXt (http://mathbook.pugetsound.edu/), and the key players creating PreTeXt all say things like “I’m not familiar with how to do this in Windows, but it’s simple in Linux.”

I set up my current (Christmas gift) pc to dual boot Ubuntu Mint and Windows10. I first practiced on my previous (and dying) pc, and I was unwilling to abandon Windows entirely.

A turning point was last week when I got onedrive-d running, so now I can have my Microsoft OneDrive directory files synced and available when in Linux.

I’d already discovered that I can create, read, and edit MS Office files using LibreOffice (provided with Linux), Google Docs, and with Chrome plug-ins, so I have no worries about collaborating with others that use MS Office formats.

Then two days ago my laptop refused to reboot in Linux.The machine hung up at the same spot with a message I could not understand (about performing fsck manually) even when I tried selecting earlier versions of Linux at startup. I could still boot in Windows, but I now think that was one mistake that led to a bigger one: I okayed an update to Dropbox, and when the update was complete, Windows had killed “grub” and Linux was nowhere to be found on my computer.

But using my still-working old laptop, I managed to find online how to restore grub, then how to perform the manual fsck (I managed to follow steps, but I still have no idea what I was actually doing). And then I still had to repair a github directory (the corruption of which had led me to try a reboot in the first place).

But I think I’m back in business. And I'm even more convinced that Linux is the way to go.

Tuesday, March 20, 2018

PreTeXt: A Promising Tool for Open Source Textbooks

Although I have been (and remain) largely unimpressed with the quality of the great majority existing open source math textbooks, I now believe there is hope on the horizon.

The model used by open software developers, namely of modifying existing materials and improving them, can apply to math textbooks. Rob Beezer’s PreTeXt tool gives authors a means of producing digital textbooks that are superior to existing popular commercial textbooks.

With PreTeXt you can convert a (properly marked up) plain text file into html format that is optimized for reading on mobile devices. It uses MathJax for the mathematical expressions, and the developers are designing to take advantage of open software like the Sage cell for interactivity and WeBWorK or MyOpenMath for homework checkers.

Another cool feature is that PreTeXt can take the same source text file and output LaTeX and hence pdf, for students who may want a printed version of the textbook.

The readability on mobile devices, the adaptability of the source files, the tools for interactivity, and the attention the developers pay to accessibility issues are huge benefits to adopting PreTeXt books (which, by the way, are not limited to math—PreTeXt has also been used in computer science, poetry, and music textbooks).

Saturday, December 3, 2016

Learning Styles and Optimizing Learning

"Learning Styles: Concepts and evidence," Psychological Science in the Public Interest, V. 9 No 3, December 2008, by Harold Pashler, Mark McDaniel, Doug Rohrer, and Robert Bjork examined whether or not there was scientific evidence to support the "learning-styles" practice of matching the format of instruction to the learning style of the individual learner.

They authors described what evidence would be appropriate to validate the learning-styles practice. For distinct learning styles A and B, learners with learning style A should learn better with an instruction/intervention with learning style format A than with instruction/intervention with learning style format B, and vice versa: learners with learning style B should learn better with instruction/intervention of format B than format A.

The authors were actually more generous in what they considered acceptable evidence. But simply finding that learners A did better with intervention A than with intervention B was not sufficient--it was also necessary to show that learners B did worse with intervention A than with B.

And at that time (December 2008) the authors found no evidence base to justify incorporating learning-style practices. And in 2016, there still has been no adequate evidence to justify the use of learning-styles practices.

Among professionals who research learning, there is consensus that it is not effective to teach towards the learning style of the individual student. Yet across the K-16 spectrum there are ardent adherents to the learning-styles practice.

From the opening of Philip M. Newton's "The Learning Styles Myth is Thriving in Higher Education," Frontiers in Psychology, 15 December 2015, "The existence of ‘Learning Styles’ is a common ‘neuromyth’, and their use in all forms of education has been thoroughly and repeatedly discredited in the research literature. However, anecdotal evidence suggests that their use remains widespread."

Both Newton and Pashler et al. acknowledge that people can readily identify with a preferred learning style and that there is validity to the identification. But there still is no evidence that learning is enhanced by identifying a student's learning style and providing instruction geared towards that style.

The promotion of the learning-styles practices is not an innocuous indulgence. Not only does the emphasis on learning styles take time and resources away from proven effective interventions, the emphasis on learning styles can potentially reinforce a fixed mindset and steer students away from certain leaning challenges and academic paths.

Tuesday, June 7, 2016

"Unjustified use of Algebra 2"

The U.S. Department of Education organized a meeting (“California Math Convening: Gateways to Access – May 31, 2016”) to discuss California's use of Algebra 2 (a.k.a. Intermediate Algebra) in higher education. The meeting was held at the chancellor's office of the California State University (CSU) system. The participants included representatives from the CSU, the University of California, the California Community Colleges, K-12 educators, and educational policy organizations.

The meeting was the DOE's response to a September 30, 2015 letter from Christopher Edley, Jr., to the Catherine Lhamon, Assistant Secretary of Civil Rights, U.S. Department of Education. The letter begins with:
“I write to request that your office investigate the educationally unjustified use of Algebra 2 as a gateway course by all three segments of California’s higher education system: the University of California system; the California State University system; and the California Community College system. There is evidence to suggest that, in varying ways, these institutions have adopted policies and practices that impose a disparate impact on protected groups in violation not only of the equal protection clause of the California State Constitution, but also in violation of federal regulations implementing Title VI of the Civil Rights Act of 1964.”
The letter cites the success of Statway, a project of the Carnegie Foundation for the Advancement of Teaching, as evidence that Intermediate Algebra is not actually necessary for success in completing math requirements for baccalaureate degrees in some majors. The letter concludes with:
“If there are villains here, they are the indifference and inertia that confirm and perpetuate unequal educational opportunity. I believe this discrimination is, for the most part, without animus. Regardless, the injury is real.” 
At the meeting, Christopher Edley Jr. explained that neither intent nor a history of practice would be considered relevant when determining if there is a violation of the Civil Rights Act. The presence of both Catherine Lhamon and also the Under Secretary U.S. DOE, Ted Mitchell, made abundantly evident that the DOE wants California's higher education community to recognize and address the issue.

Another speaker was William McCallum, mathematician with numerous distinctions including being one of the three lead writers of the Common Core State Standards in Mathematics (CCSSM). Bill explained that because College Algebra was the de facto mathematics requirement in U.S. baccalaureate granting institutions at the time of writing the CCSSM, the document needed to include the math that would lead to College Algebra, namely Algebra 2. He commented that it is  inappropriate for colleges or universities to cite the CCSSM to define what is currently needed to be college ready--it makes no sense to argue against modifying college math requirements based on the content of the CCSSM, as the CCSSM were created trying to reflect what the earlier college math requirements had been.

The U.S. DOE evidently intends to hold another such meeting in 3 or 4 months to check on what progress has been made.

Saturday, February 7, 2015

Strategies to help developmental math students

Nationally about 70% of incoming community college students are placed into developmental (a.k.a. “remedial” or “foundational”) math classes that earn no college degree credit. But only 10% of these students successfully move past developmental math to earn their degrees.


Four broad areas are being addressed to increase student success through developmental mathematics (1) Placement, (2) Pedagogy, (3) Curriculum, and (4) Student attitudes.


Improving Placement
Failing a class is not the only barrier to completion--the length of the developmental math path defeats many students. More developmental math students drop out of college without ever failing a math class than flunk out of math. One strategy to reduce the number of “exit points” is to help students place into as high a math level as reasonable.


For example, CaƱada College uses its Math Jam both as an intensive preparation for the math placement exam and also as a recruitment tool to get more students into STEM fields.


The placement instrument itself, typically a machine-graded standardized test, can be augmented or replaced.  High school GPA, recency of the previous math course, weekly work hours, and total course load could be part of “multiple measures.” Some schools have abandoned placement into developmental math courses, typically offering supplementary resources for students in credit-bearing classes.


Modifying Pedagogy
James Stigler lists three key types of learning opportunities that students need to experience to become flexible learners: productive struggle, explicit connections, and deliberate practice.

Modularized courses can allow students to spend time only on topics they need to study. The Emporium Model relies on software to do the pretest, primary instruction, and mastery testing, with human interaction largely limited to one-on-one tutoring in the computer lab (where students work lessons and take assessments). The University of Illinois uses software for placement and remediation, and California State University Northridge uses software as part of its hybrid lab remediation for students considered “at risk.”


Technology also plays a key role in both MOOCs (Massive Open Online Courses) and the “flipped” classroom. However, the “MO” aspects of MOOCs appear not to improve student success compared with the online developmental math courses that have existed for decades.


Another way to address the attrition between courses in a sequence is to offer “compressed” courses. The students take two courses during one term, but each course meets the standard number of hours per term--students are essentially immersed in math, which comprises most or all of their studies for that term.


Adjusting the Curriculum
The American Mathematical Association of Two-Year Colleges (AMATYC) has a 2014 position paper that states, “Prerequisite courses other than intermediate algebra can adequately prepare students for courses of study that do not lead to calculus.”


There are numerous “pathways” that have been created to allow developmental math students to pass a transferable math course--typically statistics or a quantitative reasoning course--that do not require many topics typically associated with intermediate algebra. The pathways normally reduce the number of developmental math courses required before earning transferable math units.
  • Path2Stats is part of the California Acceleration Project, based on a program developed by Myra Snell at Los Medanos College.
  • Statway and Quantway are projects of the Carnegie Foundation for the Advancement of Teaching.
  • The Dana Center’s Math Pathways include pathways for both STEM and non-STEM students.
  • Mathematical Literacy for College Students (MLCS) and Algebraic Literacy grew out of an AMATYC project. They can serve as alternatives to beginning and intermediate algebra classes for STEM majors, or the MLCS  can serve as prerequisite for a transferable non-STEM math course.


Addressing student attitudes
The “affective domain” includes attitudes, values, beliefs, interests, and motivation.


Carol Dweck’s research indicated that students (from grade school through graduate school) with “growth mindsets” persist and succeed better than peers with “fixed mindsets”. And importantly, students can learn to move from a fixed mindset to a growth mindset.

David Yeager’s research suggests that the performance gap in math--specifically developmental math--suffered by women and other underrepresented groups can be eliminated by specific brief interventions.

City University of New York's Accelerated Study in Associate Programs (ASAP) is an initiative that does not attempt to modify what occurs in the classroom. ASAP stipulates full-time enrollment and provides participants with academic advisement, career services, tutoring, financial supports, specially blocked or linked courses.

Sunday, December 28, 2014

Perfect numbers

"'I'll show you one more thing about perfect numbers," he said..."You can express them as the sum of consecutive natural numbers." Yoko Ogawa, The Housekeeper and the Professor, translated by Stephen Snyder

The examples that immediately follow the statement in the novel show that the perfect numbers 6, 28, and 496 are triangular numbers, that is, can be expressed as a sum of the form \(1+2+3+\cdots+n\).

This got me thinking about perfect numbers. For example,

Theorem: The reciprocals of the divisors of any perfect number sum to 2.
Proof: If the divisors of the perfect number \(N\) are \(1, d_2, d_3, \ldots, d_k\), and \(N\), then the sum of the reciprocals would be
\(1/1 + 1/d_2+ 1/d_3 \cdots + 1/d_k+ 1/N\). If we call this finite sum \(s\), then
\[Ns=N/1 + N/d_2+ N/d_3 \cdots + N/d_k+ N/N\]
\[ = N + d_k +\cdots +d_3+ d_2 + 1\]
\[= N +(\text{proper divisors of N})  = N+N = 2N\] Thus \( s=2\)

Theorem: Every even perfect number has the form \(N=(2^k -1)\cdot 2^{k-1}\), where \(2^k - 1\) is a Mersenne prime.
Proof:  Let \(N\) be an even perfect number. Then \(N\) can be written in the form \(N=2^{k-1} m\), where \(k>1 \)and \(m\) is odd.
Define \(\sigma(n)\) to be the sum of all positive divisors of \(n\).  In particular, \(\sigma(n)=2n\)  whenever \(n\) is perfect.
When \(a\) and \(b\) are relatively prime, \(\sigma(ab)=\sigma(a)\sigma(b)\)  because every divisor of \(ab\) can be uniquely written as the product of a divisor of \(a\) times a divisor of \(b\), so summing the divisors of \(ab\) can be accomplished by first computing \(\sigma(a)\)  and multiplying it by each of the divisors of \(b\), and summing those products.
Now \(\sigma(N)=2N=2^k m\) because \(N\) is perfect. By adding a finite geometric series with common ratio 2, we see that \(\sigma(2^{k-1})=2^k-1\), and we have
\(2^k m=  \sigma(2^{k-1}m) = \sigma(2^{k-1})\sigma(m)=(2^k-1)\sigma(m)\)  

Solving for \(\sigma(m)\), we get \(\sigma(m)=m+ \frac{m}{2^k-1} \)

From the definition of \(\sigma(m)\), \(m+ \frac{m}{2^k-1} \)  must represent the sum of all divisors of  \(m\), and in particular the fraction must be an integer.  But then \(\frac{m}{2^k-1}\) must itself divide \(m\), and as \(m\) clearly divides itself, \(m\) and \(\frac{m}{2^k-1} \)  must be all the divisors of \(m\).  Thus \(m=2^k-1\) must be prime.

Conversely, Theorem: Each Mersenne prime \(2^k -1\) gives the perfect number \(N=(2^k -1)\cdot 2^{k-1}\).
Proof: If \(2^k - 1\) is prime, then the divisors of \(N=(2^k -1)\cdot 2^{k-1}\) are \(1, 2, 2^2, \ldots, 2^{k-1},\) and also the product of any of those powers with the prime \(2^k - 1\). Summing all the proper divisors is the sum of two geometric series, each with common ratio 2. We get
\[\left(1+2+…+2^{k-1}\right) + \left(2^{k}-1\right) \left(1+2+…+2^{k-2}\right)\]
\[ = \left(2^k -1\right) + \left(2^{k}-1\right) \left(2^{k-1}-1\right)\]
\[= \left(2^k -1\right) \left(1+ 2^{k-1}-1\right) = N\]
We can see that every even perfect number is a triangular number, because \(N=(2^k -1)\cdot2^{k-1}\) has the form \(\frac{(n-1)n}{2}\), where \( n = 2^k\).

Euler evidently knew everything about even perfect numbers, but as far as I know, neither he nor anyone else has proven whether or not any odd perfect number exists.

I don't know whether an odd perfect number would need to be a triangular number. But the Professor only asserted that perfect numbers can be expressed as a sum of consecutive natural numbers. And of course any odd number \(O = 2n+1\) --including any odd perfect number-- can be expressed as the sum of two consecutive natural numbers: \(n + (n+1)\).