Engineered

Statement of support and solidarity for the AAPI community

Whereas diversity and inclusion are core values of Southern Methodist University, benefiting every member of our community by providing a wide range of perspectives based on our varied identities and life experiences,

Whereas incidents of verbal and physical violence targeting members of the Asian-American and Pacific Islander (AAPI) communities have surged nationally since the beginning of the COVID-19 pandemic, including the tragic shootings in Atlanta on March 16, 2021 that took the lives of eight innocent people, six of whom Asian women, and have been preceded by years of rising anti-Chinese sentiment in the United States,

Whereas a hate crime is defined by Oxford’s English Dictionary as a “crime, typically one involving violence, that is motivated by prejudice on the basis of race, religion, sexual orientation, or other grounds”,

Let it be resolved that the SMU Faculty Senate stands with the Asian-American Pacific Islander community in vehemently condemning these hate crimes and any hate crimes,

Let it be further resolved that the SMU Faculty Senate resolutely opposes bigotry, discrimination and xenophobia in all their forms, and expresses its unflinching support, staunch commitment and deep appreciation to AAPI students, faculty, staff, alumni and friends in these trying times,

And let it be further resolved that the SMU Faculty Senate calls on all SMU faculty to stand in solidarity with the AAPI community, offer support and compassion to anyone impacted directly or indirectly by these hate crimes, and report any hate crime to the appropriate authorities.

This statement was unanimously approved by the SMU Faculty Senate Executive Committee earlier this afternoon.

Demographics and the Demand for Higher Education

Today's post is about Demographics and the Demand for Higher Education by Nathan D. Grawe, which is 150 pages about a model the author created to estimate the probability of attending college by race, region and ethnicity. The book is very dry and it reminded me of a HBR cover a few years ago, which exhorted readers not to let numbers undermine their strategy. I think by this point, everyone who works in higher education knows the demographic trends are very worrisome, and I appreciated that the author quantified it and attempted to provide a fine-grained description of how the national trends will play out across the country and various groups. But I think overall all the discussion of the numbers (an increase of 3% here! a decrease of 10% there!) masks the big picture.

Here are a couple of thoughts that crossed my mind while I was reading. First, there is all this talk about how undergraduate student enrollment is projected to drop by the end of the decade, and the author does have one line buried in the book on how the end of the 2020s is definitely not going to be a good time to look for a tenure-track position, but there is no talk about what it is going to mean for the number of graduate students admitted in PhD programs, especially programs that tend to support their students through teaching assistantships and that predominantly graduate students to go into the professorate (in other words, the humanities). The size of a doctoral program matters to its ranking, for instance in the Carnegie classification, and so I don't expect graduate programs to decrease their size willingly, but if an administration provides fewer teaching assistantships to reflect decreases in enrollment, that's probably something students should know before they accept. So programs that guarantee funding for a number of years will be at an advantage because of the increased precarity in the others.

But it seems that there will be, at many institutions, an increase in contingent faculty ("not tenured track") and I am not sure if the students who are now about to pursue doctoral studies understand it will apply to them. The doctoral programs with a track record of placing their students in academia will be at an advantage, but I think most of the faculty at top institutions come from a narrow set of schools. So the upcoming demographic changes at the undergraduate level will affect graduate education in some profound ways, at least in some disciplines. Also, if we think about this a little bit longer, universities that face the most financial pressure may well discontinue certain programs (that is one way to get rid of tenured professors too) and I wonder how the rise of Zoom will play out in that respect. For instance you could imagine that local (ailing) universities will ally to each focus on their strengths and allow their students to enroll in the partner universities' courses via Zoom to take courses that are no longer offered at their institution. 

The other comment I wanted to make on this is that the rise of the importance of the high-tuition/high-aid model and in particular the importance of the full-pay student is something I find very worrisome. In academia some of us like to say the student is the product of higher education, and the consumer is society. In teaching we try to do what is best for society by giving the product the right features (right courses etc). But if students start viewing themselves as the consumer of higher education, it changes the dynamics and the time horizon, which gets shifted to the short term. If a student is full-pay, he probably has other avenues to get outstanding work/educational experiences, such as internships through his parents' companies or social networks, since said parents are clearly very successful. Then he may not attend college with the same goals as a student receiving need-based or merit-based aid. He may also get the impression his tuition buys him a passing grade by default.

The interesting part about full-pay students is that universities want to attract them without giving them money (which would go against the definition of having full-pay students) and so they need to attract them based on amenities and experiences, which are expensive to build and maintain, as well as social/alumni networks. I can't end this discussion about full-pay students without being reminded of a comment a colleague of mine made at Lehigh, and I paraphrase: we tend to think that American full-pay students come from families with a lot of money, but certain full-pay foreign students have even more money. Back at Lehigh I used to have quite a few very good Turkish undergraduate students who had been sent to the States because they hadn't done great at the entrance exam to the Turkish universities and their parents were aware of the Lehigh ISE name (often because a parent had gotten a degree from Lehigh). And of course there was the appeal of getting their children away from the country.  

Finally, the author comments on the different outlook for two-year and four-year colleges. I thought when I read this that this called for a different strategic model for two-year colleges. For instance, just like regional and national universities hope to attract full-pay students, there could be a greater emphasis for two-year colleges to attract retirees eager for continuing education that would stimulate their mind, like art history or foreign languages. Overall the book completely ignores not-for-credit education (continuing education especially online). It may represent only a small fraction of revenues today, but it could help alleviate the financial pressures down the road. If two-year colleges focus on retirees and provide an enjoyable experience for them toward the end of their life, they may increase the number of them who will include two-year colleges in their will.  

One last thought (and then I will really be done, I promise) was about the fact that elite universities are likely to respond to rising demand by increasing selectivity. The author writes (p.120) "[elite schools] may opt to pass on some highly attractive Asian American students for the sake of a more representative student body" and (p.121) "Asian Americans passed up by elite schools would almost surely be snatched up by national schools facing decreasing total enrollments." (It may also mean that national schools would see more students transfer out as the top students, aware that they deserved to have been admitted to a better school, may apply to transfer more aggressively after a year or two.) So attracting top Asian Americans may become a strategy for national schools eager to move up in the rankings. 

On rising through the rankings

The rankings in the title are of course The Rankings, aka the love-to-hate but always-so-important U.S. News rankings. I finally came across data about universities' U.S. news rankings over the years; specifically, a website called publicuniversityhonors.com has the U.S. News rankings for 57 leading universities from 1983 to 2007. It shows how the rankings for universities that were in the top 50 at least at some point during that period fared over the years. (Some rankings are not provided for the years where U.S. news only grouped universities by tiers.)  

The reason I am interested in rankings over time is that I am trying to better understand which university have been able to make a noticeable, long-lasting rise through the rankings while others have just bobbled up and down along with the tweaks U.S. News makes to its rankings every year. It is great to have as aspirational peer M.I.T. or Harvard or Penn or Notre Dame or Vanderbilt, but if those universities have been at the top for a very long time, it may be useful to also have as aspirational peers universities that have been able to rise through the rankings - in other words, aspirational peers not only in terms of the ranking they currently occupy or the quality of the education they deliver or the strategic positioning they have in higher ed, but in terms of trajectory path. 

Here is what I found for the period for 1983 to 2007:

• U.C. Berkeley dropped from #5 in 1983 to #24 in 1989 back to #13 in 1990 and stayed in the #20s between 1995 and 2007.
• University of Chicago has gone from #6 to #14 back to #9 in the time period considered.
• University of Michigan Ann Arbor was initially ranked #7 and then stayed in the #20s between 1997 and 2007.
• University of Illinois at Urbana-Champaign was at #8 and then after years of unavailable data reemerged at #45, and stayed in the #40s.
• (there are a lot of public universities that dropped in the rankings during that time frame, probably due to funding issues, so I won't go over all of them)
• Washington University in St Louis seems to have been able to slightly move up in the rankings during that time period.
• Interestingly enough, back in those days Lehigh was apparently in the high #30s, perhaps due to its strong engineering reputation, while it has been solidly in the high #40s for a few years now.
• University of Southern California, like WashU, seems to have been able to slightly move up in the rankings during that time period.
• UNC Chapel Hill slipped in the rankings during that time period. 
• U.Virginia also slipped slightly.
• Penn rose. 

Of course that data only tells part of the story, for instance UCSB used not to be in the top 50 but then managed to stay there very steadily during the time period considered, which is an accomplishment too, and now they are ranked #30. It would definitely be best if the data was available until 2021 rather than 2007.

Anyway, I have to wonder whether some private universities benefited from a widespread decrease in funding (?) of the public universities, pushing high-quality professors and students toward the private universities that were able to take advantage of the moment. Our previous provost often referred to USC and I remember him gifting faculty (maybe at the Emerging Leaders seminar?) The Contrarian's Guide to Leadership. USC continues to be ranked #24 in US News (acceptance rate: 11%, graduation rate: 92%, average cost after aid: $30k), while WashU is #16 (acceptance rate: 14%, graduation rate: 94%, average cost after aid: $27k) and Penn is #8 (acceptance rate: 8%, graduation rate: 95%, average cost after aid: $25k). (In case my readers are wondering about Lehigh's stats: acceptance rate: 32%, graduation rate: 89%, average cost after aid: $31k, and SMU's: acceptance rate: 47%, graduation rate: 79%, average cost after aid: $37k. All numbers courtesy of Google.com.) USC, WashU and Penn are universities I'd love to know more about, even if they are not necessarily aspirational peers, but to better understand how they pulled off their rise through the rankings and managed not to slip back down. 

Just like something disrupted negatively public universities in the US News rankings from 1983 to 2007, the upcoming demographic change in the high school student population promises to disrupt the rankings to the advantage of universities that can embrace this shift. The advantage, I think, will be for universities that have exceptional programs in place to welcome first-generation students and/or students of underrepresented minorities. 

Online Learning

For today’s post, I want to talk about online learning. Zoom has obviously made the continuity of instruction possible in higher education in spite of the pandemic, and for that I am very grateful. To me, synchronous education is better than asynchronous. But I also want to point out that (through no fault of Zoom’s), turning online synchronous lectures into Internet versions of on-campus ones is an implicit choice that we don’t necessarily need to make when considering online education by itself, rather than as a companion to in-person instruction.

By that I mean that a semester in a brick-and-mortar campus has always had a certain number of weeks, let’s say fourteen weeks, followed by finals. And the lectures have always had a certain duration that worked well for in-person instruction, such as fifty or eighty minutes. But if we consider 100% online instruction, it is not clear to me completion of the course and retention of the material are best achieved when the online course is also fourteen weeks long, followed by finals.

For instance, many online students are full-time professionals. They may be able to set aside four weeks or maybe even eight weeks for a course, but three months may give them pause, because their boss may wonder if they can really give the team their best. I think for online models it makes more sense to have a course split into modules and having one week or two of break in between the modules. Also, the modules could be offered repeatedly throughout the year so if a student wants to take a break from learning to focus on their full-time job, they could be off for a few weeks and then resume easily. You may also have to teach only once a week to make the schedule work for students who also have a full-time job and can’t make the commitment multiple times a week.

Further, I’d love to see more one-morning or one-afternoon online courses: something of very narrow duration and therefore very small cost that lowers the bar for interested would-be students who are not sure of what it would mean to take an eight-week course with a certain professor or at a certain school. Offering day-long (or half-day-long) courses would lower the barrier to entry into an online graduate program by giving students a better sense of what to expect.

In summary, for online education to deliver its potential we should design online courses that are best suited for the medium, and the best of them may be radically different in duration (of a lecture), length (of the module), course organization and testing, as well as size of enrollment, than their in-person counterparts.

I mention size of enrollment because part of the fun of an in-person course is to run into one’s classmates in the classroom before the start of the class and catch up. In the Zoom it is a bit harder to do since students may be reluctant to talk all over each other in trying to greet a specific classmate, it is awkward to have everyone else listen it to the conversation and it may feel invasive to send a stranger private chat messages to continue the discussion.

So from experience I would say ten students with the instructor is ideal to maximize interaction without having to overly duplicate instruction across sections. This means that some courses that had twenty or twenty-five students enrolled in person would be better served by opening several smaller online sections, if the course instructors seek that their expertise transfer well to online education. Online education is certainly more complicated than just putting an off-line course on Zoom. But whatever form online education takes next, Zoom is sure to play a critical role in delivering it and remain part of the future of online learning.

Convergence

Today’s post is about a concept I learned in a National Academies report called convergence. The term sounds a bit odd for a mathematician like me, but it won’t when you think about rivers coming together. It basically adds a layer to the now familiar concept of interdisciplinary research. Where interdisciplinary research is research at the intersection of multiple fields, convergent research is research that combines separate streams and creates something new. The National Academies report, based on a workshop held in September 2013, is titled “Convergence: Facilitating Transdisciplinary Integration of Life Sciences, Physical Sciences, Engineering and Beyond.” The report brief states that the report “focuses on the need for convergence – an approach to problem solving that integrates the knowledge, tools, and ways of thinking from multiple scientific disciplines, including economic, social and behavioral sciences.” The authors give the example of genomics, which resulted from the combination of genetics, chemistry, optics and bioinformatics. They point out the convergent projects funded by the Department of Energy’s Advanced Research Projects Agency – Energy (ARPA-E), for instance one on electrofuels. Another example provided in the full report is about how 3-D printing will bring new healthcare options.

They recommend to:

1. organize around a common theme, problem or scientific challenge,
2. implement management structures tailored to the challenges to convergence in each institution,
3. foster opportunities to interact formally and informally,
4. change existing faculty structures and reward systems,
5. work with and across existing departments,
6. embed support for convergence in the promotion and tenure process,
7. design facilities and workspaces for convergent research,
8. design education and training programs that foster convergence,
9. establish partnership arrangements across institutions,
10. explore sources of funding within and beyond government agencies.

They also make specific recommendations for different actors: a national vision-setting body, a funder of science and technology innovation, an academic leader, a government laboratory, an industry, medical or regulatory stakeholder.

One thing they might have left out, or they overlooked, is the need to create new academic journals or new departments in existing academic journals specifically for convergent research instead of trying to make convergent research fit into existing departments of existing publications.

The full report provides two examples of radical reorganizations of faculty structures: Arizona State University, which got rid of departments and instead created 23 schools centered around themes and the Janelia Research Campus (funded by the Howard Hughes Medical Institute), which was created without any departmental affiliations or tenure. Half its staff has had initial training in biology and half in physics, computer science and engineering. For universities less inclined toward radical reorganization, interdepartmental institutes of centers. As of the report’s writing, “the top 25 research universities in the US average[d] more than 100 research centers, many of which are organized in discipline-crossing ways.” Other examples of centers and institutes are Bio-X at Stanford, Wyss Institute, the Lewis-Sigler Institute for Integrative Genomics, the Christopher S. Bond Life Sciences Center at the University of Missouri. Many universities have undertaken cluster hiring.

The report also states: “A reward structure that emphasizes individual investigator-driven research and publication and questions of how credit is assigned for multi-investigator-led projects represent widely acknowledged challenges to any form of interdisciplinary or collaborative research, including convergence.”

In 2019, a workshop was held on “Fostering the Culture of Convergence in Research,” which expanded on many of the ideas above. I particularly enjoyed the section “Evaluation criteria and metrics for success”, especially enabling researcher recognition and advancement: funding proposals, publications, promotion and tenure. I also enjoyed the ideas in the “Funding convergence” section, including joint proposal solicitations and possibly supporting people rather than projects.

I think there may be a need for more detailed success stories on convergent research: what are best practices in identifying collaborators, in working with them, in generating the initial idea, in refining the research, in keeping track of everyone’s contributions and appropriate reward structures.

Finally, when I read “Finally, universities are making investments in new buildings specially designed to house convergent institutes and foster interactions across disciplines as well as new types of core facilities that serve as gathering places and resources on campus, along with their associated technical staff,“ I was reminded of SMU’s new Gerard J. Ford Hall for Research and Innovation, which does exactly that. Make sure to watch the video toward the end to see how beautiful the building is. Or better yet, come and visit!  

My paper in ANOR with Laurent Alfandari and Victoire Denoyel

 It is time I started blogging again and for my first post of 2021, I figured I would mention my paper that was published last year in Annals of Operations Research : "Solving utility-maximization selection problems with Multinomial Logit demand: Is the First-Choice model a good approximation?" with Laurent Alfandari and Victoire Denoyel. For some reason I've noticed that, after one of my co-authors asked ANOR to publish an erratum because a formula in one of the proofs had been typeset wrong, the PDFs of the article point to the erratum rather than the article itself (which also has been corrected) and it was very hard for me to access the actual PDF from my online university's library system. I also noticed that the paper isn't getting cited, and I'd like to correct that, at least for researchers who work in optimization with decision choice models. So, here is a shameless plug for our paper.  

Below, I've pasted the abstract. "We investigate First-Choice (FC) assignment models, a simple type of choice model where customers are allocated to their highest utility option, as a heuristic or starting point for the Multinomial Logit (MNL) model in the context of selection problems with a utility maximization objective. This type of problem occurs in a variety of applications, from location problems to assortment planning or transportation planning. FC assignment models are less refined but computationally more tractable than the more commonly used MNL. MNL suffers from tractability issues due to its nonlinear structure when used within a large size optimization problem with binary decision variables. We design the first comparison of the two modeling frameworks in a context of customer utility maximization for selection problems with binary variables. We provide a probabilistic analysis of the expected customer choice probabilities, document the computational challenges faced by the MNL model in our setting and show in numerical experiments that the FC model exhibits excellent performance as an approximation of the MNL model with an average gap for instance of at most 2.2% for uniformly distributed utilities and of at most 1.4% for normally distributed utilities (and below 1% in a majority of test cases). The key contribution of this paper is to build the case for the FC model as a tractable, high-quality approximation of the MNL model for binary selection problems with utility maximization."

As a sneak peek, below I am also pasting a table from the numerical experiments, obtained for normally distributed utilities, so you can see how well the first-choice heuristic performs in practice. 

Let me know if you want to see a preprint of the paper! (athiele at smu dot edu)