Chapter

The Future of Social Science & National Security Project

Bridging the Gap Between Academic Research and National Security

Examining the tension in social sciences between basic and applied research in the shifting funding landscape and research priorities for sponsored programs

By Mia Bloom

Defense Policy & Posture

November 14, 2024

This chapter examines the ways in which academic research, especially research that is focused on violence, can serve more than one purpose. Certainly, academic research must first and foremost consider whether it is grounded in the theoretical traditions of a particular discipline. It remains a paramount goal that any research on violence furthers the sciences of the study and contributes to a better understanding of politics and power. The gold standard is evident in the natural sciences where they conceive of science as “a cycle that moves from discovery to invention and back again, a model that breaks down disciplinary walls and encourages collaboration.” Political science might be the only science that disdains and dismisses applied research. On the one hand, debates rage over whether something applied is sufficiently theoretical, and a second debate blames social scientists for failing to demonstrate the utility and usefulness of their research. The chapter explores the debate.

Theory vs. Utility: Navigating the Tensions

In International Relations (IR) theory, there has been a long-standing tension between academic work that is rigorous and elegant versus that which is nuanced and useful; and by useful we usually mean research that can be applicable to addressing some problem. Michael Horowitz (2015) addressed this issue in a War on the Rocks essay when he noted how many emerging projects were attempting to bridge the divide between theoretical work and policy-relevant research, highlighting for example, Jim Goldgeier’s project out of American University and George Washington University.¹

Bridging the Gap project trains academics on how to ask policy-relevant questions and write for public policy audiences. To complement these efforts, our understanding of what policy relevance “is” needs improving. In particular, distinguishing between research with significance for policy and research that is policy actionable — promoting realistic policy actions — can bring analytical clarity to the concept of policy relevance and help enhance efforts to bridge the gap.²

Horowitz ended his seminal essay with the following concern: “A lack of precision in what policy relevance is, can encourage thinking about the issue that does not accurately capture reality and leads to confusion…” ³ Rachel Glennerster summarized what appeared to be an inherent dichotomy when she contrasted the two different research approaches between the academy and the policy worlds.

Academia rewards findings that are different and unexpected. In policy, it is more important to be right than novel… In academia, people argue a lot about the direction of an effect but very little about the magnitude: in policy, it’s the reverse.⁴

A decade ago, I witnessed this tension play out in real life. In a political science seminar at Penn State University, someone asked a senior IR scholar who was designing a “model for counter insurgency” whether their model would apply to Afghanistan. The response is burned in my memory and has stayed with me all these years, because the scholar in question responded: “why would I? Afghanistan is an N of one.”

An N of one in which the US was embroiled in warfare for two decades at the cost of thousands of American lives, and trillions of dollars of treasure, and has left Afghanistan no better than how we found it — providing safe haven for the 9/11 terrorists and a continued source of instability in the Middle East. Thus the argument for the continuing relevance of granular case studies can be made for specific cases that have great importance for national security and US foreign policy. Not all detailed case studies are equal; a single-country study of Lichtenstein might not be as significant as one on Iraq, Syria, or Afghanistan. Additionally, detailed and granular case study approaches allow researchers to examine a particular case from a variety of perspectives and consider different levels of analysis including how changing and emerging technologies will likely influence a variety of cases. Finally, this in-depth approach offers the reader the chance to see changes over time instead of a single snapshot.

As we observed from our research on ISIS, the Taliban used social media to project power, encouraging Afghan National Army to surrender,⁵ giving the (false) impression that they had already taken over several Afghan provinces. Their social media campaign likely encouraged Afghan security forces to put down their weapons without a fight. As the US withdrew on August 30, 2021, the Taliban had reoccupied province after province in days not months. Further, videos of the Taliban equipped with US weapons⁶ stolen or seized from the retreating Afghan army provide corroborations that the 20 years barely made an impact.

This chapter examines the ways in which academic research, especially research that is focused on violence, can serve more than one purpose.⁷ Certainly, academic research must first and foremost consider whether it is grounded in the theoretical traditions of a particular discipline. It remains a paramount goal that any research on violence furthers the sciences of the study and contributes to a better understanding of politics and power. The gold standard is evident in the natural sciences where they conceive of science as “a cycle that moves from discovery to invention and back again, a model that breaks down disciplinary walls and encourages collaboration.”⁸

The Misplaced Stigma on Applied Research

Being useful is not the enemy of being good. If academics act as if applied research is less valuable than theoretical work and dismiss it because this approach may have real-world applications, academia has inadvertently elevated the obscure for no reason other than it appears to be more scientific than applied work. Political science might be the only science that looks down on applied work. Certainly, in the above examples, mathematics and engineering have no such qualms about research relevance, applying their work to addressing real-world problems, and consider the long-term impact of research might result in designing something they will trademark in the future. Scientists see the value of their research transitioning and generating income and not merely counting citations.

Gary Andres and Janice Beecher acknowledged that political science appeared to be behind the curve in terms of “bridging the gap” between work that was theoretical and work which was applied. The natural and formal sciences have always done a better job of moving back and forth between the two settings without losing “their sense of professional identity, connection, or esteem.”⁹ The onus of responsibility has been repeatedly placed on academics who have allegedly failed to demonstrate the “value” of what we do compared to the natural and formal sciences such as research on COVID-19 or cancer in the STEM fields, despite the fact that the S in STEM includes the social sciences. Additionally one can safely assume that scientific researchers are working to resolve problems associated with their discipline but few political scientists aspire to work in politics (rare examples exist), and many do not weigh the benefits of potential applications.

Like many academics, political scientists have often done a poor job explaining their discipline to public officials, the media, or society in general. The public has some innate idea of what, say, cancer researchers are doing, but very little idea of what political scientists do.¹⁰

There are two opposing and inconsistent tensions at work.¹¹ On the one hand, debates over whether something applied is insufficiently theoretical, and a second debate blames social scientists themselves for failing to demonstrate the utility and usefulness of their research. The move from basic research to something more applied has been challenging across multiple areas and several US government funding agencies from the NSF to the Department of Justice.

Very few (social) science researchers can foresee the outcomes of research projects ahead of time. Following social science and human behavior, we cannot always predict outcomes that are complex, nuanced, and unpredictable. However, keeping an open mind during research, allows one to (re)imagine creative ways to apply the findings of our research projects.

Case Study: Unexpected Insights from ISIS Research

This was the case with a small add-on project that was awarded in 2016. At the time, my assistant and I thought a project dedicated to documenting ISIS’ activities would be interesting and useful. There was no such archive available for free, and the research arm of the Navy has always had the goal of creating a comprehensive shared data repository to merge the findings of the various projects where results could be replicated and validated but also so that research could be used by others. The original idea was to keep a copy of the official ISIS announcements, breaking news, and material posted to their encrypted Telegram channels.¹² However, in my zeal to join as many channels as possible, often quickly (channels were available for 20 minutes or you lost out), the research team accidentally included several chats in addition to semi-official channels.

My error ended up offering researchers the opportunity to study the effects of excessive social media usage on human behavior, highlighting the addictive properties intentionally designed to maintain user attention and foster both psychological and emotional dependence on the virtual world.¹³ ISIS achieved this by limiting the time that new content, e.g., a video, was available on its channels. Users were required to constantly monitor the channels to ensure they did not miss out on this content. This design combined the fear of missing out (FOMO) with what BF Skinner called a “variable interval schedule of reinforcement.”¹⁴ In practice, this meant content reinforced the commitment of the end-user by being varied, stimulating, abundant, constantly changing and only allowing for a limited time for response. Such qualities are present in other compelling (and addictive) activities (e.g., slot machines, online gambling, and mobile gaming).¹⁵ These dynamics not only encouraged behavioral commitment but fostered a psychological (and perhaps even physical) dependence and addiction. ¹⁶

Monitoring both channels and chatrooms offered a variety of research data we did not anticipate. We were further able to conduct experiments with the chats, and compare and contrast what was said in English, versus Arabic, versus a half dozen other languages. We could also monitor behavioral elements and detect the use of bots. At the pinnacle of the research, we had 8 different accounts, about half were identified as female by using “umm” as part of the name, and others as male by using “Abu” in the Kunya (or nom de guerre). Being able to vary gender and language resulted in original findings that allowed the research team to assess whether there were differences in recruiting men versus women, whether women were exposed to comparable graphic content, and whether messaging differed depending on language, age, and gender.¹⁷

Starting a project with modest goals and finding a myriad of ways to use the data and the methodology was an unexpected yet pleasant surprise. We could time the channels and chats to see whether time zones impacted the amount that was posted. How quickly was information disseminated via automated accounts or bots? How many channels reproduced the same content? Did we observe the same administrators across multiple channels? Our IRB created some unique challenges and explicitly instructed researchers to have very limited engagement, such that we observed but did not participate in the propaganda channels. This also led to some challenging moments when ISIS, concerned that their networks had been infiltrated, tried to contact the account owners as the group became wary of infiltrators or people who merely observed but did not contribute whom they deemed “lurkers”. These observations about posting frequency, varying the nature of the content (along an intermittent schedule of reinforcement), and timing how quickly accounts were banned or kicked off the platform all became data. To this end, graduate students were encouraged to keep a diary of the research, as one would do while conducting in-person field research.

We faced a conundrum early on when several of the accounts detected threats against a citizen journalist in Dallas who was engaged in his own research, and doing what we had done, by joining the accounts. However, he attempted to engage and “interview” ISIS supporters. Pretty quickly, the ISIS supporters decided that his inquiries were too invasive, and he might be a CIA agent. The admins doxed him, posted pictures of him from his social media accounts, published his address and cell phone number, and issued a fatwa (religious ruling) that encouraged ISIS supporters to find him and kill him.

This posed ethical challenges one never expects to encounter when conducting research. You are not supposed to play a part in the research; you are merely observing the processes at work. But we could not sit back and knowingly leave someone in harm’s way. A choice was made; we contacted law enforcement and provided screenshots. The wannabe journalist was placed in protective custody. This is where the theoretical research revealed its potential for being applied without premeditation. By having access to a platform where the occasional terrorist plot was hatched or discussed, we had multiple opportunities to preempt violent attacks. No one expected this outcome, but the team found it gratifying to be able to prevent violent attacks from happening. Several attacks never happened because the research team sent screenshots about a plot or an operation still in its infancy planning stages to relevant law enforcement in the US, UK, Australia, and Canada.

On occasion, the media contacted us to ascertain whether an ISIS claim was true. We verified for the BBC and CNN when Abu Muhammed al Adnani (real name Taha Subhi Falaha) the ISIS propagandist and spokesman was assassinated in August 2016.¹⁸ With experience and observation over several years on the project, it became obvious which attacks perpetrated against Western targets were officially ISIS and which ones were merely inspired by the group. The Western media tended to interpret any attack that mentioned ISIS as being planned and sanctioned by the group. In essence, if someone made a declaration that they were ISIS, the media and US Counter-terrorism experts tended to count it as an ISIS attack — blurring the distinction between direction and inspiration. There was no way to be on every channel all the time, and terrorist operations occurred despite our eagle-eyed monitoring. However, observing the differences in reporting on the chat rooms and channels it became obvious when an attack had been ordered and supervised by the group versus an opportunistic individual claiming attacks in the name of ISIS, for which there was little to no fingerprint.¹⁹ The differences of coordination and the content of the propaganda, for example, between Bataclan, Dhaka, and Orlando or San Bernardino made clear ISIS’s role on the former but not the latter. Added to this, conducting the research helped elucidate when ISIS made false claims for attacks, for example when they would opportunistically claim an attack that yielded casualties. This was most apparent when ISIS claimed, without evidence, that Stephen Craig Paddock, the Las Vegas shooter in 2017, had been a soldier of ISIS. A claim that proved to be blatantly untrue.²⁰

We encouraged federal law enforcement to duplicate this methodology but drew the line at sharing our accounts and access with them. We asked them why they were not monitoring these platforms themselves. Their answer was simple: our research team had the access, acumen, time, and personnel which agencies did not have. We knew there were others engaged in this activity,²¹ and we were grateful to be part of a larger ecosystem that if we saw something we would say something. By this time, private companies and think tanks were also monitoring the Telegram channels. We were happy not to be the only ones engaged on these platforms, monitoring the activities, and conducting this research. No one wanted the responsibility of being the only barrier between ISIS and a successful terrorist attack.

The project remained adaptable throughout the three years such that we could watch as ISIS accounts dabbled in other platforms or applications that would likely be less infiltrated by researchers and law enforcement. At various times, ISIS admins would express their frustration with Telegram and tried to move their activities to Hoop messenger, Tam Tam, and Rocket Chat, but they always returned to Telegram.²² We could also measure to what extent de-platforming succeeded at drawing down ISIS’ efficacy in recruiting.

While Twitter and Facebook saw early adoption by ISIS propagandists and sympathizers in their so-called “virtual caliphate,” account closures and policing of content became more aggressive, leading many ISIS online recruits to move to Telegram. That became ISIS’s preferred platform for almost four years.²³

After multiple incidents where accounts were lost or logged out permanently during an automatic application update, the burner phones could no longer reinstate our access because of university bureaucracy that did not allow government funds to be used for cell phones; only disposable phones with a “pay as you go” card could be used for sponsored research. Since the team had stretched a one-year project into multiple years using no-cost-extensions, the university refused to purchase any of the phone cards after year one to keep those budget numbers current; once we lost access for example, because Telegram actively shuttered “ISIS-following” accounts, there was no way to retrieve all of the data still associated with those accounts even after we provided Telegram with ample documentation that we were academic researchers and not ISIS fanboys. On the bright side, this small add-on project resulted in 4 years of collection and monitoring since we found ways to automate the collection and stretch the resources as we added efficiency and cut costs; we were productive in ways unexpected in the original project and ready to use the data in creative ways. During this period, we generated dozens of deliverables, including conference papers, briefings, articles, and genuinely multi-disciplinary approaches. Best of all, we had transitioned the project and built a platform that could be useful for a variety of future national security projects, as well as developing a methodology for “digital ethnography” that could be used to study other extremist groups.

Findings

Advancing AI and Adapting Research for Emerging Threats

The small project had discovered and assessed the addictive quality of social media but also generated a living archive of data. The computer scientists on our Georgia State University interdisciplinary team used the project data to improve the quality of the AI that relied heavily on existing thesauri of images and stock photos. The computer scientists had refined some of their tools in Artificial Intelligence to “see” items and recognize shapes without full information and even in a bad light, despite multiple occlusions. The AI calculation run multiple times allowed the algorithms to recognize matter in real-world settings.

The project contributed to computer science as much as it did to the behavioral social sciences and the methodology was easily adaptable to different groups as Telegram shuttered many of the ISIS’S accounts by May 2020. As the saying goes, when God closes a door, he opens a window. As social media companies clamped down on jihadi terrorist use of their platforms, other groups found a hospitable environment and Telegram became the platform for the boogaloo movement, QAnon conspiracy theorists, and a host of other Racially/Ethnically Motivated Extremist (RMVE) violent actors. The algorithms we developed to scrape Telegram became equally useful for scraping Parler before Amazon hosting services removed it from the cloud in January 2021 in reaction to the January 6, 2021, attack on the capitol. The big data tools we had developed could be adapted from our ISIS research to pivot quickly and research the far right and domestic terrorist groups that had emerged in recent years. The ability to access metadata also allowed for geotagging posts and uploads.

Bridging Theory and Practice for Real-World Change

The project also fulfilled the “gold standard” of being theoretical and useful as one of the published findings from the project, “Far-Right Infiltrators and Agitators in George Floyd Protests: Indicators of White Supremacists” was entered into the Congressional Record during the policing hearing on June 10, 2021, by Maryland Representative, Jamie Raskin.²⁴ While the basic research had generated testable hypotheses about terrorist use of propaganda, the methodologies we developed as part of this research as an interdisciplinary team — walking a tight rope between theoretical and applied research — was the best deliverable of all.

What this means for most basic research is that it is important to have a clear direction for the research and to address the various theories and hypotheses originally conceived. But it is also beneficial to be intellectually agile and adapt along the way. This advice might be a better reflection for researchers after tenure, since department committees and university promotion and tenure committees speak reems about interdisciplinary research but rarely walk the walk and reward it.

Navigating Political Boundaries in Research Funding

There is another consideration, one that is far more treacherous than the threat of conducting research in which only a handful of colleagues will ever read. It is the unfortunate reality that much of the research we conduct for Minerva might be construed as political despite the DoD’s stated preference to avoid partisan political debates. While international terrorist groups have clear legal designations such as being listed as foreign terrorist organizations, domestic terror groups do not (yet) have comparable status. Moreover, studying domestic extremism has the potential to inflame political parties who might consider the members to be part of the electoral base. It is one of the main reasons why in 2013 the National Science Foundation (NSF) stopped funding political science²⁵ as part of the Omnibus spending bill. The NSF declined “to explain its reasons for eliminating the grant call, one of two that typically take place each year. But leaders in the field are blaming Congress, which on 21 March passed a bill requiring that NSF-funded political-science research benefit either national security or economic interests.” The cancelation excluded a small tranche of security scholarship²⁶ (which was rebranded in 2019 as “security and preparedness.”²⁷

“Evidence-based policymaking is an attractive paradigm, but political considerations will almost always take precedence. Policy decision-making is bound by what is politically feasible.”²⁸

This challenge is especially salient for future Minerva researchers. The DoD, Department of Homeland Security (DHS), and National Institute of Justice (NIJ) all have very specific types of research they can fund. DHS is constrained from funding research of activities within the United States, as a program manager commented to me on what I thought was a great pitch for a research project to use the dark web to predict where attacks might happen, “We have this thing called the US Constitution and that research would violate individual rights to privacy, freedom of speech, assembly and so on.” NIJ funding is targeted at the local level, specifically beneficial for policing. So as the Secretary of Homeland Security, Alejandro Mayorkas, acknowledges threats emanating from insider groups, right-wing extremists, and domestic actors, there is no mechanism to study these issues given the current stove-piped research agendas. We know that most terrorist threats will likely come from domestic actors, but DoD and Minerva have tended to limit their funding to international terrorism even when clear connections to domestic terrorism exist. DoD does not lead work on domestic populations; it is DHS’s responsibility or another agency depending on the topic.

The Path Forward for Social Science and National Security Research

The funding landscape for social sciences is challenging. We saw this firsthand in 2020 when then Secretary Mark Esper canceled the Minerva program²⁹ at the recommendation of Michael Griffin, the then Under Secretary of Defense for Research and Engineering, who clearly did not view social science as a research priority. After a bipartisan Congressional Committee supported its revival, and with a new administration in the White House, Minerva program funding was eventually reinstated in the core budget.³⁰ For many academics, sponsored research is crucial to support not just the researchers but also graduate students and undergraduate students. It is increasingly a skill demanded by employers, and sponsored research professionalizes our students to teach them how to navigate the challenging landscape whether it is for DoD, NSF, or other agencies. At a university like mine, which is overwhelmingly first-generation and diverse, Minerva-sponsored research has offered us the chance to support nontraditional students and offer them the chance to present their own research at conferences and venues they otherwise would not have had access to. Minerva is a great program but also must consider the changing threat landscape and determine the optimal ways forward to continue research without the threat of accusations of partisanship. It is crucial that we continue to engage in social science research and the evolving challenges to national security as we face the influence of misinformation and the threats posed by generative artificial intelligence.

We also need to continuously interrogate our own research to make its usefulness obvious. We should pair our academic research with policy-relevant editorials or other publications that better translate the value of social science research to lay people. We should engage as much as possible with the military academies by writing clearly and with a minimum of acronyms or jargon. In doing so we not only move the research and the science forward but avoid accusations of ivory tower navel-gazing.

Notes

1
https://www.jamesgoldgeier.com/bridgingthegap
2
https://warontherocks.com/2015/06/what-is-policy-relevance/
3
https://warontherocks.com/2015/06/what-is-policy-relevance/
4
https://blogs.lse.ac.uk/impactofsocialsciences/2014/02/11/policy-world-versus-academia/
5
https://www.nytimes.com/2021/08/20/technology/afghanistan-taliban-social-media.html
6
https://www.cnn.com/videos/world/2021/08/20/taliban-weapons-afghanistan-tsr-todd-pkg-vpx.cnn
7
https://www.seas.harvard.edu/news/2017/01/false-choice-basic-vs-applied-research
8
https://www.seas.harvard.edu/news/2017/01/false-choice-basic-vs-applied-research
9
Gary J. Andres and Janice A. Beecher, “Applied Political Science: Bridging the Gap or a Bridge Too Far?” PS: Political Science and Politics Vol 22, No. 3 (Sept 1989), pp. 636-639.
10
https://psmag.com/news/congress-nsf-political-science-budget-54283
11
https://www.formpl.us/blog/basic-applied-research
12
https://crestresearch.ac.uk/comment/islamic-state-messaging-telegram/
13
Jean M. Twenge, IGen, NY: Simon and Schuster, 2017, http://www.jeantwenge.com/igen-book-by-dr-jean-twenge/; see also https://onlinelearningconsortium.org/an-interview-with-jean-twenge-about-the-igen/?gclid=EAIaIQobChMI847Fw5vo8gIVFI7ICh1LDQlbEAAYASAAEgLtJfD_BwE
14
https://www.verywellmind.com/variable-interval-schedule-2796011
15
Natasha Dow Schüll, Addiction by Design: Machine Gambling in Las Vegas. Princeton University Press, 2014
16
https://crestresearch.ac.uk/comment/islamic-state-messaging-telegram/
17
Mia Bloom & Chelsea Daymon, “Assessing the Future Threat: ISIS’s Virtual Caliphate,” Orbis, Volume 62, Issue 3, 2018, pp. 372-388,
18
https://www.atlanticcouncil.org/blogs/syriasource/isis-spokesman-abu-mohammad-al-adnani/
19
https://www.tandfonline.com/doi/abs/10.1080/09546553.2017.1339695
20
https://www.npr.org/sections/thetwo-way/2017/05/24/529685951/what-does-it-mean-when-isis-claims-responsibility-for-an-attack
21
For example GNET, https://gnet-research.org/2020/07/30/telegrams-anti-is-campaign-effectiveness-perspectives-and-policy-suggestions/, and GW Program on Extremism, https://extremism.gwu.edu/sites/g/files/zaxdzs2191/f/EncryptedExtremism.pdf, RESOLVE network at USIP, https://www.resolvenet.org/system/files/2019-09/RSVE_RVESeries_ResearchingJihadistPropaganda_CWinter_May2019.pdf
22
https://www.degruyter.com/document/doi/10.7312/krau19300-021/html
23
https://www.justsecurity.org/67605/no-place-to-hide-no-place-to-post-lessons-from-recent-efforts-at-de-platforming-isis/
24
https://twitter.com/just_security/status/1270811076448354304?s=20
25
https://www.vox.com/mischiefs-of-faction/2019/3/6/18252793/why-republicans-hate-nsf
26
https://psmag.com/news/congress-nsf-political-science-budget-54283
27
https://politicalsciencenow.com/an-update-on-political-science-at-the-nsf/
28
https://blogs.lse.ac.uk/impactofsocialsciences/2016/05/25/should-academics-be-expected-to-change-policy-six-reasons-why-it-is-unrealistic/
29
https://www.insidehighered.com/news/2020/03/05/pentagons-social-science-research-program-chopping-block
30
https://www.defensenews.com/congress/2021/08/02/defense-top-line-will-probably-go-up-key-dems-see-gop-boost-as-path-to-a-deal/

Recent & Related

Op-Ed

The United States Is Self-Destructing Amid Empire Collapse

Julia Gledhill

Op-Ed

USS Gerald Ford limps out of hot war and into embarrassment. Why?

Dan Grazier

Commentary

Five Things to Know About Iranian Minelaying

Kelly A. Grieco • Marie-Louise Westermann