In 2015, as the U.S. military prepared to support Iraqi forces in their push to retake Anbar Province from the Islamic State (ISIL), a group of Psychological Operators (PSYOPers) from the U.S. Army came up with an idea – they should rehearse the work they were preparing to do. Elementary as this might sound, the reality is that the military provides very few opportunities to actually test-run operations in the information environment. This is a story about how a scrappy group of us tried it out, and what we learned that might help us all do it better in the future.
One of the critical functions the U.S. provided to the Iraqi Army during the 2015-2016 counter-ISIL campaign was one called Military Information Support Operations (MISO), which are designed to support the kinetic outcome of a particular activity through influence. In this particular case, the objective was to encourage the civilian populations of major urban areas in Anbar Province to rise up against ISIL, or at least provide tacit support to the national forces as they fought to regain territory.
The group of us who came together to help the PSYOPers test the messaging they planned to use was eclectic. There were political scientists, sociologists, engineers, anthropologists, historians, linguists and advertising executives. I was there as a wargame designer, to help build the sandbox environment in which we would conduct the trials and give feedback on what worked – and what didn’t.
We quickly realized that we had a fundamental problem: how were we supposed to evaluate the impact of messages on urban residents of Anbar without having access to any? We were not going to fix this, and we had a short timeframe, so the collective wisdom in the room was that we would just do our best to model them. But model them how? There was some thin polling data and survey work about attitudes, but it all (obviously) predated the ISIL occupation. So we came up with a gut-instinct schematic. We would recruit our fake audience from people who resembled urban Anbaris in what we felt were key respects for the information campaign in “tiers” and evaluate their responses accordingly. The top tier was expatriate Iraqi Sunnis from Anbar – they were few and far between in the weeks we had – next came any expatriate Iraqi Sunnis, then any Sunni Arabs, and other subject matter experts who spoke Arabic and had worked in Iraq, regardless of their ethnicity, and so on…
The good news is our results ended up being helpful. The bad news is we winged it. The science on how best to approximate the responses of a specific demographic audience is far more nascent than it should be at this stage in the development of both practical military operations and the academic disciplines associated with MISO. Red Teaming practitioners need to keep this constantly in mind, and we should all be looking for opportunities to improve and expand our knowledge base through scientifically valid projects. Existing statistical techniques are good at extrapolating inference from relatively small samples, but they cannot overcome fundamental differences in cultural bias. For the sake of the validity of all of our work – and the lives and security of the people we all work to protect – we need to elevate basic science on this issue to the top of the community research agenda.
This article originally appeared in the SUNY Albany Center for Advanced Red Teaming (CART) Red Siren quarterly newsletter.