First 100 Days: The Ethics of Big Data

By Maria Christoff

Assistant Editor


During the 2016 primary election season journalists and mental health professionals indulged in armchair psychiatry with abandon (for example, here, here, and also a piece by personality researcher McAdams). How could they resist? The material Trump provided was explosive and rich. After November 9th, there was a lull. What use or fun was it to dissect his character when no amount of previous analysis had an effect? He won the election anyway. As a therapist and a citizen, I understand that Trump has supporters, and in both roles I strive to understand the perspectives of those who support him, despite my own concerns. After the initial shock of the election, Trump began his tenure as president, continuing to tweet, tweet, and to sign a number of executive orders, usually penned by Bannon. My concerns grew.

On February 9th, Dr. Lance Dodes of the Boston Psychoanalytic Society released a statement signed by him and 34 colleagues in the mental health profession, voicing their grave concerns about Trump’s presidency, and breaking the Goldwater Rule. Allen Frances, of DSM revision fame, retorted along these lines: It’s not about what is wrong with Trump, it’s more about what he does as president. I agree with Frances. After all, mental illness itself does not act on anyone’s behalf. Plenty of mentally ill people are good and moral, whereas plenty of sane people cause a great deal of harm to others.

Dodes’ statement did not gain much traction. Other theories, more health-than-mental-health related, have emerged. Is he senile? Does he have neurosyphilis? Returning to the statements made by mental health professionals, I have begun to wonder what underlies the impulse of Dodes, McAdams, and others to take this form of social action. What if, on some level, the mental health field feels a sense of responsibility regarding Trump’s election, and these statements attempt to undo its guilt? In fact, that may not be too far from the mark. Psychology did play a role. However, the manner in which psychology helped elect Trump has nothing to do with clinical work, and everything to do with personality research.

The University of Cambridge Psychometrics Centre and the private company Cambridge Analytica utilize Facebook data to create personality profiles based on the Five Factor Model of personality. During the 2016 election cycle, Cambridge Analytica used these personality profiles to create targeted ads, which were deployed on Facebook to influence people both to vote for Trump and to opt out of voting for Clinton. These targeted ads, combined with the increasing isolation of Trump supporters in a social media bubble, make a plausible case for big data having had at least some influence on the 2016 election. Based on their website, the company would certainly like to have everyone believe their part in electing Trump was quite large, and that they also influenced the U.K. Brexit vote. The question is not whether politicians paid private companies to use the big data techniques developed by psychologists to sway the election. They did. The question is whether it worked.

Many of the individuals targeted based on their personality profiles and demographic data were rural, poor, and in despair. Many were wealthy, educated, and implicitly or explicitly racist. Facebook CEO Mark Zuckerberg appears somewhat guileless in his calls for users to live their entire lives on the platform, while also giving inadequate thought to social responsibility. Nevertheless, information bubbles, fake news, and targeted ads create profit. The colonization of minds by neoliberalism is neither covert nor unconscious. It is carried out in the open, while its designers brag of its success and implore users to use more. On the other hand, many argue that big data’s role is vastly overrated or even conspiracy-theory-level bonkers. After all, this perspective asserts, candidates win elections, not campaign strategies or data companies. Through the muddle, some facts remain. These include that, regardless of the nature of psychological manipulation techniques (personality profiling and ad targeting), they do not equate to total mind control. An ad on a screen, perhaps nestled beside an image of a beloved family member, registered below the level of consciousness, might not entirely strip one of critical thinking abilities, or reduce free will to negligible amounts. Nevertheless, it can have tremendous influence. Whatever role big data played, the responsibility will fall back onto the voters, as will the consequences, albeit unfairly distributed, with the poor and minorities receiving the greatest burden.

This would hardly be the first time that the methods and tools of psychological research have been used towards morally questionable ends. The field of psychology was integral to the 20th century eugenics movement in both the United States and Germany. Gould (1996) described psychology’s role supporting biological determinism and scientific racism throughout the 20th century in detail. More recently, APA accredited psychologists were found to be complicit in the development and implementation of torture techniques at post-9/11 Guantanamo Bay. The question that emerges is to be ethical, let alone activist, how must psychology respond to this role? How must psychology respond to its beloved Five Factor Model being used in this way? Clinical psychologists, psychoanalysts, and non-researcher psychologists may wish to disavow any connection to big data companies. However, as accredited psychologists, they are all covered under the same ethics code (or a similar code) as personality researchers. Yet, save the work of reporters who have reached out to those parties involved, the psychological field on the whole remains largely silent on the issue. A special issue of the journal Psychological Methods published in November 2016 regarding big data in psychology neglected to include an article solely devoted to the ethics of big data.

The larger crux for practitioners of psychotherapy does not exist only in the overlap between being a therapist and a private citizen activist. Surely, that area is difficult enough to navigate, and well worth exploring in its own right. Rather, the problem presented here is the political positioning of the field of psychology itself, which its practitioners represent. Psychologists must consider how the use of big data methods to help Trump win the 2016 election reflects on the field of psychology (regardless of the effectiveness). They should consider what kind of responsibility they must admit to, and what steps should be taken to repair whatever damage has been caused. Clearly, the historical record of psychology positing itself as apolitical or even super-political has long ago crumbled. It is time for psychology to condemn the use of its methods by private companies paid by politicians and working with the biggest of big business. It is time for psychology to work harder to safeguard its techniques and adhere to its code of ethics. Clinicians can diagnose Trump as much as they wish, but as long as psychology refuses to turn its critical gaze back on itself, it will remain complicit in the same neoliberal system it condemns.