myPersonality Project

CITATION:

Stillwell, D. J., & Kosinski, M. (2015). myPersonality Project website.

@misc{stillwell2015mypersonality, title={myPersonality Project website}, author={Stillwell, DJ and Kosinski, M}, year={2015}}

May, 2018

Note: In April 2018, we decided to stop sharing the data with other scholars. Maintaining the data set, vetting the projects, responding to the inquiries, and complying with various regulations has become too burdensome for the two of us.

myPersonality in a nutshell:

  • myPersonality was a Facebook App that allowed its users to participate in psychological research by filling in a personality questionnaire. It also offered them feedback on their scores. It was created in 2007 by David Stillwell (then a grad student at University of Nottingham and now a lecturer at the University of Cambridge). In 2009, David was joined by Michal Kosinski (then a grad student at the University of Cambridge, now a professor at Stanford).

  • myPersonality was closed in 2012 due to a lack of time to maintain it.

  • Our users could opt in (but did not have to!) to donate their Facebook data to our research; about 40% of them volunteered to do so.

  • Our research resulted in dozens of peer-reviewed papers, many in top scientific journals, driving progress in the understanding of human psychology, health, and behavior.

  • Much of our research focused on exposing privacy risks. Our findings informed new EU and U.S. privacy laws (e.g., here) and Facebook privacy policies.

  • Following the principles of open science and replicability, we provided other scholars with anonymized data they could use for non-commercial academic research. This has allowed academics from many fields to make significant discoveries advancing our understanding of human behavior.

  • We never shared the data of any myPersonality user for any commercial purpose.

  • myPersonality had nothing to do with Cambridge Analytica or SCL. They have never had access to our data, our algorithms, our expertise, etc. Moreover, unlike Cambridge Analytica, we only collected data from volunteers that gave us and opt-in consent to do so. Also, see this statement from the University of Cambridge Psychometrics Centre.

Longer story:

myPersonality was a Facebook App that allowed its users to participate in psychological research. It was created by David Stillwell in the summer of 2007. David had just finished his undergraduate course in psychology and was accepted for his PhD at the University of Nottingham. David’s plan was to share a personality questionnaire with his 50 Facebook friends. He was rather surprised when the friends of his friends, and then the friends of the friends of his friends, joined the study. Soon, thousands of people were participating every day for fun, self-insight and to donate their data to psychological research.

In 2009, David met Michal Kosinski, a psychology grad student at the University of Cambridge, and invited him to collaborate on this project. David and Michal co-authored dozens of peer-reviewed papers based on myPersonality data, many in top scientific journals.

Six million people took one of the myPersonality questionnaires until we decided to close the app in 2012, due to a lack of time to maintain it.

myPersonality participants could (but did not have to!) opt in to donate their Facebook data to our project. To do so, they would have to read and opt in to a consent statement, and then confirm their intent again in a pop-up dialog window shown by Facebook. About 40% of our users decided to generously donate their data to our research. We only recorded the data that users gave us access to, and never recorded any data of their friends.

myPersonality was one of many similar projects where volunteers could participate in psychological research. Following the principles of open science and replicability, we shared some of the data donated by our users with other scholars for the purpose of non-commercial, academic research. The data was anonymized.

In 2018, we decided to stop sharing the data with other scholars. Maintaining the data set, vetting the projects, responding to the inquiries, and complying with various regulations had become too burdensome for the two of us.

Many startups approached us with ideas of products and services that could benefit from what we learned from the myPersonality project. We considered some of the ideas and even built some prototypes, but soon realized that a career in academia was more appealing to us. Importantly, we never shared the data of any myPersonality user for any commercial purpose.

Also, a number of individuals and institutions have approached us with lucrative offers to purchase the myPersonality app or our data. We have turned all of them down. Most famously, we turned down Alexander Kogan when he approached us in 2014 with an offer to access our data for his work with SCL/Cambridge Analytica. Again, we never shared the data of any myPersonality user for any commercial purpose.

Instead, our research and experiences have made us more keenly aware of the privacy risks faced by individuals in a world that is becoming increasingly digitized and we have set out to warn users, businesses and policy makers about these privacy risks. To this end, we have given hundreds of talks (and here) warning against the privacy risks; written op-eds (e.g., for the Financial Times); and spoken with journalists (e.g., here), artists, and privacy activists who have in turn created hundreds of articles, podcasts, movies (e.g., here), and even a theater play discussing such issues.

We have also published empirical, peer-reviewed research aimed at discovering the extent of these privacy risks.

In 2013, at a time when the Facebook Page Likes of all users were publicly available, we published a paper showing that intimate traits can be accurately predicted from such Likes. In our paper we warned that this “… may have considerable negative implications, because it can easily be applied to large numbers of people without obtaining their individual consent and without them noticing. Commercial companies, governmental institutions, or even one’s Facebook friends could use software to infer attributes … that an individual may not have intended to share. ... such predictions, even if incorrect, could pose a threat to an individual’s well-being, freedom, or even life. Importantly, given the ever-increasing amount of digital traces people leave behind, it becomes difficult for individuals to control which of their attributes are being revealed. For example, merely avoiding explicitly homosexual content may be insufficient to prevent others from discovering one’s sexual orientation."

This paper started an important conversation; it was the third most widely discussed scientific paper published that year. Facebook science and legal teams reviewed it and followed up with us with additional questions. Previously, liking a Facebook Page was always a public action, but soon after the publication of this paper, Facebook altered its privacy policies to make it private by default. Policymakers in the EU and U.S. also took notice (e.g., here).

More recently, we investigated the efficiency of psychologically tailored advertising, of the kind used by Cambridge Analytica (but without their involvement; again, we had nothing to do with them). We found that it is, indeed, effective, boosting the efficiency of online advertising by up to 50%. We warned that it “... might be used to exploit “weaknesses” in a person’s character,” and that it “... challenges the extent to which existing and proposed legislation can protect individual privacy in the digital age. While ... having direct access to an individual’s digital footprint makes it possible to accurately predict intimate traits, … such inferences can be made even without having direct access to individuals’ data.

Frequently asked questions:

Are you still sharing data with others?

No. In 2018, we decided to stop sharing the data with other scholars. Maintaining the data set, vetting the projects, responding to the inquiries, and complying with various regulations had become too burdensome for the two of us.

Did myPersonality violate any Facebook policies?

No. We have always tried to comply with Facebook's policies, and Facebook has never told us that we have broken any of their rules. To the contrary, Facebook has been supportive of the myPersonality project for many years. In 2011, Facebook invited David to a workshop in Silicon Valley aimed at promoting the use of Facebook data for academic research. Since then, Facebook has remained keen to use our research and to work together with us, making further inquiries in 2013 and 2014. In 2015, Facebook invited Michal to present results from the myPersonality project as part of an academic symposium organized and chaired by Facebook at the annual meeting of the Society for Personality and Social Psychology.

Moreover, our publications exposing the privacy risks of Facebook data (and this one in particular) seem to have inspired positive changes to Facebook's privacy policies.

Did you have anything to do with Cambridge Analytica?

No. We had nothing to do with Cambridge Analytica or SCL. They have never had access to our data, our algorithms, our expertise, etc. Quite the opposite: As mentioned above, we have been vocal about the privacy risks exploited by Cambridge Analytica long before they came into existence.

What about Dr. Kogan / Spectre?

Dr. Kogan (aka Spectre) was a professor at the University of Cambridge at the time when both Michal and David were students there. Dr. Kogan and his students obtained access to myPersonality data for their academic projects in 2013 and 2014. He was asked to delete the data in 2014 and provided assurances to University of Cambridge both that he did so and that he never used it in his activities with SCL or Cambridge Analytica.

Was there a “leak” of data from the myPersonality database?

In April 2018, it was brought to our attention that one of the scholars who had access to our anonymized data (a professor at a prestigious American university) put their login credentials in a file intended to be shared with their students on Github, an online code repository. However, the file was publicly available on the Internet, which is clearly a breach of the terms that academics agreed to when requesting a collaboration with myPersonality. Once we learned of this, we closed their account.

In nine years of academic collaborations, this is the only such instance where something like this has occurred. myPersonality collaborators have published more than 100 social science research papers on important topics that advance our understanding of the growing use and impact of social networks. We believe that academic research benefits from the properly controlled sharing of anonymized data among the research community.

I am a young academic interested in myPersonality and “big data”. Do you have any recommendations?

While psychology research is often conducted on undergraduates in a lab setting, we feel that the digital revolution has opened new vistas for scholars interested in understanding human beings and improving people’s lives. Though myPersonality is no longer maintained, there are still a tremendous number of big data sources available to academics (see here). Working with big data and computational methods does pose some unique risks and challenges, but on the whole, we believe such approaches have a lot to contribute to the social sciences. We wish all the best with your research!