Draft:MyPersonality

Disclosure: I was a collaborator of myPersonality project so I may have a conflict of interest. I am submitting this draft through AfC for independent review per WP:COI guidelines.

 ----

myPersonality was a Facebook application and research platform developed by David Stillwell at the University of Cambridge's Psychometrics Centre in 2007.^[1] It allowed users to take psychometric tests — primarily the Big Five personality questionnaire — and receive immediate feedback on their results, while optionally consenting to share their Facebook profile data for academic research. Over four years, the application attracted more than six million participants, making it one of the largest psychology studies ever conducted.^[1] The resulting dataset was used by nearly 200 researchers from over 100 academic institutions and contributed to more than 1,700 peer-reviewed publications.^[2]

myPersonality became central to public discussion of data privacy following the Facebook–Cambridge Analytica data scandal, although the creators were not involved with Cambridge Analytica and the platform's data was not used in Cambridge Analytica's operations.^{[3] [4]}

myPersonality was created in 2007 by David Stillwell, then a PhD student at the University of Cambridge's Psychometrics Centre.^[1] The application was initially shared with Stillwell's approximately 150 Facebook friends and grew through snowball sampling, eventually reaching six million participants without paid advertising.^[1] By 2012, approximately 150,000 users had "Liked" the myPersonality Facebook page, enabling the researchers to recruit tens of thousands of participants for new studies within hours.^[1]

The application offered 25 psychometric instruments, ranging from the 300-item IPIP proxy for the NEO-PI-R to the Satisfaction with Life Scale and Raven's Progressive Matrices.^[1] Participants received immediate feedback on their results as the primary incentive for participation, a design choice that the creators found produced higher-quality data than financial incentives.^[1]

Michal Kosinski, then a graduate student at Cambridge, joined the project and collaborated with Stillwell on developing the research program around the collected data.^[1]

myPersonality was discontinued in 2012 because of Stillwell's inability to continuously update the application to keep pace with Facebook's changing technical requirements and API specifications.^[1]

The myPersonality dataset combined self-reported psychometric test results with Facebook profile data that participants had voluntarily shared. The combined database exceeded 50 gigabytes, with some tables containing hundreds of millions of rows.^[1]

The psychometric data showed reliability comparable to or exceeding traditional laboratory samples. The 100-item IPIP NEO-PI-R measure achieved a Cronbach's alpha of 0.91, compared with 0.89 in the instrument's standardization sample.^[1] Facebook-reported gender matched self-reported gender in 98.8% of cases (n = 28,628), and age matched in 95% of cases (n = 18,321).^[1] The IPIP NEO-PI-R was retaken over 1.15 million times, with more than 5,000 participants completing it more than ten times, enabling extensive test-retest analyses.^[2]

The sample was international, younger, and more female than the general population, consistent with Facebook's user demographics at the time.^[1] However, it included substantial age diversity — over 100,000 participants were over 55 years old.^[2]

The myPersonality dataset enabled a series of influential studies demonstrating that personal attributes could be predicted from digital footprints. Kosinski, Stillwell, and Graepel (2013) showed that Facebook Likes could predict personality traits, sexual orientation, ethnicity, political affiliation, and other attributes with high accuracy, a finding that raised significant privacy concerns.^[5] That paper has been cited more than 4,400 times.^[6]

Youyou, Kosinski, and Stillwell (2015) demonstrated that computer models using Facebook Likes achieved personality judgment accuracy (r = 0.56) that exceeded average human judges (r = 0.49) and approached the accuracy of spouses (r = 0.58), requiring only 10 Likes to outperform a work colleague and 300 to outperform a spouse.^[7]

Park et al. (2015) used myPersonality data to show that language-based assessments from Facebook status updates converged with self-reported Big Five personality traits at r = 0.38, comparable to the accuracy of informant reports.^[8]

Matz, Kosinski, Nave, and Stillwell (2017) used personality profiles inferred from Facebook Likes to conduct three field experiments reaching 3.5 million users, demonstrating that advertisements matched to recipients' personality traits produced up to 40% more clicks and 50% more purchases than mismatched advertisements.^[9]

The dataset was also used in research on musical preferences and personality,^[10] personality similarity in couples and friends,^[11] the geographic distribution of personality traits in the United States,^[12] and crowd intelligence.^[13]

• Big Five personality traits
• Digital footprint
• Computational social science
• Facebook–Cambridge Analytica data scandal
• Psychometrics

Category:Meta Platforms applications Category:Psychometrics Category:Personality tests Category:University of Cambridge Category:2007 software

• AfC submissions by date/27 May 2026

• CS1 maint: DOI inactive as of May 2026
• AfC submissions declined as a large language model output
• Declined AfC submissions