Vetting job applicants’ social media profiles long ago became standard practice and new services help applicants screen and scrub their profiles – but can your social media activity really reveal details of your personality? A new Australian analysis suggests that it can.

The two-year research project – a collaboration between volunteer data analyst and statistician Alice Matthews and Reputationaire founder Andrew Hine – found a strong correlation between the content of participants’ social-media posts and their performance on standardised personality tests.

The project included 348 volunteers from 41 countries who were asked to provide their LinkedIn, Twitter, Reddit, and Stack Overflow handles and complete the standard 50-item Big 5 OCEAN questionnaire.

OCEAN – a personality test that measures ‘big 5’ traits including openness to experience, conscientiousness, extraversion, agreeableness, and neuroticism – has been widely used for decades for purposes including screening of job candidates.

A proprietary Reputationaire algorithm scraped each participant’s public content, then fed it to IBM’s AI-powered Watson Personality Insights to parse that content and derive a profile of the user against a broad range of personality traits.

When comparing these scores against the user’s OCEAN results, the researchers found statistically significant correlations between the two measures in three of the five OCEAN traits – Openness, Conscientiousness, and Extraversion.

Previous research has associated all three traits with both an increased number of promotions throughout a person’s career, and increases in their gross income.

Openness, conscientiousness and agreeableness are also associated with the number of promotions a worker gets within a particular organisation.

By confirming that those traits can be evaluated using a candidate’s social-media profiles, Hine told Information Age, the results “allow organisations to spot these diamonds in the rough rather than trying to search for those elusive senior developers.”

“Individuals have massive value locked away in their online profiles”, he said, “and we’re trying to educate individuals about that to help them prove ownership of their data – and to help organisations realise that they can get real business advantages if they do consider evaluating candidates using non-traditional data sources.”

Quantifying work experience

The idea for Reputationaire was born five years ago after a colleague of Hine’s migrated here from India, but had trouble securing a rental property because all of her references were Indian contacts who were hard to reach and evaluate.

Eventually, she secured accommodation by printing out Airbnb reviews that showed them to be a good house guest – leading Hine to realise the value of using social-media content as testament to workers’ value as an employee.

Heavy participation and strong engagement in a Github developer forum, for example, could prove that a developer is a good team player and productive coder in ways that a checklist of certifications could never do.

The goal of that work – which produced a provisional patent funded by RMIT University – was to build a verifiable, standardised representation of online activities “to help individuals overcome a lack of local references, work experience, or credit scores, especially when they’ve moved,” Hine said.

“Migrants and fresh graduates find it very difficult because they don’t have local references or work experience, and this is all about trying to prove their characteristics by giving an organisation peer-reviewed data to prove that this person is maybe worth a second look.”

The firm’s TechieRank provides such a metric for developers – although, Hine said, many people had questioned whether candidates’ broader online profiles really do reflect their real-world character.

The latest research, he added, confirms that they do.

That’s not always a positive for job seekers, many otherwise qualified candidates have found as employers investigate potential hires – one survey of 1000 hiring managers found 70 per cent admit doing so – and dig up old Instagram nightclub shots and late-night Twitter rants.

Fully 57 per cent of those looking online had found a reason not to hire a potential candidate due to content like provocative or inappropriate photos or videos (40 per cent), drinking or drug use (36 per cent), discriminatory comments about race, gender and religion (31 per cent), poor communication skills (27 per cent), and other red flags.

Little wonder a cottage industry has emerged as startups like Filtari, Scrubber, BrandYourself and Redact use AI to screen customers’ social-media accounts for potentially problematic content.

Many hiring managers also reported positive results from their searches, with 31 per cent saying the candidate’s posts provided a good feel for their personality, 28 per cent recognising the candidate’s communications skills, and 23 per cent noting “great references” from other people.

Such findings are corroborated by the latest research, Hine said, noting that “a strong relationship exists between online content and psychometric survey measures of personality… [and] confirms that online content can function as a useful estimator for personality.”