Little or no social media
RemoveThe vendor says: “Insufficient data is treated as elevated risk and lowers the score.”
The six questions
- Fail1. Is it related to the job?
Choosing not to use social media says nothing about ability or reliability.
- Fail2. Does it stand in for who someone is?
Tracks family income and background (strong), gender (strong), age (weak). Rural and lower-income candidates are less active online.
- Pass3. Is it the right person?
No content is involved.
- Concern4. Did the candidate publish it, and would they expect an employer to use it this way?
The absence of content is the signal.
- Fail5. Does it work equally for everyone?
Penalises precisely the candidates the tool cannot see.
- Pass6. Is there a less intrusive way?
No less intrusive way to learn the same thing was identified.
If a signal stands in strongly or directly for a personal characteristic (question 2 fails), remove it.
None. A removed signal is not used under any condition.
Evidence
Where the Act applies, personal data may be processed only for a lawful purpose, with consent or for a listed legitimate use. Consent must be free, specific, informed, unconditional and unambiguous, limited to the data necessary for the purpose, and preceded by a notice.
- US Uniform Guidelines: the four-fifths ruleReference only
A selection rate for any group that is less than four-fifths (80%) of the rate for the group with the highest rate is generally regarded as evidence of adverse impact.
European regulators' guidance on employee and applicant data. Employers should not assume that a public profile may be processed simply because it is public; screening must be necessary and relevant to the job, and applicants must be told.