created using Envato Reference checks have a validity of .26 for predicting job performance. That's barely better than flipping a coin (which would be .00). Yet 96% of organisations still conduct them. Here's why the paradox exists: → Selection bias (candidates choose friendly references) → Legal fear (references say little beyond dates and titles) → Memory limitations (recalling specific behaviours is hard) → Halo effects (overall impression drowns out specifics) → Due dilig