Three categories of evidence
The IRS evaluates: (1) Behavioral control — does the business control or have the right to control how the worker performs the work? (2) Financial control — does the worker have a meaningful investment, opportunity for profit/loss, ability to make services available to others? (3) Type of relationship — written contracts, employee-style benefits, permanency, integral to the business?
Common red flags
Worker has a regular schedule set by you. Worker can't hire substitutes. Worker uses your tools and equipment. Worker is reimbursed for expenses. Worker is integrated into your business operations. Engagement is open-ended, not project-based. Worker provides services to no other client. Each red flag is not dispositive — but several together suggest employee status.
Section 530 safe harbor
Section 530 of the Revenue Act of 1978 protects an employer from reclassification if (a) the worker was consistently treated as an independent contractor, (b) all required information returns (1099-NEC) were filed, and (c) there is a reasonable basis for the classification (industry practice, prior IRS audit, judicial precedent, professional advice).
Form SS-8
Either the business or the worker can request a formal IRS determination by filing Form SS-8. Takes months to receive. Not binding on third parties (state unemployment agencies, workers' comp boards) but persuasive evidence.
State complications
Many states use stricter tests than the IRS — California's ABC test, Massachusetts's ABC test, app-based-worker rules in several states. A worker may be a federal contractor and a state employee simultaneously. Coordinate carefully.
Common mistakes worth avoiding
The recurring mistakes filers make on this topic cluster in three patterns: (1) optimizing for current-year tax at the expense of multi-year tax, (2) treating the choice as binary when the IRS framework actually allows nuance (partial elections, hybrid methods, year-by-year reassessments), and (3) deferring the analysis until the return is due rather than running it during the year when the result can still influence behavior. A short annual review — even thirty minutes — catches all three failure modes and replaces vague intuition with documented reasoning.