Behavior-Based Safety in a Coal Mine: The Relationship Between Observations, Participation, and Injuries Over a 14-Year Period

A safety team ran a behavior-based safety program in one coal mine for 14 straight years.

Workers watched each other, filled out short checklists, and turned them in. A steering committee tallied the cards and posted weekly charts.

The mine kept daily injury logs, so the team could track if hurts went up or down as participation rose or fell.

Injuries dropped a large share after employee observation cards hit about a large share of the crew each week.

When participation dipped below a large share, injury counts crept back up the next quarter.

Geller (2005) sketched the same playbook—workers, not bosses, spot risks—years earlier, but gave no numbers. Hagge’s 14-year graph fills that gap with hard injury counts.

Bachman et al. (1988) also used model-practice-feedback to teach parents, and skill jumps stayed big. Hagge shows the same training style keeps working when the topic is hard-hats, not head-colds.

Shih et al. (2012) looked unrelated—stroke patients hitting switches—but both studies push adult learners to emit more correct responses and both hit large gains, hinting that behavioral activation works across very different arenas.

If you consult in factories, plants, or warehouses, aim for a large share weekly participation in peer safety observations. Track cards turned in, not just training hours. Post the percent every Monday so crews see the goal. When the number slips, schedule a quick booster and watch injury logs for the next month—small dips can be fixed before hurts climb back up.