What the Evidence Says About Safety Interventions — And Why Most Fall Short

---

In Brief

• Behavior-based safety programs produce meaningful results under specific conditions — participative goal-setting, sustained management engagement, attainable targets — and consistently fail when those conditions are absent.
• Theory-informed educational interventions produce more durable behavioral change than generic training, particularly when the theoretical framework is matched to the specific barriers operating in the target population.
• The most effective program in the 39-study review combined peer-to-peer feedback, joint labor-management barrier removal, and leadership development — producing approximately 85% reduction in at-risk behaviors.
• Most safety programs target individual worker behavior while leaving unchanged the organizational conditions — poor safety climate, production pressure, peer norms — that make unsafe behavior likely in the first place.

In 1995, researchers walked onto a construction site in the United Kingdom with a clipboard and a simple proposition: what if workers collaboratively set their own safety targets and received regular, visible feedback on how they were progressing? The study showed meaningful improvements in scaffolding safety, housekeeping standards, and safe access to heights (Marsh et al., 1995). And then, to test the mechanism rather than simply measure outcomes, the researchers removed the feedback component. Behavior deteriorated within weeks.

The lesson was not merely that goal-setting works. It was that it works only as long as it keeps working — and that the organizational conditions sustaining a behavior change program are not separable from the program itself. Remove the conditions, and you remove the effect. That finding, now three decades old, prefigures most of what the evidence on safety interventions has continued to show since.

---

About This Research

This post draws on Safety Behaviors and Behavioral Science Approaches in High-Hazard Industries: A Systematic Review Report by Dr Khaleda Ahmadyar (April 2026). Conducted in accordance with PRISMA 2020 guidelines, the review synthesizes 39 peer-reviewed studies — including 11 intervention studies — across more than 20 countries and six high-hazard industries: construction, mining, manufacturing, agriculture, oil and gas, and transportation. It is the only cross-industry synthesis that evaluates behavioral science intervention effectiveness in occupational safety contexts.

Download the full report free →

---

What Has Actually Been Tested — and What Three Approaches Tell Us

The 39-study review included 11 intervention studies evaluating structured behavior change programs across construction, manufacturing, agriculture, and transportation. Three broad approaches appear in the evidence: behavior-based safety (BBS) programs, theory-informed educational interventions, and nudge-based approaches. Each has something to teach us, and the pattern of their success and failure is not random.

---

Does Behavior-Based Safety Work? What the Evidence Actually Shows

Behavior-based safety programs (BBS) use systematic observation, structured feedback, and participative goal-setting to reinforce safe behavior. They are the dominant intervention type in occupational safety, and their track record is mixed in ways that are instructive rather than merely discouraging.

The 1995 UK construction study described above works precisely because the conditions were right: workers set their own targets collaboratively, feedback was regular and posted visibly, and management was actively and demonstrably engaged throughout. A 2023 Iranian construction study added a further refinement: rather than delivering a standard BBS program, researchers first assessed workers’ weakest psychological factors using baseline questionnaire data, then tailored the training and normative modeling to address those specific deficits. Unsafe behavior dropped from 49.5 percent to 33.7 percent, and those effects were maintained at three-week follow-up after supervision was withdrawn (Mohajeri et al., 2023). The diagnostic design — assess first, then intervene — made the difference.

BBS programs fail consistently when those conditions are absent. A Singapore construction study found no statistically significant overall improvement in unsafe behavior despite implementing a full program; qualitative data revealed that production pressure, peer norms, and workers’ perception that monetary incentives were unattainable counteracted the program’s reinforcing loops (Guo et al., 2018). A second Iranian study found that safety harness use did not improve because workers lacked adequate anchor points — a structural constraint that no motivational program could address (Mohajeri et al., 2023). Behaviors driven by structural constraints show less improvement than those driven by motivation or habit. The intervention could reach the individual but not the system around them.

Which intervention type fits your context and industry?

The systematic review maps program effectiveness against industry context, organizational conditions, and the behaviors targeted — giving HSE teams the evidence to make informed decisions before investing.

Download the report →

---

Why Theory-Informed Educational Programs Outperform Generic Training

Theory-informed educational programs produced more consistent results than generic training across the 11 studies, and the pattern of success is traceable to the theoretical frameworks applied.

The strongest behavioral effects in the entire dataset came from an Iranian agricultural intervention designed around the Health Action Model — a framework that integrates motivational, normative, and environmental systems together rather than targeting cognitive knowledge alone. Safety behavior scores among pesticide handlers nearly doubled post-intervention and remained significantly elevated at six-month follow-up compared to an active control group (Nasrabadi et al., 2025). The active comparator design — comparing against standard education rather than no intervention — gives this study the most credible evidence base in the review.

Motivational interviewing outperformed lecture-based training in a glass manufacturing context at 12-week follow-up — not because it conveyed more information, but because it engaged a different mechanism: rather than telling workers why safety mattered, it helped them articulate why it mattered to them (Navidian et al., 2015). The distinction between information transfer and motivation activation is not theoretical. It is measurable in observed behavior.

A consistent limitation appeared across all three educational interventions: peer and group norms were notably resistant to individually targeted programs. Communication reaching workers one at a time does not shift what the group does together. That requires engaging the social environment, not bypassing it.

---

The Most Effective Program in the Dataset — and What Made It Work

The strongest results in the entire review came from a multifaceted safety culture program in the US freight rail sector combining peer-to-peer behavioral feedback, joint labor-management barrier removal, and safety leadership development (Zuschlag et al., 2016). The program produced approximately 85 percent reduction in observed at-risk behaviors. A dose-response relationship across implementation sites — stronger implementation yielding greater reductions in objectively verified safety incidents — provided compelling evidence for a genuine causal effect.

What distinguished this program from weaker ones was not primarily the observation component. It was the integration of structural barrier removal through a continuous improvement mechanism, and the sustained, active involvement of management throughout. It treated behavior change as a system-level challenge rather than an individual one. And management was not a passive sponsor — it was a participant, and its engagement was the condition on which effectiveness depended.

Why Most Safety Programs Fall Short: The Fundamental Gap

The most consequential finding from comparing the observational and intervention evidence in the 39-study review is the degree of mismatch between what drives safety behavior and what most programs target.

The observational literature — detailed in the barriers post and the drivers post in this series — consistently identifies safety climate, leadership quality, peer norms, and psychosocial working conditions as the most powerful determinants of safe behavior. The intervention literature is dominated by programs that operate at the level of individual worker behavior and largely treat the organizational and social environment as a fixed backdrop rather than a target for change.

Placing the burden of safety performance on individual workers — those with the least power to alter the conditions generating risk — while leaving organizational climate largely untouched, is both inefficient and, as the data shows, often ineffective. The most effective programs treated the conditions themselves as the target. The behavioral science frameworks explored in the previous post offer the diagnostic structure to identify which conditions matter most in a specific context. The systematic review provides the evidence base to act on that diagnosis with confidence.

Frequently Asked Questions

What is behavior-based safety (BBS) and does the evidence support it?

Behavior-based safety is an intervention approach that uses systematic observation of worker behavior, structured feedback, and participative goal-setting to reinforce safe practices. The evidence for BBS is mixed: programs produce meaningful improvements when goal-setting is genuinely participative, feedback is regular and visible, and management is actively engaged throughout. They consistently fail when production pressure counteracts the motivational mechanisms, when incentives are perceived as unattainable, or when the underlying cause of unsafe behavior is structural rather than motivational.

Which type of safety intervention produces the most durable behavioral change?

The evidence points to theory-informed programs that match their approach to the specific barriers operating in the target population. The most durable behavioral effects in the 39-study review came from a Health Action Model-based educational intervention that produced lasting changes at six-month follow-up, and from a multifaceted safety culture program combining peer feedback, barrier removal, and leadership development. Generic training programs — those not grounded in a diagnostic assessment of which mechanisms are driving behavior — consistently produced less sustained results.

Why do most occupational safety training programs fail to produce lasting behavior change?

Most training programs operate on the assumption that knowledge deficits are the primary barrier to safe behavior. The research shows this is rarely the case. The more powerful barriers — poor safety climate, peer norms, production pressure, psychological strain, and structural constraints — are not addressed by knowledge transfer alone. Additionally, most programs target individual workers in isolation rather than the social and organizational conditions that shape behavior at the group and system level. Lasting change requires both: individual capability and the organizational environment that supports using it.