Observational learning and workplace safety: The effects of viewing the collective behavior of multiple social models on the use of personal protective equipment

Abstract

Introduction

The current project evaluated the effects of the collective behavior of multiple social models on the use of personal protective equipment (PPE).

Method

Prior to completing a simulated baggage-screening task, participants (N = 64) watched a scripted training video that included three confederate trainees. Participants were randomly assigned to one of four manipulations, where different proportions of confederates were shown putting on over-ear sound mufflers before starting the task (0, 1, 2, or 3). White noise played at 70 decibels in the test room, and PPE use was observed unobtrusively through a lab window at five time intervals.

Results

The mean intervals of PPE use generally increased as the number of positive social models increased (0 = 0.63, 1 = 0.50, 2 = 1.25, 3 = 3.06), and differences between groups were significant [χ² (3, N = 64) = 14.92, p < .01, η² = 0.24]. The results suggest that the aggregate prevalence of safety behavior within work groups may be an important determinant of initial PPE use by new employees.

Impact on Industry

Results suggest that new hires are likely to use PPE at a rate that is proportional to the collective PPE use observed among their peers. Safety leaders should regularly measure the collective level of PPE use at job sites and encourage majority usage through appropriate interventions such as increasing the availability or quality of PPE, training, or positive reinforcement for compliance.

Introduction

Occupational injuries are a significant social problem. In 2006 there were over 4 million occupational injuries and illnesses in the United States, with 1.2 million requiring time away from work (Bureau of Labor Statistics, 2007). Occupational injuries cause substantial human suffering and create a significant financial burden for society. The direct costs of compensated injuries in 2005 were $48.3 billion dollars (Liberty Mutual Research Institute for Safety, 2008), with additional indirect costs estimated at $120 billion annually (Occupational Safety & Health Administration [OSHA], 2008). Given the magnitude of the problem, it is important to understand the psychological factors that contribute to occupational injury and to design appropriate interventions.

In the field of professional safety, interventions are implemented according to a hierarchy of safety controls. The first priority, and most effective prevention method, is to completely remove hazards from the work environment through engineering or job design controls. When hazards cannot be completely eliminated, the next priorities are to engineer physical barriers or safeguards, encourage technical competence and hazard awareness through personnel selection and training methods, and to promote appropriate prevention behaviors, such as the use of personal protective equipment (PPE). Although hazard removal is the first priority, it has been estimated that human error is a contributing factor in 84% to 94% of industrial injury cases (Salminen & Tallberg, 1996). One of the most common errors is the failure to wear appropriate PPE. To illustrate, one third of all construction fatalities are falls, and 34% of fatal falls are due to a failure to wear PPE that was available at the relevant worksite at the time of the incident (National Institute for Occupational Safety & Health [NIOSH], 2000). An additional 13% of fatal falls are due to improper use of PPE. Therefore, large reductions in occupational injuries and fatalities may be achieved through increasing the prevalence of relatively simple discretionary safety behaviors, such as PPE use.

There is evidence that frequent safety feedback (e.g., Cooper, 1994, Guastello, 1993, Krispin and Hantula, 1996, Sulzer-Azaroff and Austin, 2000, Tuncel et al., 2006) and safety-related interactions with supervisors (e.g., Smith, Anger, & Uslan, 1978, Zohar and Luria, 2003) are associated with increases in discretionary safe behaviors at work, but relatively little attention has been paid to the social-cognitive effects of observing co-workers’ safe or at-risk behaviors (Bandura et al., 1961, Bandura, 1986). The current paper explores the relationship between social modeling and relatively simple motor behaviors. However, observational learning is also relevant to the development of cognitive skills, rules for generative and innovative behavior, and efficacy beliefs across a variety of social domains (Bandura, 1997). There are clues that observational processes are important in the occupational safety domain. For example, laboratory and field studies show that both serving as a safety observer and being observed increase an individual's subsequent safe behavior (e.g., Alvero and Austin, 2004, Sasson et al., 2007). Moreover, the work group is likely to function as a powerful modeling “association network” for individuals (Bandura, 1997, p. 93) because peers provide a primary source of social support and solidarity in the workplace. Given the importance of relationships with peers for success and enjoyment at work, differences between personal behavior and group norms would be likely to function as motivating operations for safe behaviors and their associated consequences (Laraway et al., 2003, Olson et al., 2001).

Social motivational variables may be especially salient for new hires as they observe coworkers to view examples of appropriate behavior during acculturation. In such situations, the collective level of PPE use among coworkers would indicate behavioral norms, predict social consequences for imitation or deviation, and provide symbolic evidence of organizational safety priorities. The learning experiences reported by nursing students during acculturation illustrate the importance of social modeling during acculturation. Barrett and Randle (2008) conducted qualitative interviews with 10 nursing students about their hand hygiene practices during internships, and an emergent theme was a reported tendency to observe and try to “fit in” with group norms. For example, one student reported, “It's amazing how much you do copy what you see to try and fit in, especially if it's your first placement you don't want to upset anyone. You mix in with the crowd and do what they're doing” (p. 1854). Students also commented about being nervous about negative social consequences for speaking up about observed violations of recommended hand hygiene practices, and on how they picked up habits based on what they saw others do.

A few studies have directly evaluated the effects of social modeling on PPE use, with results showing that the probability of safety behavior generally increases after viewing a positive behavioral model (deTurck et al., 1994, deTurck et al., 1999, Racicot and Wogalter, 1995, Wogalter et al., 1998, Young and deTurck, 1995). For example, deTurck et al. (1994) constructed a laboratory scenario where participants were asked to test a kitchen cleaning product for a manufacturer. Student participants were exposed to one of four conditions prior to cleaning an oven (positive model, negative model plus a negative consequence, negative model, and control). PPE use was 95% for the group that observed a positive social model, 56% and 36% for the groups that observed negative social models (with and without the confederate experiencing a pretended “chemical burn,” respectively), and 78% for participants in the control condition. deTurck et al. (1999) conducted two more experiments with similar procedures and results, with positive model conditions resulting in high PPE use relative to negative model conditions, and control groups showing levels matching or approaching effects of positive model conditions. Using a different experimental approach, Racicot and Wogalter (1995) showed participants one of three instructional videos for a chemistry task prior to task completion. All videos showed a warning sign indicating that PPE was recommended due to possible skin and lung irritation, but two videos also showed a person completing the task while wearing PPE. Participants who viewed the videos with a positive social model wore PPE 92% to 100% of the time when completing the chemistry task, while the instruction-only video resulted in just 50% PPE use.

Prior safety social modeling experiments related to PPE use have produced relatively large effects, but all have employed single confederate models instead of groups of models, as would be common in workplaces. Although the characteristics of individual social models are important factors influencing imitation (Bandura, 1986, Bandura, 1997), the collective behavior of individuals within a group is likely to function as a special type of dynamic stimulus event, where imitation is most likely if a critical threshold of collective behavior is reached (Conradt & Roper, 2003). In addition to limitations in modeling conditions, the relatively high rates of PPE use observed in prior control groups suggest that experimental methods may have created experimenter demands for safety behaviors (Orne, 1962). In some studies with an explicit safety focus, such as Racicot and Wogalter (1995), effects could also be partially attributed to the salience of instructions rather than the presence or absence of social models. In other words, instructional videos with models may simply garner more attention than those without.

In order to address the gaps in the literature we designed an experiment that (a) manipulated aggregate PPE use within a group of social models, and (b) controlled for experimenter demands to engage in safety behaviors. Participants worked at a simulated work task in a testing room with an ambiguous noise hazard. Prior to the testing session, each participant was exposed to an instructional video for the task where three confederate trainees modeled use of over-ear sound mufflers. The safety focus of the study was hidden from participants, and sound mufflers were never referenced in scripted communications. Our hypothesis was that the use of PPE in the sample would increase as the proportion of social models wearing PPE in the training videos increased.