|
|
|
|
|
|
|
The authors ask you:
This study used a between-participants design to assess discrimination and demonstration of safe and unsafe behavior. The participants were enrolled in an undergraduate psychology course and assigned in randomized-blocks of four to the following groups: no training, safe training, unsafe training, and a combination of safe and unsafe training. The training incorporated the following components: instructions, modeling, practice, and feedback. Immediately following training the participants viewed a video for an assessment of discrimination between safe and unsafe behavior. Following the discrimination assessment, the participants were asked to demonstrate the trained response. The data suggest that participants in the combination safe and unsafe training group showed greater discrimination and correct demonstration than the no training, safe only, and unsafe only training groups. The results advocate that training should focus on both safe and unsafe behavior in order for workers to appropriately learn and perform safe behavior.
Recent literature suggests that there may be a two-fold purpose to the behavioral observation process. Alvero and Austin (2004) found that persons who conduct safety observations (observers) are not only necessary for collecting data used for the feedback process and improvement of goals, but also that the act of conducting observations (of another person's safety performance) results in the observer improving his or her safety when engaging in the same responses. In other words, an observer will perform more safely as a result of conducting safety observations. Due to the novelty of this line of research, Alvero and Austin (2004) recommended that future research measure the accuracy of observers' scores to determine if a relation exists between accuracy of observations and safety performance. In other words, how can observation training affect observer's own safety performance?
There has been little research on how to achieve high rates of observer accuracy. Without empirical support, Krause (1997) suggested that observation training use the following components: model, rehearsal, feedback, and re-practice. Krause does not comment on whether both safe an unsafe behavior should be trained. In other words, to what extent does training of either or both safe and unsafe behavior affect discrimination and demonstration of the response. Animal research suggests that training should include stimuli correlated and uncorrelated with reinforcement (Jenkins & Harrison, 1960).
This study will assess the extent to which training of one or two stimulus classes affects discrimination and demonstration of two stimulus classes. The experiment will analyze the effects of training safe, unsafe, or safe and unsafe behavior on discrimination and demonstration of safe and unsafe behavior. Furthermore, this study will bridge animal literature with OBM (Organizational Behavior Management - the application of Behavior Analysis to a work environment) literature through assessment of discrimination and expanding upon Alvero & Austin (2004) by examining the relation between accuracy of observation and performance of safety behavior.
Two-hundred undergraduate Queens College students (25 per group) participated within a laboratory.
Independent Variables
The three independent variables were training of safe, unsafe, or both safe and unsafe lower leg behavior. Each independent variable used the following components: instructions, modeling, rehearsal, and feedback.
Safe was defined as "lower legs must be nearly perpendicular to floor." The safe pictures and 5-s clips in model and feedback showed the lower leg response between 77° and 100° according to a protractor.
Unsafe was defined as "lower legs must be angled substantially forward or backward." The unsafe pictures and 5-s clips in model and feedback showed the lower leg response below 77° and above 100° according to a protractor.
Dependent Variables
Discrimination. The assessment that followed training presented thirty 30-s clips of a confederate. The confederate performed the lower leg response safely in 15 of the 30 clips used in the assessment. The participants used a 30-s whole interval recording procedure during the assessment. The assessment was presented at the end of the PowerPoint slide show. The dependent variable was the number of correctly and incorrectly identified safe and unsafe 30-s clips in the assessment. The 30-s clips were randomly assigned to the order used in the assessment.
Demonstration. The participants engaged in the lower leg response following the discrimination assessment. The participants demonstrated where safe and unsafe met with the right leg in the forward and backward positions. The participant's lower leg behavior was recorded by a picture and then measured using a protractor. The dependent variable was the lower leg position quantified in degrees.
The order of production of the lower leg response was either order one (i.e., leg forward and then backward) or two (i.e., leg backward and then forward). The participants were instructed to engage in order one or two through a script read by an experimenter (see design and condition assignment for more detail on order assignment).
Design and Condition Assignment The study applied a between-participants design to assess four conditions for lower leg behavior: 1) no training; 2) training of safe behavior; 3) training of unsafe behavior; and 4) training of safe and unsafe behavior. Conditions 2-4 were comprised of the following components: instructions, modeling, rehearsal, and feedback. The participants were assigned to one of the four conditions using randomized blocks of four based on their chronological order of participation. Additionally, participants within each group were assigned to demonstration order one (i.e., leg forward and then backward) or two (i.e., leg backward and then forward) in randomized blocks of two.
The above graphs show data that were recently collected. Due to a time constraint a comprehensive statistical analysis has yet to be conducted. Visual display of the data show promising results for both the discrimination and demonstration variables. We plan to use signal detection theory d' to analyze the discrimination variable. The authors predict that a trimming of outliers will show greater differences between the safe and unsafe combination group and the other groups for the demonstration variable. Overall, the data for both the discrimination and demonstration variables clearly suggest that training should incorporate both stimulus classes, combination of safe and unsafe. Training in either safe only or unsafe only results in a substantial generalization to non-target stimuli. Participants without training show greater discrimination than the safe and unsafe only groups. Nevertheless, the safe and unsafe only groups show greater correct demonstration of the target response than the no training group. Again, the results show that a combination of safe and unsafe stimuli should be used in training.
Future research should examine a manipulation of the proportion of safe and unsafe stimuli used in training. Additionally, future research should examine variations of the sequence of stimulus classes and the necessity of stimuli mediums (picture and/or video).
Alvero, A. M., & Austin, J. (2004). The effects of conducting behavioral observations on the behaviorthe observer. Journal of Applied Behavior Analysis, 37, 457-468.
Krause, T. R. (1997). The behavior-based safety process (2nd edition). New York: Van Nostrand Reinhold.
Jenkins, H. M. & Harrison, R. H., (1960). Effect of discrimination training on auditory generalization. Journal of Experimental Psychology. 59 (4), 246-253.
The findings and conclusions in this poster are those of the author(s) and do not necessarily represent the views of the National Institute for Occupational Safety and Health. Citations to Web sites external to NIOSH do not constitute NIOSH endorsement of the sponsoring organizations or their programs or products. Furthermore, NIOSH is not responsible for the content of these Web sites.
View Comments on this Poster