Abstract:
Training can improve safety in occupational settings. Pesticides represent an occupational hazard to a large number of workers in agriculture. Many of the affected workers in the Pacific Northwest come from or relate to Spanish-speaking, migrant communities. Literacy, language and cultural issues may affect their ability to fully engage in English-only pesticide safety training. This study evaluated tools and techniques to improve the training and assessment among mixed-literacy, mixed-language participants of Washington State Department of Agriculture (WSDA) and Washington State University (WSU) sponsored pesticide handler training. The study represents the second year of a collaborative effort by WDSA, WSU and the Pacific Northwest Agriculture Safety and Health Center (PNASH). The collaborators worked to systematically incorporate and measure the effect of fluorescent tracer on knowledge gained and reported behavioral intent within pesticide safety training. Specific aims were to:
Evaluate the effect of incorporating fluorescent tracers as training aids on pesticide safety knowledge and knowledge of appropriate pesticide safety behaviors among participants in WSDA/ WSU sponsored pesticide handler training events using curriculum-based tests. Compare numbers-based and faces-based icon tools to answer scaled responses among participants in WSDA/WSU sponsored pesticide handler training events. Measure perceived self-efficacy using a faces-based icon scale.The Fluorescent Tracer Project evaluated the impact of incorporating fluorescent tracers into pesticide safety training by assessing knowledge gain changes in reported behavioral intent. The study participants were trainees recruited during WSDA/WSU sponsored pesticide handler training. The participants were divided into two groups- ‘tracer’ and ‘non-tracer.’ The study participants complete pre and post-evaluations immediately before and after training. Also included in the evaluations were scaled response- from strongly agree to strongly disagree- questions intended to evaluate icon-based scales and measure reported self-efficacy. The scaled response questions evaluated the performance of a Wong-Baker Faces Pain Scale derived icon scale verses a numerical scale. The data were analyzed using non-parametric statistics and linear regression.
There were 104 participants in the fluorescent tracer study. The results of the study were mixed. No statistically significant differences were found between tracer and non-tracer groups on overall test performance. Performance on the post-evaluation was better for those with nine or more years of education verses those with eight or less (p < .03). The findings suggest the unintentional inclusion of sources of item bias in the test instrument.
The faces based icon scale did not perform significantly better than the numerical scale. Statistically significant linear relationships between education and improper responses (e.g. failing to respond or responding multiple responses for a single question) were identified in both the fluorescent tracer and icon evaluation studies. This suggests differences in test taking skills existed among study participants.
While the test development and administration did not identify significant differences between the tracer and non-tracer groups, it did add value to the trainers and the pesticide handler training. WSDA and WSU personnel were involved in the question development and validation process. The process of refining questions helped surface substantive differences between the instructors’ beliefs of what the training should emphasize. The back and forth exchange of ideas during the question development and validation improved the standardization of training. The test development and validation helped calibrate the training, trainers, and curriculum.
The study demonstrated the inherent difficulty of measuring knowledge gain and behavior change in mixed-language and mixed-literacy populations. It also illustrates the need for further research in training effectiveness and training design for audiences with diverse backgrounds. Further research is needed to validate the usefulness of icon-based tools to measure scaled responses.