Agreement between self-reported use and observed presence of pesticides and its relationship to the odds ratio.
Sieber-W; Williams-L; Wilkins-J
Joint Statistical Meetings, August 11-15, 2002, New York, New York. Alexandria, VA: American Statistical Association, 2000 Aug; :301534
Obtaining accurate information about the types of agents an individual may be exposed to is important in studying the association between disease and exposure. The agreement between two methods of obtaining exposure data and its relationship to the odds ratio was investigated. Information on the number and type of pesticides used on crops was collected using a questionnaire and compared to the presence of pesticides noted during site visits to 71 farms. Kappa statistics for agreement between self-reports and observation varied between 0 and 0.37, indicating poor to fair agreement. Odds ratios for the use of pesticide and presence of headache showed a positive relationship with the percent of concordant positive pairs (agent indicated both on questionnaire and observed). Odds ratios were reduced as additional positive responses from either the questionnaire or site visit were incorporated. Such results could be due to factors such as recall bias, misclassification in completing questionnaires, or differing pesticide application and storage practices on farms, and indicate the difficulty often met obtaining valid exposure data from a self-reported questionnaire.
Pesticides; Pesticides-and-agricultural-chemicals; Agriculture; Agricultural-chemicals; Agricultural-industry; Questionnaires; Occupational-exposure; Exposure-assessment; Statistical-analysis
National Institute for Occupational Safety and Health, 4676 Columbia Parkway, Cincinnati, Ohio, 45226
Abstract; Conference/Symposia Proceedings
Research Tools and Approaches: Surveillance Research Methods
Joint Statistical Meetings, August 11-15, 2002, New York, New York