572 ABSTRACTS (ACE) AEP Vol. 13, No. 8 September 2003: 559–596 duration of pesticide exposure among either controls (Spearman R = 0.11, p = 0.30) or cases (R = -0.12, p = 0.36). There was no association between risk of prostate cancer and duration of pesti- cide exposure. However, after adjusting for age, body mass index, and other confounders, two individual OCs showed an elevated risk of prostate cancer (oxychlordane OR = 2.5, 95% CI = 1.0- 6.0 and PCB 180 OR = 2.4, 95% CI = 1.0-5.8) within the middle vs. the lowest concentration categories. CONCLUSION: The findings using serum levels of OCs did not correlate with those based on self-reported pesticide exposure. Self- reports are subject to recall, case response and misclassification biases. Serum levels avoid these types of biases and provide a more direct measure of dietary and occupational exposures to pesticides that are resistant to metabolism. doi: 10.1016/S1047-2797(03)00170-4 P037 FEAR OF TERRORISM IN NEW YORK ON THE BRINK OF WAR JA Boscarino, CR Figley, RE Adams, New York Academy of Medicine, New York, NY PURPOSE: To assess the prevalence and impact of terrorism fears in New York after the September 11 attacks. METHODS: We undertook two statewide telephone surveys of New York adults – one in September 2002 (N = 1,001) and one in March 2003 (N = 484). RESULTS: The September survey indicated that 46.1% (95% confidence interval [CI] = 42.7-49.5) of residents were fearful of future terrorist attacks. Fear was higher on all indicators among downstate residents (New York City and Long Island) compared with the rest of the state (p  0.001). A multivariate model pre- dicting higher overall fear indicated that down-state residents (odds ratio [OR = 1.9, p = 0.001), women (OR = 1.6, p = 0.012), those 45-64 years old (OR = 2.0, p = 0.034), African Americans (OR = 2.4, p = 0.003), Hispanics (OR = 2.6, p  0.001), those with high school educations or less (OR = 2.2, p = 0.002), those who would flee the area (OR = 2.1, p  0.001), were more likely to be fearful of attacks. However, compared with the March survey, overall fear decreased among New Yorkers (26.0% vs. 34.1%, p  0.05). Conversely, the number of residents reporting they made preparations for future attacks increased (14.7% vs. 6.6%, p  0.05). In addition, 61.0% (95% CI = 56.1-65.8) of New Yorkers indicated they would likely re- ceive the Smallpox vaccine, if available. Furthermore, while about 30% of New Yorkers indicated they would likely flee in the event of a biological or nuclear threat, 69.0% (95% CI = 64.2-73.4) re- ported they would have confidence in information provided by police or fire department officials in making evacuation decisions. While this confidence was lower for government information (54.9%, 95% CI = 44.7-54.8), trust in this information actually increased with the onset of the Iraq War (54.9 vs. 45.1, p  0.05). Finally, 37.9 (95% CI = 33.1-43.0) of New Yorkers reported they were aware of bio-terrorism information on the New York Health Department’s web site. CONCLUSION: Terrorism preparedness has increased among New Yorkers and fear of terrorism has decreased. The disparity in terrorism fears by gender, race, ethnicity, and education level, however, may be reason for concern. doi: 10.1016/S1047-2797(03)00171-6 P038 POTENTIAL SOURCES OF BIAS IN A STUDY OF INFERTILITY AND ENVIRONMENTAL EXPOSURES WR Carl, P-H Chyou, J Herr, DL Kempf, AR Greenlee, Marshfield Clinic Research Foundation, Marshfield, WI PURPOSE: To examine self-report and other data for potential systematic error that could impact results of the Fertility Risk Factor Study. METHODS: ZIP Code data were analyzed for all females enrolled in the study, 322 cases and 322 matched controls, to test for the presence of potential confounding in fertility status-environmental exposure relationships. Reliability of self-report was assessed for a sample of 64 case females, their matched controls, and respective partners of each (256 total subjects). Demographic and health data were abstracted from the medical chart and compared to survey responses. RESULTS: Female cases were more likely to live in towns with paper mills (adjusted O.R. 3.02; 95% C.I. 1.86, 4.91), reflecting the association of town with a larger population to generate cases and also with potential of exposure sources (paper mill chemicals) that may impact reproductive outcomes. In sampled groups, 37% of both case and control women underreported at least 1 of 14 unique items recorded in the medical chart, 10 being unique to or with potential to impact reproductive health. All case women reported 60 of 74 and control women reported 11 of 25 total diagnoses related to reproductive health. Male reproductive health information was documented for 48% of cases and 12% of controls. All case males reported 33 of 55 and control males reported 4 of 10 total reproductive health diagnoses. Concordance analyses were conducted when 75% of a sampled group had specific medical record entries. Agreement varied from 50-100%, depending on sampled group and item. CONCLUSION: In this study, living in a paper mill town is a significant confounder in the fertility status-environmental expo- sure relationship. Also, self-report data of conditions specific to reproductive health have potential to introduce bias in statistical models of infertility and environmental exposure. doi: 10.1016/S1047-2797(03)00172-8 P039 HAIR DYE USE: CHOOSING BEST QUESTIONS AND BEST METHOD USING CONFIDENCE, CONCORDANCE AND RELIABILITY SM Pinney, K Helzlsouer, A Correa, Min Yuan-I, Department of Environmental Health, University of Cincinnati College of Medicine Cincinna `ti, OH PURPOSE: Case-control studies have investigated a possible asso- ciation between cancer and personal use of hair dyes but informa- tion on patterns of use has been insufficient in detail for exposure characterization. Responses to detailed questionnaire items, designed to elicit detailed history of hair dye use, were evaluated for internal consistence and reliability, by method of administra- tion and self-reported degree of confidence in response. METHODS: Structured questions developed to elicit information about periods of use of hair coloring products, color, process (per- manent, non-permanent, highlighting, toning, etc.) and product type (permanent, semi-permanent, temporary), place of applica- tion, and frequency and duration of use. For selected questions,