Debilitating test anxiety is a general threat to validity if it biases assessment scores. Moreover, if bias differs between demographic groups, anxiety also raises concerns about test fairness. This study applied structural equation modeling to investigate possible measurement bias due to anxiety on the ACT® assessment and relationships among contextual factors, anxiety, and estimated proficiency. Results indicated that ACT scores were not biased by anxiety--in general and for female, male, minority, and White examinees. Anxiety was primarily related to examinees' feelings of being unprepared for the test. Health factors such as hours of sleep and eating breakfast were related to reduced anxiety and greater test performance. Results support practical guidance for students, parents, and counselors who want to minimize the emotional and physiological symptoms of test anxiety, even if it is not likely to affect test performance.