Standard errors in mixture studies of tuberculous infection prevalence

Tuberculous infection status in individuals assessed by skin injections of 5 TU PPD-S tuberculin are subject to error. Some truly infected persons may respond with a small skin induration, and some not infected with Mycobacterium tuberculosis can respond with a large induration. Rust and Thomas used data from 701,317 white US Navy recruits seen during 1961-1968 to describe induration size distributions among those infected and not infected, and to provide prevalence estimates. Here, standard errors of prevalence and induration size probabilities are presented through use of the jack-knife methodology.