In the event of a short-term, large-scale toxic chemical release to the atmosphere, shelter-in-place (SIP) may be used as an emergency response to protect public health. We modeled hypothetical releases using realistic, empirical parameters to explore how key factors influence SIP effectiveness for single-family dwellings in a residential district. Four classes of factors were evaluated in this case-study: (a) time scales associated with release duration, SIP implementation delay, and SIP termination; (b) building air-exchange rates, including air infiltration and ventilation; (c) the degree of sorption of toxic chemicals to indoor surfaces; and (d) the shape of the dose-response relationship for acute adverse health effects. Houses with lower air leakage are more effective shelters, and thus variability in the air leakage of dwellings is associated with varying degrees of SIP protection in a community. Sorption on indoor surfaces improves SIP effectiveness by lowering the peak indoor concentrations and reducing the amount of contamination in the indoor air. Nonlinear dose-response relationships imply substantial reduction in adverse health effects from lowering the peak exposure concentration. However, if the scenario is unfavorable for sheltering (e.g. sheltering in leaky houses for protection against a nonsorbing chemical with a linear dose-response), the community must implement SIP without delay and exit from shelter when it first becomes safe to do so. Otherwise, the community can be subjected to even greater risk than if they did not take shelter indoors.