Knowledge of how mosquitoes respond to insecticides is of paramount importance in understanding how an insecticide functions to prevent disease transmission. A suite of laboratory assays was used to quantitatively characterize mosquito responses to toxic, contact irritant, and non-contact spatial repellent actions of standard insecticides. Highly replicated tests of these compounds over a range of concentrations proved that all were toxic, some were contact irritants, and even fewer were non-contact repellents. Of many chemicals tested, three were selected for testing in experimental huts to confirm that chemical actions documented in laboratory tests are also expressed in the field. The laboratory tests showed the primary action of DDT is repellent, alphacypermethrin is irritant, and dieldrin is only toxic. These tests were followed with hut studies in Thailand against marked-released populations. DDT exhibited a highly protective level of repellency that kept mosquitoes outside of huts. Alphacypermethrin did not keep mosquitoes out, but its strong irritant action caused them to prematurely exit the treated house. Dieldrin was highly toxic but showed no irritant or repellent action. Based on the combination of laboratory and confirmatory field data, we propose a new paradigm for classifying chemicals used for vector control according to how the chemicals actually function to prevent disease transmission inside houses. The new classification scheme will characterize chemicals on the basis of spatial repellent, contact irritant and toxic actions.