To reduce nonresponse bias in surveys, it has been suggested that researchers allocate additional fieldwork efforts to cases with low estimated response propensity. If these efforts are successful, nonresponse bias may be reduced by changing the variance of response propensities and hence the covariance between response propensities and variables of interest. The present study provides experimental evidence about how to intervene on low-propensity cases through interviewer selection and allocation in telephone surveys. The results suggest that this intervention can successfully increase cooperation rates for the low-propensity group and reduce nonresponse bias without increasing field costs.