We use a web-and-mail survey on attitudes towards and use of marijuana to demonstrate how a web-recruited cohort could be integrated into an address-based sample using a calibration-weighting procedure in the software language SUDAAN 11 (TM). A Holm-Bonferroni procedure is employed to test whether a pivotal assumption underlying the integration is supported by the data for individual survey items as well as for the survey as a whole. Delete-a-group jackknife weights for the integrated sample are then developed.
