We consider the problem of point and interval estimation for the risk ratio using double sampling with two-sample misclassified binary data. For such data, it is well-known that the actual data model is unidentifiable. To ...