Some problems of nonresponse and nonresponse adjustment in the survey of income and program participation