9See http://www.bls.gov/opub/hom/pdf/homch16.pdf for a description of the changes associated with the introduction of CAPI interviewing. Cho and Pickering (2006) analyze whether CAPI interviewing improved data quality. First, the amount of spending in (ex ante thought to be) under-reported categories rises with CAPI and the amount of spending in (ex ante thought to be) over-reported categories falls, although the net effect is an overall fall. At the same time, CAPI increases the number of items reported – higher counts (but not higher expenditures). Finally, there are more high-spending levels for households.