Most large-scale ongoing face-to-face surveys which began using pen and paper interviewing
(PAPI) face an eventual shift to computer-assisted personal interviewing (CAPI). In
preparation for such a shift in the Household, Income and Labour Dynamics in Australia
(HILDA) Survey, a trial of the CAPI collection mode was undertaken in the 2007 test
sample. This involved a split-sample test of 764 households, where interviewers rather than
households were randomly assigned to the PAPI or CAPI mode. This paper reports on the
findings of this split sample test, both in terms of the fieldwork operations and the quality of
the data collected. Apart from some concerns about the length of the interview, few
differences were identified in the data collected by the two modes. Where CAPI differed
from PAPI, it was generally in the direction thought to enhance data quality rather than
reduce it.