Over the past few decades, the computer has played an increasingly lar
ge role in the collection of survey data. The primary focus of compute
rs in survey research, however, has been in computer-assisted telephon
e interviewing. The use of computers in the elicitation of responses d
irectly from the respondent has been the focus of increasing efforts i
n recent years but still remains relatively undeveloped. Although ther
e have been previous attempts at investigating the effects of introduc
ing computerized self-administered surveys among adolescents, no such
investigation has been attempted in a school-based survey to our knowl
edge. The authors examine methods and issues from the Minnesota Studen
t Survey Mode Effects Experiment - an on-line versus paper-and-pencil
comparison from a large school survey of adolescents. Some of the issu
es include comparability with paper-and-pencil versions and data colle
ction issues, such as computer literacy, security, and case management
.