This paper considers new computer methodologies for assessing the impa
ct of different types of public health information. The example used p
ublic service announcements (PSAs) and mass media news to predict the
volume of attempts to call the CDC National AIDS Hotline from December
1992 through to the end of 1993. The analysis relied solely on data f
rom electronic databases. Newspaper stories and television news transc
ripts were obtained from the NEXIS electronic database and were scored
by machine for AIDS coverage. The PSA database was generated by compu
ter monitoring of advertising distributed by the Centers for Disease C
ontrol and Prevention (CDC) and by others. The volume of call attempts
was collected automatically by the public branch exchange (PBX) of th
e Hotline telephone system. The call attempts, the PSAs and the news s
tory data were related to each other using both a standard time series
method and the statistical model of ideodynamics. The analysis indica
ted that the only significant explanatory variable for the call attemp
ts was PSAs produced by the CDC. One possible explanation was that the
se commercials all included the Hotline telephone number while the oth
er information sources did not.