Results from a longitudinal survey of sexual behaviour and HIV-relevan
t cognitions amongst 258 sexually-active adolescents are reported. Dem
ographic characteristics, previous sexual experience, prior condom use
, beliefs specified by the health belief model (I-IBM), peer norms reg
arding condom use and condom use intentions were measured using a conf
idential postal questionnaire. Measures of sexual behaviour and condom
use consistency were included in a follow-up questionnaire one year l
ater. A discriminant analysis revealed that demographic and HEM measur
es did not predict reported sexual activity over the study year. Path
analysis revealed that intentions and HEM measures did not account for
significant proportions of variance in consistency of condom use or m
ediate the effects of prior sexual experience or demographic measures;
Those reporting more frequent sexual intercourse were less likely to
use condoms consistently while those who had used condoms previously r
eported more consistent use. Gender differences suggested that young w
omen's good intentions were less likely to be translated into subseque
nt, consistent condom use. Theoretical and applied in;plications are c
onsidered.