Public touch screen information systems or kiosks have become a popular way
of delivering consumer information and much of the user activity is automa
tically recorded in electronic log files. Fortunately for the information p
rofessional, log files are a significant source for understanding the infor
mation seeking behaviour of consumers., Surprisingly, however, there has be
en no discussion in the literature on what metrics are available for their
evaluation. This paper seeks to address this problem and examines the metri
cs that can be generated from touch screen kiosk logs. The research reporte
d here forms part of a Department of Health funded study, which is evaluati
ng the use and impact of 70 health kiosks located in medical and other loca
tions throughout the UK [1].
The way that kiosks function and the contents of their log files are descri
bed. Metrics are discussed in the context of the data generated from a touc
h screen health kiosk in Wakefield. Aggregate use, user numbers and descrip
tive statistics generated on age, gender and pages viewed are examined. An
examination of the distribution of time-based variables, such as session ti
me, found them to be highly skewed; in consequence, robust estimates were g
enerated and compared. A metric for depth of session was developed by group
ing users according to their patterns of use and a method for assessing kio
sk 'utilisation' is proposed.