Measuring multimedia exposure of users based on psychophysiological measurementASLAN OGUZ, EVIN (Avtor) Košir, Andrej (Mentor) multimedia exposurepsychometric instrumentpsychophysiological signalsmultimedia communicationuser behaviorad exposureThe aim of this research was to measure short-term multimedia ad exposure and its impact on the consumer behaviour of young adults (18-24). The current state of the art lacks in providing a reusable instrument in the context of advertis­ing. Additionally, when estimating ad exposure from psychophysiological sensor signals, they miss out on providing a weak ground truth and attempt to indi­rectly predict ad exposure from these signals. We aim to overcome the various shortcomings of traditional methods by providing an instrument for multime­dia ad exposure measurements and to use the proposed instrument as a weak ground truth, while partially estimating user exposure from the psychophysio­logical sensor signals. We also aim to evaluate the contribution of each sensor by the proportion of that sensor's explained variance to the selected multimedia exposure aspect. We conducted an observational crowdsourcing study to develop a scale called the Multimedia Ad Exposure Scale (MMAES) - an instrument de­signed to measure the short-term effects of multimedia ad exposure in terms of engagement, psychological reactance, awareness and attitude, and purchase in­tention. The developed MMAES consists of four subscales with which each user is scored. It has good validity and reliability, and is internally consistent. We used MMAES in another observational study where users are measured with a set of psychophysiological sensors while watching the ads and answered scale questions after watching them. We found that 45 features of certain signals are significantly correlated with the scores of the MMAES subscales, after Bonferroni correction there are 16 such signal features significantly correlated. Basic modelling of the MMAES subscale scores from the sensor signals via ML classification algorithms can distinguish low- and high-scored users at a maximum of 65%. The signif­icance of this study is that it provides an instrument for measuring short-term advertising exposure. It also provides relevant sensors and signals to be used when estimating ad exposure from psychophysiological signals.20232023-06-16 12:55:00Doktorsko delo/naloga146947sl