Reliability of EEG Metrics for Assessing Video Advertisements

Abstract

Electroencephalography (EEG) is well suited to examine the effectiveness of television (TV) commercials because of its high temporal resolution. EEG metrics have been linked to market-level success of advertisements (ads) and to distinctive psychological processes relevant to ad testing research. However, evidence for reliability of these metrics is scarce. We assessed the reliability of six EEG metrics (alpha, beta, gamma, theta, alpha-asymmetry, intersubject correlation [ISC]) commonly used in advertising research in relation to sample size and repetitive ad viewing. We analyzed EEG data from 116 participants viewing 13 TV commercials that promoted various brands. For each metric, reliability was examined by computing the extent to which commercials scored similarly across independent samples of participants. We also assessed the reliability of the temporal pattern of peaks and troughs of each metric within a commercial. Results indicated that the reliability of EEG metrics differed markedly and ranged from poor (alpha-asymmetry) to good (alpha, beta) to excellent (ISC), while repeated viewings of ads improved reliability. Sample sizes between 30 and 40 were required for most metrics to reliably assess overall performance of an ad, whereas more participants were needed to reliably track the temporal pattern. Guidelines have been provided for the neuromarketing industry.