INTRODUCTION: In the classical approach, we assume that when designing the experiment in the way described in the literature, we can observe specific components of the event‑related potential (ERP) induced in specific areas of the brain with specific latencies. Using these standard methods of data analysis, i.e., looking for activity changes only in components commonly known from the literature, there is a risk of not noticing new, interesting effects. METHOD(S): In order to check if the data-driven approach gives the opportunity to verify the classical approach and whether it allows to better match the analysis, we compared it with the classical analysis, using data from emotional experiments. We investigated the electrophysiological correlates of execution of an ambiguous task under the influence of emotionality of words stored in working memory. RESULTS: The analysis of variance (ANOVA) classical analysis of ERP was compared with an exploratory approach using GFP (Global Field Power), calculated as spatial standard deviation. Analysis of the GFP curve was used to determine the time periods in which we performed a 4‑factor ANOVA with repeated measures. CONCLUSIONS: In the present case, we were able to find significant effects related to the valence and origin consistent with classical analysis while maintaining control of the statistical significance. Phenomena were shifted in the time domain and with a tilted pattern in the spatial distribution.