Ear-EEG has emerged as a compact, low-energy alternative to conventional EEG, enabling brain activity monitoring in naturalistic settings. This technology shows potential for applications such as auditory attention decoding, sleep monitoring, epilepsy detection, and brain-computer interfaces. However, ear-EEG signals are characterized by low spatial resolution, high susceptibility to artifacts, and considerable interindividual variability, which challenge reliable signal interpretation. In this study, we investigated the feasibility of using Pearson’s linear correlation between the acoustic envelope and EEG signals as a simple, low-complexity method to discriminate attended from ignored auditory stimuli. While individual trials occasionally exhibited clear separation between attended and ignored conditions, global patterns were modest, with strong overlap between conditions and high variability across participants and trials. Comparisons with non-parametric metrics (such as Kendall and Spearman) revealed no consistent improvement over Pearson correlation, suggesting that more complex approaches may not necessarily lead to substantially better results. These findings align with previous studies, where methods like canonical correlation analysis (CCA) and multivariate Temporal Response Functions (mTRF) yielded similar accuracy levels. Although linear correlation can capture markers of auditory attention in certain cases, robust decoding of attention in ear-EEG requires methods capable of modeling temporal dynamics and integrating information across multiple channels. Despite these limitations, Pearson correlation remains a computationally efficient tool for preliminary analyses, particularly in scenarios with limited resources, providing a useful baseline for future methodological advancements.