This study aims to examine the relationships between linguistic features in online reviews and two well-known phenomena in digital consumer behavior: opinion polarization and the bandwagon effect. Opinion polarization is assessed through the skewness of review rating distributions, while the bandwagon effect is measured using lagged regression analysis to determine how earlier ratings influence later ones. The analysis uses approximately 50,000 customer reviews from 50 products and services collected from three major U.S.-based e-commerce platforms. By applying regression analysis, this study empirically explores the statistical connection between specific linguistic features in review texts and both polarization and the bandwagon effect. The results show that lexical density, a measure that combines lexical diversity and review length, has a statistically significant negative relationship with both polarization and the bandwagon effect. Meanwhile, readability was not significantly linked to polarization but was positively related to the bandwagon effect. Unlike previous research, this study emphasizes the role of linguistic features in contributing to information distortion within online review environments, providing a more systematic perspective. These findings offer empirical insights for businesses aiming to establish fair and trustworthy review systems.