1. In linear regression, homoscedasticity assumes that the variability of the residuals is constant across all levels of the independent variable.
2. Violations of homoscedasticity can lead to incorrect inferences and biased estimates in statistical analysis.