The dN/dS ratio, a measure of selection pressure on proteins, is often used to compare sequences from divergent species. However, it is frequently applied to sequences from a single population, where differences represent polymorphisms rather than fixed substitutions. This study shows that the dN/dS ratio is relatively insensitive to selection pressure within a population and does not follow a monotonic relationship with selection. As a result, the hallmark signature of positive selection (dN/dS > 1) is not observed within a population. Instead, strong positive selection may lead to dN/dS < 1. These findings challenge the interpretation of dN/dS in population-genetic samples and suggest that dN/dS values near 1 in microbial populations may reflect strong positive or moderate negative selection. The study also highlights that the relationship between selection and dN/dS differs significantly between divergent lineages and within a population. Theoretical and simulation analyses confirm that dN/dS is not a reliable indicator of selection pressure within a population. This has important implications for interpreting genetic variation in population samples, particularly in microbes and viruses. The results suggest that dN/dS values should not be used to infer selection pressures in population samples without considering the time-scale of the sequences being compared.The dN/dS ratio, a measure of selection pressure on proteins, is often used to compare sequences from divergent species. However, it is frequently applied to sequences from a single population, where differences represent polymorphisms rather than fixed substitutions. This study shows that the dN/dS ratio is relatively insensitive to selection pressure within a population and does not follow a monotonic relationship with selection. As a result, the hallmark signature of positive selection (dN/dS > 1) is not observed within a population. Instead, strong positive selection may lead to dN/dS < 1. These findings challenge the interpretation of dN/dS in population-genetic samples and suggest that dN/dS values near 1 in microbial populations may reflect strong positive or moderate negative selection. The study also highlights that the relationship between selection and dN/dS differs significantly between divergent lineages and within a population. Theoretical and simulation analyses confirm that dN/dS is not a reliable indicator of selection pressure within a population. This has important implications for interpreting genetic variation in population samples, particularly in microbes and viruses. The results suggest that dN/dS values should not be used to infer selection pressures in population samples without considering the time-scale of the sequences being compared.