To evaluate the threat of the environmental dissemination of antibiotic resistance associated with wastewater treatment plants (WWTPs), the removal efficiency of antibiotic resistance genes (ARGs) during wastewater treatment needs to be assessed. The sample collection strategy is one factor that is often overlooked in study design and most studies on ARGs in wastewater perform grab sampling. Here, we hypothesized that wastewater sampling (i.e., grab and composite sampling) influences the observed ARG concentrations and calculated removal rates across WWTPs. We compared the removal rates calculated based on the two different sampling methods for several genes, including some clinically relevant ARGs (blaNDM-1, blaOXA-1, MCR-1, MCR-5, MCR-10, and qnrA). We conducted summer and winter 24 h sampling campaigns where grab samples were collected every 2 h from the influent, secondary effluent, and final effluent. The snapshot removal rate of each target gene calculated based on the 12 grab samples fluctuated by 0.5–1.6 log in the winter and 0.9–2.7 log in the summer, indicating diurnal variation. Overall, for each target gene, the removal rates calculated based on 24 h composite samples were approximately equal to the median of the 12 snapshot removal rates. Our study confirms the importance of using composite samples to monitor ARGs in wastewater.