Climate models differ in their responses to imposed forcings, such as increased greenhouse gas concentrations, due to different climate feedback strengths. Climate feedbacks in NCAR's Community Atmospheric Model (CAM) are separated into two components: the change in climate components in response to an imposed forcing and the "radiative kernel," the effect that climate changes have on the radiative budget at the top-of-the-atmosphere (TOA). This technique's usefulness depends on the linearity of the feedback processes. For the case of CO₂ doubling, the sum of the effects of individual clear-sky components (water vapor, temperature, and surface albedo) on the TOA clear-sky flux is similar to the clear-sky flux changes directly calculated by CAM. When monthly averages are used rather than values from every time step, the global average TOA shortwave change is underestimated by a quarter as a result of intra-month correlations of surface albedo with the radiative kernel. The TOA longwave flux changes do not depend on the averaging period. The zonal averages are within 10% of the model-calculated values, while the global average differs by only 2%. Cloud radiative forcing (ΔCRF) is often used as a measure of cloud feedback strength. The net effect of the clear-sky feedbacks on ΔCRF is -1.6 W m^-2, based on the kernel technique, while the total ΔCRF from CAM is -1.3 W m^-2, indicating that clear-sky feedbacks contribute significantly to ΔCRF and make it more negative. Assuming linearity of the ΔCRF contributions, these results indicate that the net cloud feedback in CAM is positive.