The detrimental effects of item nonresponse on survey samples and question outcomes have been studied for many years. However, the difference between don't know answers and refusals—two kinds of item nonresponse—has not been examined empirically. This study used the theoretical constructs question sensitivity and cognitive effort to distinguish between don't knows and refusals and examined 242 questions from various national surveys. The results show that more sensitive questions get more refusals while questions that require more cognitive effort to be answered receive more don't knows. However, cognitive effort also correlates significantly with refusals. Thus, to decrease refusals, researchers should pay special attention to cognitive effort as well as sensitivity. To decrease don't knows, researchers should focus mostly on cognitive effort.