In spite of numerous health efforts to reduce cancer prevalence and mortality, current estimates place this condition as the second most deadly disease in the U.S., falling only behind heart disease. To assist health professionals in their attempt to alleviate the burden associated with cancer, this study examines communication, demographic, rational and affective factors that predict involvement in cancer-related behaviors among U.S. adults. In addition, this study explores gender differences among the proposed relationships and tests the applicability of risk-as-feeling hypothesis to cancer-related phenomena. The study uses the Health Information National Trends Survey data set collected in 2005. The data contains measures related to cancer knowledge, attitudes and behaviors, access to and use of communication channels, risk perceptions, worry and demographic characteristics. A path analysis is employed to test the relationships identified by three hypotheses. These analyses are specific to certain cancers including lung, colon and skin cancers. In addition, these analyses feature a multiple group comparison technique (males versus females). The main findings include no support for a gender effect on the proposed relationships. Even though several gender effects are encountered for some relationships, the patterns are not common across cancers or predictors. In addition, the results suggest partial support for the risk-as-feeling hypothesis. The hypothesized impact of emotions on reason and the influence of affect on cancer-related behaviors are supported by some of these analyses. Finally, based on the results, this study is able to advance suggestions for the development of a single model across various types of cancers. In addition, several individual results are of noticeable importance. The control variables, age, race and education are consistent predictors of knowledge and online health information seeking. Worry consistently predicts risk perceptions, regardless of the cancer type. Moreover, worry has the highest coefficient among all predictors of risk perceptions. Almost a third of the variance in cancer-related knowledge, online health information seeking behavior and risk perceptions is attributed to the proposed predictors and the control variables. Predictors, while significant, account for a small portion of the variance in cancer worry or cancer related behaviors