This study examines how solution exemplars that seekers provide in crowdsourcing ideation contests affect the scanning, shortlisting, and selection of ideas by solvers; these three ideation activities are essential layers in the Knowledge Reuse for Radical Innovation model. Specifically, we consider the role of confirmation bias in solvers' behaviors and their use of seeker exemplars in the ideation activities. We posit that solvers' use of exemplars is affected by the extent to which the exemplars are consistent with solvers' prior belief and by different considerations in different ideation activities. The results from a crowdsourcing ideation contest experiment largely support our theorizing, as we find that problem-related and problem-unrelated seeker exemplars affected different ideation activities differently. This research contributes insights into seeker involvements and solver behaviors in crowdsourcing ideation contests and offers practical implications for seekers.