In the first part of my dissertation, I address the puzzle of how majority parties can create a legislative reputation that appeals to two very different constituencies -- primary and general election voters. I argue that the majority party leadership strategically sets the agenda to send different informational signals about party members prior to each election. The findings show that there is variability in the types of bills that are scheduled at different times during the year. This suggests that the majority party can send different informational signals about its members in order to appeal to either the primary or general election constituency. Furthermore, it suggests that if voters are myopic, they are not getting the entire picture of the legislator's behavior and voting patterns by only looking at recent votes, which could ultimately affect democracy in the long run.

In my second paper, I argue that party labels can act as information shortcuts to help voters more reliably guess the issue positions of a representative, even when voters have no direct information about the representative's voting record. I isolate the effect of an individual being "treated" with knowledge of a representative's party label as a heuristic using an instrumental variable design. I characterize each state's Senate delegation to be either mixed or unified, with mixed meaning one Republican Senator and one Democrat Senator and unified meaning two Senators from the same party. I find that knowing a Senator's party makes voters 19.5% more likely to correctly guess their position on major votes. This finding suggests that party labels can independently improve citizen inference about their representatives, even when other forms of knowledge are held relatively constant.

In the last part of my dissertation, I investigate the effect of varying the level of information conveyed by party brands using an experiment. I argue that cheaper information has not increased voters' knowledge about individual candidates because voters possess an even cheaper and increasingly informative cue: party id. I propose an experimental design to test how varying levels of information affect individuals' willingness to pay for additional information when the goal is to make a correct decision. I theorize that as the parties become increasingly distinguishable, individuals will be less willing to pay the cost of seeking out information on their own individual legislators. I test this hypothesis using a between-subjects design. Subjects are randomly selected to receive a party label or not. I systematically vary the informativeness of the distribution of votes that they receive according to real-world information level and measure their willingness to pay for extra information -- in this case, a signal that will help them make their final decision. I use willingness to pay as the main dependent variable and calculate the average treatment effect for each informational level (JSD). Over time, willingness to pay for information has decreased by 30% as the party label has become more informative.