No, such data doesn't exist, because, from a methodological point of view, it would be very difficult to isolate the causal effect. On the other hand, there is documentation that proves the notion of electoral fatigue beyond a doubt.
It seems reasonable to me that some people would be confused about the date of the poll and the location of the polling station. My work with Élections Montréal confirms that too much overlap prevents certain things from happening. For example, if institutions like Élections Montréal wanted to run major advertising campaigns to remind people of the importance of voting in municipal elections, they would probably refrain from doing so, or would wait to do so until only municipal elections were underway, to make sure there was no confusion. There may be no direct effect, but many indirect effects can accumulate.