The sampling size of the age groups is somewhat concerning. I don’t think 43/57 is likely to be accurate, though it’s apparently a toss up among the older folks...
If I understand the way polling works the use the results of the polls and the demographics of the responders and normalize the responses across the spectrum of likely voters.
Sometimes this normalization captures the true voter turnout but sometimes it doesn't.
The detailed poll results are not saying that x% will vote only that "of the people we polled x% were women over 45."