Most of these polls are not measuring the opinions of all adult Americans.
If they did, Drump's approval rating might be closer to 30 percent or so.
These polls are using methodologies to poll only a subset of Americans - "likely voters" is Rasmussen's method. That subset of Americans likely to vote in the next election.
A subset of Americans classified as "likely voters" is going to disproportionately represent conservative activists. Because conservatives - especially old white men - are most likely to vote on a consistent basis and also face the fewest barriers to voting.
I am not sure which poll is actually polling "all adults", but they would surely be consistently lower than Rasumussen