The core problem is that economics cannot use falsifiable experiments, and it is very highly politicized.
I agree that falsifiable experiments are difficult in economics, which is why we must rely instead on logic and reason---a strategy seemingly abandoned by modern economists.
I remember the ridiculous (and since discredited) argument by Card & Krueger that increases in the minimum wage increased employment, at least according to their data. And yet, when you raise prices on normal goods, the quantity demanded falls; if you force companies to pay employees more, they will hire less. If the data say otherwise, the data are either wrong or too noisy to be trustworthy. (Unfortunately, their paper wasn't merely an academic curiosity---it gave Bill Clinton and Congress political cover to raise the minimum wage in 1996.)
And now we hear economists fret that Americans, who have been under-saving and over-spending for years, may yet ruin the economy---by saving too much and spending too little. Someone needs to tell them that wagging your finger and saying "liquidity trap" doesn't magically give the emperor clothes.
N.B. I predict the rise of quantum decoherence theory will change this picture in fairly short order.
For what it's worth, the only economists who consistently avoid setting off my bullshit meter are the Austrians. And yet, their views border on lunatic fringe according to mainstream economists. In a world where those same economists set policy, that is an alarming combination.