Ryan Avent is the source of the conjecture that economics is a science.
Read his post.
I tried to be polite at his site and wrote "I think you are totally absolutely wrong" but I honestly think that he is psychotic. I don't know how any sane person could write what he wrote. The only explanation that I can think of is that he has decided to evaluate the economics profession without reading the top ranked journals.
That is I think he only looks at economic research which he considers scientific and then asserts that it is typical of economic research in general out of lazy ignorance.
The rest of my comment.
I think that your claim that economics is based on data is nonsense. For one thing, your claim that economics is one field is nonsense.
Yes there are many economists who test theories with data. Many of them, by the way, perform experiments. Others look for natural experiments. Many change their views because of the data. Many admit they were wrong when their predictions are contradicted by the data.
But that is only one part of the profession. There are also economists who consider economics a branch of mathematics. They do not speculate as to whether their economic results correspond to the real world. That's not the point.
You can look at top economics journals and find many articles which make no reference to data. They are contributions to a theoretical literature whose origin is not based on empirical success. This is simply a fact. I have mystified economists by suggesting that theories should be testable.
Then there is a whole huge literature based on drawing testable implications from a model, testing them, rejecting them and sticking with the model. This is always a problem -- it is impossible to test a hypothesis of interest without auxiliary hypotheses. But in this case, the false model is used and its implications are used to advise policy makers. That is the whole model (with auxiliary hypotheses) which has been rejected by the data.
Rejection of a model leads many (probably most) economists to say that models are false by definition and that a false model can nonetheless be useful. What economic hypotheses have been abandoned because they did not fit the data ? I'd like a complete list. It won't be long. I claim it is empty.
Notice I have avoided the topic of fresh water macroeconomists. They are another breed apart. Their views are much much further from those of Krugman than the average news consumer would ever guess. Their views are so extreme that it is not possible for non economists to believe that they mean what they say (and they rarely say it in public).