Is economics a science?
I understand the word science lends some credibility to its subject. Supporters of phony baloney theories and medicine often paper over bad/missing research with the word science and vapid non-peer-reviewed articles. But I don't think that's comparable to economics broadly speaking, so I'm going to bite anyway.
Of course it depends on how you define science. If science is defined as forward looking, repeatable, and fully controlled experiments, then no, I don't think economics is a science by and in large. But neither is psychology, sociology/anthropology, history, or literally anything dealing with humans or uncontrollable events. And that's fine, we usually call these fields social science.
Fundamentally, I think of the field is misunderstood. People read Krugman or libertarian ramblings or watch CNBC and think that's what economists do. There is more to econ than broad macro stuff - philosophical or otherwise, and a lot more diversity of methods & interests than most people think. Personally, I think economics is more of a descriptive behavioral science than a prescriptive or prognosticating one.
One criticism I see is about controlled experiments. Some economists just use modeling or data mining, but plenty use controlled experiments. For example, Esther Duflo, and most of the MIT anti-poverty lab, has been running real non-contrived experiments on reducing poverty. Vernon Smith & others pioneered an entire branch of experimental economics. Dan Ariely has become somewhat famous for using experimental economics in behavioral econ (for example). Like any other study of human behavior, it is difficult to ethically run experiments and have control, but sometimes you can do it.
I've seen criticism because of the different schools of economics and the disagreements among those schools. There is disagreement only in the broadest contexts. First, not all economists are macro economists. We have lots of micro fundamentals fairly well figured out - all the econ 101 stuff about prices, s&d, elasticities, etc. We're starting understand the way people make specific decisions - both in aggregate and individually. Second, on the macro stuff, economists nearly always agree about what effect results from most changes, they just disagree about which effects predominate. Disagreement in a field doesn't mean no one knows what they're talking about, it characterizes healthy debate and critical thought.
I've read criticism about not having accurate testable predictions. The word accurate is a bit loaded, every field of study learns by making good & bad predictions. Of course you can make predictions which are testable, but in many cases (short of the controlled experiments above) you'll lack a counterfactual. Good economists know the limits of their models, so doing experiments or forecasting, they always have a slew of assumptions and limitations. Economists essentially test hypotheses for a living, those with untestable assertions about the world are more like philosophers or pundits than economists.
The more broadly you try to predict - e.g. for the entire economy - the less accurate you're going to be, all else equal. Economists working in micro areas, say someone from visa predicting HD TV sales for Christmas, can be quite accurate. Even in macro areas economists have had some pretty remarkable successes, like the brazilians taming inflation with conversion to the real, taiwan & singapore becoming major industrialized economies despite having virtually no natural resources.