When making the conditional predictions economics allows us to make, economists use the assumption ceteris paribus, aka “all else held equal.” Non-economists (and even some economists!) interpret this to mean “this result will hold only if we assume the world is a vacuum.” This leads to many objections such as “minimum wage may cost jobs in theory, but this is the real world and businesses will still hire!” or “I agree with free trade in principle, but China manipulates their currency!” While these statements may be true, they are irrelevant to the question asked.
Rather than assuming a vacuum, ceteris paribus means “Looking at only the effect of this change.” In other words, we want to know how the change in something (minimum wage, opening trade, etc) will affect another thing. The fact there are other variables that may swamp the variable effects is wholly irrelevant.
Let me explain by way of example: A small town is considering passing minimum wage legislation. The town economist predicts it will result in the loss of 20 low-skilled jobs in the town. The town passes the legislation anyway. Shortly thereafter, Wal-Mart opens a store in the town and hires 50 low-skilled workers at the new minimum wage. Does this mean the economist’s prediction was wrong? Not at all! Wal-Mart’s decision to open the store in the town was done independent of the minimum wage legislation; it would have occurred regardless. The fact that Wal-Mart opened the store swamped any negative consequences of the hike doesn’t mean the effects wouldn’t have occurred (I’d be remiss as an economist to not point out there could be many unseen effects going on here not captured by the statistics).
This misunderstanding of economic analysis leads to many ineffective, and often counter-productive, policy arguments.