First of all, minimum wage increases aren't binary. You can't divide states into "above" and "at or below", count them, and prove a point, because states that are higher above the federal minimum than others have to be weighted more heavily when determining the correlation. Your logical error is akin to if we tried to measure the correlation between ionizing radiation exposure on leukemia by simply recording if someone was exposed to any level of ionizing radiation above the mean instead of recording the actual level they were exposed to.

This is why we have statistical analysis -- to measure actual correlations. These do not prove or disprove causation, but they can be used as part of an inductive proof or refutation of a question. Your facile binary analysis cannot.

And, anyway, your numbers looked like bullshit to me, so I checked them, and they're indeed bullshit. Here's what a spreadsheet of current state minimum wage laws and current state unemployment rates looks like. The numbers on the right side show the rankings within the top 10, 20, etc. Note the complete absence of a pattern. Five of the top ten have minimum wages higher than the federal minimum, and five are lower. It continues with 10 of the top 20, 13 of the top 25, etc. I don't know what kind of numbers you were looking at, but they're just not even close to accurate, and, again, you can't just look at above vs. below, you have to look at how much above.