For the seasons unit, I’ve done a fair amount of giving students data sets to graph, looking for patterns, similarities, difference. We have been doing so in order to build evidence for or against various claims about what could cause the seasons. I think we’ve learned a lot along the way.
Anyway, there are two observations that have driven student a fair amount of engagement, and I don’t want to forget them:
#1 McCurdo Station in Antarctica has the sun shining on it for 4 straight months, but its average temperature is still below freezing during that time. [If duration was only factor, then we’d expect McCrudo Station to be very hot]
#2 In June, Murfreesboro, TN is 20 degrees hotter than Quito, Ecuador. [Shouldn’t the equator always be hotter?]
Teach. Brian. Teach. 
I’ve been experimenting this year with when I give students data sets and when I expect them to measure and share their data sets. I don’t have a clear way to decide when one is more advantageous than the other. Any thoughts? It’s a real dilemma when measuring electrical quantities, since it’s difficult to even use the instruments correctly if you don’t have a solid grip on the ideas behind them. But it’s hard to get a grip on the ideas without getting hands-on with a circuit… still calibrating this.
By the way, thought you might enjoy this tidbit about the philosophy and nature of science, courtesy of Michael Doyle and Richard Feynman.