Part of what makes science hard is that data, even the best, most reliable data, are often extremely difficult to interpret.

Interpreting results involves describing, categorizing, and organizing data. These are not necessarily straight-forward matters. Often too there are earlier key decisions about what data to collect and what to ignore. Not everything that seems relevant at first really is.

Then there are the stats. Almost any data set can be reasonably analyzed in more than one way. This can be very helpful. It can yield deeper insight into what they do, and do not, have to say. But there is a risk too. Along the way you may discover that you are more like to get a result that appeals to you in some ways, or less likely in other ways.

This item is one of the best presentations I’ve seen addressing these issues for a non-specialist audience. It includes an excellent do-it-yourself demo.