Calculating how many degrees of freedom I will have after I finish my doctorate

While I wait to hear if my dissertation manuscript has been approved, I am relearning statistics. It's either that or crossword puzzles. I am adept at running statistical tests in Excel—any trained monkey can do that, once it figures out that installing the Data Analysis Toolpack results in beaucoup bananas. I can compare the scores of two groups to see if perhaps their differences are due to the random chance we all face as we flit about our day, or due to the fact that they are in fact really different in some significant way. Like they root for rival soccer teams or something, I don't know, I'm just making this up. Click the button, whoosh, Excel performs its magic, and voilá, you have output! It's like a statistical meat grinder. Of course, like a meat grinder, what you get out of it depends a lot on what you put into it. I didn't collect this data, so I have to accept what I have. (Have you noticed that I've used two French words in one paragraph? Zut alors!)

Today I discovered the equation that calculates the degrees of freedom needed to conduct a t-test on two independent samples. (No, I don't mean I discovered the equation. I mean I figured out how to type it into Excel. I feel like I imagine Columbus felt when he discovered India, that is to say, like an ignoramus.)

There are so many ways to go with this topic, it's hard to pick just one. Like, are you wondering what degrees of freedom are? Tantalizing, isn't it? We like freedom, it's one of our national values, although it hasn't always been applied fairly, but still, we live and die for it, so it must mean something to us, freedom. Degrees of it sounds a little uncertain, but we can always use more freedom, right? Can there ever be too much freedom? Hmmm. Ask any kid who doesn't get a lot of parental attention. Maybe too much of a good thing, like eating ten maple bars when one or two would do? Food for thought.)

The degrees of freedom I'm talking about actually have to do with calculating a specific statistical test to see if two independently collected samples are significantly different from one another. Does that sound like a foreign language to you? Mais non, if you are a statistician, which I'm not. I love statistics, but no matter how many times I study statistics, I can barely grasp the concepts before they slip away. Like anything to do with numbers, statistical concepts just don't stick in my brain, and the older I get, the less they stick, along with phone numbers, birthdays, and what I had for lunch yesterday. It's like my brain is hardening into a slick marble ball. I look on the bright side: When I'm dead, the morgue attendants can go bowling. Miniature bowling. (Do they have such a thing?)

The statistical equation for calculating degrees of freedom in Excel, in case you were wondering, is this:

Where S1 and S2 represent the standard deviations of the two samples, respectively, and N1 and N2 are the sample sizes of the two samples. That's all you need, plus a buttload of parentheses in exactly the right places. (Oh, and don't type the DF part.) No sweat. I guarantee after you type this in successfully, you will feel a strange tingling sensation that can be interpreted as a frisson of freedom. A successful outcome, by the way, will be obvious when you generate a value that is just slightly smaller than the sum of the two sample sizes. Clear signs that you have erred would include a negative number or a ridiculously high number like, oh say, 16,345.2345 when your combined sample size is 40. Ha! Logic prevails, eventually.