It's funny. I sat in front of my laptop working for most of the day and still didn't manage to get this written while I had a working wifi connection.
I need to look into signal boosters or something like that. This is starting to make me crazy.
anyway, I partook in a perfectly lovely meal with some lovely interfaith church goers at FOY today. Those ladies can cook, man. Well, imagine 100 grandparents in a room and you've got one hell of a potluck. Yummy. I had scalloped corn today that was so good I couldn't stand it. Fluffy textures appeal to me.
The sweet potatoes could make you cry, I swear.
Speaking of sweet potatoes, I was in a store last week and saw that they were selling sweet potatoes - 5 lb for a buck. A lady in line at lunch said she'd seen them at walmart for 18 cents a pound. I think I pay a couple of dollars a pound at home.
I mentioned above that I worked most of today - it's true. I started really getting into my tables - I'd done some talking about interactions to this point because I was avoiding working with the tables. It's not that they're hard - just time consuming.
In the course of things, I realized I hadn't calculated a couple of things, so I wrestled with the best way to do that. In my orginal excel data files (10 to fix, which is a pain) or do the calculations in my stats program (it required ATAN and square roots and so on and I had no idea how to do that).
I had the calculation formulae in another excel spreadsheet, so I went with the tedious 10 file fixing solution. Except - after I'd finished the 10 files I realized I'd done some editing inside the original data files to remove egregious outliers. More than a few std dev and I drop the outliers. ANYWAY, I have changed how I drop outliers - I do it inside my stats program so I don't have to fiddle fart around with original data sets and the originals stay pristine.
So I fixed those problems and in the course of doing that, realized that my formulas weren't calculating properly. They showed correctly but the values didn't make any sense.
I got fed up and decided that this was God's way of making me learn something new about my stats program, so I went back to all 10 files, deleted the calcuations within them and made the calculations inside the stats program. It's tidier AND I learned something new. Yay!
I'm glad I had to go back into the source data files though, because I had forgotten that I'd made those edits and had subsequently dropped more outliers from the already trimmed dataset. It's not like were talking about dozens of drops, maybe 4-5 from a dataset of 500 individuals, but still. The upshot was that when I dropped my outliers properly it changed some of my means (I wasn't dropping as many outliers as before, so that makes sense). The problem was that it appeared that my outlier drops had resulted in a pretty significant change - not terrible, not inexplicable but not as tidy as it had been orginally.
I'd been working so long that I thought I'd really screwed something up. I felt a real idiot, and wrestled with the dataset for ages until it was time to go watch a movie with some friends. I left the work behind me though, and watched the river wild with no thoughts of my data - how could you with that creepy Kevin Bacon in that movie? Blech. I came back and checked my stats output - I'd left a model running when I left and wanted to see what had happened.
The numbers were essentially the same - and coming back fresh was when I realized that the numbers I thought were wrong? Yeah, weren't wrong or different at all. I just can't read. I was so relieved that I hadn't been going around telling people the wrong result for the past 6 months I tried to explain my new realization to my mom.
My mom is not dummy, but a statistician she is not. Even if she didn't have any idea what I was talking about, she smiled and nodded in all the right places and sometimes that is all you need from a listener.
There's a couple lessons to be learned here. a) A well timed break can be worth its weight in gold - whether 2 hours or 2 weeks. b) Listen well enough to know what response it required.
Now I must go to bed - we are hitting up the early sales tomorrow (no 4am for me, though) and I should get a little sleep. I've printed my shopping lists - not that I will buy 4 pages of stuff, but there are things I want to look at if nothing else. I want to see what all the hullabaloo about black friday is about. By the time I load this I expect the shopping will be over with so I shall have to write about that and the stuff I said I would talk about yesterday. I'm not quite sleepy yet so I shall have to read my book for a while.
I am loving the Outlander series by Diana Gabaldon. In case I haven't mentioned that yet.