Journal 10-20-2010
Today I went to talk to the Bio statistician about the data I have collected. A couple of days ago he had given me a formula to determine the minimum sample size needed to achieve statistical significance. I had had some trouble with the formula because I didn’t understand what he wanted for a variable. Today I went in to talk to him in the morning and had a hurried confusing conversation with him. He had me write down one row of averages from one of my groups. Then he told me he would calculate the standard deviation for it. This is not what I needed help with at all. But he was in a hurry to make it to a meeting so I didn’t have time to clear up the misunderstanding. I have all my data in Excel and can type in STDEV( … and get the standard deviation in less than a minute. After having this conversation I went and sat at john’s physiotherapy padded table “desk” and worked out the equation. It turned out it wasn’t hard at all. But the results were a little bit startling. As it turns out I only need a sample size of 14 people for my study to be statistically significant. I had anticipated that I would need something like 50 people per group. So this was good news and means that I will be able to wrap up the data collection much earlier than I thought. I will of course get more than 14 people but I wont have to bust my butt forcing villagers to let me poke at their feet for another week (haha). I will also be able to do the data analysis and hopefully start writing before I leave Karigiri.
As far as data analysis goes, it is a HUGE help to have a computer with Excel along on the trip. It makes everything so much easier! Everyone on the trip is convinced that bringing a laptop is definitely the way to go.
When I met up with the Statistician again later in the day I was able to show him that I did in fact know how to calculate standard deviation and that I had figured things out on my own. After we talked about the small sample size (14), which turned out to be correct, I showed him the T test on groups 1 and 2. I think that he was genuinely interested in the results. When I get a couple more samples and am ready to finish up the data collection I am going to go back and talk to him again.
No comments:
Post a Comment