When controlling for age and geographic location, when a person who goes to 4 or more seminars on obesity prevention has a lower BMI than a person who goes to less than 4 seminars on obesity prevention.
2006-12-11 14:07:46 · 2 answers · asked by A M 2 in Science & Mathematics ➔ Medicine
If you "pair" the age and geographic locations BMIs, you'd use a paired student's t-test.
If you don't "pair" (the age and/or geographic location isn't the same between people), an unpaired t-test.
I do stats all day.
2006-12-11 17:22:50 · answer #1 · answered by Griffin 2 · 0⤊ 0⤋
If you have alredy controlled for age and location, all you need is a simple student's t-test. You can do it on excel, just put your two groups (more than 4 and less than 4) of BMI's in seperate columns and then go to tools, statistics, t-test with equal or unqueal distributions. You don't need to pair it becuase you've already controlled for age and location. Most people set thier p value to 0.05. Which is just your chance of making a type 2 error.
2006-12-11 17:34:31 · answer #2 · answered by jason e 2 · 0⤊ 0⤋