An example: you try testing the extent to which caffeine pills of different strengths affect how awake people feel. You have weak and strong concentration pills. That's your independent variable, the thing you control: how strong the pills you're giving people are, and whether you give them at all.
You ask people to say how awake they feel as a score out of ten. Then you ask them to take either a weak pill or a strong pill and ask again 20 minutes later and compare the two to see if on average there is a difference and if so how much. This is the dependent variable, the thing you measure to see what influence your independent variable has.
So you find that strong pills make people say they're more awake by 4 points and weak pills by 2 points. But you don't know if that's because of the caffeine or because they feel they should be feeling more awake after taking the pill, or they may just be humouring you.
So you run a *control* at the same time with pills that are decaffeinated versions of the original pills so that you can compare the main experiment to that, and rule out all of these things in the comparison. Ideally the test should be done by someone who also doesn't know which pills are which during the test. That's called 'double-blind'.
Sorry that's a bit long-winded.
No comments:
Post a Comment