The Hypothesis
All good experiments start with a hypothesis. A hypothesis is an educated guess about what should happen under a certain set of conditions. For example, “stem cells can grow in the eye and replace worn out RPEs”.
The way I was taught, this statement is then ‘turned around’ to be the negative. “Stem cells cannot grow and replace worn out RPEs”. This ‘turned around’ statement is called the null hypothesis. This is what will or will not be disproven with our experiment. The idea is to see if we can generate a result that will be significantly different from the starting point. The null hypothesis predicts no change. We want to see if we can affect a change that is big enough to matter. Good so far?
The Plan
After deciding what we want to figure out, we need a plan. Our plan is our experimental design. We want to design our experiment so we are reasonably sure there are no stray influences that may have caused the results.
The groups
If we are comparing treatments or treatment versus no treatment, we need to make sure our groups are as similar as possible. If we want to see if our new exercise program is better than another training program we really would (ethically) not want 20 year olds in one group and 80 year olds in another.
This is part of the reason good experiments on clinicaltrials.gov list exclusions which means reasons why a person couldn’t be in the trial. They are looking for similar subjects. Another reason for exclusions may be related to the treatments they are using and people’s sensitivities to them. You don’t want a subject who has an allergic reaction to your treatment!
The Sample
Once we have our sample (all possible examples or cases of something is called the population), we should be reasonably sure we have controlled for most of the important variables. Variables are factors that can affect the outcome of our study. However, there are other variables we have to control for in our experimental design.
The Treatments
One of these is the Hawthorne Effect. Anyone who has ever taken intro to psych or industrial/organizational should remember the Hawthorne Effect. In a plant they were trying to increase productivity. The problem became, no matter what they did, production improved! These people could have been working in the dark with the heat turned up to 90F but they were going great guns.
The researchers discovered just the fact that they were noticing the workers and trying to do something for them was enough to increase productivity. It did not matter that working conditions worsened. The workers felt noticed. They wanted to please the researchers.
Placebo effect happens when subjects believe the ‘treatment’ is effective even though it is not. They may show improvement as a result of their belief.
You will see attempted control for the Hawthorne and Placebo Effects in studies that list themselves as ‘double blind’ or ‘double mask’. They will perform fake (sham) procedures on people in the control so they feel noticed and loved, too.
Double blind? The double is because doctors and nurses treat people getting the actual treatment differently. They may not realize it but they may be more encouraging to those actually getting the treatment than they are towards the people who are the controls. When a study is double blind neither the patients nor the people caring for them actually know.
That is my 500+ words for this one. One or two more pages and I should be done. Hang in there!
For more information, click here for Designing Experiments Using the Scientific Method from dummies.com.