GER1000: Tutorial Class 1

GER1000: Tutorial Class 1

How do we know the effectiveness of a medicine/treatment?
> control group
> treatment group
We need a comparison.
How can we be sure that the treatment group is getting better because of the medicine?
> they must be the same in the beginning
What do we mean be similarities?
Health situation must be more or less the same, the average age is the same and the proportion is the same.
How can we construct two groups that are similar?
> Randomly assign people
But for this to work, we need to start with a large group of people.
In conclusion -> Randomised control experiment
This is to minimize confounding, to ensure that there’s no other factors that will be confounded with the effect of the treatment.
Double blind experiement:
To ensure even the evaluator doesn’t know who is an control or treatment.
____________________________________
Power:
We don’t have the power to control the variables of our subjects.
>smoking
>tall
We can not control but we can observe.
E.g
Heart disease is more common among the smokers.
We cannot assume that smoking cause heart diseases. Causation is not association.
Rate(A|B) -> rate of A among B
Rate(A|not B) -> rate of A among not B
I’m comparing B and Not B.
Rate(A|B) >Rate(A|not B) positively associated
Rate(A|B)< Rate(A|not B) negatively associated
But if we negate all
Rate(Not A| NotB )> rate(not A|B) positively
Confounder:
A third variable. Associated with An independent and dependent variable.
Slicing method: breaking comparison into smaller factors