Lecture 5 Flashcards
Purpose of analytic studies
To test hypotheses regarding, for example,
causes of disease;
methods for prevention of disease; the effects of treatments such as vaccines
experimental =
researcher actively intervenes
observational =
testing a predefined claim, no researcher intervention
replication
Responses vary among people. Replication allows us to separate out
true effects from chance effects.
do what you are doing to multiple people, increasing the number of samples
control
Provides context for evaluating the effect of interest.
need to be able to compare without doing it
What would have happened without the intervention or exposure of interest?
trying to compare the thing we are interested in with people who have no been exposed to the thing as you are unable to do it to the same group
placebo group is a type of control group
how easy/hard is causality to prove?
causality is very difficult to prove, especially impossible with a single study
Designing a study to evaluate the effectiveness of something?
make sure the exposure group and control group are similar and control for all other sources
Experimental studies
The researcher manipulates the conditions (intervenes in a natural process) and records the results.
The aim is to control all other factors to isolate the effects of the intervention.
Best way to study causation.
Randomisation can be used to ensure that effects of unmeasured factors are equalised across the intervention and control groups.
Experimental studies are common throughout science
Randomisation
Randomisation can be used to ensure that effects of unmeasured factors are equalised across the intervention and control groups.
ensure we are randomly allocating individuals to treatment and control group to ensure both groups are similar/comparable in number of each group in the treatment and the control group
What is the best way to study causation?
Doing a experimental study ]
Types of analytic study
Experimental
Observational
Observational studies
The investigator does not intervene, simply observes a naturally occurring process, and collects information.
The idea is to get as close as possible to the information that would have been obtained if the experimental study could have been done.
Still aiming to test a particular claim, get as close to an experimental without doing one. Cannot always complete an experimental study because of money, ethics, time etc.
Gold standard analytic study in research =
randomised control trials
what do you compare in an RCT?
compare the 2 rates of a variable in the two different groups
cohort study selection
participants are identified from population, questioned about variable, participants are effectively self selecting therefore the potential for confounding is a real danger (randomisation usually minimises this)