Stats class 2 Flashcards
What is an experiment?
Randomisation, differences between the groups are because of our manipulation. Experiments let you infer causation.
Bradford Hill criteria
- Strength
- Consistency
- Specificity
- Temporality
- Biological gradient
- Plausibility
- Coherence
- Experience
- Analogy
Case control
Find a group of people exposed to a disease (oral cancer) and then find a group of people who haven’t been exposed to the disease and then see if they have been exposed to something to cause this (tobacco)
Spurious correlations: the third variable problem
This graph makes it look like these two variables are linked. However it’s not considering the ‘third variable’ that is causing the increase in both of these factors.
The third variable in this case is ‘heat’.
Disinformation vs misinformation
Disinformation = lying
Misinformation = being wrong and telling people
Deductive reasoning
General to specific
“All bacteria need oxygen, this sample is bacteria, therefore it needs oxygen.”
Falsifiability
Disprove deductive reasoning
“all bacteria need oxygen”, finding a sample of bacteria that don’t need oxygen.
Inductive reasoning
Specific to general
“This bacteria needs oxygen, all the bacteria I’ve observed need oxygen, therefore all bacteria need oxygen”
These arguments are widely used in science, inductive arguments cannot be falsified and so they are intrinsically less robust.
Gold standard in primary research
Experiments