Inferential Statistics Flashcards
What do inferential statistics do?
they make inferences (conclusions) about the population
what are the 2 types of inferential statistics?
Frequentist (uses sample data to draw conclusions by empathising frequency / probability)
Bayes
What are the 4 features of frequentist inferential statistics?
Null hypothesis significance testing
P-values
Effect size
confidence intervals
What is Null hypothesis significance testing (NHST)?
when an experimental factor is tested against a hypothesis of no effect or relationship
Null Hypothesis is H_?
H0
no difference
Alternative Hypothesis is H_?
H1
difference
has 2 types
what are the 2 types of Alternative (experimental) hypothesis?
Directional (one tailed)
Non-directional (two tailed)
What is the P-value rule?
if p-value is below 0.05 then result is ‘statistically significant’
the level we accept results is also known as ‘alpha level’
can use 1% to be stricter
How do P-values work in frequentist inferential statistics?
probability result found occurred by chance assuming null hypothesis is true
tells us if there’s false-positives
what does it mean if the p-value is high?
more likely results are ‘statistically non-significant’
How do you write up a p-value that’s written as e-16 (for example)?
means p-value is really small
p<.001
What are Effect Sizes?
tell us about the magnitude of the difference or similarity of a correlation
What is the most popular effect size?
Cohen’s d (measures size of difference between groups)
Pearsons r (measures strength of relationship between 2 variables)
When is cohen’s d often used?
often used in independent samples t-tests
continues data with large samples
Chi Square
used to discover if there is a relationship between 2 categorical (nominal) or sometimes ordinal variables