Tests of significance Flashcards
what is the difference between sampling with / without placement
with - outcome doesn’t depend on the one before eg. coin toss
without - outcome influenced by previous eg pulling cards out of a deck - there are less hearts as you go on so the probability changes
what does a value below the 5% threshold mean
the probability of the data being different from the null hypothesis is below 5% - therefore the result is significant and we reject the null hypothesis - have significant evidence against it
what are the two forms of non - random sampling
bias - more likely to pick one thing over another
non-independence - same thing over and over again in different individuals
what is a type 1 statistical error
false positive - reject the null hypothesis when true
what is a type 2 statistical error
false negative - accept when actually false
when is type 1 statistical error particularly a problem
when you do lots of tests
what is statistical power
have a big enough sample so that if the hypothesis is wrong you can show its wrong
what is nominal data
data falls into a number of defined classes - m/f or coinnh/t
what are the two distributions for Non-parametric statistics
Binomial, Poisson
what type of data would you use non-parametric tests on
non-normal
what is a binomial distribution
the probability of a success or a failure - two outcomes for n
what are the two non-parametric tests
chi-squared test + chi squared test of independence
when can you use a non-parametric test for data in linear form
if their distribution is not normal
what is a chi-squared test
test against a prior expectation, have to look up your value in a chi squared table
If your chi-squared value is above the threshold on the table, do you accept or reject the null hypothesis
If the value is above the threshold at 5% then you reject the null hypothesis
