hypothesis testing Flashcards
how do you do a binominal test for hypothesis testing?
- binominal test is a statistical test that concerns whether a proportion observed in your data is different from a known proportion
what can a statical test systematically test?
- tests whether a given scientific claim is valid or not
what are answers like in most cases in scientific research?
- no definitive answers to any kind of questions so scientists prefer to discuss based on probability
when do we reject the hypothesis?
- if the probability is very low
when do we accept the hypothesis?
- if the probability is not so low
what are p -values?
- probabilities used to reject hypotheses
what are the threshold levels for p- values? when are they defined?
- alpha level
- normally 0.05
- needs to be defined prior to a study
what p - value do you reject the hypothesis at?
- if p- value < 0.05
what p -value do you accept your hypothesis at?
- if p- value > 0.05
what is a null hypothesis?
- hypothesis against the research question, claiming that there is no difference in the result
- only differences observed are just noise/ error
what is the symbol for null hypothesis?
H 0
what do scientifistic often have to struggle against?
- struggle against the probability that the observed difference could be chance due to noises
what is the research/ alternative hypothesis?
- claims that there is a difference in the result
what is the symbol for research hypothesis?
H a
in statistics what do we test?
- test the probability that the null hypothesis is true
why don’t we test the research probability?
- because you can never prove something is true
- can prove the null hypothesis is false
what do we do in hypothesis testing?
- first define H0
- make an attempt to reject this by showing there is a statistically significant difference
how is significance estimated?
- by the probability that the difference occurred by chance (p- value)
what always occurs when dealing with probability?
- errors
- two types; 1 and 2
what is type 1 error?
- false positive
- reject the null hypothesis when it is true
what is type 2 error?
- false negative
- not to reject the null hypothesis when it is false
do we want to minimise one of the types of errors more?
- researchers need to minimise both but based on your research question you may want to focus on one type more than the other
what are the three typical types of tests?
- observed proportion< expected proportion
- observed proportion > expected proportion
- observed proportion = expected proportion
described observed proportion> expected proportion
- 1 - cumulative probability from observed to max
describe observed proportion < expected proportion
- cumulative probability from 0 to observed
describe observed proportion = expected proportion
- two tailed cumulative probability same distance from the mean
how do you conduct a binominal test in hypothesis testing?
- describe null hypothesis with expected proportion
- report observed proportion from data
- report probability (p- value) that the null hypothesis is true
- optionally report the confidence interval
what is the confidence interval?
- range of plausible values associated with a confidence level
what is the normal confidence interval and what does this mean?
- normally 95%
- means you are 95% certain that true proportion falls within CI
what does confidence interval include?
- often used to the range in which true proportion may lie
