chapter 2 - Probability models

Question 1

what are the 3 ways to create a sampling distribution from a single sample?

Answer 1

bootstrapping, exact approach and theoretical approximation

Question 2

what is bootstrapping?

Answer 2

it is when we draw a 5000 samples (usually) from one sample, and we use that to create the sampling distribution. it works just if originial sampling is rappresentative of the population (so make big and random…like my men) and if it work for replacement + remember that bootstrap sample should as big as the original sample

Question 3

with/without replacement calcolo of probability

Answer 3

without replacement = if the sample was 200 and the probability was , the next time the probability was 20%, the second time the sample will be 19/199 = 19% –> can be ingnored if the population is large

with replacement =

Question 4

advantages of bootstrapping

Answer 4

we can get a sampling distribution from every sample we want, condisering every sample statistics we are interested in + only way to calculate median

Question 5

limitation of bootstrapping

Answer 5

it does not always reflect the true sampling distribution, as the original sample is not always represenatative of the population (pero remember that the bigger the better

Question 6

the exact approach

Answer 6

when we know the true probability in the population, so we are working with the true sampling distribuition. we can calculate probability of all sample results, but it works just for CATEGORICAL VARIABLES
+ also called ‘computer intensive’

Question 7

SPSS exact value

Answer 7

nonparamentric test, you can calculate two categorical variables with fisher exact test

Question 8

theoretical approximation

Answer 8

when we do not want to use computer, ex. bell shape of normal distribution is a sample mean

Question 9

indipendent sample test

Answer 9

ex. heads and tails, when comparing 2 samples that are statistically indipendent