Chapter 11 Statistics

3 main parts of sampling

examine parts of a whole, randomize, sample size

census

sample of an entire population, selects ALL subjects in a population

census pros and cons

population constantly changing, difficult to complete (can’t find everyone), impractical

why we randomize

to reduce influence of all features in population, to enhance likelihood of average of population

why sample size matters

should be representive of whole population, too small: miss subgroups, too big: difficult practically

parameter

used for populations, greek variables

statistics

used for samples, english variables

simple random sample (SRS)

every individual has equal chance of being selected, every combo of individuals of certain size has equal chance

examples of SRS

numbering individuals and using random number generator to select ones, IGNORE REPEATS, drawing from a deck of cards

Stratified random sample

division of population into seperate groups (strata), based on shared characteristics (men, women, age, etc), within stratum SRS conducted

pros and cons of stratified sampling

reduces sampling variability, reduces sampling error

Cluster sampling

individuals pregrouped in some way, select entire groups to sample, SRS of clusters

cluster pros and cons

efficiency in cost or practice

Systematic Sampling

select every nth person, start from random point

mulitstage sampling

any samples methods combined, determined by context of problem, refer to which METHODS using

bias

when certain responses are favored over others, over/underemphasize characteristics of the population

voluntary response

given people an option to respond, use data from those to CHOSE to respond, not representive of population

convenience sample

only surveying those most convenient, (asking the class next doors where theyre going to college, not whole school)

under coverage bias

a portion of population is unrepresented or excluded from survey

nonresponse data

individuals chosen for the sample for whom data cannot be obtained (or who refuse to respond) may differ from those whom the data can be obtained, (restaurant reviews, super happy or mad)

response bias

something about the design which alters the response, including wording bias

pilot study

trial run of a survey to see how it reads and improve it

how to describe bias

describe the direction of bias, (…will result in lying, more likely to respond no b/c…, this will inc/dec percentage of no’s, etc)