stats vocab sampling Flashcards
stratified
sampling where population is divided into groups based on some characteristic, then within each group a probability sample (likely srs) is selected (groups= strata), not all individuals within sample strata used
cluster
every member of the population is assigned to 1 and only 1 group (cluster) and then a sample of clusters is chosen with likely srs and all individuals within sample clusters are surveyed
- often economical, not as precise as strata
multistage sampling
- select sample with different sampling methods
systematic random sampling
1) list every member in pop
2) randomly select first sample element from first “k” elements
3) select every kth element in set
voluntary
people who SELF SELECT, often strong negative opinions
SRS
give each individual an equal chance of being selected (avoiding bias) and gives every possible sample an equal chance to be chosen)
convienience sample
made up of easy to reach people
census
actually obtaining data from every member in population but isn’t pratical because of cost and time required, is a parameter
setting up srs
- label each of kids and give them a unique # from 1-whatever parameter
use a rng and produce __ # s, ignoring repeats, and the kids with those numbers are the samples
srs response
this is not srs because not every group of n (Name) because not every group of (name and sample size) is equally likely to be chosen. for example, you could not have ___ (one type) and ____ (other type) in a group
stratified vs cluster
stratified sample is internally homogenous and externally heterogenous and cluster sample is externally homogenous and internally heterogenous
%
think: probability, relative frequency, data is probably categorical
of people selected
statistic
definitely quantitative and univariate
histogram
suppose we have 300 kids, we want to select 2 0kids but we want them to be 10 boys and 10 girls
this is not srs because not every group of 20 kids is equally chosen. for example, you could never get all girls in a sample.
when labeling people in sample for srs, do 01-15 for example, not 1-15
this is most likely a voluntary sample situation, so my sample is made up of mostly strong and negative opinions. Thus, I am not convinced that the majority of people disagree and the sample result most likely overestimates the number of people in opposition
voluntary response sample response