Sampling Flashcards
(18 cards)
Simple random sampling description
Assign a number to each member of sampling frame and generate random number
Simple random sampling pro
Representative
Simple random sampling Con
Time consuming, impractical, may not get minority subsets
Systemic sampling description
Chose your sample by picking every nth number of the sampling frame
Systemic sampling Pro
Fast, should be representative if order is random enough
Systemic sampling Con
Risk you will choose an unrepresentative sample eg. Cyclic patterns
Stratified sampling description
If population clearly divides into different categories and will affect variable being measured. Need enough
Measurements from each category
Stratified sampling Con
How subgroups are chosen would affect outcome
Cluster sampling Description
Populations clearly divides into different groups but no reason to suppose variables wil be significantly differently distributed . Chose one or more of the clusters and sample only from those
Cluster sampling Pro
Representative if sampling within clusters is good,fast
Cluster sampling Con
Could be unkownn differences in subgroups
quota sampling description
population is divided into different categories but you dont care aboutt the proportions. decide how many of eshc categories you are going to measure
quota sampling pro
easy to do and can collect response from a large range of group
quota sampling con
quotas chosen could cause bias
tan(2a)
(2tan^2(a))/(1-tan^2(a))
why might change in sign method for finding root not work
1) there is a repeated root
2) there is a discontinuity, change in sign but no root. Eg TanX, 1/X
3) 2 roots squeezed between consecutive integers
why might Newton-Raphson method fail
1) when a starting value leads to fâ(x)=0 ( a stationary point). The sequence becomes undefined
2) The sequence converges but not to the root near the starting value
3) 3 roots between 2 consecutive integers, the middle one cannot be found
conditions for binomial distribution
1 conduction trials on random samples of certain sizes(n)
2 On each trial the outcomes can be classified as either success (p) or failure (q)
3 The outcomes of each trial is independent of the outcomes of any other trial
4 The probability of success p is the same on each trial