Sample reflects a trait until the quota for that trait is met
+: small sample still representative, no sampling frame, quick, easy,cheap, compare groups in a population
-:potential bias, groups can be costly or inaccurate, increased scope= more costs, non responses are not recorded
Opportunity Sampling
Sampling whoever is available and fits the criteria
+:easy, cheap
-: unlikely to be representative, dependant on researcher
Census
Observes every member of a population so is very accurate, but time consuming, expensive and can't be done when testing destroys the item
Sample
Observations taken from a subset of the population. Cheaper, faster, less data to process, but may not be accurate or large enough to include subgroups of the population
Simple Random
+: Bias-free, easy, cheap for small populations, equal chance of selection
-:needs sampling frame, not suitable for large populations
Systematic
1st person is random then regular intervals
+:simple, quick, suitable for large populations
-:needs sampling frame, introduces bias if the sampling frame is not random
Stratified
no. in stratum/no. in population x desired sample size
+: reflects population structures, proportional representation of groups
-:must make distinct strata, sampling frame, unsuitable for large populations
Histograms
Frequency= frequency density x class width
Frequency polygon = joined the middle of each bar
Shows continuous grouped data and gives a good idea of distribution
Comparing Data
Use the mean and standard deviation OR the median and IQR
Regression
+b = positive correlation -b=negative correlation
using x to find y in the know range = interpolation
using x to find y outside the range = extrapolation