Introductory Sampling Theory
Various types of distributions zPopulation zSample zSampling z(Normal)
Measures of central tendency zMode zMedian zMean
Symbols zN = size of sample zX = a “score” zX = mean z = summation
Mean z“Average” zFormula
Formula for mean
Variability zHow measures spread out zRange zDeviations yDifference between the mean and the score
Variability zHow measures spread out zRange zDeviations yDifference between the mean and the score
Variance z“The mean of the squared deviations”
Variance
Standard Deviation
Various types of distributions zPopulation zSample zSampling z(Normal)
Population Distribution zDistribution of the attributes of a population or universe. zMay have any shape. y“Skewed” left or right yFlat or peaked
Sample Distribution zDistribution of the attributes of a sample drawn from a specified population or universe zShape will approximate the population or universe distribution zThe larger the sample size, the closer the approximation, in all likelihood.
Sampling Distribution zDistribution of the means (could be other statistics) of all possible samples zTheoretical distribution since all possible samples cannot be drawn zWill always be normal, because of the laws of probability
Normal Distribution zSymmetrical zDefined by standard deviations (standard errors) zCan predict what proportion of cases will fall within a specified range of values
Relation among distributions zNever know the population characteristics yPopulation characteristics are “parameters” yThat’s why research is done zSample distribution shows characteristics yCan guess at what the population characteristics are yLarger sample size give greater precision and confidence
Standard error of the mean
Five types of sampling zRandom (or simple random) zStratified random zCluster sampling zSystematic zArea probability
Random zEvery subject is known zEvery subject has equal or know probability of selection
Random zAdvantages: yDon’t have to know the characteristics of a population yTends to be completely representative zDisadvantages: yComplete list is difficult to obtain yAlways a chance of drawing a misleading sample yNeeds a larger sample size
Stratified random zPopulation classified into two or more strata zSample drawn from each one zCases drawn in proportion to representation in population zCases can be oversampled, if needed
Stratified random zAdvantages: yCan be sure no relevant group is omitted yGreater precision possible with lower sample size zDisadvantages: yNeed to know about the population yProportions must be known yDifficulty in locating cases
Systematic random zSelection of every nth name zUsually quicker
Cluster zDone for efficiency zPopulation is broken down into smaller groups zUseful when no sampling frame is available
Area zCombines cluster and systematic zBased on geography