Sample+collection+(typologies)

Collecting a sample is searching for a portion of the total population that can be representative for the whole population.

// Example: to measure average age of a university, ask a randomly selected sample of students //
 * EPSEM ** = Equal Probability of Selection Method (i.e. every case in the population has the same chance of being selected to be part of the sample)

Simple Random Sampling (SRS) Systematic Random Sampling Stratified Random Sampling Cluster Sampling
 * Some EPSEM sample techniques **


 * Simple Random Sampling **

__ Requires __ // A list of all cases in the population // // A method for selecting cases from the population so each case has the same probability of being selected. Different techniques: // // Flipping coins // // A set of random numbers // // Randomly generated list //

__ Challenges __ // Simple Random Sampling is costly when there are many cases // // Choices come from a list; who makes the list? // // If the list is long, there will be a lot of paper shuffling and confusion may occur //

Are samples collected according to a rule; // Example: Only the 1st case is randomly selected. There after, every k’th case e.g., every fifth person // __ Steps __ 1. Number the population (1,…,N) 2. Decide on the sample size, n 3. Decide on the interval size, k = N/n 4. Select an integer between 1 and k 5. Take case for every kth. unit
 * Systematic sampling **


 * Stratified sampling **

Stratified sampling in cases where population is not very uniform

// Example: UMB students //

Stratified sampling breaks the population down to various uniform subgroups or strata and draws samples from these groups

It is useful as it ensures the sample is an accurate representation of the population with respect to the selected statistics

__ Requires __

Include representation by members from key sub-groups for achieving more precise results Successfull stratified sampling requires good knowledge of the population

Selection of groups of cases, called clusters, rather than single cases. Clusters are selected randomly from the larger population of clusters. Practical strategy with very big populations.
 * Cluster sampling **

__ Challenges __ // Less likely to be as representative as a SRS of comparable size; usually multi-stage and each has a probability error. //