Chapter 14 Sampling PowerPoint presentation developed by: Jennifer Manuel & E. Roberto Orellana
Overview Introduction Nonprobability Sampling Selecting Informants in Qualitative Research Probability Sampling Sampling and Bias Probability Sampling Designs Multistage Cluster Sampling
Introduction Sampling is the process of selecting observations Probability Sampling Nonprobability Sampling Sample: a subset of a population that is observed for purposes of making inferences about the nature of the total population
Nonprobability Sampling Used when probability or random sampling is not possible or appropriate (e.g., homeless individuals) Generally less reliable, but often easier and cheaper 4 types: Reliance on available subjects Purposive or judgmental sampling Quota sampling Snowball sampling
Types of Nonprobability Sampling Reliance on Available Subjects Sampling from subjects who are available (e.g., how much an agency’s services help a particular client or group of clients) Purposive or Judgmental Sampling When a researcher uses his or her own judgment in selecting sample members (e.g., handpick community leaders or experts known for their expertise on target population)
Four Types of Nonprobability Sampling Quota Sampling A relative proportion of the total population is assigned for the target population’s characteristics (e.g., gender, ethnic groups), grouped into strata or cells, and the required number of subjects from each stratum or cell (given set of characteristics) is then selected Snowball Sampling Process of accumulation as each located subject suggests other subjects
Selecting Informants in Qualitative Research Informants are members of the group or other people knowledgeable about it who are willing to talk about the group When informants are used, they should be selected in such a fashion as to provide a broad, diverse view of the group under study
Probability Sampling: The Logic Chief criterion of the quality of a sample: Degree to which a sample is representative – that is the extent to which the characteristics of the sample are the same as those of the population for which it was selected Probability sampling methods are one approach to selecting samples that will be quite representative
Probability Sampling: The Logic Basic principle is that all members of population will have an equal chance of being selected in the sample, known as equal probability of selection method Even the most carefully selected sample will almost never represent the population from which it was selected There will always be some degree of sampling error, which can be estimated
Probability Sampling: Some Definitions Element: a unit about which information is collected and that provides basis for analysis (typically people or certain types of people in survey research) Population: theoretically specified aggregation of study elements Study population: aggregation of elements from which the sample is actually selected
Probability Sampling: The Ultimate Purpose To select a set of elements from a population in such a way that descriptions of those elements accurately portray the total population from which elements are selected The key to this process is random selection, where each element has an equal change of selection independent of any other event in the selection process
Probability Sampling: Sampling Frames and Populations A sampling frame is a list or quasi-list of members of a population (e.g., student roster, list of census blocks, telephone directory) Examples of populations that can be sampled from a sampling frame include elementary school children, high school students, church members, factory workers, and members of professional associations
Probability Sampling: Biases to Avoid Overgeneralization occurs when sampling frames are not consonant to which we seek to generalize Nonresponse bias occurs when a substantial number of people in a randomly selected sample choose not to participate
Probability Sampling: Biases to Avoid Cultural bias is the unwarranted generalization of research findings to the population as a whole when one culture or ethnic group is not adequately represented in the sample Gender bias is the unwarranted generalization of research findings to the population as a whole when one gender is not adequately represented in the sample
Probability Sampling Designs Simple Random Sampling Systematic Sampling Stratified Sampling
Probability Sampling Designs: Simple Random Sampling Each element in sampling frame is assigned a number A table of random numbers is then used to select elements for the sample (See Appendix B for Table of Random Numbers) Most fundamental technique in probability sampling, but laborious
Probability Sampling Designs: Systematic Sampling Involves the selection of every kth element or member of the sampling frame Similar to simple random sampling except: Elements chosen based on sampling interval First element selected at random to avoid bias Important to carefully examine the nature of the list and whether a particular order of the elements will bias the sample selected
Probability Sampling Designs: Stratified Sampling Involves the process of grouping members of a population into homogeneous strata before sampling (e.g., by ethnic group or gender) Stratified sampling improves the representativeness of a sample by reducing the degree of sampling error
Multistage Cluster Sampling More complex sampling technique frequently used in cases in which a list of all members of a population does not exist An initial sample of groups members is selected and then listed. The listed members are subsampled, which provides the final sample of members
Multistage Cluster Sampling Probability proportionate to size is a special efficient method for multistage sampling If members have unequal probabilities of selection into the sample, it is necessary to assign weights in order to provide a representative picture of the total population
Probability Sampling in Review May be extremely simple or extremely difficult, time-consuming, and expensive However, it remains the most effective method for selecting study elements: Avoids conscious or unconscious biases in selecting elements Permits estimates of sampling error