A sample is some part of a larger body specially selected to represent the whole. Sampling is the process by which this part is chosen. Sampling then is taking any portion of a population or universe as representative of that population or universe.
For a sample to be useful, it should reflect the similarities and differences found in the total group.
The main objective of drawing a sample is to make inferences about the larger population from the smaller sample.
A poll is a type of sample survey dealing mainly with issues of public opinions or elections, or peoples attitudes about candidates for political office, or public issues.
Polls are conducted by large polling organizations such as the Roper poll, the Harris poll, the American Institute of Public Opinion, and the National Opinion Research Center.
A census is a survey in which information is gathered from or about all members of a population
Population/Target population: This is any complete, or the theoretically specified aggregation of study elements. It is usually the ideal population or universe to which research results are to be generalized. For example, all adult population of the U.S.
Survey population: This is an operational definition of the target population; that is target population with explicit exclusions-for example the population accessible, excluding those outside the country.
Element (similar to unit of analysis): This is that unit about which information is collected and that provides the basis of analysis. In survey research, elements are people or certain types of people.
Sampling unit: This is that element or set of elements considered for selection in some stage of sampling (same as the elements, in a simple single-stage sample). In a multi-stage sample, the sampling unit could be blocks, households, and individuals within the households.
Sampling frame: This is the actual list of sampling units from which the sample, or some stage of the sample, is selected. It is simply a list of the study population.
Sample design: This refers to a set of rules or procedures that specify how a sample is to be selected. This can either be probability or non-probability.
Sample size: The number of elements in the obtained sample.
Sampling error: This is the degree of error to be expected for a given sample design or the difference between the sample mean and the population mean.
Sampling bias: This refers to the notion that those selected are not "typical" or "representative" of the larger populations that have been chosen from.
Margin of error refers to the precision needed by the researcher. A margin of error of 5 percent means that the actual findings could vary by as much as 5 points either positively or negatively.
Confidence level (or level of confidence) is a statement of how often you could expect to find similar results if the survey were to be repeated, or the degree of certainty of obtaining the same results. It often informs about how often the findings will fall outside the margin of error.
Confidence interval is a range in which we are fairly certain that the population value lies.
Parameter-the summary description of a given variable in a population.
Statistic-the summary description of a given variable in a sample.
Statistical inference:The process of reasoning by which information about a population is extracted from sample data.
In 1920, Digest editors mailed postcards to people in six states, asking them who they were planning to vote for in the presidential campaign between Warren Harding and James Cox. The editors selected the names from telephone directories and automobile registration lists. Based on the response to the postcards, the Digest correctly predicted that Harding would be elected. The Digest expanded the size of its poll and made correct predictions in 1924, 1928, and 1932, but failed in 1936. Voters gave Roosevelt a third term in office by a landslide victory (61%). Landon won 8 electoral votes to Roosevelt's 523. The main problem was with the sampling frame used by the Digest-telephone subscribers and automobile owners. It selected a disproportionately wealthy sample.
The 1936 election saw the emergence of George Gallup who correctly predicted that Roosevelt will beat Landon. He used quota sampling to successfully predict the outcome of the election is 1936, 1940, and 1944 presidential elections, but suffered the embarrassment of picking New York Governor Thomas Dewey over incumbent President Harry Truman.
i. Probability sampling is one in which each person in the population has the same probability/chance of being selected. In addition, the selection of persons from the population is based on some form of random procedure. Samples that have this quality are often labeled as EPSEM (Equal probability of Selection Method).
i. Simple Random Sampling- is a sampling scheme with the probability that any of the possible subsets of the sample is equally likely to be the chosen sample. A way of selecting the sample is by means of a table of random numbers. Once a sampling frame is available, each person in the population is assigned a number. SRS can be with or without replacement.
ii. Systematic sampling (interval random sampling) is an EPSEM strategy which gives each element in the population the same chance of being selected for the sample. We would proceed down the sampling frame selecting for the sample every Kth person, starting with a person randomly selected from among the first K persons and choosing systematically form inclusion in the sample.
Two terms are often used in connection with systematic sampling are sampling interval (the standard distance between elements selected in the sample) and sampling ratio (the proportion of elements in the population that are selected ).
iii Stratified sampling is where we begin by grouping elements that share certain characteristics, or dividing the population into several large groups, or clusters. Its purpose is to classify populations into subpopulations or strata based on some supplementary information and then a selection of separate samples from each of the strata.
The two types of stratified sampling are proportionate stratified (is where the strata sample size are made proportional to the strata population size) and disproportionate stratified (where a varying sampling is used).
iv. Cluster (Area) sampling-may be one-stage, two-stage or multi-stage cluster/area sampling (eg.-studying blacks' attitudes toward transracial adoption). It is where all the elements in selected clusters are included in the sample. Usually the sampling unit contains more than one population element, eg., sampling households as sampling units only a sample of elements is taken from each selected cluster, this is called two-stage sampling. The whole technique is referred to as multi-stage sampling.
Types of Nonprobability Sampling are:
i. Availability/Accidental sampling is where the first available appropriate sample are used.
ii. Quota sampling begins with a matrix describing the characteristics of the target population. The goal is to select people to reflect characteristics found in the population.
iii. Purposive/Judgmental sampling is where the sample is selected on the basis of knowledge of the research problem to allow selection of "typical" persons for inclusion in the sample.
iv. Snowball sampling is appropriate when the members of a special population are difficult to locate.
v. Dimensional sampling is a sampling technique for selecting small samples in a way that enhances their representativeness. There are two steps to dimensional sampling. First, specify all the dimensions or variables that are important, and second, choose a sample that includes at least one case representing each possible combination of dimensions.
Sample size relates to how many people to pick for the study. The question often asked is: How big a sample is necessary for a good survey?
This depends on: factors such as a) the researcher hypotheses or questions; b) level of precision, c) population homogeneity, d) sampling technique used; e) monetary and personal resources; and f) the amount of time available
According to the law of large numbers, the larger the sample size, the better the estimates, or the larger the sample the closer the "true" value of the population is approached.
i. Sampling saves time and money
ii. Sampling saves labor.
iii. A sample coverage permits a higher overall level of adequacy than a full enumeration.
iv. Complete census is often unnecessary, wasteful, and the burden on the public.
Types of probability sampling