Quiz 7: The Logic of Sampling

Which of the following statements about informants is FALSE?

The terms informant and respondent are interchangeable. An informant is someone who is well versed in the social phenomenon that you wish to study and who is willing to tell you what he or she knows about it. It is different from a respondent. For instance, you want to study gang activities in different neighborhoods. It is a good idea to get an informant or an insider first to get information first.

According to "Longitudinal Survey of Migrants to the U.S. from Taiwan-2004 Respondents' Handbook," the sampling frame is constructed from a 0.2% random sample of the census held in Taiwan in 2000.

True Census is often an important information source or benchmark for social surveys.

Findings based on a sample can be taken as representing only the aggregation of elements that compose the sampling frame.

True In social science research, we seldom collect information from all members of the study population. In most cases, it is either impractical or too costly to cover the whole population. Instead, we collect a sample that represents the individuals of the sampling frame. Recall, a sampling frame is the source material or device from which a sample is drawn. It is a list of all those within a population who can be sampled, and may include individuals, households or institutions. The sampling frame should be closely resembling the study population. Findings based on the sample should represent the individual elements of the sampling frame.

A population is the complete enumeration of all elements in a closed system, while a sample is a subset of a population.

True It is important to identify the population of study before we work on sampling strategy.

In practice, the size of the population should almost always be taken into account when deciding on sample size.

True It is true that the choice of sample size is related to the population size. Even though Babbie mentioned that "Notice that nowhere in this discussion of sample size and accuracy of estimates did we consider the size of the population being studied. This is because the population size is almost always irrelevant," Babbie also said that "This is not literally true in practice." Researchers must look at the population size first before deciding on the sample size, especially when the population size is small.

Multistage cluster sampling is used when it is difficult to compile an exhaustive list of the elements comprising the target population. Instead, subpopulations are sampled.

True Multistage cluster sampling is a complex form of cluster sampling. Cluster sampling is a type of sampling which involves dividing the population into groups (or clusters). Then, one or more clusters are chosen at random and everyone within the chosen cluster is sampled. See the textbook.

Generally, the more heterogeneous the population, the more beneficial it is to use stratified sampling.

True Stratified sample strategy is used to better reflect the diversity of the population and/or geography. In the case of a political survey, the researchers would make sure to include participants from different minority groups. If there are large geographical variations (think about U.S. presidential primary), then it is important to ensure that the sample represents different part of the country. The bottomline is that, if there is a large variation in the population, a stratified survey is superior to a survey of simple random sampling.

The ultimate purpose of sampling is to select a set of elements from a population in such a way that descriptions of those elements accurately portray the parameters of the total population from which the sample is selected.

True The key is to make the sample representative of the population through which it was drawn.

In systematic sampling, the sampling interval is the standard distance between elements selected in the sample.

True The sampling interval equals population size divided by sample size. The higher the interval, the fewer the number of observations will be included in the sample.

Stratification represents a modification to rather than an alternative to simple random sampling and systematic sampling.

True They are all based on the same assumption about sampling.

Findings based on a sample can be taken as representing the elements that compose the sampling frame.

True This is the purpose of sampling and to reflect the sampling frame.

Probability sampling is always more desirable than nonprobability sampling in quantitative research because nonprobability sampling cannot guarantee that the sample observed is representative of the whole population.

True We have little information about the extent to which a nonprobability sample represents the population. Here, the keyword is quantitative. If you want to carry out a quantitative research instead of a qualitative one, the representativeness of the sample is a key issue. We are unable to know how representative of the sample is using nonprobability sampling.

A stratified sample is more likely to be representative on several variables than is a simple random sample.

True We need to select several variations from which the strata can be identified and selected. For instance, age and race are important to educational attainment. We can use these two variables to carry out stratified sampling.

The chief aim of probability sampling is to be able to select

a sample whose statistics will accurately portray an unknown population parameter.

Nonprobability sampling

denies the researcher the use of statistical theory to estimate the probability of correct inferences

The unit about which information is collected and that provides the basis of analysis is called a(an):

element Sample element That unit about which information is collected and that provides the basis of analysis. Typically, in survey research, elements are people. Other kinds of units can be the elements for criminal justice research-- correctional facilities, gangs, police beats, or court cases, for example.

Rebecca determined that the mean age of all students at her community college, the population she wished to study, was 22.3 years old. This value is known as a/an:

parameter. A parameter refers to a numerical or other measurable factor forming one of a set that defines a system or sets the conditions of its operation.

One of the most visible uses of survey sampling lies in _____________.

political polling

John set up a matrix representing his population of residence hall students, using gender and class level. He then started interviewing students and continued until he had accomplished the percentages he set out in the beginning. Which sampling design does this example reflect?

quota

In general, as sample size increases:

the standard error decreases in size

Disproportionate sampling and weighting are used by the researcher:

to ensure a sufficient number of cases in each of the sample subpopulations, to give a proportionate representation to each sample element, to handle situations involving the errors and approximation that are often inherent in complex, multistage designs

Tammy wants to do a telephone survey and she turns to you for help. Which of the following statements would mislead her?

Telephone surveys should never be used in political polling because the sample is biased. Telephone surveys are widely used in political polling. In fact, nearly all political polls are done through telephone surveys. However, we need to understand the limitations of this approach and make necessary adjustments to the result.

A study population is:

that aggregation of elements from which the sample is actually selected

With respect to the cost of data collection, simple random sampling is always superior to cluster sampling.

False Cluster sampling is likely used when "natural" groupings are evident. In other words, there are sub-populations which are internally more homogeneous. So we can randomly select clusters and then randomly select samples within the randomly selected clusters. Given a fixed budget, we are able to reduce sample errors by using the cluster sampling method. Therefore, with the same budget, cluster sampling is more cost effective than simple random sampling. Simple random sampling has one sampling error, while cluster sampling has more sampling errors.

In systematic sampling, the sampling ratio is the standard distance between elements selected in the sample.

False In practice, systematic sampling is virtually identical to simple random sampling. In contrast to simple random sampling, system sampling collects data following a certain routine, say every 10th element for your sample. To reduce the human bias, systematic sampling should have a random starting point. The sampling ratio is the sample size divided by the population size. Given the same population size, a higher ratio will result in a larger sample and a smaller sampling error. The sampling interval equals population size divided by sample size.

There are several basic rules about sampling. One of them is that the more heterogeneous the sample population is, the smaller the sampling error will be.

False It should be less heterogeneous or more homogeneous

Systematic sampling is seldom used in practice, with simple random sampling being the preferred approach.

False Systematic sampling is more widely used than simple random sampling in social science research. In general, simple random sampling is less efficient than systematic sampling.

Stratified sampling is based on the principle that the larger the sample, the lower the sampling error.

False The principle applies to the simple random sampling method. Stratified sampling is used to increase the representativeness of the sample and decrease sampling errors. It is based on the principle that a homogeneous population produces smaller sampling errors than does a heterogeneous population.

Random sampling error will always produce an unrepresentative sample.

False There will also be random sampling errors. One counter measure is to increase the sample size. There are many other procedures that can reduce the error. But it does not necessarily produce an unrepresentative sample.

If two samples of the same size are drawn from the same population using simple random sampling, it follows that they will always have the same statistics.

False There will be some statistical variations. It is nearly impossible to draw the same sample.

Stratified sampling does NOT require prior knowledge of population proportions.

False Without prior knowledge of population distribution, it is nearly impossible to create strata properly.

What is the basic process in employing probability proportionate to size sampling?

Give bigger clusters a greater chance of being picked but then take the same number of elements per cluster. When clusters are of varying sizes, it is important to vary the procedure by employing probability proportionate to size sampling (PPS). In the PPS method, each cluster should be given a chance of selection proportionate to its size. Then the same number of elements should be taken from each of the selected clusters.

A researcher discovers that 40% of the households in City X are single person households and 60% are husband-wife households. The researcher tells interviewers to conduct 80 interviews and that 40% of the interviews should be with households headed by a single person and 60% with husband-wife households. Refer to Exhibit 7-1. This research uses:

quota sampling Quota sample is a type of NONprobability sample in which units are selected into the sample on the basis of pre-specified characteristics, so that the total sample will have the same distribution of characteristics as are assumed to exist in the population being studied. In other words, quota sampling is a method for selecting survey participants. In quota sampling, a population is first segmented into mutually exclusive sub-groups, just as in stratified sampling. Then judgment is used to select the subjects or units from each segment based on a specified proportion. For example, an interviewer may be told to sample 200 females and 300 males between the age of 45 and 60. This means that individuals can put a demand on who they want to sample (targeting).

In telephone surveys, unlisted telephone numbers creates a special problem for accurate:

sampling frames. There is the difference between the study population and the sampling frame. If there are substantial and systematic differences between the two, there are potential for biased results.

_____________ sampling is often employed in field research whereby each person interviewed may be asked to suggest additional people for interviewing.

snowball Snowball sampling should be the answer. It is a type of non-probability sampling. Researchers have no idea about the probability someone may be included in the sample. In other words, it is difficult to know whether the results can be replicated in the general population.

Lisa wanted to do a study of women who had participated in extramarital affairs. Since there is no sampling frame listing all such women, she visited a women's group and asked for volunteers among those who had participated in an affair. She then asked each of those women for the names of other possible study participants. She was using which design?

snowball sampling

You are doing research on hospital personnel--orderlies, technicians, nurses, and doctors. You want to be sure you draw a sample that has cases in each of the personnel categories. You want to use probability sampling. An appropriate strategy would be:

stratified sampling In statistical surveys, when subpopulations within an overall population vary, it is advantageous to sample each subpopulation (stratum) independently. Stratification is the process of dividing members of the population into homogeneous subgroups before sampling. The strata should be mutually exclusive: every element in the population must be assigned to only one stratum. The strata should also be collectively exhaustive: no population element can be excluded. Then simple random sampling or systematic sampling is applied within each stratum. This often improves the representativeness of the sample by reducing sampling error. It can produce a weighted mean that has less variability than the arithmetic mean of a simple random sample of the population.

After taking a random start between 1 and 20 and then taking every 20th element from the sampling frame, Smith learned that 40% of the sample believed the company's president was doing a good job. The calculated standard error was 3 percent. Refer to Exhibit 7-2. The sampling scheme used is:

systematic

Periodicity is particularly important in:

systematic sampling Periodicity is a potential concern for Systematic Sampling. The process of selection can interact with a hidden periodic trait within the population. If the sampling technique coincides with the periodicity of the trait, the sampling technique will no longer be random and representativeness of the sample is compromised.

Every kth element in a list is chosen for inclusion in the sample in:

systematic sampling Systemic sampling is similar to simple random sampling. To avoid human interference, researchers should randomly choose the starting point.