Chapter 5 (Sampling)

What are some things to keep in mind about sampling?

-It is important to always be alert to ways in which the population may have been narrowed by the sample selection procedures. -If we pull a representative sample from a population, we can generalize the sample results from which the sample was selected, but we should be cautious in generalizing to another setting or population. SAMPLE GENERALIZABILITY IS EASIER TO ESTABLISH THAN CROSS-POPULATION GENERALIZABILITY (EXTERNAL VALIDITY)

How does sampling error relate to a sample generalizability?

-The larger the sampling error, the less representative the sample, less generalizability -Sampling error reduces as sample size increases -Sampling error decreases if population from which sample was selected is homogenous -Sampling error is not affected by the proportion of the population that is sampled (unless pop. is large)

What are three questions to consider to assess sample quality when planning or evaluating a study?

1) From what population were the cases selected? 2) What method was used to select cases from this population? 3) Do the cases that were studied represent, in the aggregate, the population from which they were selected? **The obtain sample (not the desired sample) determines sample quality

What are the 4 circumstances when non probability sampling methods are useful?

1) When random sampling is not possible 2) When you don't have a population list 3) When you are exploring a research question that doesn't concern a large population 4) When you are doing a preliminary or exploratory study

How can generalizability be increased in qualitative research/nonprobability samples?

1) choosing sites on the basis of their fit with a typical situation 2) performing multisite studies (heterogeneous sites)

What are the 4 major types of nonprobability sampling?

1. Availability (convenience) sampling 2. Quota sampling 3. Purposive or judgment sampling 4. Snowball sampling

What are the two problems of quota sampling?

1. Lacks random selection 2. You must know the characteristics of the entire population to set the right quotas

Give examples of when systematic bias is likely.

1. Sampling frame is incomplete (i.e., city, state, nation is charaterized by migration; unavoidable omissions) 2. Nonresponse (over 30%=not generalizable)

What are the 4 major types of probability sampling?

1. Simple random sampling 2. Systematic random sampling 3. Stratified random sampling 4. Multistage cluster sampling

What are the two conditions in which sampling is not necessary?

1. When the elements that would be samples are identical 2. When you have the option of conducting a complete census of a population

Sampling distribution

A hypothetical distribution of a statistic (i.e., mean) across all the random samples that could be drawn from a population --Can used the properties of this to calculate the amount of sampling error that was likely with the random sample used in a study.

Sampling Frame

A list of the elements of a population from which a sample is actually selected. i.e., a list of residential addresses **Required in most probability sampling methods **The adequacy of this is an important determinant of sample quality

Stratified random sampling

A method of sampling in which sample elements are selected separately from population strata that are identified in advance by the researcher. Can be proportionate or disproportionate. i.e., race may be the basis for distinguishing individuals in some population of interest. Within each racial category selected for the strata, individuals are then sampled randomly. Gender can also work.

Cluster

A naturally occurring, mixed aggregate of elements of the population. eg., schools could serve as clusters for sampling students, blocks for city residents, businesses for employees

Representative sample

A sample that looks like the population from which it was selected in all respects that are potentially relevant to the study. The distribution of characteristics in this sample are the same as the distribution of those characteristics among the total population. -If a small sample has been studied in an experiment or a field research project, the study should be replicated in different settings, or, preferably w/ a representative sample.

Periodicity

A sequence of elements in a list to be sampled that varies in some regular, periodic pattern. **When it is inappropriate to use systematic random sampling

Target population

A set of elements larger than or different from the population sampled and to which the researcher would like to generalize study findings.

Sample

A subset of the larger set of individuals or other entities in which we are interested.

Normal distribution

A symmetric distribution shaped like a bell & centered around the population mean, with the number of cases tapering off in a predictable pattern on both sides of the mean.

A researcher might conclude from a sample of 200 college students that he can be 95% confident that the true mean number of hours spent studying each week in the total population of college students is between 6 and 9 hours. The range from 6 to 9 hours represents which of the following? a) confidence interval, b) sampling distribution, c) confidence limits, d) population parameter

A. Confidence interval

I am interested in attitudes of students toward armed guards in all UD buildings. I conduct a survey of students using random digit dialing of both cell and land line phone numbers registered to UD students. This is an example of a: a) simple random sample, b) purposive sample, c) quota sample, d) Multi-stage cluster sample

A. Simple random sample

A researcher obtains a sample using stratification by gun ownership. He randomly selects an equal number of cases from each group, those who own guns and those who do not own guns. This means that 50% of his sample are gun owners and 50% are not gun owners. However, of the actual population from which he drew his sample 25% are gun owners and 75% are not gun owners. What type of sampling method did he use? a) snowball sampling, b) disproportionate stratified sampling, c) simple random sampling d) multistage cluster sampling

B. Disproportionate stratified sampling

In order to collect a sample of residents within a city that would be representative of the city's population, we might first collect a sampling frame of all neighborhoods in the city, then randomly select blocks within each neighborhood, then select dwelling units within each block, and contact the residents of the unit to see if they would like to participate in the research. What type of sampling method does this best represent? a) purposive sampling, b) multistage cluster sampling, c) proportionate quota sampling, d) systematic random sampling

B. Multistage cluster smapling

Janice wants to examine the motivations for people to join environmental organizations like the sierra club. She uses an acquaintance she knows who belongs to such a group, and asks him to recommend others to interview, and they, in turn, recommend others. This is an example of a: a) Stratified Random Sample, b) Snowball Sample, c) Quota Sample, d) Multi-stage Cluster Sample

B. Snowball sample

Sampling error declines as a function of several features. Which of the following is not one of these features? a) a larger sample size b) the population is heterogeneous c) the population is homogeneous d) the use of a random selection procedure

B. The population is heterogeneous

A graduate student in psychology desires to give a survey but is short on both time and funds. Instead of giving the survey to members of the local community, she arranges with a professor to have it distributed to a class of 101 psychology students. This type of sampling is known as: a) probability sampling, b) corrupted sampling, c) availability sampling, d) community sampling

C. Availability sampling

Which of the following is a probability sampling method? a) availability sampling, b) quota sampling c) cluster sampling d) snowball sampling

C. Cluster sampling

A researcher is interested in studying a population for which it would be nearly impossible to create a sampling frame because they are hard-to-identify individuals, such as prostitutes. What type of sampling method would be best for her to use? a) systematic random, b) cluster, c) snowball, d) stratified random

C. Snowball

I am a researcher interested in what happens to nonprofit agencies that provide services to homeless people after a natural disaster. To investigate this issue, I locate a homeless shelter near the New Jersey shore that has re-established their operation after Hurricane Sandy and interview the people who work there. After the interview, I ask them to recommend people who work in other shelters along the coast to be interviewed. In doing this, I am engaging in what sampling strategy? a) Stratified Sampling b) Quota Sampling c) Snowball Sampling d) Random Sampling

C. Snowball Sampling

If martha conducts a survey in each of her intro sociology courses and collects 1,000 surveys, the sample she obtains is a: a) random sample, b) quota sample, c) purposive sample, d) availability sample

D. Availability sample

Instead of drawing a sample from the population of a small town, a local polling agency simply decides to give the survey to every resident of the town. This is known as a/an: a) population survey, b) population sample, c) thorough examination, d) census

D. Census

I am a representative to the Delaware State Legislature and I want to sponsor a bill that makes assisted suicide legal for those with a terminal disease. I want to make sure the voters in my district agree with the bill so I call people who have given money to my campaign and ask them their opinions about assisted suicide. In doing this, I am conducting a _____________________. a. Simple random sample b. Stratified random sample c. Systematic sample d. Nonprobability sample

D. Nonprobability sample. This is a availability sampling.

Systematic bias is eliminated from a sampling method when: a) only random samples are drawn b) chance is eliminated as an influence on the selection of elements c) the obtained sample perfectly mirrors the population from which it was drawn d) nothing but chance affects the selection of elements

D. Nothing but chance affects the selection of elements

Random selection

Every element of the population has a known and independent chance of being selected into the sample. **Even when sample is random, it can still have some sampling error due to chance (can mathematically determine)

True or false, Nonprobability samples are typically more representative than probability samples.

False.

True or false, Nonrespondents do not affect the representativeness of a sample; it is another part of the randomness of random sampling.

False.

True or false, Proportionate stratified sampling is a probability sampling method in which elements that are selected from strata in different proportions form those that appear in the population.

False.

True or false, Simple random samples produce the most representative samples.

False.

True or false, Standing on a street corner in a city and giving a survey to everyone who passes by will give you a random sample that is representative of the city population.

False.

True or False, I am on the city council of Newark and I think there are too many businesses on Main Street that serve alcohol. To find out what other Newark residents think, I put a small survey on the Newark city website and get 10,668 responses. The majority of this sample does not think there are actually too few businesses that serve alcohol on Main Street. From these results, I can infer that the majority of Newark residents also think there are too few businesses that serve alcohol on Main Street.

False. Cannot make generalizability from non-random sample. This is a form of convenience sampling.

True or False, I want to know whether student athletes at UD get enough support for meeting the demands of both course work and collegiate sports. I have a list of all student athletes at UD and randomly select the number 7 to begin sample selection from the list. I select the 7th student on the list and then select every 7th student thereafter. In doing this, I am collecting a stratified random sample.

False. It is systematic random sampling.

Why is stratified random sampling more efficient than drawing a simple random sample?

If you plan to draw a sample of 500 from an ethnically diverse neighborhood and you used simple random sampling, you might end up with disproportionate numbers of each group. But if you created a sampling strata based on race and ethnicity, you could randomly select cases from each stratum: 75 AAs (15%), 50 Hispanics (10%), 25 Asians (5%), 350 Whites (70%).

Probability of selection

Likelihood that an element will be selected from the population for inclusion in the sample. In a census of all the elements of the population, the probability that any particular element Will be selected it is a 1.0, because everyone will be selected. If half the elements in the population will be sampled on the basis of chance (e.g., tossing a coin), the probability of selection for each element is 0.5. **When size of desired sample as a proportion of the population decreases, so does the probability of selection.

Inferential statistics

Mathematical tool for calculating sampling error; how likely it is that a statistical result based on data from a random sample is representative of the population from which the sample is assumed to have been selected.

Simple random sampling

Method in which every sample element is selected only on the basis of chance, through a random process; EQUAL probability of selection method. e.g., flipping a coin, rolling a die, lottery, random number table, or random digit dialing **Can be done with or without replacement sampling

Respondent-driven sampling

Most sophisticated version of snowball sampling. Involves giving financial incentives, also called gratuities, to responds to recruit peers. e.g., "We'll pay you $5 for 3 names, but only 1 can be from your own town." **Can reduce potential for bias

Snowball sampling

Nonproability sampling method in which elements are selected as they are identified by successive informants or interviewees. --When you identify one member of the population, speak to him or her, and then ask that person to identify others in the population so that you can speak to them. Process repeats.

Purposive (judgment sampling)

Nonprobability sampling method in which elements are selected for a purpose, usually because of their unique position (e.g., gang members). **Researcher uses his judgment about whom to select into a sample, rather than drawing sample elements randomly

Quota sampling

Nonprobability sampling method in which elements are selected to ensure that the sample represents certain characteristics in proportion to their prevalence in the population. Involved dividing the population into proportions of some groups that you want to be represented in your sample. **Similar to stratified probability samples, less rigorous and precise as they do not involve random selection (e.g., you decide you want 50% male and 50% female, but it is not random; thus representative of gender distribution, but not race distribution)

Systematic bias

Overrepresentation or underrepresentation of some population characteristics in a sample resulting from the method used to select the sample; a sample shaped by systematic sampling error=biased sample. **Random selection reduces this (e.g., flipping a coin)

Nonresponse

People or other entities who don't participate in a study although they are selected for the sample. **undermines sample quality

Sample Components in a Two-Stage Study

Prisons are the elements and the primary sampling. Inmates are the secondary sampling units; they provide information about the prisons.

Pros and cons of snowball sampling

Pro: useful for hard-to-reach or hard-to-identify populations for which there is no sampling frame, but members are somewhat interconnected Con: the initial contacts may shape the entire sample and foreclose access to some members of the population of interest

Pros and cons of cluster sampling

Pros -useful when sampling frame is unavailable -useful for large populations spread out across a wide area or among many orgs. -requires less prior information Cons -less efficient than simple random sampling -has more sampling error -the more clusters selected, higher the travel costs

Which of the following is true regarding the main difference between quota sampling and stratified probability sampling?

Quota samples are generally less rigorous and precise in their selection procedures as compared to stratified probability samples.

Random digit dialing

Random dialing by a machine of numbers within designated phone prefixes, which creates a random sample for phone surveys.

Random sampling error vs. systematic error?

Random: differences between sample and population that are due only to chance Systematic: error in method used to select the sample; resulted in biased sample in which there is overrepresentation or underrepresentation of some population characteristics in the sample.

Census

Research in which information is obtained through the responses that all available members of an entire population give to questions. -Not often used; too expensive & time-consuming; hard to get people to complete a survey

Replacement sampling

Sample elements are returned to the sampling frame after being selected, so they may be sampled again.

Systematic random sampling

Sample elements are selected from a list or from sequential files, w/ every nth element being selected after the first element is selected randomly w/in the first interval; EQUAL probability of selection method. **Almost always yields a simple random sample, except when the sequence of elements are affected by periodicity

Multistage cluster sampling

Sampling in which elements are selected in two or more stages: 1) the random selection of naturally occurring clusters 2) random selection of multilevel elements within clusters

Availability/Convenience Sampling

Sampling in which elements are selected on the basis of convenience (e.g., professor giving survey to his class, person-on-the-street interviews)

Nonprobability sampling methods

Sampling methods in which the probability of selection of population elements is unknown. **Cannot be certain selected sample is generalizable **QUALITATIVE

Probability sampling methods

Sampling methods that rely on a random selection procedure to ensure there is no systematic bias in the selection of elements. The odds of selecting elements are independent, equal, and known, and the method of selection is carefully controlled. **More desirable when goal is to generalize to a larger population **QUANTITATIVE

Unrepresentative sample

Some characteristics are overrepresented or underrepresented, and sampling error emerges.

Random number table

Table containing lists of numbers that are ordered solely on the basis of chance; used for drawing a random sample.

Sample generalizability

The ability to generalize from a sample, or subset, of a larger population to that population itself. **Most common meaning of generalizability.

Cross-population generalizability (external validity)

The ability to generalize from findings about one group, population, or setting to other groups, populations, or settings

What is the difference between a desired sample and an obtained sample?

The desired sample is the sample we wish to have (the elements that we actually want to sample); the obtained sample is the sample we actually get.

Sampling error

The difference between the characteristics of a sample and the characteristics of the population from which it was selected.

Elements

The individual members of a sample

Sampling interval

The number of cases between one sampled case and the next in a systematic random sample.

Confidence interval

The range defined by the confidence limits for a sample statistic.

Population

The set of individuals or other entities to which we want to be able to generalize our findings.

Population parameter

The statistic computed for the entire population. INFERENTIAL STATISTIC

Sampling units

The units actually selected in each stage of sampling (obtained sample) i.e., the adult individuals in the households

Confidence limits

The upper and lower bounds around an estimate of a population parameter based on a sample statistic. **Shows how much confidence can be placed in the estimate. Basically, it is the degree of confidence we can have in our findings estimated from inferential statistics (have to be at least 95% confident to make inferential statement).

Sample statistic

The value of a statistic (e.g., mean) computed from sample data. It is an estimate of an population matter. INFERENTIAL STATISTIC

Why might disproportionate stratified sampling sometimes be more efficient than proportionate stratified sampling?

To ensure that cases from smaller strata are included in the sample in sufficient numbers because you want to get a valid representation. -Also efficient if the costs of data collection differ markedly b/w strata or if the variability (heterogeneity) of the strata differs.

What is the purpose of sampling?

To generate a set of individuals or other entities that gives us a valid picture of all such individuals or entities.

Enumeration units

Units that contain one or more elements and that are listed in a sampling frame i.e., a sample of households

When might availability sampling be useful?

When a field researcher is exploring a new setting and trying to get some sense of prevailing attitudes or when a survey researcher conducts a preliminary test of a questionnaire.

Disproportionate stratified sampling

When elements are selected from strata in different proportions from those that appear in the population; oversampling. e.g., Race: 25% White, 25% Black, 25% Hispanic, 25% Asian **Probability of selection of every case is known, but unequal b/w strata

Proportionate stratified sampling

When elements are selected from strata in exact proportion to their representation in the population. e.g., Race: 70% White, 15% Black, 10% Hispanic, 5% Asian