Chapter 6 Flashcards
Secondary Data Sources
I. Simmons Market Research Bureau, Inc.,
provides data on product, brand and media
usage
II. Mediamark Research, Inc. (MRI) provides demographics and media use by product category and brand
III. Standard Rate and Data Service
IV. Nielsen provides demographic data for television programs
V. Arbitron offers demographic breakdowns for radio programs
What do the terms population or universe mean?
all the people who possess a particular characteristic,
or in the case of texts–all the messages that share a characteristic of interest
What is a cross-sectional sample?
sample drawn at one point in time
What is a panel sample?
select a sample and follow its members over time, returning to study it more than once
What is a cohort-trend sample?
researchers go back to the field to gather data, but draw new samples each time
What are the probability sampling methods?
Random Sampling, Systematic or Ordinal Sampling, Stratified Sampling, Cluster Sampling.
What is a target population/public?
the population/public of interest
What does a sample mean?
a sub-set of the population chosen by the researcher to study because it would be difficult, if not impossible to contact all members of the population
Sampling Procedure
Define the target population carefully/determines who is in or out of the study.
Random Sampling -
Chance alone determines who is selected. Every person in a population has an equal chance of being included.
Systematic or Ordinal Sampling -
Often used with large populations, because it’s easier to use with long lists
Stratified Sampling -
Categorizes a population with respect to a characteristic that a researcher considers
to be important, creating two groups then samples randomly from each group. Results in an “overly large sample” of minority group.
Cluster Sampling -
Sampling frame is made up of lists of clusters. Clusters are pre-existing natural or administrative groups in the population *can be geographic locations *other common groupings, like universities. Researcher first samples clusters, then individuals from the clusters chosen. Often used when lists of individuals can’t be identified.
Sampling
Population or universe, Target population/public, Sample, Sampling Procedure,
Population or universe
- all the people who possess a particular characteristic,
- or in the case of texts–all the messages that share a characteristic of interest
Sampling Procedure
- Define the target population carefully/determines who is in or out of the study.
- The researcher cannot study everyone in a population.
- He/she selects some individuals to examine, arguing that the sample will represent the entire population closely.
- Obtain a sampling frame (a “list” of everyone in the target population). -use only a “fresh” list from a reputable source
- Prepare the list.
- purge duplicates
- examine the structure of the list, consider randomizing if structure creates a bias - Use the margin of error to determine how many names to draw from the sample.
- remember, not everyone contacted will participate in the data gathering effort (non-response rate)
- draw enough names so that enough completed questionnaires are returned to meet the requirement of the margin of error
Cross-sectional sample
sample drawn at one point in time for a cross sectional study
Panel sample for longitudinal studies
select a sample and follow its members over time, returning to study it more than once
Cohort-trend sample for longitudinal studies
researchers go back to the field to gather data, but draw new samples each time
Random Sampling
Chance alone determines who is selected, research does not make any decision as to whom is selected. Every person in a population has an equal chance of being included.
Simple Random Sampling
- Randomly selecting individuals from a sampling frame
2. Like pulling names from a hat
Systematic or Ordinal Sampling
- Choose every nth person from a complete list of the population after a randomly selected starting point
- Often used with large populations, because it’s easier to use with long lists…
- Avoid biases in the structure of the list
Stratified Sampling
1.Categorizes a population with respect to a
characteristic that a researcher considers
to be important, creating two groups
2. Then samples randomly from each group
3. Results in an “overly large sample” of minority group
4. Weights scores to compensates for oversampling
Cluster Sampling
Sampling frame is made up of lists of clusters
- Clusters are pre-existing natural or administrative groups in the population
* can be geographic locations
* other common groupings, like universities - Researcher first samples clusters, then individuals from the clusters chosen
- Often used when lists of individuals can’t be identified
Probability Sampling
Random Sampling, Simple Random Sampling, Systematic or Ordinal Sampling, Stratified Sampling, Cluster Sampling
Nonprobability Sample
chance does not guide selection. Researcher selects subjects for some particular reason or because of opportunity.
Types of Nonprobability Samples
Convenience Sample, Volunteer Sample, Purposive Sample, Quota/dimensional Sample, Network/snowballing Sample
Convenience Sample
based on availability of respondents
Volunteer Sample
respondents choose to be a part of the study
Purposive Sample
respondents are non-randomly selected on the basis of a particular characteristic
Quota/dimensional Sample
respondents selected non-randomly based on their known proportion of a population
Quota=one criterion of sample chosen and a predetermined number of subjects selected to match each
dimensional=several criteria of sample chosen and a predetermined number of subjects selected to match each combination of those criteria
Network/snowballing Sample
respondents refer other respondents to participate in the study
Sampling Frame
a “list” of everyone in the target population
