Flashcards in Chapter 4 Deck (33)
When the unit of data is an individual, we refer to that individual as a participant
______________ means the same thing as participant
_________________ always use the term participant
is the collection of participants from whom data have been collected in a study
population (or target population or universe)
is the group of individuals to whom we wish to generalize the results of our study
the characteristics (e.g., gender, grade, subject matter) that define the population
Survey population or Sampling frame
the list of actual individuals you will choose your sample from. While the population may be theoretical, the survey population must be finite
A technique in which each member of the survey population has an equal chance of appearing in the sample. Random sample will represent their populations if the sample size is large enough.
Margin of Error
An interval within which the true population value lies. Typical values are 95% and 99% confidence intervals.
Simple Random Sampling
Every member of the population is enumerated. The sample is randomly selected from this list.
Every nth element of the sampling frame is selected for inclusion in the sample.
________________ can be biased if some characteristic of the sampling frame is related to the sampling cycle. So if you choose every 30th student, and class sizes are just over 30, you may always choose a student in the back of the class.
Certain characteristics (i.e., strata) of the population are identified for representation in the sample. Once the strata are identified, they are represented in the sample at the same percentage as they exist in the population (proportional stratified sample). The strata are filled through a random sampling procedure. It is also possible to represent strata in percentage different than they exist in the population (e.g., whites and blacks at 50% each; disproportional stratified sampling). This might be done when certain subgroups are to be compared in later analyses.
The sampling of naturally occurring groups (i.e., clusters) of participants. Participants can then be drawn randomly from the sampled clusters (e.g., students from within classrooms). This procedure works, but special statistical techniques (e.g., hierarchical linear modeling) must be used to analyze the data, since participants from within a cluster are more similar to each other than participants from different clusters.
In non-probability sampling . . .
the probability of each individual in the survey population being included in the sample is not equal.
________________ typically use non-probability sampling techniques
The sample is selected because of its easy availability. Typically, convenience samples will be biased.
Basically, this is a stratified sampling approach in which the strata are filled with convenience samples rather than random samples.
Purposive Sampling (or Judgment or Judgmental Sampling)
The research selects participants because they will be particularly informative about the topic under investigation.
Participants are selected on the basis of identified characteristics that will provide needed information
Typical Case Sampling (Model Instance Sampling)
Participants are selected on the basis of being “typical” or “representative”
Extreme Case Sampling
Participants are selected on the basis of being “unique” or an “outlier”
Maximum Variation Sampling (Maximum Heterogeneity Sampling)
Participants are selected to represent both ends of a continuum under investigation
Snowball Sampling (Network Sampling)
One begins with a few participants, and these participants nominate other participants, and these recommend still other participants. This creates an obvious lack of independence among the participants.
Critical Case Sampling
Participants are chosen on the basis of a dramatic example of the phenomenon being studied
Purposeful Random Sampling
A random sample of small size from the population is used as a purposeful sample
Two independent samples are taken concurrently: a random sample for the quantitative part of the study, and a purposeful sample for the qualitative part
Samples are taken that are representative of different levels of aggregation that comprise the overall population (e.g., district, school, and classroom). This is sometimes called nested, because groups are contained within one another.
Stratified Purposeful Sampling
Type I: First collecting a stratified probability sample, and then taking a purposeful subsample from this original stratified sample
Type II: First collecting a purposeful sample, and then stratifying this sample on a variable of interest, and finally taking a subset of the original sample using the stratification variable as part of the selection process