U9 Flashcards
Population
Entire aggregation of cases that meet specified criteria
Eligibility Criteria
Inclusion Criteria
Specified criteria that delimit the study
Target Population
The entire population which a researcher is interested
Accessible Population
Comprises cases from the target population which are most accessible
Sample
A subset of a population
Elements
Entities that make up the samples and populations
Sampling bias
systematic overrepresentation or underrepresentation of some segment of the population in terms of characteristics relevant to the research question
Strata
mutually exclusive segments of a population based on a specific trait
Sampling Designs
Quant vs Qual
Quant plan ahead, Qual plan on the go
Qualitative Sampling Designs
1 - Nonprobability Sampling -selection of elements by non-random methods. can’t estimate probability of including each element and every element does not have a chance for inclusion.
2 - Probability Sampling - random selection of elements from a population, where the random assignment is after population selection.
Nonprobability Sampling
selection of elements by non-random methods. can’t estimate probability of including each element and every element does not have a chance for inclusion.
CONVENIENCE Sampling risks sampling bias, available subjects may be atypical of the population.
SNOWBALL Sampling - sample members are asked to refer others with the eligibility criteria.
QUOTA Sampling - researchers identify strata of the population and then determine how many participants are needed from each stratum to meet a quota
PUPOSIVE Sampling - based on belief that researchers knowledge about the population can be used to hand pick the cases to be included in the sample
Evaluation of Probability Sampling
avoids conscious/unconscious biases
- allows for estimation of sampling error
- expensive and demanding
- usually beyond scope of most researchers to draw a probability sample without a narrowly defined sample.
Probability Sampling
SIMPLE RANDOM Sampling - most basic probability sampling design. Population is defined, then researchers establish a sampling frame (actual list of population elements), the elements in the frame are numbered consecutively, table of random numbers or computer is used to randomly draw numbers till desired sample size is reached.
No researcher bias, no guarantee sample is representative of population, but guarantee it is by chance. larger the sample the better.
Laborious process
STRATIFIED RANDOM Sampling - population divided into homogenous sub sets from which elements are randomly selected.
-Aims to enhance samples representativeness, usually group the elements of the startup and randomly select the desired number of elements, may select proportionate samples or disproportionate (weighting can be used to analyze results)
CLUSTER Sampling - most common for national surveys.
SYSTEMATIC Sampling - involves selection of every Kth case from a list or group.
Evaluation of Probability Sampling
avoids conscious/unconscious biases
- allows for estimation of sampling error
- expensive and demanding
- usually beyond scope of most researchers to draw a probability sample without a narrowly defined sample.
Sample Size in Qualitative Studies
Larger sample always = more representative of the population
Power Analysis
aids researchers to estimate how large their sample should be to test their research hypotheses adequately.
a large sample can not correct for a faulty sampling plan
Sampling in Qualitative Sampling
typically small and nonrandom also an emergent design Convenience and Snowball Sampling Purposive Sampling Theoretical Sampling
Sampling in the 3 main Qualitative Traditions (Ethnographic, Phenomenological, and Grounded Theory)
Ethno- big net approach, mile with as many members of the culture as possible, yet often rely on few key informants
Phen- typically less than or equal to 10 participants, all must have experienced the phenomenon with ability to articulate it
Grounded Theory - 20-30 participants with theoretical sampling to develop and refine categories and theory.
Critiquing Sampling Plan
Particular scrutiny with Quantitative as can lead to bias and decrease confidence in results
1-did researcher describe sampling strategy adequately?
2- did researcher make good sampling decisions?
Quantitative - response rates, number of participants relative to number samples, nonresponse bias - difference btw participants those who declined. if target population defined broadly, there may be missed opportunities to control extraneous variables and gap btw accessible population and the target population too great
Qualitative - adequacy and appropriateness (sufficient and quality data and the methods to gain the data)
-good to see participants with unusual or negative cases
fittingness - degree of congruence btw study sample and ppl at another site
-transferability