Ch 7: Selecting Sample and Setting
What are 3 threats to external validity?
- interaction of selection & tx - interaction of history & tx - interaction of setting & tx
What are the types of settings?
- natural settings (field settings) - controlled environments (partially or highly)
natural setting (field settings) are used with what type of Research studies?
- often in qualitative - quantitative (descriptive or correlational)
Non probability major disadvantages?
- rarely representative of population - sampling bias -results could be misleading
How do you determine the adequacy of the sample?
- sample size - external validity (interaction of selection and treatment, interaction of history and treatment, interaction of setting and treatment)
What are the 4 types of Probability sampling?
1. Simple Random Sampling 2. Stratified Random Sampling (Proprortional, Disproportional) 3. Cluster (Multistage) sample 4. Systematic sample
1. The larger the sample size, the ___ the sampling error. 2. The larger the sample size, the__ the chance of detecting statistical significance. 3. The __ confidence in the relationship tested, the larger the sample size should be.
1. smaller 2. greater 3. less
T/F Population is restricted to humans.
False - not restricted to humans It can include: hospital records, blood, cases of occurrences, animals, objects, events, etc.
How large should my sample size be?
it depends... - large enough to be an accurate representation of populaiton - large enough to achieve statistically significant resutls
The __ the sample size, the greater the ability to detect significant results.**
larger
What is a representative sample?
main characteristics closely approximate those of the population
loss of study subjects during course of data collection
mortality (threat to internal validity)
A type of setting in which the researchers go where the participants are, in their environment. The researcher does not manipulate the setting in any way.
natural setting (field setting)
This type of sampling plan has no way to ensure that every member of the target population has a chance for inclusion.
nonprobability
Which is easier to obtain- probability sampling plans or non probability sampling plans?
nonprobability
What is the major limitation with purposive sampling?
nonprobability - limits ability to generalize the findings
This is a type of sampling plan in which subjects are chosen by nonrandom methods.
nonprobability sampling
This type of sampling does not involve section of elements at random.
nonprobability sampling
ICU floor - setting
partially controlled
A type of setting in which the researcher manipulates/ modifies the setting in some way, such as using an ICU to collect data.
partially controlled setting
"All U.S. Nursing students enrolled in BSN programs" is an example of __
population
A well-defined set that has specified properties or characteristics
population
The aggregate of cases in which a researcher is interested
population
The purpose of ____ is to estimate how large the sample size should be to adequately test the research hypothesis. Must consider attrition (jail, death) and missing data.
power analysis - most researchers increase sample size by ~15% to account for missing data and attrition.
Formulas based on number of variables and study design to determine best sample sizes***
power analysis = estimates sample size
A sampling plan that involves random selection of elements; each element has an equal, independent chance of being selected.
probability
Sampling strategies are generally grouped into either ___ sampling or ___ sampling.
probability sampling or nonprobability sampling
What type of sampling is the "gold standard" for quantitative research?
probability sampling - involves random selection of elements from population
Would you be more likely to get a representative sample from probability sampling or non probability sampling?
probability sampling (involves random selection)
What type of sampling? - Reseracher's knowledge about population can be used to hand pick sample members - researcher selects (the "judgement" from researcher) subjects thought to best represent the phenomenon being study and are typical of a population - used when a highly unusual group is being studied - limits ability to generalize findings
purposive sample (non probability)
What type of sampling would be best to do a study on a rare genetic disease or a certain type of cancer?
purposive sampling (nonprobability) bc. highly unusual group is being studied so need research knowledge to hand select subjects
Nonprobability sampling plans are appropriate for what type of methods? quantitative or qualitative
qualitative
Purposive sampling can be used in quantiative or qualitative research?**
qualitative - it is the backbone of qualitative research
What type of sampling is this an example of? Nursing school has 10% males and 90% females. Research obtains 10 males and 90 females for study.
quota (non probability)
What type of sampling? - need some knowledge about the population in order to build a representative sample - researchers identify population strata and determine how many participants are needed from each stratum - researchers guide selection of subjects sot hat the sample includes an appropriate number of cases from each strata - increases representativeness - not randomly selects; subjects are recruited until the certain amount of subjects is filled
quota (non probability) - recruit until quote is filled
What is the difference between quota sampling and stratified random sampling?
quota = nonprobability = subjects are not randomly selected for each stratum
When it is not possible to randomly select, then you apply ___ to help strengthen and control bias. This helps to enhance representativeness to the population, because when you have representativeness, you have generalized ability to apply****
random assignment
In cluster sampling, selection of individuals from within the clusters may be done by ___ or ___ sampling.
random or stratified sampling - may be impossible to obtain the sampling frame and perform a true random sample.
What is the difference between random selection vs. random assignment?
random selection = equal, independent chance of being selected (refers to how individuals are chosen to participate in a study). NOT a prerequisite for random assignment Random assignment = random allocation of subjects to either experimental or control group.
What is the main characteristic of probability sampling?
random selection of subjects
Probability involves _____ of elements. Each element has an equal, independent chance of being selected.
random selection.
Major consideration in quantitative studies is _____, sample whose characteristics closely approximate those of the population. **
represenativeness
When sampling is conducted properly, researchers can draw inferences and make generalizations about the population through examination of the sample. This is called a ___ sample.
representative
This is a sample in which key characteristics closely approximate those of the population.
representative sample
This is a type of sample in which the main characteristics closely approximate those of the population
representative sample
Setting is based on ___.
research design
A set of elements that make up the study population
sample
The goal of Stratified Random Sampling is to enhance representation, thus reducing the risk that___
sample might be underrepresented of population
This is the number of study participants in the final sample (the final number; not the number asked to participate bc does not include those who started but stopped)***
sample size
What is a key determinant of sample quality in quantitative research?**
sample size
The process of selecting an individual for a study in such a way that individual represents the larger group from which they are selected.
sampling
This is the process of selecting portion or subset of the designated population to represent the entire population
sampling
Distortions that arise when a sample is not representative of the population from which it is drawn
sampling bias Data may not accurately represent the target population.
In Simple Random Sampling, the researcher uses a _____, which requires them to make a list of all population elements (must define population and list all of the elements of the population).
sampling frame
What are two threats to internal validity?
selection bias mortality instrumentation maturation history testing
This can result from the way subjects were selected; can occur if sample selected in way the sample is not representative of target population
selection bias (threat to internal validity)
The site or location used to conduct a study (recruit participants, collect data, etc.)*
setting
What type of sampling? - early sample members are asked to refer other people who meed the criteria - used when located people who meet the inclusion criteria is difficult
snowball/ network sampling (non probability)
What type of sampling would be best for a study of undocumented immigrants?
snowball/ newtork sampling (nonprobability)
What type of sampling? - Populated divided into strata. Each strata group should be homogenous; then random selection from the stratified sampling frames
stratified random sampling (probability)
Inclusion and exclusion criteria are established to control for extraneous variability or bias that would limit the __ of the evidence.
strength
The entire population of interest is called the ___ population.
target
The ___ population = all children with asthma in the US. The __ population = children with asthma in NOLA clinics
target accessible
The entire set of elements about which the research would like to make generalizations
target population
What is the theoretical populaiton?
target population
Where do we get our accessible population from? **
target population
T/F Probability sampling is more rigors than non probability sampling.
true - probability sampling involves random selection
These errors are called "underpowered" because in reality, intervention did have an effect, but it was reported as no difference or no effect because insufficient sample size.
type 2
Researcher may commit a ___ error if sample size is NOT large enough.
type II error
What is a big advantage of Strafed Random sampling?**
- enhances representativeness - comparison can be made among groups
The portion of the target population that is accessible to the researcher. This is the population form which a sample is drawn.
Accessible population
Where do we draw our sample from? **
Accessible population
The extent to which study results can be generalized from the study sample to other subjects, populations, measuring instruments, and settings.
External validity
Is it feasible to study an entire target population?
No
What are some advantages and disadvantages of Simple Random Sampling?
Probability Pro = selection is not subject to conscious biases of researcher; representativeness of sample in relation to population characteristics is maximized; differences in characteristics of sample and population are purely a function of chance and the probability of choosing a non-representative sample decreases as the size of sample increases con = time consuming; may be impossible to obtain an accurate or complete listing (sampling frame) of every element of accessible population
What is the only viable method of obtaining representative samples? ***
Probability why? - if all elements in the population have an equal probability of being selected (RANDOM SELECTION), then the resulting sample is likely to do a good job of representting the sample.
In Stratified Random Sampling, you can sample ___ or ___ from the strata.
Proportionally or disproportionally Ex: 100 studies: 10% AA, 5% hispanic, 5% asian, 80% white proportionate = 10, 5, 5, 80 disporportionate = 20, 10, 10 , 60
The process of selecting representative units of a populations in a study to represent the entire population
Sampling
Why is setting important?*
Setting is based on design, and can have important consequences (controlling extraneous variables). Settings interact with participants and can influence results.
What type of sampling? - Uses a sampling frame (a list of all population elements) - Involves random selection of elements from sampling frame - Cumbersome - Not used in large, national surveys - very controlled process
Simple Random Sampling
What type of sampling is this an example of? The researcher wants to examine reasons for nursing turnover among employed RNs in LA. She determines the desired sample size. She obtains list of 40,000+ employed RNs from LSBN. The computer program generates subjects randomly selected.
Simple Random Sampling (Probability) - sampling frame - involves random selection
What type of sampling is this an example of? - LSUHSC student strata would include: nursing, allied health, medicine, dentistry, graduate studies and public health. These subjects would then be randomly selected from each of these strata.
Stratified Random Sampling (Probability)
What type of sampling is this an example of? The researcher selects every 3rd patient on a list of HF patients.
Systematic Random Sampling (probability)
What type of sampling? - First subject randomly selected - Then selection of every Kth case from a list. - some may argue this is not entirely random
Systematic Random Sampling (probability)
I want to conduct a study of coping with depression in person with HIV. What would my target population be? What would my accessible population be? Where do I get my sample from?
Target = all persons with HIV infection who have depression (impossible to contact them all for study) Accessible = those attending ambulatory HIV clinic, psychiatric clinics, hospital units anywhere I could access person tin HIV who are depressed Draw sample from my accessible population sites.
What is the list of all the element that you would be able to select in order to do random sample?
The sampling frame
T/F You don't have to be randomly selected to be in an assignment that is randomly assigned.
True
T/F Simple Random Sampling is not the same as random assignment**
True simple random sampling uses random selection of ELEMENTS from sampling frame random assignment is when researcher randomly assigns participants to groups in experiements
T/F You can never know the entire characteristics of a population.
True - it is always changing, dynamic, too big. This creates sampling bias- never having the entire population.
T/F Sampling is a complex process, and it is NOT feasible to study whole population.
True - we study a segment of the whole population
What are the 4 types of Nonprobability Sampling?
a) Convenience (accidental) b) snowball/ network c) quota sample d) purposive sample
Set of elements readily available to the researcher and represents the target population as closely as possible
accessible population
What is the study population?
accessible population
In Simple Random Selection, why is the sample size important?
because the probability of choosing a non-representative sample decreases as sample size increases
Where can you find sample and setting of research in the article?
both should appear in methodology section - description of sample (demographics, sampling strategy) - description of setting
What type of sampling is this an example of? In drawing a sample of nursing students in the US, the researcher first draws a random ample of nursing schools. Then, he draws a random sample of students from those schools.
cluster sampling (probability)
What type of sampling? - You first randomly chose the groups form the population. - Then you random choose individuals from the groups.
cluster sampling (probability)
What type of sampling? - multistage sampling - GROUPS, not individuals, are randomly selected. - used for convenience when the group is very large or spread out geographically - impossible to et a listing of all elements - widely used in national surveys - feasible and economical
cluster sampling (probability)
What type of sampling is this an example of? Distribute questionnaires to the first 100 people who walk into the store
conveience/ accidental (non probability)
What type of sampling has the greatest risk of bias?
convenience/ accidental (non probability)
What type of sampling? - use of the most readily available subjects/ people - easy to obtain - subjects may be atypical of the population - greatest risk of bias - most common type of sampling used in nursing research
convenience/ accidental (non probability)
What is the most common type of sampling used in nursing research?
convienence / accidental (non probability)
What type of research studies is partially controlled settings used with?
corerlational quasi experimetnal experimental
Random selection means... and is associated with this type of sampling...
each element in the population has an equal, independent chance of being selected. probability sampling
The study must specific _____, which are the characteristics or attributes that define the population.
eligibility/ inclusion criteria
inclusion or exclusion criteria? - those characteristics or attributes that restrict the population from the study; deliminations
exclusion
A representative sample gives us confidence in _____, meaning we have the ability to take results from sample and refer those to general population.
generalizability
The most important relationship among Target population, Accessible population and Sample is to get a representative sample in order to have _____.**
generalizability
It is important that sample represents the population that is being studied in order to make ___ about the findings. If it is not representative, can arrive at false conclusions.
generalizations
___ settings have the ability to control extraneous variables allowing the research to examine the effect of IV on DV.
highly controlled
A type of setting described as an artificially constructed environment developed sole for the purpose of research such as research units, research centers and laboratories.
highly controlled setting
Louisiana cancer research consortium - setting
highyly controlled
We want a ____ group of subjects.
homogenous
inclusion or exclusion criteria? - Must speak English and be between the ages of 20-40.
inclusion
The researcher needs to identify what characteristics are needed to be included in the study that define the population. These form the basis for ___ and ___.
inclusion and exclusion critera
This is a threat to external validity: - are findings from a study on cholesterol done 20 years ago generalizable to today's population?
interaction of history and treatment
This is a type of threat to external validity: - can you generalize results to different periods in the past or future? - can answer this question with replication studies that corroborate an establish relatioshpi btw results form previous studies and current/ future studies - always consider WHEN a study was conducted, and the context.
interaction of history and treatment
This a type of threat to external validity: - subjects are sampled based on specific characteristics (age, race, diagnosis). - cannot generalize results to those without characteristics - those who participate may be different from those who refuse, so report numbers invited and refused.
interaction of selection and treatment
This is a threat to external validity: - Can you granulize results of an exercise program from a rehabilitation hospital to nursing home setting?
interaction of setting and treatment
This is a threat to external validity: - differences in settings can limit geenralizability - can answer this question through replication studies in different settings
interaction of setting and treatment
Threats to ____ asks whether the IV really made the difference in the DV
internal validity