Week Three Flashcards
What is a sampling frame and a sampling unit?
Sampling frame: the list from which the elements of the population are selected
Sampling unit: any single unit sampled from the population, primary sampling units - entitles selected in the first stage of the sample, secondary sample units - entities selected in the second stage of the sample if there is one
What are the two types of generalisability
Sample generalisability: ability to generalise from a subset of a larger population (I.e. Sample) to that population itself. This often occurs through inferential statistics
Cross- population generalisability (external validity): ability to generalise from the findings about one group, population, or setting to other groups, populations or settings
What is a sampling error
Any difference between the characteristics of a sample and the characteristics of the population from which it was drawn
The larger the sampling error, the less representative the sample is of the population
What is probability
Sample methods that allow us to know in advance how likely it is that any element of a population will be selected
Random selection: everyone in the population has a know / equal chance to be picked
Non-probability: sample methods in which the likelihood of selection is not know in advance
What does probability sampling do
Allows researchers to select study subjects to be statistically representative of population they want to learn about (generalisability)
What is simple random sampling?
Identifies cases strictly on the basis of chance eg random number tables
Typically every subject has equal probability of being selected
CANNOT do this without a sampling frame that contains every element of the population
Can be done with or without replacement sampling
What is replacement sampling
Each element is returned to the sampling frame from which it is selected so that it may be sampled again
What is systematic random sampling?
The first element is selected randomly from a list, and then every nth element thereafter is selected
What are the three steps of systematic random sampling?
Calculate sampling interval: the total number of cases in pop is divided by the number of cases required in the sample
Identify the first case to be selected: a number within the sampling interval is randomly selected and used to select the first case
Selection of subsequent cases: every nth case is selected, wherein is the sampling interval
What is stratified random sampling
Uses information known about key characteristics of the population prior to sampling to make the process more efficient
Often used to ensure that subjects who have a Rare or uncommon characteristics are adequately represented in the sample
What are the two steps of stratified random sampling?
- Distinguish all elements in the populations (I.e. In the sampling frame) according to their value on some relevant characteristic.
- that characteristic forms the sampling strata
- each element must belong to one and only one stratum - Sample elements randomly from within each strata
What is proportionate and disproportionate stratified sampling
Proportionate ratified sampling: every group is equally proportional to its size in the population
Disproportionate stratified sampling: proportionate of each group is unite nationally varied from population - useful to include small or underrepresented groups
What is multistage cluster sampling
Sampling in which elements are selected in two or more stages
- Clusters: preexisting groups of elements in a population - a random sample of clusters is selected
- Within clusters: within each cluster a random sample of elements is selected
What is non probability sampling
Each member of population is an unknown probability of being selected
What are the four types of non probability sampling
Availability or convenience
Snowball
Quota
Purposive/judgment
What is convenience sampling
Elements are selected because they are available or easy to find
What is quota sampling
Elements are selected to ensure that the sample represents certain characteristics of the population
Similar to stratified probability sampling but generally less rigorous and precise in selection procedures
What is purposive sampling
Each sample element is selected for a purpose. Usually because if the unique position of the sample elements
May involve studying the entire population of some limited group or a subset of a population
What is snowball sampling
Sample elements are selected as they are identified by successive informants
Useful for hard to reach populations in which members are interconnected
What is random sampling error
Refers to the difference between a sample statistic and the true value in the population due solely to chance
What is sampling distribution
Statisticians estimate the probability that a sampling result is representative of a population using this
Estimates the margin of error and confidence interval around the sample statistic
Sampling distribution for many statistics is a normal or bell shape
What is systematic sampling error
Refers to differences between the sample and population due to some problem with the sampling method
What is ecological fallacy and reductionism
Eco: a researcher draws conclusions about individual level processes from group level data
Reductionism: a researcher draws conclusions about group level processes from individual level data
What are the two types of causal explanations
Quantitative (nomothetic)
Qualitative (idiographic)
What is nomothetic causal explanation?
Quantitative
Involved the belief that variations in an independent variable will be followed by variation in the dependent variable, when all other things are equal (ceteris pariaus)
The value of cases on the dependent variable differs from what their values would have been in the absence of variation in the indecent variable (counterfactual)
What is the idiographic causal explanation
When a series of concrete events, thoughts or actions result in a particular event or individual outcome
Includes statements of inital conditions and then relates a series of events st different times that led to outcome
Explanation is concerned w context and deterministic
Focus on time order and causal mechanism
What is the five criteria for causation in research design
Empirical association (required)
Time order (required)
Nonspuriousness (required)
Mechanism
Context
Why is empirical association
Variation in one variable is related to variation in another
Relationship must be observed - cannot be assumed or believed
The independent variable and the dependent variable must vary together
Same direction of variation - positive associations
Opposite direction - negative associations
What is time order
The change in the independent variable must occur before the change in the dependent variable
What is nonspuriousness
A relationship between two variables that is not due to variation in a third variable
When this third variable (an extraneous variable) cause the variation, it creates a spurious relationship between the independent and dependent variable
What is causal mechanism
Process that creates the connection between variation in an independent variable and the variation in the dependent variable it is hypothesised to cause it
What are the two experimental research designs and two non-experimental?
Experimental:
True experiments
Quasi experiments
Non experimental
Cross-sectional
Longitudinal
What are the three requirements of true experiments
- An experimental group that is reviewing condition and one control group that is not
- Must have random group assignment
- Must have assessment of change (pre and post) in both groups both change is elected
Research design is quasi-experiment when all but one requirements are met
Explain cross sectional designs
Observations are made at one time point
Independent variable is fixed at some point prior to the variation in the dependent variable
You believe respondents can give you reliable reports of what happened to them or what they thought at some earlier time
Measures are based on accurate records that contain information in cases in earlier periods
You know that the value of the dependent variable was similar for all cases prior to the treatment
What are the three types of longitudinal design
Repeated cross sectional design
Fixed sample panel designs
Event based designs
What is repeated cross sectional design
Trend
Data is collected at two or more points in time from different samples of the same population
Appropriate design when the goal is to determine whether a population has changed over time
What is fixed sample panel design
Data are collected from the same individuals (the panel) at two or more points in time
Issues: cost and subject attrition - leaving
What is event design
Event or cohort based design
Data are collected at two or more point in time from individuals in a cohort
Types of cohorts: birth, seniority, school etc
Can be repeated cross sectional design or a type of panel design
What is individual and aggregate matching
Two selection methods for quasi experimental
Individual: individual cases in the treatment groups are matched with similar individuals in the control group
Aggregate: using group level data, the treatment group is matched to similar control group
What is internal validity and it’s five sources
Ability to yield valid conclusions is determined by comparability of experimental and control groups.
Sources:
Selection bus
Endogenous change
External events
Contamination
Treatment misidentification
What is selection bias
Treatment and comparison group are different at beginning or end of study - one group has more experience, education etc
Differential attrition is one type of selection bias: groups become different because for various reasons subjects are more likely to drop out of groups
A pretest helps determine whether selection bias exists and controls for it
What is endogenous change?
Occurs when natural developments in the subjects, independent of the experimental treatment, account for some or all of the observed changes between pretest and post-tests
Three specific threats:
- testing: pre-test influenced post-test scored because subject learn something earlier and react different
Maturation: part of natural maturational experience, subjects respond differently on the post-test than they did in the pre- test
Regression to the mean: subjects experience cyclical or episodic changes that result in different posttest scores
What is contamination
Occurs when the control group is in some way affected by or affects the treatment groups
Two types:
Compensatory rivalry: control group member works harder to make up for not having treatment or intervention
Demoralisation: control group member feels left out and becomes worse than he/she would have been without study
What is treatment misidentification
Occurs when the treatment itself does not cause the outcome but does cause some intervening process the researchers is not aware of
Sources:
The halo effect: staff too excited toward certain group
Placebo effect: improve because of expectation not treatment
Hawthorne effect: subjects may change because more attention is being paid to them
What is external validity
The application of a treatment across subgroups within an experiment and/or across different populations, times or settings
Sample generalisability
What is factorial survey design
Can increase generalisability
Embed the features of true experiments into a survey design in order to maximise generalisability
Randomly selected subsets of respondents are asked different questions or asked to respond to different vignettes
Allows researchers to determine the causal effect of the variables represented by these differences
What is the Institutional review board?
Committee that review research proposals to ensure that research is abiding by ethical principles
Required for every institution that seeks federal funding for biomedical or behavioural research on human subjects