Chapter 1 Flashcards
Continuous variables
Numbers that are continuous can be broken down l,
Discrete
Number that has a jump between it & the next - a count
Nominal
Categorical variable without order
Ordinal
Categorical variable with ordering
Explanatory variable
The variable that is suspected to affect another variable
Response variable
The variable that is effected by the explanatory variable
Observational study
Observes data without intervention. We cannot infer causation from an observational study
Experimental study
Individuals randomized into groups, tested for interventions vs placebo – principals of experimental design are: controlling ( for differences between the groups), randomization, replication (collecting sufficiently large samples and/or repeating the entire study ), blocking ( grouping individuals based on suspected blocks & then randomizing)
Sample statistic
Subset of cases of target populations
Population vs sample.
.population is the entire group - sample is a subset of cases & usually a small fraction of the population
Blocking variable /
Variable that maybe correlated with both explanatory & response variables so the sample groups are broken up to ensure representation in control E intervention
Cluster sample
Population broken into Clusters with a fixed #e sampled and observations collected from all sample
Simple Random sample
Randomized selection from 1 large group-most basic random sampling technique. Everyone has an equal chance of being selected
Stratified sampling
Population divided into groups (strata) usually into similar group, then a second randomizing method is performed - useful when the cases is each stratum are very similar with respect to the outcome of interest
Multistage sampling
Cluster sampling but from a random sample from chosen clusters
Convenience sampling
Biased method of easier data collection - more accessible individuals are more likely to be selected
Double-blind
Neither patients nor researches that interact with them know which group they are in ( treatment vs placebo) to reduce bias. The people interacting with the individuals being studied can’t know who is in which group to be double blind
Summary statistic
Single number summarizing a large amount of data
Associated vs independent variables
Associated variables are related in some way, independent are not. Variables can not be both
Confounding variable
A variable that is correlated to both the explanatory & the response variables - no guarantee that all confounding variables can be accounted for in observational studies which is why we can’t make causal conclusions from them
