methods Flashcards
Case control survey
a survey design where the statisitcal population is divded into a case and control group based on a response variable to study how other facotrs might be assoicated with the observed response
cluster survey
- a survey done to average over unwanted heterogenity by seperating the sampling unit (cluster) from the observation unit
- Google definition: researchers divded a population into smaller groups known as clusters, then randomly select among these clusters to form a sample
cohort survey
a survey design where sampling units are selected at random from the statisitcal population and then followed over time, looking for a change in the response variable
Confounding variable
(typically) an observed variable that can affect both the explantory and response variable causeing an incident or spurious relationship
cross sectional study
a snapshot of the statisitcal population at a single point in time
experimental study
a study involving a statisitcal pipulation where the investigators can control the experimental variables
explanatory variable
a variable that an investigator beleives may explain the response variable
longitudinal study
follows the selected sampling units over time
Observational study
a study using observations from a statisitcal population where the investigator has no control over the explanatory variables
prospect study
selects sampling units before the respinse of intrest has occurred
response variable
a variable that the investigator is intrested in studing as a was to answer a research question
retrospective study
selects sampling units after the response of intrest has occurred
simple random survey
a survey design where the sampling units are selected at random from the statisitcal population
spurious relationship
when a relationship between the explanatory and response variables is largely driven by the influence of a confounding variable
strata
the name given to a subgroup within the statisitcal population in a stratified survey
stratified survey
survey design in which the researchers first defines strata that represent nautral grouping within the statisitcal population and the proceeds to do a simple random survey within each strata
finding 1st quartile
creating a subset of the data then finding the median for lower half
3rd quartile
creating a subset for the data then finding the median for the upper half
Interquartile range
difference between 3rd and 1st quartile and gives the range at the innermost 50% of a numerical sample