STAT 121 Test Study Questions Verified Answers
observational study
values of variable(s) of interest are recorded as they naturally occur, NO TREATMENTS
ASSIGNED,
almost impossible to establish causation (ex: sample survey)
experiment
researchers control explanatory variable to see how it will affect the response
variable; TREATMENTS ARE ASSIGNED; if valid, it can attribute causation
explanatory variable
the variable that is manipulated, the method/treatment
response variable
the characteristic measured (ex: success/failure)
lurking variable
outside variable (other than explanatory) that may affect the response variable -> create a
control or comparison group
open questions
subjects write whatever they want to answer questions; unlimited responses, hard to analyze
but subjects can answer more freely and specifically
closed questions
subjects choose from possible answers specified by the researcher; limits responses, easier
to analyze; must be sure to consider all reasonable options and provide an "other" option
causes of bias (sampling)
, - undercoverage
causes of bias (non-sampling)
- misleading response
- non-response
- interviewer effect
- question order
- question wording
factor
planned explanatory variable (includes treatment and placebo groups)
treatment (ttt)
different values of the explanatory variables administered to subjects
control group
subjects who received no specific treatment
replication
apply each treatment to more than one subject of each treatment group -> reduces the chance
of variation from the population in the results (NOT the same as reproDUcing)
observational study
values of variable(s) of interest are recorded as they naturally occur, NO TREATMENTS
ASSIGNED,
almost impossible to establish causation (ex: sample survey)
experiment
researchers control explanatory variable to see how it will affect the response
variable; TREATMENTS ARE ASSIGNED; if valid, it can attribute causation
explanatory variable
the variable that is manipulated, the method/treatment
response variable
the characteristic measured (ex: success/failure)
lurking variable
outside variable (other than explanatory) that may affect the response variable -> create a
control or comparison group
open questions
subjects write whatever they want to answer questions; unlimited responses, hard to analyze
but subjects can answer more freely and specifically
closed questions
subjects choose from possible answers specified by the researcher; limits responses, easier
to analyze; must be sure to consider all reasonable options and provide an "other" option
causes of bias (sampling)
, - undercoverage
causes of bias (non-sampling)
- misleading response
- non-response
- interviewer effect
- question order
- question wording
factor
planned explanatory variable (includes treatment and placebo groups)
treatment (ttt)
different values of the explanatory variables administered to subjects
control group
subjects who received no specific treatment
replication
apply each treatment to more than one subject of each treatment group -> reduces the chance
of variation from the population in the results (NOT the same as reproDUcing)
