Stats M. 5 Flashcards
Census
obtain information from the entire population of interest
sampling
obtain information from a subset of the population of interest
designed experiment
researchers assign subjects to a certain group, intentionally change the conditions of each group, and record the value of the response variable for each group
-establishes causation
observational study
researchers observe characteristics and take measurements
-recorded in natural settings
-only reveal association or correlation (NOT causation)
sampling frame
list of all of the members of a population that can be sampled
sampling method
the process used to obtain a sample
random sampling
employs a random device to select the sample
Simple random sampling
each possible sample of a given size has the same chance of being selected
-with or without replacement
-randomly select 10 employees from a company of 250 employees
Stratified random sampling
split the population into subgroups that share a similar characteristic and take a sample from each subgroup
-splitting band up into woodwinds, brass, percussion, and string sand taking a random sample from each subgroup
systematic random sampling
use a pattern to determine which members of the population are included within the sample
-to check the accuracy of an ice cream machine, every 500th. container is selected a weighed
cluster random sampling
we split the population into diverse sections and take a census from randomly selected sections and nothing from others
-in a medical research study, a researcher randomly selects hospitals and then interviews all the patients at the selected hospitals from that day
treatment
experimental condition (manipulated by the researcher)
experimental unit
subjects within the experiment that receive treatment (including no treatment from the control group)
treatment group
groups that receive the treatment (experimental condition)
control group
group that participates in the experiment and whose treatment is “nothing”
placebo
a “fake” treatment that looks just like the treatments being tested
-keep subjects who are in the control group from knowing they are in the control group
placebo effect
subjects treated with a placebo sometimes improve even though they are not receiving an actual treatment
single blind
subjects don’t know which group they are in
double blind
subjects and collectors don’t know which groups the subjects are in
randomization
randomly assign subjects to treatment groups
-eliminates bias
replication
apply each treatment to a number of subjects
-repeat the entire experiment at different levels of the treatment
control
make conditions as similar as possible for all treatment groups
blocking
group similar individuals together and then randomly assign treatments within each of the blocks
convenience sample
include individuals who are easy to sample and therefore may not represent population
-standing outside Armstrong and asking students questions
volunteer sample
made up of people who self-select into the survey
-not representative of the population
-follow up questionnaires received via email after visiting doctors
under coverage
occurs when the sampling frame foes not represent all parts of the population f
over coverage
occurs when members that are not in the population of interest are included in the sample
non response bias
some sampled subjects cannot be reached or refuse to participate
response bias
wording of a question is confusing or misleading or subjects lie because they think their response is socially unacceptable
